Quarterly journal of microscopical science

so
iw
4
,

Digitized by the Internet Archive
in 2009 with funding from
University of Toronto

htto://www.archive.org/details/nsquarterlyjourn65comp

) ae

~

we my Shay a:
(2 ke
7 ie y Ci 7 : ae fn) ,
See a
i by f i

¥' y y ! | !
4 yi j
J

i y : }
pa OP ia! ih }
ie iy if; iy ; ‘
TRA
ian Vy 4 j

ae aay He ak ¥) " wea |
ie lite iN i |
ue i an iB?
, UT iet s Wl.

: 5 A sero tel nth a:
ia ai vi AN AMT - ve
hy ar Pea ator kaa ye ale | a iy

a
RG

hf ’ 4 Mi yy un nae ‘ ¥
sit RA i ae tk
; ie Mit ft

hy AA ba ih ai ‘ svi me

: al a a

QUARTERLY JOURNAL

MICROSCOPICAL SCIENCE.

HONORARY EDITOR:

Sir RAY LANKESTER, K.C.B., M.A., D.Sc., LL.D., F°R.S.

EDITOR:

EDWIN S. GOODRICH, M.A., F.RS.,

PROFESSOR OF COMPARATIVE EMBRYOLOGY IN THE UNIVERSITY OF OXFORD}

WITH THE CO-OPERATION OF
SYDNEY J. HICKSON, M.A., F.R.S.,

BEYER PROFESSOR OF ZOOLOGY IN THE UNIVERSITY OF MANCHESTER ;

GILBERT C. BOURNH, M.A., D.Sc., F.R.S.,

LINACRE PROFESSOR OF COMPARATIVE ANATOMY IN THE UNIVERSITY OF OXFORD;

J. GRAHAM KERR, M.A., F.R.S.,

REGIUS PROFESSOR OF ZOOLOGY IN THE UNIVERSITY OF GLASGOW ;

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

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

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

VOLUME 65. New Series.
WITH LITHOGRAPHIC PLATES AND TEXT-FIGURES

OXFORD UNIVERSITY PRESS,
HUMPHREY MILFORD, AMEN CORNER, LONDON, E.C. 4.
192]

. ‘ ,
5 ‘fe 5 rf Pot wis eat

r) AC

gis Cr, ; . “i
Co ye ty if, Zs | a ie. e
SE ae

~ UN “ A ' midst, | _
\Yako \v {>

ma \\b eee ;

J 4 8

7 /4] 4" a

rR oe ee

“a

Q H : a

’ PRA) Trill) An
Reh bet
figs

CONTENTS

CONTENTS OF No. 257, N.S., December, 1920.
MEMOIRS :

PAGE
The Structure of certain Chromosomes and the Mechanism of their
Division. By Arruur Bottes Ler, Hon, F.R.M.S. (With
Plates 1-2) - : : : 1
® On the Pharyngeal or Salinas Gland < ie Biarihworst, Dy D.
Kerr, Sc.D., Beit Memorial Research Fellow. From the Quick
Laboratory, Univ ersity of Cambridge. (With Plate 3 and 7 Text-
figures) . : - : 33
Some fibaeraations on Gadel Auiioes ad eee in the
Gecko (Hemidactylus flaviviridis, Riippel), with Notes
on the Tails of Sphenodon and Pygopus. By W. N. F. WoopLanp,
D.Sc. (Lond.), Indian Educational Service, Senior Professor of
Zoology, May Central College, Allahabad, India. ave ith 6 Text-
figures) . 63
¢ On the Etenornics hl Develoy mene a Lygocer us Pei aee
manus, Kieffer, and Lygocerus cameroni, Kieffer, (Procto-
trypoidea- Ceraphronidae) parasites of Aphidius (Braconidae).
By Maup D. Havmanp, Fellow of Newnham College. Saas
18 Text-figures) ’ 101
On the Terrestrial Spee fom the Talend of Mawritins nd
Rodrigues ; with a Note upon the Canal connecting the Female
Genital Organ with the Intestine. By Toxto Kasurakt, Zoological
Laboratory, The Museums, fies (With Plate 4 and 6 Text-
figures) . 129
Gonospora wee Hinti. n. Sp., a Gieeame habits fe. egg of
Arenicola, By Epwin 8. Goopricu, F.R.S., and H. L, M. Prxetn
Goopricu, D.Sc. (With Plates 5-6) . : 157

CONTENTS OF No. 258, N.S., March, 1921.
MEMOIRS :

The Eye of Peripatus. By Wiit1am J. Dakin, Derby Professor of
Zoology, University of Liverpool ; late Professor of Biology in the
University of Western Australia. (With Plate 7 and 3 Text-figures) 163

®* On the Development of Cucumaria echinata v. Maren-
zeller. By Hrrosnt OnsHma, soe Plates 8, 9, and 11 Text-
figures) . é : 173

* Observations on the Breton paraeatie in eae LaRTHio as Seton
Desn, Part III. Pseudo-trichonympha pristina, By D. Warp
Cutter, M.A. Cantab. (With Plate 10 and 8 Text-figures) . . 247

The Cytoplasmic Inclusions of the Germ-Cells. Part IX. On the
Origin of the Golgi Apparatus on the Middle-piece of the Ripe
Sperm of Cavia, and the Development of the Acrosome. By
J. BRrontéE GATENBY, B.A., B.Sc., D.Phil. (Oxon.), Lecturer in
Cytology, University College, London, and Senior Demy, Magdalen
College, Oxford; and J. H. Woopcer, B.Sc, (Lond.), “Assistant in
Zoology and Comparative Anatomy, University eo London.
(With Plates 11, 12, and 2 Text-figures) . 265

Further Studies on Restitution-bodies and free Tissue- huitans in
Sycon. By Junian S, Huxtey. (With Plates 13 and 14) . .. 293

CONTENTS OF No. 259, N.S., August, 1921.

MEMOIRS :

The Proboscis of the Syllidea. Part I. Structure, By W. A. Has-
wett, M.A., D.Sc., F.R.S., Emeritus Professor of Biology, Univerg”
of Sydney. (With 1 late aT 5) :

The Life-history of Melicertidium oc toe oniibipaies (Sars),
a Le »ptomedusan with a theca-less Hydroid Stage. By Prof. JAMES
F. Geman, University College, Dundee, (With Plate 16)

On the Blood-Vascular System of the Earthworm Pheretima, and
the Course of the Circulation in Earthworms. By Karm Narayan
Bani, D.Se,, of the Muir Central College, Allahabad, India, (With
11 Text-figures)

The Development of fie be ary shalt Oven Ege otis a Mostjuitet
Anopheles maculipennis, Meig. By A. J. NicHoLson,
M.Sc, (Birmingham), (With Plates 17-20) .

On the Bionomics and Post-Embryonic Development of cater
Cynipid Hyperparasites of Aphides, By Maup-D, Havmanp,
Research Fellow of Newnham College. (With 11 Text-figures)

Notes on the Larval Skeleton of Spatangus purpureus. By
Hrrosut Onsuima, Asst. Professor in the Department of Agricul-
ture, Kyushiu Imperial University, Fukuoka, Japan, (With
Plate 21) . ; : : : : c c A

CONTENTS OF No. 260, N.S., November, 1921.
MEMOIRS :

On the Classification of Actiniaria, Part II, Consideration of the whole
group and its relationships, with special reference to forms not
treated in Part I. By T. A. SrepHEnson, M.Sc., University meee
of Wales, Aberystwyth, (With 20 Text- figures) : :

he Development of the Sea Anemone Bolocera tuediae (gennat: )
By Prof, JAMEs F, Gremmity, University College, Dundee. (With
Plate 22) .

Observations on the Sings a the Nucleus sal as Determinaiiaat
3y CHRISTIAN CHAMPy, Professeur agrégé & la Faculté de Médecine

de Paris, and H. M, Carteron, Demonstrator in Histology, Uni-

versity of Oxford, (With Plates 23 and 24 and 11 Text- figures)

On the calcium carbonate and the calcospherites in the Malpighian
tubes and the fat body of Dipterous larvae and the ecdysial elimina-
tion of these products of excretion, By D, Kemun, Se.D., Beit
Memorial Research Fellow, (With 5 Text-figures)

The Early Development of the summer egg of a Cladoceran (Simo -
ce phalus vetulus), By H. Granam Cannon, B.A., Demon-
strator in Zoology, Imperial College of Science, South Kensington,
(With Plate 25 and 1 Text-figure)

Studies in Dedifferentiation, II, Dedifferentiation and Toren
in Perophora. By Junttan 8, Huxtey, New ee Oxford.
(With Plates 26-28 and 1 Text- figure) :

LEVIEW,

“The Microtomist’s Vade-Mecum ’, by A. Boies Ler, edited by

Prof. J. B. Garenspy
InDEX OF VOLUME 65

4 sh

PAGE
323
339
349
395

451

479

493

577
089
611

627
643

699
701

The Structure of certain Chromosomes and the
Mechanism of their Division.

By
Arthur Bolles Lee, Hon. F.R.M.S.

With Plates 1 and 2.

Part I. Structure.
(a) Historical.

THE first suggestion of any structure at all observable in chro-
mosomes seems to be due to Pfitzner (‘ Morph. Jahrb.’, vii,
1882), who suggested that a chromosome is made up of a row of
eranules of chromatin embedded in an achromatic or less
chromatic thread. Belief in these granules—later dignified by
the names of ‘chromomeres ’, ‘ chromioles ’, and the like—long
held sway, and still lmgers in many minds. I do not think it
necessary to enter into a detailed discussion of this view ; for
I think it is now indubitable that the supposed granules are
nothing but the misinterpreted images of twists of the chromo-
some, or of bulges in it. The figures illustrating this paper
afford abundant instances of bulges caused by twists of the
chromosomes ; and those illustrating my paper on the chromo-
somes of Paris quadrifolia (‘La Cellule’, xxviii, 2,
1912, p. 265) of bulges caused by alveoles in them ; either
of which, if indistinctly seen, may lend themselves to an
erroneous interpretation as granules.'

At the present time two other theories are in the field: the
chromonema theory, and the alveolation theory.

1 The chromomere theory seems to have been given up even by Fl em-
ming, who at one time accepted it. For in his paper, ‘* Neue Beitriige zur
Kenntniss der Zelle”’, IT. Th. (‘Arch. mikr, Anat.’, xxxvii, 1891), whilst dis-
cussing the division of chromosomes, no mention is made of the granules,
which he had formerly taken to be active agents of the division ; and his
figures no longer show any such granules, but in many places show instead
more than hints of the bulges of a twisted thread.

NO, 257 B

9 ARTHUR BOLLES LEE

The chromonema theory conceives of the chromo-
some as composed (at least at a certain stage) of a continuous
filiform chromatic element—oftten spirally coiled supported
or contained in an achromatic cylindrical

on an achromat ic core,

matrix.

This notion is due to Baranetzky, who in 1880 (‘ Bot.
Zeitung’, p. 241) deseribed and figured, in the pollen mother-
cells of Tradescantia virginica, a fine chromatic fibre
spirally coiled, at the surface of the chromosomes, round an
achromatic core.

In 1901 Janssens (‘ La Cellule’, t. xix, pp. 55 and 58) de-
scribed similar chromatic spirals uncoiling themselves from the
chromatin clumps of the restmg spermatogonia of the newt,
and even figured similar filaments coiled with the chromosomes
of the telophase, closely applied to an enveloping membrane.
Later (‘ La Cellule’, t. xxii, 1905, p. 413 and figs. 42 to 50 and
52 to 55) he figured achromatic membranes clearly existing
around the ‘ pachytene’ chromosomes of the auxocytes of
Batrachoseps attenuatus, and concluded that in the
stages of the bouquet and the strepsinema all the chromo-
somes are in contact with their neighbours by means of these
membranes—' les chromosomes se touchent tous ’.

Bonnevie (Arch. Zellforsch.’, i, 1908, p. 450, and particu-
larly pp. 471, 473, 477, 479, 509 ; i, 1908, p. 201, and particu-
larly pp. 266-70; 1x, 1913, p. 483) from a study of chromo-
somes of Ascaris, Allium, and Amphiuma, deduces
the following conclusions: A prophasic chromosome consists of
an achromatic core on the surface of which is spread a con-
tinuous mantle of chromatin (I find no mention of a membrane).
In the telophase this mantle becomes differentiated into a
spirally coiled thread, whilst the achromatin is cast out into the
new nucleus. The spiral threads of chromatin then put forth
lateral processes which anastomose with those of neighbouring
threads, and so form a nuclear network. At the next prophase
the anastomoses are withdrawn, the chromatin threads shorten
and thicken, and differentiate mto chromosomes showing a
newly formedachromatic core with a continuous mantle of

STRUCTURE OF CERTAIN CHROMOSOMES 3

chromatin derived from the persisting chromatin of the telo-
phasic spirals. These spirals are therefore the rudiments of
anew generation of chromosomes.

K. C. Schneider (‘ Festschr. f. R. Hertwig ’, i, 1910, p. 215)
also describes the chromosomes of the anaphase as consisting
of a chromatic spiral enveloping an achromatic core; but
finds this spiral become double in the telophase. He does
not find in the quiescent nucleus a network formed by ana-
stomosing processes of the spirals, but only a tangle formed by
the attenuated and elongated spirals themselves. But these
spirals are differentiated into chromatic granules united by an
(apparently) achromatic thread. The prophasic chromosomes
are formed by the condensation of the granules into (two) new
chromatic spirals enveloping this thread.

Vejdovsky (‘ Zum Problem der Vererbungstriiger ’, Prag,
1912) also finds that a ‘ripe’ chromosome consists of an
achromatic¢ core round which is wound a chromatic fibre. To
this fibre he gives the name of ‘chromonema’. He finds no
membrane. At the telophase, the achromatic core is cast out,
and, swelling, forms the nuclear enchylema. But the chromo-
nema differentiates into a new achromatic thread with chroma-
tie granules (‘ chromioles ’) imbedded in it. The threads thus
constituted anastomose into the network of the quiescent nu-
cleus. At the prophase the anastomoses are withdrawn, and the
chromioles fuse into a new continuous chromonema, spirally
coiled round the persisting threads. In the later prophase the
chromonema segments into ‘chromomeres* which undergo
bipartition, and so bring about the division of the chromosomes.
So that Vejdovsky, though a supporter of the chromonema
theory in so far as he recognizes the chromatic thread as a
chief constituent of the chromosome, does not entirely discard
the granule theory of Balbiani and Pfitzner. Like Bonnevie,
he conceives of the chromonemas as the rudiments (A niag en)
of a new generation of chromosomes (op. cit., p. 171. et
passim).

The alveolation theory was foreshadowed by some
observations of van Beneden’s, but has only been worked up

B 2

4 ARTHUR BOLLES LEE

into a theory of the quiescent nucleus lately, by Grégoire and
his pupils (Grégoireet Wygaerts, “Tha pine du

Noyau et la Formation des Chromosomes ” *La Cellule’, xxi,
1903, p. 7; Grégoire, ~ La structure de élément ance

somique au repos et en division”, ibid., xxil, 1905, p. 311;
and other papers by himself and his pupils). According to this,
the homogeneous chromosomes of the prophase become during
the telophase honeycombed with numerous vacuoles or alveoles,
which end by splitting each of them up into a mere network of
chromatin. These networks then anastomose by lateral pro-
cesses, and there is thus formed a network of networks, the
reticulum of the quiescent nucleus. At the next prophase the
anastomoses are drawn in, and homogeneous chromosomes are
formed anew from the remaining reticular tracks by the
obliteration of their alveoles and condensation of their honey-
combed chromatin into a homogeneous thread.

I have already (‘ La Cellule’, xxviii, 1918, p. 265) published
a study of the essential points at issue between Grégoire and
Bonnevie, as exemplified in the pollen grains of Paris qua-
drifolia. I there found the chromosomes to be alveolated
as described by Grégoire; but I did not find their alveolatioa
to progress in the telophasic chromosomes to the point of
breaking them up into networks. On the contrary, I found their
alveoles to disappear, and the chromosomes to condense mto
thin spiral threads. But I did not find these threads to ana-
stomose into a network in the resting nucleus, as described by
Bonnevie. I found nothing worthy of the name of a network,
but only a tangle of the much elongated and attenuated spiral
chromosomes. J found these persisting throughout the mter-
phase, and at the next prophase forming typical chromosomes
hy shortening and thickening and at the same time again
becoming alyeolated. Fig. 1+ represents a typical group of

’ This is a drawing of the anaphase shown in fig. 6 of my paper, amended
by the addition of the sheath and lateral processes round the axis of the
chromosomes, which had escaped me when the original drawing was made.
I think it quite likely that there may be also a very fine periaxial spiral, in
correspondence with the lateral processes, round the axis of the chromo-

STRUCTURE OF CERTAIN CHROMOSOMES 5

chromosomes honeyeombed by easily perceptible alveoles,
of the existence of which there can be no doubt. For a detailed
description of the characters of these alveoles, the reader will
do well to refer to the paper quoted. Fig. 2, which is a slightly
corrected copy of fig. 13°° of the same paper, shows the solid
spiral threads into which these alveolated chromosomes become
transformed during the telophase.

Later, [ have extended this study to the chromosomes of the
nuclei of the pollen cells and of some tissues of Liliuim
croceum and L. martagon, and obtained exactly the
same results. Combining these results with those of Grégoire
and Wygaerts for Trillium grandiflorum and T.
cernuum,of Grégoire for Allium cepa, A. asca-
lonicum, and A. porrum, and of Sharp (‘La Cellule’,
xxix, 1913, p. 297) for Vicia faba, and rejecting as erroneous
the statements of those writers who have described in plant
chromosomes a spiral fibre instead of alveoles,’ we find that all
the plant chromosomes that have been successfully studied
hitherto possess an alveolated structure in the prophases,
equatorial phases, and anaphases.

The present paper deals with certain anim a | chromosomes.
Only one recent writer, Kowalski, has described any of these
as alveolated. Kowalski (‘La Cellule’, xxi, 1904, p. 349),
studying divers nuclei of the larval Salamander, arrived at the
conclusion that their chromosomes all conform to the alveolation
theory. I have carefully examined all the chromosomes studied
by Kowalski, and many other of the Salamander larva,

somes ; and that if this spiral cannot be made out with certainty (1 think
I sometimes catch glimpses of it), it is because the image of it is obscured
by that of the walls of the alveoles. But this, if it exists, is certainly not
the spirally coiled thread described by Bonnevie. I intend to return to this
point in another paper.

1 Bara necki’s observations may safely be rejected, because they have
been controlled by Carnoy and by Stras burger, who didnot find the
alleged fibre ; and those of Bonne vie on Allium, because they are con-
tradicted by the everyday experience of botanical cytologists. Both these
writers have apparently misinterpreted images of walls of alveoles, or of
torsions of the whole chromosome, as images of a spiral fibre.

6 ARTHUR BOLLES LEE

and find that neither these nor any other of the animal chro-
mosomes that I have studied do so; but that on the contrary,
at one period of their existence, they all do possess a cer-
tain spiral differentiation answermg, to some extent, to
Vejdovsky’s ‘chromonema’. The following pages set forth
the evidence for this, but will, as I think, also show that the
advocates of the chromonema theory have pushed it too far ;
for the spiral differentiation im question does not constitute
an independent fibre, and does not form the germ of a new
chromosome.

The chromosomes described are chiefly those of prophases,
equatorial phases, anaphases, and telophases; but I have
touched on those of some interphases in which certain of their
characters are demonstrable. I do not attempt in this paper to
describe the nuclein elements of completely ‘ resting’ nuclei.
The results set forth are based on the study of chromosomes of
the Amphibia (chiefly Urodela). Careful investigation of the
nuclei of the other classes of the Vertebrata has shown that their
chromosomes, though conforming apparently in all respects with
those of the Amphibia, are mostly too small to afford trust-
worthy images of the details im question. The same is the cage
with most of the Invertebrata, only certain nuclei of the Ortho-
ptera being found to possess chromosomes which, though
smaller than most of those of the Amphibia, yet afford images
which are often clearer. The majority of the figures are of
chromosomes of spermatogonia, the most favourable kind for
study. Those of spermatocytes and odcytes are excluded from
the survey, because in them the details are obscured by the
complications due to the processes of conjugation. Most of
the images described are from paraffin sections: surface
preparations show nothing more than these. The most trust-
worthy fixing agent has been found to be picro-formol (Bouin’s
formula). [ron haematoxylin has been found to be incomparably
the best stain; but it should not be used quite as laid down
in the books, which give excessive times and strengths. You
should mordant (sections of 7-5 microns, or less) for not more
than 23 minutes in a solution of iron alum of 4 per cent. or

STRUCTURE OF CERTAIN CHROMOSOMES ‘4

weaker ; and stain in a half per cent. (or weaker) solution of
haematoxylin till the sections appear dark grey, not black
(about twenty-five minutes in a virgin solution, or not more
than four in one which has already had several slides passed
through it); and differentiate in the iron solution for at least
a couple of minutes after the sections, examined in water,
seem sufficiently extracted. For the stain always appears much
lighter in water than m balsam. For the study of the sheath,
mount in Gilson’s camsal balsam or euparal, rather than in
balsam.

(b) Descriptive.

It will be best to begin with the study of some chromosomes
taken at the anaphase, the most favourable moment, figs. 3 to
18.1 The chromosome of fig. 6, which may be taken as typical,
is from a spermatogonium of Salamandra maculosa.
It shows the following two (not three) constituents, namely
a chromatic (basophilous) axis, and an ‘achromatic’ (i.e.
acidophilous) sheath enveloping this. The chromatic axis
is by far the more conspicuous of the two ; so much so that, as
the sheath is seldom conspicuous enough to compel attention,
the axis alone is all that is usually seen, and is therefore generally
taken as the whole of the chromosome. But the sheath (which
is none other than the achromatic membrane described by
Janssens, ‘La Cellule ’, xxii, 1905, p. 413 and figs. 42 to 50 and
52 to 55, as found in the auxocytes of Batrachoseps
attenuatus), though it is a difficult object on account of
its great tenuity, can generally be made out in well fixed
specimens.

The axis has approximately the form of a cylinder, showing
a circular section. But it is not a cylinder of regular calibre,
for it is generally somewhat dilated at the ends, as seen in
figs. 6, 7, 14 (and to a slighter degree in figs. 3 and 4), thus
becoming somewhat claviform. And it is generally notably
narrower at the polar bend than elsewhere, figs. 3, 4, and

1 For the objects from which these figs. are taken, see the Explanation
of the Plates.

5 ARTHUR BOLLES LEE

especially 14 ; and at this point is generally somewhat flattened.
At its ends (where not sectioned by the knife) it terminates
in & smooth dome-shaped surface, from the sumnut of which
there can frequently be seen to emerge a tiny tag, the vestige of
its union with its late sister chromosome, figs. 6, 7, 14, 5, 12,
all of which show the tag; and 3 and 4. — It is undoubtedly
solid, not hollow. Surface views (see the figs. quoted) show no
lumen, nor any trace of the alveoles found in plant chromosomes ;
but they may seem to show a border darker than the Innmermost
part, as in one or two of the chromosomes of figs. 3, 4, and 5.
But in these cases it is generally possible to see that this border
is not continuous, but consists of a series of elongated dots.
Transverse sections frequently show as disks with a dark
border and lighter centre, fig. 15, which may give rise to the
unpression that there exists an axial lumen. But I have satis-
fied myself that the axis is in reality solid, and that the dark
border is due, for the most part at least, to the periaxial spiral,
about to be described, showing there. It is frequently pos-
sible, by very careful focusing, to see that this border is darker
at one side of the disk than the other, which I take to be due
to a sector of the spiral being in sharpest focus there. Thus in
fig. 15 a, at the top left it is darker to the right ; at the top right,
darker at the bottom ; and in the lowest disk darker at the top.
And the darker sector can be seen to turn round the disk with
every change of focus ; which is Just as a spiral viewed end-wise
ust behave." Similar images are shown, more clearly, by three
of the less darkly stained chromosomes of fig. 15¢. Those of
fig. 15b show the darker border as an apparently entire ring,
not a mere sector ; and the fourth chromosome of 15¢ shows
as a disk with a mere hint of a darker border.

Further, in the lighter-coloured centre of the disk there can
sometimes be seen a darker comma-shaped dot. One of these
is Seen as a mere dot in the two upper disks of fig. 15 a, and as

' Por this spiral to be demonstrated it is imperative that the chromosome
be not overstained, for if it is the axis will appear as dark as the spiral, and
the spiral will not beseen. Vejdovsky’s figures grossly exaggerate the
distinctness of the spiral at the best of times.

STRUCTURE OF CERTAIN CHROMOSOMES 9

a comma in the lower one. This I have no doubt is nothing but
an out-of-focus portion of the periaxial spiral coming into view
from a lower depth, in a somewhat tilted chromosome.

I think the utmost that can be admitted in the way of
any hollowness of the axis is that this may possibly possess
a cortical layer somewhat denser than the rest. But I think
the appearances are sufficiently accounted for by the periaxial
spiral.

On the surface of this otherwise homogeneous cylinder there
runs a spiral of somewhat denser substance than the rest,
figs. 3 to 14. This periaxial spiral is evidently some-
what denser than the rest, because it resists decoloration in
regressive staining more strongly ; but it is evidently of the
same composition, for its affinities for stains are the
same. It is not something separate from the rest of the
cylinder, but is continuous with it. It is not fittingly described
as a fibre woundround a core: for there is no space between
the spiral and the rest of the axis; there is no hint of a dis-
continuity between the two either in surface views or in section.
Nor should it be described as a fibre countersunk or partially
embedded in the axis: for if it were a fibre its section would
show as a small circle (or other figure) having a definite limit all
round ; but these spirals only show a definite limit outside
the general surface of the core; inside, they merge in its
substance indistinguishably. Vejdovsky’s term of * chro-
monema ’ is a misnomer: the thing is not a fibre, but a rib or
ridge. It must therefore be taken to be a mere spiral con-
densation of the cylinder substance.

It is true that cases such as that shown in the left-hand
chromosome of fig. 3 are not very infrequent. At the middle
of the longer limb of this chromosome there is a break ; and
the spiral is seen to bridge over the gap between the two parts,
But I take it that that is only because its toughness has enabled
it to resist where the rest yielded : just as when you break a
twig you frequently get the two parts hanging together by a
strip of bark.

The periaxial spiral sometimes seems to course wunterruptedly

10 ARTHUR BOLLES LEE

the whole length of the chromosome (with the exception of
the extreme tips). But often, as shown in fig. 14, it seems to
be interrupted at the polar bend, the bend only showing an
attenuated tract of the core without any perceptible mdge on
it. At the tips, the spiral ceases at the base of the dome-shaped
surface, and is not continued up to its summut, figs. 6, 7, 14.

It seldom shows a regular pitch throughout, for its turns are
sometimes very widely spaced, as in figs. 6 and 7, but often
so closely approximated that they almost touch one another,
as shown at the tip of the right-hand lmb of fig. 14. The
drawings, in which the spacmg between each turn has been
reproduced with scrupulous care, will give a better idea of
this than any description.

It has been said that the spiral shows no definite limit
inside the general surface of the axis; but outside this
it does. Its optical section there shows as a series of minute
conical elevations, giving, in inferior images, the appearance of
a row of minute thorns. These elevations are figured in several
of the drawings of recent observers, and are by their authors
considered to be in effect minute thorn-like processes. But
careful observation of well-preserved specimens (with good
objectives and a first-class condenser) shows that the two out-
lines of each of these apparent cones do not terminate at
the apparent apex shown under inferior definition, but merge
there into a single line which is continued outwards, generally
in a perceptible curve, till it reaches the membranous sheath.
And it can often be seen to insert on this by means of a delicate
conical enlargement. All the drawings, figs. 2 to 18, show some
of these lines, and the enlargement is shown very clearly in
figs. 7 and 23, and less clearly, but still recognizably, in several
parts of the remaining figures. These enlargements, then,
show as a row of minute cones having their bases applied to
the iner surface of the sheath, and their apices continuous
with the line which springs from the cones on the axis. ‘There
is always one of these cones on the sheath for each one on the
core. Those on the sheath can often be seen to be situate, not
diametrically opposite to those on the core, but a little higher

STRUCTURE OF CERTAIN CHROMOSOMES 11

up or lower down, at the extremity of a line which prolongs the
course taken by the spiral across the axis. This is shown in
fig. 14; but in the remainder of the figures is not shown clearly
on account of the frequent derangement of the symmetry of
the disposition caused by stretching or other displacement of the
sheath. But there can be no doubt that the relations of the
two sets of cones are as described.

The line that joins the elevations on the axis to the sheath,
including its aponeurosis thereon, is very faint, but it can
sometimes be seen to be stained. In that case, it stains in the
same tone as the axis; for instance, I have obtained it un-
mistakably red with safranm. This ligament, then, is a
prolongation of the substance of the spiral. And, taking all
these facts together, we must come to the conclusion that each
of these apparently filiform hgaments is nothing but the optical
section of a flange-lke or pterygoid membranous ex-
pansion of the spiral. This cannot be seen as a membrane,
full face, because it winds round the axis in such a way as always
to present its edge to the observer; and also because it is so
thin (I should think anything under a twentieth of a micron)
that if ever a portion of it should come to lie full face it would
still be invisible through its thinness."

We may, if we like, call the optical sections of this membrane
lateral processes of the axis; which well describes the
optical image. But then we must bear in mind that there
is in reality only one of them, which courses contmuously
round the axis like the lamina spiralis cochleae round the
modiolus. And we can make a rough model of a chromosome
of this type by taking a carpenter’s screw and inserting it into
a quill into which it will just fit.

The whole of the chromatic axis, the innermost part as
well as the spiral and the lateral processes, is most decidedly
basophilous: no part of it is achromatic nor acidophilous
(which is what the authors quoted in the Introduction mean when

1 The aponeurosis of this membrane on the sheath can sometimes be seen

as a spiral line running along the sheath. I have abstained from drawing
it on account of the difficulty of showing it clearly.

12 ARTHUR BOLLES LEB

they say ‘achromatic ’). It stains energetically in the fresh
state with acid methyl green ; and in the fixed state it stains
energetically and selectively with safranin, gentian violet,
and the other usual basic stains. The only ground that I
can discover for the belief in an ‘achromatic’ core in it is
the fact discussed above, that the periaxial spiral generally seems
more darkly stained than the rest of the cylinder round which
it winds. But that does not in the least pomt to a difference of
chromatophily between the two. The inner part of the axis
stains (generally) Jess darkly than the spiral because it is less
dense. And that is all; for the two stain, qualitatively, with
exactly the same selectivity for staims.

The sheath isa continuous tubular membrane, of a thick-
ness of the order of about one-twentieth of a micron. It is of
irregular calibre, but roughly of a diameter of about three times
that of the axis (see figs. 2 to 18 and others). It is very fre-
quently seen to be indented where the lateral processes insert
on it, as though it were held down at these points, but blown
up between them. It is sometimes seen to be continued round
the tip, as in most of the figures given; but sometimes seems
only to reach to the base of the dome-like surface, as in fig. 14.
It is absolutely structureless. Itis decidedly acidophilous,
staining readily though somewhat feebly (to about the same
degree as spindle fibres, for instance) with Siurefuchsin,
Sdureviolett, or Lichtgri ; and not staining with basie dyes.
‘The space between this membrane and the axis is filled with a
substance of glassy clearness, which is free from all trace of
granules or other differentiations, and entirely achromatic,
not staiming in any way. If it appear to be tinted, as it some-
times may, that is due to the staining of the membrane. This
substance may be liquid, or may be gelatinous.

I find the sheath on all anaphase chromosomes of which
I can obtain sufficiently good images ; and have concluded that
it is as universal an attribute of all chromosomes of this stage
as the axis and the periaxial spiral.

These, then, are the features which can be detected on
favourable specimens of animal chromosomes at the anaphase.

STRUCTURE OF CERTAIN CHROMOSOMES 13

We have now to inquire to what extent they are present in
other phases ; and this with special reference to the assertion of
Bonnevie and Vejdovsky that at the telophase one part
of the chromosome axis is cast out mto the new karyoplasm,
whilst another persists as a spirally coiled thread which forms
the rudiment of the new chromosome.

At the end of the anaphase the ‘ daughter-star ’ of chromo-
somes contracts into a figure which is called by some the
‘tassement polaire ’,a term which we may translate by polar
elump. Inthis clump (figs. 29 to 34) the chromosomes become
so densely crowded, and even agglutinated together, that it is
impossible to follow out their minute details with accuracy
throughout (im the Amphibia: in some other groups the case
may be different). Still, enough can be seen in suitably fixed
clumps, such as those of figs. 30 and 31, to warrant the assertion
that the essential features of the chromosomes persist. In
fig. 30, for instance, the chromosome axes can in many places be
made out, appearmg as thin threads (therefore considerably
shrunken) collocated in pairs (an important detail, the dis-
cussion of which is best reserved for Part II). The periaxial
spirals can just be detected on some of them; and on others,
where they cannot be seen as lines wound round the shaft,
their presence is made probable by the lateral processes which
can be seen on their edges. And towards the ends of the chromo-
somes, wherever they stand clear, the sheath membrane can
generally be made out as a fine line bridging over the tips of
the processes. The sheath can indeed generally be seen round
the edges of even highly-agglutinated clumps, figs. 32, 33, 34.
In fig. 31 (Bombinator) these details can only just be
elimpsed here and there, on account of the smaller size of the
elements ; but indubitably exist there as described for fig. 30.
We may conclude that at the height of the clump stage the
chromosomes—though generally much shrunken, compressed,
crumpled, and otherwise distorted—have more or less retained
all their essential features.

This stage is of short duration, the clump soon passing by
a process of expansion (to be explained in Part II) into the

14 ARTHUR BOLLES LEE

telophase. This next stage will be most conveniently studied
in the spermatogonia and oogonia of the Amphibia. For
here, as the clump passes into the telophase, it expands into
a wide ring, on the surface of which the chromosomes are set
on widely spaced meridians, figs. 48, 44, 45, 48, 49, 50, and
others. Owing to this arrangement they show only a minimal
amount of overlapping, and, standing out on a clear back-
sround, can be studied with sufficient accuracy.

In the earliest stages of this process of expansion (figs. 35 to
38) we find much the same state of things as m the denser clump.
The paired chromosome axes can be more clearly distinguished :
periaxial spirals can be just detected on some of them, and
on others their existence is placed beyond all reasonable doubt
by the lateral processes visible on the edges of the axes. And
the sheath can be made out on many of them (same figs.). In
later stages such as figs. 89 to 47, the demonstration of these
details becomes more difficult, mainly on account of two com-
plications which here ensue. One of these is the formation
of trabeculae (‘ anastomoses’ of some authors) between the
chromosomes. These trabeculae obscure the lateral pro-
cesses, with which they are easily confused, and so deprive us
of an important guide for the detection of the periaxial spirals.
The other is, that as the clump expands, the chromosomes
elongate ; and as they elongate their duplicate axes t wine
round one another, figs. 85, 39 to 47.1 This involves

1 This gives us the key to Kowalski’s assertion (op. cit.)that the chro-
mosomes of the salamander larva are at certain periods alveolated. Thirteen
of his figures purport to show the alveoles in question. Eight of these are
of telophases. On comparing them with my figs. 39 to 51 it becomes
evident at once that Kowalski has interpreted images of doubled and
entwined chromosome axes as borders of alveoles—which is very natural,
for a thus doubled chromosome easily gives the impression of an alveolated
cylinder if you are not able to obtain a sufficiently sharp focusing of its
entwined axes, The remaining five of Kowalski’s figures of ‘alveolated’
chromosomes are of spiremes, such as my figs 25 to 27, and manifestly only
show that the chromosomes he had before him were double, transverse
trabeculae uniting their two moieties being taken for transverse walls of
axial cavities in an undivided cylinder or riband,

STRUCTURE OF CERTAIN CHROMOSOMES 15

a continual displacement of the direction of the axes, making it
extremely difficult to follow them accurately for more than very
short distances, and thus making it next to impossible to dis-
tinguish the periaxial spirals running across them. Still, at
this stage, it can be inferred with certainty that these exist at
least to some extent ; for indubitable lateral processes can be
made out in some places ; and the sheath can be observed with
certainty in favourable places, as shown in figs. 39 to 45 (in some
places of these, where not sufficiently evident in the drawings,
T have marked it with a cross).

When the expansion of the clump has attained its greatest
extent, we have the telophasie ring, figs. 48 to 51,
and others. The chromosome axes are here about as distinct
as before; but the periaxial spirals, lateral processes, and
sheath seem to be waning. The spirals can no longer be
seen as lines running across the shaft; and the lateral pro-
cesses can only be distinguished from the interchromosomal
trabeculae here and there. But this does not necessarily imply
that they have diminished in number. For at this stage the
chromosomes have elongated considerably ; and since by their
elongation the periaxial spirals and their processes must be
pulled away from one another, we naturally find far fewer pro-
cesses than before on any given length of an axis. But this is
probably not all that happens. The chromosome of the
anaphase and early polar clump is a very tightly twisted
cylinder ; and there is nothing forced in the supposition that
the spirals on its surface, and their lateral processes, are mere
effects of the torsion it has undergone. And it appears natural
that as the axis elongates at the telophase, it should unt wist ;
and that m consequence of this untwisting the spirals come to
subside into the shaft, carrying their processes down with them.
Not that the substance of the spirals and processes degenerates
or dissolves ; but that it undergoes a change of configuration :
as when I extend a finger, wrinkles start up on its surface ;
and when I flex it these wrinkles are smoothed down. But be
this as it may, it is certain that in the telophase the periaxial
spirals and processes begin to wane out of sight, till im the

16 ARTHUR BOLLES LEE

interphase it is seldom possible to detect even a vestige of them
with certainty.

As to the sheath at this stage, the appearances are similar.
In the nucleus of fig. 47 (Bombinator) (which shows one
half of a ring such as that of fig. 50), I am not able to see it,
except (possibly) on the chromosome at the extreme left. In the
nucleus of fig. 48, a later stage, also Bombinator, I have
not been able to detect it. In that of fig.49 (Triton) I think
I can see it in the two places marked with a cross, and glimpse
it in one or two others. In that of fig. 50 (Salamandra)
I have been able to see it in a fragmentary way in half a dozen
places, as marked. In that of fig.51 (Triton, follicle nucleus
of testis) I have been able to detect it m only three places
(also marked). It is certainly less abundantly evident in these
nuclei than in the earlier stages. And this can hardly be
accounted for by greater difficulties in the way of observation ;
for the chromosomes are now more widely spaced than before,
and observation of their edges should therefore be easier. Add
to this that the sheath when detected can only be made out
in a fragmentary way; can only be followed for very short
distances ; 1s less regular than in earlier stages, being frequently
distinctly dilated ; and can in some places be seen distinctly to
be ruptured (details which it is not possible to render satis-
factorily in a drawing). It may be stated as certain that
towards the end of the telophase the sheath has generally to
a great extent disappeared. And this disappearance seems to be
due to a process of real disintegration ending in destruction,
rather than to a mere change of configuration or relation of
parts. For in completely ‘ resting’ nuclei, even if these are
such as to offer every facility for observation, not a trace of it
can be detected.

The periaxial spirals and sheath thus lost to view at the
telophase come into view again gradually at the next prophase.
In the earliest stages in which the spireme is recognizable as
being indubitably such (figs. 24 and 25) it seems to consist
merely of tortuous naked threads (often clearly double, same
figs., and especially fig, 25). These may be united by inter-

STRUCTURE OF CERTAIN CHROMOSOMES 17

chromosomal trabeculae, but show no other lateral processes nor
sheath, though they may show in considerable abundance
minute nodes or varicosities. And the appearances suggest
that these are nothing but nodes of contraction and torsion
which may well be the first visible stage of the formation of
periaxial spirals and processes. In more advanced stages of the
spireme, such as that of fig. 26, lateral processes and a sheath
can often be made out with certainty, though with extreme
difficulty. At this time (when the loops of the chromosomes are
still so closely crowded together that almost all the sheaths are
in contact with their neighbours) the lateral processes are
sometimes so abundant that when fairly well visible they
give the image of a dense network spread over the whole of
the ground of the nucleus, as shown in fig. 26. Periaxial spirals
cannot be made out on the axes at this time ; but since we have
found that lateral processes are signs of the existence of the
spirals—being in fact only lateral expansions of these outwards—
we must admit that by this time the spirals are in course of
formation, if not completely formed, even when we cannot so
much as glimpse them.

As the chromosomes contract, they become more widely
spaced, and by the time they have contracted into the state
known as the ‘segmented’ spireme the lateral processes and
sheath have come into evidence as clearly as in the anaphase,
figs. 27 and 28. In fig. 27 the periaxial spirals cannot be made
out, the moieties of the chromosomes being here especially thin
(as I invariably find to be the case in endothelium nuclei).
In fig. 28 they can just be glimpsed in some places. But not
till we come to the chromosomes of the equatorial plate,
figs. 19 to 23, do we find the axis clearly differentiated into a
shaft with regular spirals on its surface. In equatorial plates
whose chromosomes have not entirely assumed the form which
they show when definitively arranged on the spindle, the
aspect of the axes is still rather that of a structureless though
twisted thread than that of a shaft with spirals on it (fig. 19).
In the entirely completed and regularized plate the spirals
certainly exist throughout, see figs. 20 to 25. If they do not

NO. 257 C

18 ARTHUR BOLLES LEE

at this time show with all the vigour and distinctness with which
they show at the anaphase, this may be sufficiently accounted
for by the greater difficulty of observing them in the closely
collocated moieties of the equatorial chromosomes. But it
may equally well be that they only attam their complete
development at the anaphase. We find, then, that the periaxial
spirals are only temporary formations. The assertion of B on-
nevie and Vejdovsky that they persist after the telophase
as rudiments of a new generation of chromosomes is contrary to
the facts. For we have found that the chromosomes of the
late telophase are for the most part without periaxial
spirals and sheath ; and that that which persists and passes
into the interphase is nothing but the thus simplified a x es of the
chromosomes. These, on passing into the interphase, frequently
become coiled into very regular spirals, such as have been de-
scribed and figured by many observers (for instance, Bonne vie
for Ascaris and Allium, Vejdovsky for Ascaris
and other objects, Schneider for Salamandra, and
myself for Paris quadrifolia); but these do not consist
of periaxial spirals set free from the shaft of the axis, but of
the entire axis in a simplified state. The chromonema
theory is a mare’s nest.

We may nowsumup. There are two types of chromo-
somes: one (hitherto only found m plants) which is alveo-
lated from the prophase to the telophase ; and one (hitherto
only found in animals) which is not alveolated at those stages
or any other. This last consists (at those stages) of a solid
basophilous axis, possessing a certain spiral sculpturing of its
surface, which we have called the periaxial spiral, and enclosed
in an acidophilous sheath. But this sheath is perhaps common
to both types ; and if the suggestion thrown out in the note on
p- 4 should prove correct the periaxial spiral would also be
common to both. Then the only important difference between
the two would be that the plant chromosomes have an alveo-
lated, i. e. more or less hollow, axis, whilst the animal chromo-
somes have an entirely solid one,

STRUCTURE OF CERTAIN CHROMOSOMES 13)

Part II. Division.
(a) Historical.

Tt was made out by Flemming in 1880 that the chromosomes
of the equatorial plate are double, that is, composed of two
similar longitudinal halves, closely approximated. The parallel-
ism and close approximation of these halves naturally suggested
that they arise by a longitudimal splittmg of a previously
undivided mother chromosome; and this suggested inquiry
as to the means by which the supposed splitting could be
brought about.

In 1881 Pfitzner’ put forth a schema of this splitting which
seemed plausible and met with general acceptance. According
to this, the mother chromosomes are composed either of a single
row of globular granules of chromatin, of a diameter exactly
equal to that of the chromosome and embedded in an achro-
matic matrix ; or of a double row of such granules, of only
half the size of those of the simgle row. These double rows
are sometimes very closely approximated, sometimes less
so; and finally separate from one another as daughter
chromosomes. The ‘splitting’ of the mother chromosome
would thus seem to be brought about by the binary division
of each of its constituent ‘ granules ’.

This theory won ready acceptance; and the supposed
‘ oranules ’, under the names of ‘ Pfitzner’s granules ’, * micro-
somes ’, ‘ chromomeres ’, ‘chromioles’, and the like, are still de-
scribed and believed in and made the basis of much fanciful
explanation.

According to my own very extended observations, this notion
of the ‘ splitting ’ of chromosomes being brought about by the
splitting of their component ‘ chromomeres ’ is baseless. For
no such granules exist at any time. It is abundantly clear to
me that all the appearances that have been described as

1“ Uber den feineren Bau der bei der Zelltheilung auftretenden faden-
formigen Differenzirungen des Zellkerns ”’, in ‘ Morpholog. Jahrbuch ’, vil,
p- 289—a much quoted but rather wretched performance,

C2

20 ARTHUR BOLLES LEE

‘ Pfitzner’s granules ’, ‘chromomeres ’, and the like, are, as
already explained, nothing but ill-seen and faultily interpreted
images of bulges and twists of the axis of the chromosomes
(figs. 8 to 23 and many others of this paper should make this
sufficiently clear). It therefore only remains to be seen whether
any other mode of division can be made out.

To settle this point, the first step must be to make out at
what stage chromosomes can first be seen to be double. Accord-
ing to Flemming (‘“ Neue Beitriige zur Kenntniss der Zelle”’,
ii, in ‘ Arch. mikr. Anat.’, xxxvul, 1891, pp. 787, 744, and 745) the
supposed splitting takes place in the spireme stage. And
this is apparently the view still taken by the great majority
of cytologists.

I am not aware that any observer has asserted a division of
chromosomes during the interphase. A longitudinal splitting
at the telophase has been asserted by several writers,
and with especial insistence by Dehorne. ‘This writer even
maintains (in his ** Recherches sur la division de la cellule ”’, in
‘ Arch. f. Zellforschung’, vi, 1911, p. 613) that it may take place
as far back asthe anaphase. This is indubitably erroneous.
For beyond all doubt at this stage the chromosomes show no
hint of duplicity. But as regards the telophase I find
that—in some cases at least—at that stage the chromosomes
are certainly double—in a sense; and I acknowledge the
essential correctness of Dehorne’s clever figs. 7, 9, 10, 11, 12,
and 18 (his fig. 6, which corresponds to my fig. 48, I think has
been imperfectly understood by him). But I find no trace of
any evidence that this duplicity is brought about by a
longitudinal splitting.

A division of the chromosomes at the telophase has also been
maintained by K.C. S¢hneider. In his‘ Lehrbuch der ver-
gleichenden Histologie’, 1902, pp. 10, 118, 848, and 989, he
states it as a probable inference. He suggests that at this stage
the chromosomes segment transversely at the polar
bends ; and that the two moieties thus formed grow past one
another so as to become parallelly approximated throughout
their lengths. I have duly investigated this point, and find no

STRUCTURE OF CERTAIN CHROMOSOMES Al |

signs of such a process. I need not enter into further details, as
Schneider himself seems to have abandoned his supposition.
For in a later work (his “ Histologische Mittheilungen”’, ii,
“Chromosomengenese ”’, in ‘ Festschr. f. R. Hertwig’, i, 1910,
pp. 218, 219, 221) he maintains his view that a division of the
chromosomes probably takes place at the telophase (or ana-
phase), but now supposes it to bea longitudinal one.*

Of this also I find no evidence. But I do find evidence of
another and simpler process by which the observed images of
duplicity are brought about. To the consideration of this we
may now proceed.

(b) Descriptive.

We have already seen incidentally, in Part I, that in the
Amphibia the chromosomes of the later telophase are double
structures, that is, that they consist of two chromatic threads,
longitudinally collocated and more or less entwined.

This is by no means peculiar to the Amphibia. In smaller
chromosomes than theirs the images are more difficult ; and in
much smaller ones it may be impossible to obtain satisfactory
resolution. But enough can be made out to leave no doubt
that it is a very widespread phenomenon. In the Mammalia
T have found it fairly clear in Homo, fig. 54. In some of the
Insecta (notably the Orthoptera) it is as certain as in the
Amphibia, see figs. 62, 66, 67. I think we may take it as the
invariable rule that in animals all the telophase chromo-
somes are thus doubled, that is, possess already the duplicity
observed in the chromosomes of the prophase. ‘This relieves us
from the necessity of looking for any process of splitting in the
phases between the telophase and the prophase ; and it only
remains for us to make out in what way the telophasic doubling
is brought about.

1 The reason he gives for this is a strange one. He admits (p. 218) that
the daughter chromosomes of the metaphase only show one spiral; but
thinks (without asserting it positively) that in the anaphase and telophase
they contain two, because * the coils they show are so closely set that they
could hardly be the expression of a single spiral’. How about a reel of
cotton ?

22, ARTHUR BOLLES LEE

To ascertain this we must return to the study of the earlier
telophase, or polar clump. In the daughter-star of the
anaphase (figs. 8, 4, 5, 61) we have a loose assemblage of
chromosomes, radially arranged in a rmg. These contract into
short staves; and as they contract the whole figure shrinks
(figs. 29 to 34), so that the staves become closely huddled
together and come into contact by their margins. They
generally seem to agglutinate there, and their outlines become
hardly distinguishable, indeed very often quite mdistinguish-
able. The clump then appears (figs. 80, 31, 38) as"an almost
homogeneous ribbed disk, with a central pore, generally
obturated by a perforated membrane or web formed (as shown
by profile views) by the confluent remains of the polar spimdle
fibres. The mutual contact or agglutination of the chromosome
staves takes place first in the region of the clump that is nearest
to the pole, their more distal portions remaining longer free :
so that at this stage we get the image of a compact ring with
digitiform processes depending from it—the ‘ figures pectini-
formes’ of Henneguy (figs. 32 and 34). In badly fixed cells
the clumping results in a formless mass, in which the chromo-
somes seem to have become completely fused together. This
state is shown in fig. 34. But, as I gather from the study of my
most favourably fixed specimens, this is an artefact ; and there
is not at any time a real fusion of the chromosomes, but only
intimate contact to the point of indistinctness, or possibly
superficial agglutination.’ Fig. 83 seems to me to show the
utmost degree of agglutination that should be taken to be
normal; and the real state of things to be fairly well repre-
sented by fig. 30 or 31.

Careful examination of the staves of the clump at this
stage seems to show that they are always in reality double
structures ; for in favourable cases they show unmistakable
indications of a longitudinal duplicity. In fig. 29 there are
four staves, marked with a cross, which show this. In the left-
hand one (near the top) the tip is distinctly bifid ; and this is

1 Cf. Janssens, ‘La Cellule’, xix. 2, 1901, p. 86, and Janssens et
Dumez, ibid., xx.2, 1908, p. 450 and fig. 15, who have arrived at the same
conclusion.

STRUCTURE OF CERTAIN CHROMOSOMES 3

also the case with the one at the bottom. In the two right-hand
ones the tips are distinctly double ; and by careful focusing
it can be made out that each of these staves is composed of two
longitudinal moieties, superposed and to a slight extent twisted
round one another. And in three or four of the short dark
staves of the ner tier there can be seen a light longitudinal
dividing line (not sufficiently clear in the drawing).

In fig. 33 nearly one-half of the twenty-one staves drawn are
seen to be notched at the periphery, and two of them show a
longitudinal dividing line continuing the notch mwards. In
fig. 30 I find three cases similar to these, and in fig. 32 two.
Thave no doubt that with better fixation these nuclei would have
shown several more such cases. In the clump of fig. 31 I think
I can detect three or four similar cases, though doubtfully.

The clump does not long remain in this state of dense ag-
glomeration, but soon begins to expand into the telophasie
ring. The manner of this expansion is as follows. Amongst
the staves of the clump—but never on their outer surfaces—
there appear certain hyaline globules which, growing, push the
staves apart and so loosen the clump. In fig. 38 are shown two
such globules, one to the right, and one to the left ; in fig. 35
three (on the left ; one very indistinct) ; in fig. 37 five ; in the
nucleus of fig. 36 there are a dozen or so, of which only a portion
of one (at the left) could be shown in the drawing, the rest bemg
too much masked by the sheaths. In fig. 62, to the right, are
seen three ; in fig. 67 two can just be glimpsed (at the left and
middle). These globules are entirely hyaline and uncolourable.
Their outlines are generally quite smooth. They are, as I think,
ovoid in shape, not spherical : they may show a circular outline,
as in the left-hand ones of figs. 38 and 43, and other places ; but
that is the expression of a transverse section of them. I suspect
that there is formed at first one of them for each chromosome.
If that be the case it is a likely hypothesis that they consist of
the clear contents of the sheaths of the chromosomes, expressed
from them by the pressure of the clump. But it is difficult to
ascertain the number formed, because they soon fuse with one
another into a small number of large globules, see figs. 43, 44, 45.

24 ARTHUR BOLLES LEE

They ultimately all fuse, apparently, into a single homogeneous
ring, as shown in figs. 49, 50, and others.

As soon as these globules have attaimed a certain size,
figs. 48, 44, 45, 49, and less clearly yet still indubitably in
figs. 36, 37, 38, the chromosomes, which in the clump appear
as straight staves, now appear as more or less sharply
curved staves, set on the surface of the globules or ring,
that is, outside them and not embedded in them, see
particularly the profile views figs. 48, 44, 45. Their outer
surface is irregularly convex; but their inner surface is
flattened on to the curvature of the globule or rmg. They are—
at the stage we are considermg—of a length equal to about
that of one of the limbs of the V-shaped chromosomes of the
anaphase (see figs. 3, 4, 17, 61). They do not form complete
hoops round the ring, but arcs that embrace about half a
meridian of it. They thus show two ends, a polar end and an
antipolar end. The polar ends, abutting on the lumen of the
ring, are generally closely huddled together and sharply curved
downwards, so that it is impossible to get clear images of them.
But their antipolar ends are generally widely spaced (figs. 48, 44,
45), and here their two component threads may frequently be
seen, with certainty, to be widely divaricated, figs. 48 (in the
middle), 44, 45, which is not the case with the polar ends.

As soon as the process of expansion has set in, the images of
the clump become less indistinct, and the chromosome staves
appear as shown in figs. 30, 38, 35, 36, 37; that is, they are seen
with certainty to contain or consist of the thin chromatic
threads running in pairs, which in our study of the clump in
Part I we recognized by them structure as shrunken chromo-
some axes, without discussing the fact of their collocation in
pairs. The members of these pairs run very close together and
in the main parallel to one another, as shown in figs. 30 to 35.
Images such as these may suggest, strongly, that during the
earlier stages of the clump the chromosomes have contracted
into short staves, each of which has undergone a longitudinal
division ; so that the threads would be the cleavage products
of such a division. Now there is no sign of any such division

Lo

STRUCTURE OF CERTAIN CHROMOSOMES 5

taking place at any time; but there is evidence that each of
these threads represents an entire limb of the anaphase V
from which it is derived ; and that their parallelism in pairs is
brought about by the folding together of the two
limbs of that V. This evidence is contained in the following
considerations.

In the daughter-star of the anaphase the chromosomes are
indubitably V-shaped, with equal limbs diverging to an angle
of some 45 degrees,’ figs. 3 and 5 (the apparent shortness of
some of the limbs in these figures, and the apparent hook shape,
is due partly to unequal degrees of contraction, partly to fore-
shortening). But as the star passes into the clump stage this
divergence becomes less pronounced, and in the completed
clump we find no such open V’s, but in their place a bundle of
short straight staves, figs, 29 to 33, each of which shows the two
thin chromatic threads mentioned above. The observer’s first
impression naturally is that each of these staves represents
one limb of a V, the relation of this one to the other being
masked by the crowding of the elements. But consideration
shows that this can hardly be. For the staves are only present
in a far smaller number than the limbs of the anaphase V’s—
in the completed clump in only half that of the limbs. Take
for instance fig. 29. This clump, a very early one, contains, as
IT make it, thirty-two seeming staves, of which twenty-nine
are shown in the drawing. Now the anaphases of Salaman-
dra atra, from which this is taken, have twenty-four V's,
therefore forty-eight limbs. Manifestly, therefore, not all the
staves of the clump can represent single limbs ; but some of
them must represent entire chromosomes. Let us suppose that
sixteen of them are in this case; then these will account for
thirty-two limbs ; and the remaining sixteen staves will represent
sixteen single limbs, thus making up the required tale of forty-
eight. Now take fig. 30, a completed clump. I make out
twenty staves shown fairly distinctly (not all drawn), and the
unanalysable portions of the clump may account for a very

1 This for the nuclei of the Amphibia. As we shall see, it is not the case
for those of all groups of animals.

26 ARTHUR BOLLES LEE

few more. So here we have about twenty-four staves, repre-
senting forty-eight limbs. Or take fig. 33, also a completed
clump. It shows twenty-one staves, and may contain a very
few more. Therefore here again about twenty-four staves for
forty-eight original limbs. Now take fig. 31, a nearly com-
pleted clump from Bombinator igneus. The diploid
number of chromosomes in this species is sixteen, showing
therefore thirty-two limbs at the anaphase. The clump con-
tains twenty staves. Therefore not all of these can represent
limbs of V’s; but twelve of them probably represent twelve
whole V’s, and the remaining eight represent single limbs of
such ; total, thirty-two.

It is therefore certain that in any polar clump some of the
staves—and highly probable that in the completed clump all of
the staves—must represent each of them two limbs of
a V. And the conclusion follows, that each of those of the
completed clump is in fact a V whose limbs have folded together.
So that the observed duplicity of the staves is not due to the
chromosomes having undergone a cleavage after having in some
other way assumed the shape of staves, but to their consisting
of the two limbs of an anaphase V—or what remains of these.
For the folding fully accounts for the duplicity.

In the Amphibia the postulated folding of the V’s takes place
as a rule only during the formation of the polar clump, not
before. But exceptionally it may take place during the early
anaphase. Fig. 41s a case in point. In this anaphase the limbs
of the V’s are im several instances closed in to a distance of only
about half a micron (as measured by the drumhead of the fine-
adjustment), and so accurately superposed on radii of the figure
that it is only by the most careful attention that the elements
can be seen to consist of two superposed moieties.

But this, which in the Amphibia seems to be the exception, is
in some other animal groups the invariable rule. For instance,
in the spermatogonia of the Acridian Oedipoda cothurna
(Arecyoptera variegata) I imvariably find the state
of things represented in fig. 61. This is a sagittal section of
a mid-anaphase, the chromosomes being not yet half-way to the

STRUCTURE OF CERTAIN CHROMOSOMES 27

pole. ‘They consist, all of them, of tightly-folded V’s, appearing
as short staves with the spindle-fibre insertion at the end. But
they are certainly folded V’s with the insertion at the apex :
the two limbs can be made out with certainty at the tips of
four of them; and a longitudinal duplicity can be at least
glimpsed in all of them.’ I find the same state of things exactly
in Oedipoda germanica, Oe. coerulescens, Oe.
(Mecostethus) parapleura,Gomphocerus rufus,
Stenobothrus morio, St. biguttulus, and some
other species of St eno bothrus which could not be determined
with certainty. So that in all the Acrididae I have examined
the folding takes place not later than the early anaphase. And
as at this stage the images are not obscured by the crowding
of the chromosomes which takes place in the polar clump, there
can be no doubt about the folding actually occurring.

So also in the Locustidae. Fig. 64 shows an anaphase of
a spermatogonium of Decticus verrucivorus. The
chromosomes are here smaller than in the Acrididae, and appear
for the most part as short rods with the spindle-insertion at the
end. But it can be made out in favourable instances that they
are in reality folded V’s ; and where this cannot be done, the
analogy with those of the Acrididae puts it out of doubt that
they are in the same case. Similar images are afforded by
Decticus griseus, Locusta viridissima, L. can-
tans, and Pterolepis aptera. Im Gryllotalpa
vulgaris and Gryllus campestris I find apparently
the same state of things, the anaphase chromosomes (with
the exception of the monosome in Gryllus) appearing
as short rods inserted by one end on the spindle. These
apparent rods are too small to be analysed with certainty ; but
judging by the analogy of those of the other Orthoptera
mentioned there can be no doubt that they are in reality

1 The drawings figs. 12 and 13 (Dissosteira carolina), and 18
(Steiroxys), of the paperof Davis, ‘ Spermatogenesis in Acrididae’’,
in ‘ Bull. Mus. Comp. Zool. Harvard’, with the interpretations given,
pp. 69, 70, 71 of the text, should, as I conceive, be corrected in the
sense indicated above.

28 ARTHUR BOLLES LEE

tightly-folded V’s.1. And this is also doubtless the case with the
very short thick chromosomes of the Hemipteron Pentatoma
(Carpocoris) nigricornis.

We find, then, that in the nuclei we have been studying the
chromosomes become doubled at the telophase, or before,
through a folding-in of their limbs. This brings
those limbs into a state of parasyndesis or close juxtaposition
throughout their length, so that little change (other than the
elongation due to their growth during the interphase) is required
in order to bring them into the state in which they are found
at the commencement of the spireme stage. This is illustrated m
figs. 55 to 59. But this process is perhaps not followed exactly
in all nuclei. I have evidence that the folding, or at all events
the definitive parasyndesis, of the limbs may be deferred, and

' In the Orthoptera the folding takes place not only as early as the early
anaphase, but sometimes as early as the equatorial phase. In the equa-
torial figures shown in figs. 60 (Oedipoda cothurna) and 63
(Decticus verrucivorus) all the chromosomes are tightly folded
into the stave shape. The same is the case in Oedipoda germanica,
Oe. coerulescens, and Oe. (Mecostethus) parapleura. In
Gomphocerus rufus the majority of the chromosomes appear in the
stave form; but there may be some open V’s. In Stenobothrus
biguttulus I suspect that the equatorials have always exactly two
large chromosomes of the open V shape, all the others being tightly folded
into the stave shape. It is perhaps not rash to conclude that all the cases
of chromosomes described by authors as straight rods with a terminal
spindle insertion are in reality cases of tightly-folded V’s with an apical
spindle insertion,

Fig. 63 (Decticus verrucivorus) shows sixteen large autosomes,
fourteen small ones, and a monosome, therefore thirty-one in all. This is
as it should be: for in this species I find in all unobjectionable images
either sixteen large autosomes and fourteen small, or fifteen large and
fifteen small, and a monosome ; the difference resulting from the fact that
it is sometimes difficult to decide whether a chromosome is an unusually
small ‘large’ one or an unusually large ‘small’ one. Buchner (‘ Arch.
Zellforsch.’, ili, p. 342, and fig. 82 of Taf. xix) correctly gives the number
as thirty-one in all. Vejdovsky (op. cit., pp. 33 and 44), notwithstand-
ing that he had this description before him, insists that there are only
twenty-three in all. Reference to his figs. 65 to 69 shows that he has
mistaken entire chromosomes tightly folded into the stave shape, and
fortuitously approximated at their apices, for mere limbs of open V’s.

STRUCTURE OF CERTAIN CHROMOSOMES 29

take place only at the moment of the formation of the spireme,
or even at an advanced period of its evolution. In this ease, the
limbs pass through the interphase in a more or less widely
divaricated state, which gives to the interphase a facies very
dissimilar to that of the interphase of nuclei in which the
parasyndesis has taken place at the telophase. A description
of this is reserved for a future paper. But in either case the
mechanism of the division of the chromosomes is the same in
principle. There is no longitudinal splitting.
The division is a transverse one, brought about by the
folding of the chromosomes at their middle, and their ultimate
segmentation at the bend there formed. The moieties which
Separate at the metaphase are the two limbs of the
chromosome thus folded, therefore metameric, not antimeric,
moieties

HXPLANATION OF PLATES 1 anp 2.

Illustrating Mr. Arthur Bolles Lee’s paper on ‘ The Structure
of certain Chromosomes, and the Mechanism of their
Division ’.

Magnification 1,500 diameters throughout.

PuateE 1.

Fig. 1.—Anaphase of pollen grainof Paris quadrifolia. Chromo-
somes alveolated, with sheath.

Fig. 2.—Early interphase of pollen grain of P, quadrifolia. Chromo-
somes without sheath, not alveolated, elongated into spirals.

Fig.3.—Triton alpestris. Anaphase of spermatogonium. The
chromosomes as open V’s, showing the chromatic axis and periaxial spirals
and sheath.

Fig. 4.—The same, a somewhat later stage, showing the chromosomes
_ folded into very narrow V’s.

Fig.5.—Bombinator igneus, spermatogonium. Portion of ana-
phase, showing the chromosome axes and periaxial spirals, but not the
sheath.

Fig.6—Salamandra maculosa. One limb of an anaphase
chromosome, spermatogonium. Chromatic axis, periaxial spirals (very
widely spaced), lateral processes, and sheath,

30 ARTHUR BOLLES LEE

Fig. 7.—Salamandra atra, do., do, Shows same details ; also the
terminal tag on the dome-shaped end of the axis.

Fig.8.—Salamandra maculosa, oogonium. Anaphase chromo-
some, entire. Same details.

Fig. 9.—Do.,epiderm. Anaphase chromosome, one limb. Same details,

Fig. 10.—Do., epidermal gland ; anaphase ; one limb of a chromosome,
Spiral with very wide pitch.

Fig. 11.—Do., kidney cell. Same details.

Fig. 12.—Do., cornea, Spirals much flattened on to axis.

Fig. 13.—Do., retina of larva, rod and cone layer. Details as last.

Fig. 14.—Triton alpestris, larva, pulmonary epithelium. Entire
anaphase chromosome. Note the spiral very closely coiled at tip of right-
hand limb, and not continued round the polar bend.

Fig. 15.—a, Triton palmatus, spermatogonium; b,Salamandra
maculosa, spermatogonium ; c, do., epiderm. Transverse sections of
anaphase spermatogonia. See text.

Fig. 16.—H 0 m 0, pus corpuscle from ulcerated skin. Two chromosomes
from an equatorial division figure. Sheath and lateral processes shown,
periaxial spirals invisible, though doubtless existent.

Fig. 17.—Gallus domesticus, embryonic cartilage. Portion of an
anaphase. Periaxial spirals just visible, sheath strong.

Fig. 18.—Anecylus lacustris, buceal epithelium. Tangential
section of anaphase, Spirals, lateral processes, and sheath just visible.

Fig. 19.—Salamandra maculosa, epiderm. Chromosome from
a not completely regularized equatorial figure. Spirals indistinct, giving
an impression of * granules ’,

Fig. 20.—Do., from a completed equatorial figure of a spermatogonium.
Details as last.

Fig. 21.—Do., portion of equatorial chromosome of an oogonium.
Details as last two figs.

Fig. 22.—Do., renal epithelium, One limb of an equatorial chromosome.
Spirals distinct on each of the two moieties.

Fig.23.—Oedipoda cothurna. Equatorial chromosome of
secondary spermatogonium. Details as for fig. 19, but sheath stronger.

Fig. 24.—Triton palmatus, spermatogonium. Spireme, early
stage. Chromosomes double, no sheath or other detail.

Fig. 25.—Salamandra maculosa, larva, epithelium. Spireme
somewhat more advanced than last. Moieties of chromosomes varicose
(dawn of periaxial spirals).

Fig. 26.—Do., pulmonary epithelium. Spireme, later stage. Moieties
very varicose, with abundant lateral processes and sheath.

Fig. 27.—Do., pleural endothelium. ‘Segmented’ spireme. Moieties
with large varicosities (Pfitzner’s * granules’), and lateral processes and
sheath,

STRUCTURE OF CERTAIN CHROMOSOMES 31

Fig. 28.—Triton palmatus, spermatogonium. Later spireme.
Periaxial spirals can just be glimpsed.

Fig. 29.—Salamandra atra, Spermatogonium. End of anaphase.
Chromosome V’s folded into the stave form,

Fig. 30.—Triton palmatus. Spermatogonium., Polar clump.
Chromosomes tightly folded, much contracted.

Fig. 31—Bombinator igneus, Spermatogonium, Polar clump.
As last.

Fig. 32.—Triton palmatus. Spermatogonium. Clump showing
chromosomes coalesced, Wholly or in part an artefact.

Fig. 33.—Do., do., do. Clump in polar view.

Fig. 34.—Triton alpestris. Do.,do.,do. Profile view.

Fig. 35.—Triton palmatus. Do. Clump expanding, early
stage.

Fig. 36.—Do., do., do. Later stage.

Fig. 37.—Do., do., do. Later stage of expansion, clump passing into
telophase. i

Fig. 38.—Do., do., do. Same stage, profile view.

Fig. 39.—Salamandra maculosa. Oogonium. Same stage, or
early telophase. Axes of limbs of chromosomes closely entwined round
one another.

Fig. 40.—Do., do., do. Somewhat later stage, chromosomes elongating,
Fig. 41.—Bombinator igneus, spermatogonium. Clump in stage
of figs. 38 and 39.

PLATE 2.

Fig. 42._Salamandra maculosa, spermatogonium, Telophase,
early, showing telophasic ring in profile (section).

Fig. 43.—Triton palmatus. Do.,do.,do. Note the chromosomes
flattened on to the outside of the hyaline globules, which are in course
of fusing into a ring.

Fig. 44.—Do., do., do. Tangential section of ring. As last. Two large
hyaline globules shown in the middle. Note the ends of the chromosome
axes showing divaricated at the antipolar ends.

Fig. 45.—Do., do. Profile view of a ring at a slightly later stage. Chromo-
some moieties looser ; chromosomes longer.

Fig. 464.—_Salamandra maculosa. Renal epithelium. Telo-
phasic ring, same stage as last, same details.

Fig.47.—Bombinator igneus. Spermatogonium. Section of
ring, same stage as last, and same details.

Fig. 48.—Do., do. Later stage of telophasic ring, polar view.

Fig. 49.—Triton palmatus. Polymorph spermatogonium, Mid-
telophase, ring beginning to close. Chromosomes elongated,

32 ARTHUR BOLLES LEE

_

Fig.50.—Salamandra maculosa, oogonium (primary). Telo-
phasic ring, about same stage as last, chromosomes more elongated and
taking on an erratic course.

Fig. 51.—Triton palmatus. Large endothelium nucleus from
follicle of testis. Late telophase, ring almost closed. Nucleus very flat ;
almost all the chromosomes drawn ; chromosome axes distinctly doubled
and entwined.

Fig. 52.—Do., do., a smaller nucleus, somewhat later stage.

Fig. 53.—Bombinator igneus. Endothelium nucleus, entire,
testicular peritoneum. Polar view (not a section) of telophase of same
stage as last. All the chromosomes have been drawn, though not through-
out all their length.

Fig. 54.—H o mo. Endothelium of vein of cutis. Section of telophase,
about the stage of fig. 51 or 53.

Fig. 55.—Triton palmatus, Spermatogonium, early interphase.

Fig. 56.—Do. Late interphase, or dawn of spireme.

Fig. 57.—Do., do. Early spireme. Karyoplasm browned by osmium.

Fig.58.—Bombinator igneus, Peritoneal endothelium. Early
rest stage.

Fig. 59.—Do., do. Later rest stage.

Fig. 60.—Oedipoda cothurna. Spermatogonium. One halfofan
equatorial figure. Chromosomes all of them as tightly-folded V’s.

Fig. 61.—Do., do. Sagittal section of anaphase. Chromosomes so
tightly folded that they appear as stout curved staves.

Fig. 62.—Do., do. Early telophase, tangential section of ring. Shows
three hyaline globules (to the right).

Fig.63.—Decticus verrucivorus. Spermatogonium. Equa-
torial figure. All the chromosomes drawn. All are tightly folded into the
stave shape ; m is the monosome.

Fig. 64.—Do., do, anaphase, polar view. Chromosomes folded into the
shape of wedges ; m, monosome.

Fig. 65.—Do., do. End of anaphase. Chromosomes as before.

Fig. 66.—Do. Primary spermatogonium. Mid-telophase: m, the
monosome. Some of the chromosomes seem to have their moieties divari-
cated at bot h ends, asif a transverse segmentation had taken place at the
polar ends.

Fig. 67.—Do. Nucleus of connective tissue enclosing cyst of testis. Early
telophase,

Quart. Pourn.NMicor Sci. Vot.65,NS. 4.1. .

os

—

60.
nat.del

4..B.Lee ad

Huth London

9
m
a3

67

66.

Quart. fourn. Mier Sci Vol. 68,NS. PA. 2.

65.

64.

34 D. KEILIN

larger specimens, even with the naked eye. It is richly vas-
eularized and its surface is irregular and lobulated. In longi-
tudinal median section the pharyngeal bulb is seen to be
composed of the three following portions: (1) an external
epithelial sheath, (2) a median mass of musculo-vascular tissue,
and (3) an internal portion composed of aggregates of deeply-
staining cells.

Almost all zoologists who have dealt with the anatomy
of earthworms have given more or less attention to this
organ, but, unfortunately, their opmions as to the nature
and function of the deeply-staiming cellular aggregates are
either unsupported by observations or contradictory. I do
not intend to give here a complete account of the previous work
on this subject, as this has already been done by Vejdovsky
(1884, pp. 101-6) and Stephenson (1917, pp. 253-60). I shall
therefore confine myself to a brief indication of the maim
views held on this subject by previous authors, classifying them
under the four following groups :

(1) Several authors, without paying special attention to the
structure of the pharyngeal bulb, accorded to it the function of
a salivary gland; in this category come the observations of Leo
(1820) and Clarke (1856) (cited by Vejdovsky), Lankester (1864,
p. 264), Vogt and Yung (1888, pp. 461-8), and Beddard (1895).

(2) Vejdovsky (1884, pp. 101-6), Willem and Minn (1899),
de Ribeaucourt (1900, pp. 246-7), and others, sueceeded in
tracing ducts which led from the deeply-staining cellular
ageregates, through the muscular portion, but, although they
could not detect any continuity of these ducts with the pharyn-
seal lumen, they nevertheless accorded to these cells a secretory
function similar to that of a salivary gland.

(3) Michaelsen (1886, cited by Hesse), Hesse (1894, pp. 10-12
and Pl. 1, fig. 24), and especially Eisen (1894-6), found the
ducts of the deeply-staining gland cells to pass through the
muscular portion, penetrating between the cells of the pharyn-
geal epithelium and opening into the pharyngeal lumen.

(4) Finally, Stephenson (1917) completely denied the existence
of any communication between the deeply-staming cells, which

PHARYNGEAL GLAND OF THE EARTHWORM 35

he calls ‘ chromophile cells’, and the pharyngeal lumen. The
function of these cells, according to this author, remains
unknown.

Of all the above-mentioned views, those of Eisen and
Stephenson are specially interesting, as bemg diametrically
opposed, though both based upon the study of the detailed
structure of this organ. They deserve, Ne to be examined
in greater detail.

Hisen (1894-6), in his series of papers on the Oligochaetes of
the Pacific Coast of America, describes and figures the pharyn-
geal or salivary glands of almost all the earthworms he studied,
and especially those of the following five species: Phaenico-
dnrilus tasie (1894. pp. 66-7, Pl. xxx, figs. 1, 2, and
Pl. xxxii, fig. 18), Pontodrilus Michaelseni (1894,
pp. 77-8, Pl. xxxiv, fig. 36), Benhamia nana (1896, p. 129,
Pl. xlvu, figs. 15-18), Sparganophilus Benhami
(1896, pp. 104-5, PI. lin, figs. 112-13),and Sparganophilus
Smithi (1896, p. 157).

To demonstrate the views of this author, we shall quote from
his paper the following descriptions which concern respectively
the salivary glands of the first two species mentioned above.

Phaenicodrilus taste (pp. 66-7): ‘The narrow
ducts from the gland penetrate the pharyngeal epithelium and
form, at its outer edge, small ovoid pockets for temporarily
storing a small amount of the salivary secretion. These ducts
end with the pharynx, the oesophagal epithelium neither being
furnished with ducts nor storage pockets. . . .’

Pontodrilus Michaelseni: ‘The ducts lead directly
to the pharyngeal epithelium ; arrived here they branch out,
sending numerous discharge-tubes between the epithelial cells
(fig 36, gl. dt.), discharging the salivary mucus in the pharyn-
geal cavity. These ductules are frequently, though not generally,
branched while in the epithelial layer. Hach ductule 1s
farnished at the distal end with a small storage-chamber
(86, A Pl. 34) of oblong form and considerably smaller than the
nucleus of the epithelial cells.’

According to these observations, the pharyngeal cells, which

D2

36 D. KEILIN

exist probably in all earthworms, form a salivary gland which
pours its secretion into the pharynx. This has been denied,
however, by Stephenson, in a paper specially devoted to this
subject.

After a careful critical examination of the work of all the
previous authors, Stephenson writes (loc. cit., p. 260): “The
authors who have seen ductules and their ending in the pharyn-
geal epithelium have, I believe, been misled by preconceived
ideas due to the transformation of the deeper cells into connec-
tive tissue.’ Earlier (p. 259) he says: ‘ It will save repetition to
state that in none of my sections, which were taken in all the
three planes, have I seen structures that could be interpreted as
ductules.’

He passes then to the description of these cells and their
gradual transformation into the ‘ fibrillar or reticular packing
tissue (** Fiillesewebe’’) between the muscles’ in several species
of earthworms belonging to the genera Pheretima and
Helodrilus (Allolobophora). His study is concluded
by the following statements: ‘ The “ pharyngeal gland-cells”’
of earthworms are not gland-cells in the usual sense, and do not
communicate with the pharynx ; the term “chromophile cells ”’
is proposed for them because of their intense coloration by
haematoxylin and similar stains. The so-called “septal glands”’
of earthworms are aggregations of similar cells at a more poste-
rior level.” . . . ‘ While most of the cells form a more or less
compact aggregate on the surface of the pharyngeal mass, a
number penetrate inwards towards the pharyngeal epithelium,
and become progressively metamorphosed into fibrillar con-
nective tissue.’

As to the function of the chromophile cells, he writes (p. 281) :
Though in the light of what has gone before we may reject
the usual supposition that the cells pour a secretion into the
pharynx (or oesophagus, in the case of the smaller more
posteriorly-situated aggregates), it is not easy to propose
another hypothesis to take its place.’. ..‘ That the main
function of the cells is metabolic is, though only a vague state-
ment, perhaps as far as we are justified in going.’

‘

Es

PHARYNGEAL GLAND OF THE EARTHWORM 37

During my research on Pollenia rudis, a Calliphorine
fly, the larvae of which live as parasites n Allolobophora
chlorotica, I often had oceasion to study sections of the
pharyngeal bulb of several species of earthworms, and I always
believed that I was dealing with a salivary gland as described by
Hisen. The recent paper of Stephenson came therefore as a
surprise to me. It mduced me to re-examine more closely my
previous sections, and to prepare fresh ones, using this time
special methods, which, as we shall see further on, enable us to
solve finally the questions as to the nature, and, consequently,
the functions of the deeply-staining cell-aggregates.

This seems to me to be very important, for two reasons :
(1) the pharyngeal bulb is an organ of conspicuous size and
appears to exist in all earthworms, and (2) the common earth-
worm being generally used as a type for the purpose of class
dissection, it is very necessary that all observations concern-
ing its anatomy should be accurate, in order to avoid a wide
dissemination of erroneous information.

2. MarertaL AND MeruHops.

The earthworms used for this study comprise three species :
Allolobophora chlorotica Sav., Allolobophora
foetida Hisen, and Lumbricus sp. For the study
of the general structure of the pharyngeal bulb I used as
fixatives: Boum and Schaudinn with 3 per cent. of acetic
acid, followed by staiming in P. Mayer’s Haemalum or Glychae-
malum with Eosin or Orange, or in Magenta-red and Picro-
Indigo-carmine. For the more delicate structures of the
sland and pharyngeal epithelium small pieces were fixed in
Champy’s chromo-osmic solution and staimed with Iron
Haematoxylin and Eosin. The protoplasmic inclusions were
examined in sections prepared by Champy’s (1911) method
(fixation in Champy’s solution, post-chromization with potas-
sium bichromate, and staining in Tron Haematoxylin).

For the study of the glandular secretion, which I naturally
supposed to be mucin, I had to apply several methods. Since
Langley’s important research on salivary glands and their

38 D. KEILIN

secretion (1889) a fairly large literature on mucin glands has
accumulated, and several good methods now exist which
enable us to detect the smallest amount of mucin in very fine
ductules. For a eritical account of these methods, the reader
is referred to the papers of Hoyer (1890 and 1903) and Michaelis
(1903).

The methods of staining which I have used in connexion with
this study are of two kinds :

(a) A purely mucin stain: Mucihaematein of P. Mayer
(1896).
(b) Metachromatic stains: Thionin and Toluidin blue.

(a) Muecin stain: Anterior portions of earthworms are
fixed for twenty-four hours in Bouin’s Picro-formol or in
a modified solution of Bouin’s Picro-sublimate formol (Corro-
sive sublimate, saturated sol. 20 ¢.c., Picric acid, saturated sol.
20 ¢.c., Formol, 20 ¢.c., Acetic acid, glac. 5 ¢.c.). After fixation
they are well washed in Alcohol (70 per cent.) and embedded by
the ordinary method. The sections (4-6 in thickness), having
been freed from paraftin, are stained from two to five minutes in
a 10 per cent. solution of Mucihaematem. They are then either
mounted without any supplementary staining, or staimed with
the Magenta-red and Picro-Indigo-carmine. I have obtained
good results by staining the sections with Iron Haematoxylin
(twelve hours in Iron alum and twelve hours m 1 per cent.
solution of Haematoxylin) and counterstaining for five minutes
in Mucihaematein, and for a few seconds in Orange G.

(b) Metachromatic stain. Slightly modified methods
of Hoyer (1890, 1903) and Hari (1901) give very good re-
sults. Portions of earthworms are fixed either in 5 per cent.
solution of corrosive sublimate, or, with much better results,
in the above-mentioned Picro-sublimate formol, from two to
eight hours. ‘The sections, freed from paraffin, are passed
through the series of alcohols into the distilled water and then
for ten minutes into 5 per cent. solution of corrosive sublimate.
They are then washed rapidly in strong alcohol and distilled
water and stained in an aqueous solution 0-1 per cent. of Thionin
(Lauth’s violet), or Toluidin-blue. In about one to two minutes

PHARYNGEAL GLAND OF THE EARTHWORM 39

all the mucin appears red ; in two to seven minutes the mucin is
stained red, while all the rest of the tissue is stained blue. It is
better to examine the sections while they are still in the solution
of Thionin, as it is very difficult to mount them without des-
troying the metachromasy. There are, however, several ways
of mounting the slides in Canada balsam, by which the meta-
chromatic effect may be retained for at least seven days. I shall
mention only the following few methods which have given me
very satisfactory results.

(1) Very rapid passage through absolute alcohol, xylol, and
mounting in Canada balsam.

(2) Sections stained in Thionin, washed rapidly in distilled
water, fixed in a 10 per cent. aqueous solution of Potassium
ferrocyanide (Krause’s method), rewashed in distilled water,
and then passed rapidly through the graded alcohols, absolute
alcohol, and xylol, into Canada balsam.

(3) The sections are stained by the previously described
Thionin method, before freeing them from paraffin, washed
rapidly in distilled water, dried thoroughly with filter paper,
and then freed from paraffin and mounted in Canada balsam.

(4) Instead of alcohol, Acetone is used for dehydration, and
xylol for clearing ; and the sections are then mounted in Canada
balsam (method recommended to me by Dr. W. H. Harvey).

Mounting the sections in levulose syrup, or syrup of Apathy,
is not advisable, for even when it preserves the metachromasy,
sections thus prepared do not show clearly the cytological
structure, particularly under examination with high magnifica-
tions. I did not succeed in differentiating the sections with
Hari’s mixture (1901). Finally, the use of artificial light for
examination of the sections is strongly recommended, as it
shows a more striking contrast between the red and the
blue colours of the stained sections.

3. THE STRUCTURE OF THE PHARYNGEAL OR SALIVARY BuLB.

The pharyngeal bulb has been already morphologically de-
seribed by several authors who have dealt with the anatomy ot
earthworms. In almost all species of earthworms, it has the

10 D. KBEILIN

sume general form and the same relations with the surrounding
organs, Varying only in the size and the nwnber of the glandular
lobules. The general structure of this organ is sufficiently

Text-Fic. 1.

ms

ia

cn

=;
ph. #
es
at
£5
Fa)
v.70: ea , “S yb.

Gu

«

Longitudinal median section of All. foetida. c.g.=cerebral
ganglion; ¢. 7m. v.=conductive or musculo-vascular portion of
pharyngeal bulb; /. c. = mass of coelomic cells containing droplets
of fat (ef. Text-fig. 7, p. 57 of this paper); oe. = oesophagus ;
p. ¢. = ciliated pharyngeal epithelium ; ph, = pharnygeal lumen ;
s. gl. = deep or glandular portion of the pharyngeal bulb, composed
of basophile, salivary cells ; v. 2.=ventral nerve cord. x 26.

clearly shown by ‘Text-figures 1 and 2, which represent longi-
tudinal median and submedian sections of the anterior portion
of the earthworm.’

| For the morphological variation of this organ the reader is referred to
the published papers on the anatomy of earthworms.

PHARYNGEAL GLAND OF THE EARTHWORM 4]

As to the histological structure of the pharyngeal bulb, we
shall, for the sake of clearness, examine separately the structure
of its three portions: (a) the deep glandular portion, (b) the
conductive or musculo-vascular portion, and (c) the superficial
or epithelial portion.

(a) The deep or glandular portion.

The deep or glandular portion of the pharyngeal bulb is
composed of a certain number of lobules of various sizes,
suspended in the coelomic cavity of the earthworm and extend-
ing backwards as far as the fifth or the sixth segment of the body
(Text-figs. 1 and 2, s. gl.). These lobules, as well as the entire
bulb, are surrounded by a thin peritoneal membrane (* capsule ’
of Stephenson) composed of flattened cells with elongated
nuclei. The peritoneal membrane penetrates between the
lobules, and in some places into the lobules, especially where the
latter are traversed by muscular bundles, or by the blood-vessels,
which are directed forwards and ramify in, and form the main
part of, the musculo-vascular portion of the bulb (Text-figs.
1 and 2, c. m. v.).

The cells which compose the glandular lobules are very poly-
morphie, being either spherical or elongated, or even semilunar.
Sections derived from well-fixed material (im Champy’s fixative,
for instance) do not show clearly the boundaries between the
cells, while on the other hand, a less perfect fixation, which
slightly contracts the cells, defines their contours, and demon-
strates that, in some places, the protoplasm of these cells is con-
tinuous. The size of these cells varies as much as their form ;
in Allolobophora chlorotica, for instance, they are
from 204 to 804% long and 18 wide. Each cell contains a
large spherical nucleus of 7-8 in diameter which is provided
with a large nucleolus of 8-4 in diameter (PI. 3, fig. 4, m. gl.).
The peripheral chromatin of the nucleus is generally much
reduced, but its quantity seems to depend upon the activity of
the cells. The protoplasm, as was shown by Stephenson, 1s
very basophile, for which reason he called these cells ‘ chromo-
phile’. When stained by Haemalum, ron Haematoxylin, or

12 D. KEILIN
Magenta-red, the perinuclear protoplasm of these cells is often
so deeply stained that it decolorizes more slowly even than the
nucleus. Nearer the border of the cell the basophile proto-
plasm is very irregularly distributed, and this gives to the

TEXT-FIG. 2.

ieee

yy 7)
Mira Beeve

SISAL

WWVy
Thal

Be

Se EN
Pe 4s S
FENN
BSE eR |
Re ;

S
iS

53308
TE
FNe

5
Sis

aa

Ha

Sasa avETG

pai

Longitudinal submedian section of All. foetida: ph. d.= dorsal
or salivary chamber of pharynx; ph.v.=ventral chamber of
pharynx. Other letters as in Text-fig. 1. x 26,
stained cells a very peculiar spotted appearance (Pl. 3,

figs. 2 and 4).
The clear areas of the protoplasm have a very granular

structure, the nature of which we shall examine later. The

PHARYNGEAL GLAND OF THE EARTHWORM 43

basophile protoplasm does not show any special structure, and
it appears to contain a diffused chromatic substance (extra-
nuclear chromatin). In sections of the glandular cells of Lum.
bricus sp. prepared by Champy’s method (fixation in
Champy, postchromization followed by Iron Haematoxylin)
the protoplasm is seen to contain a number of bodies which
are probably mitochondria (Text-fig. 3). These protoplasmic
bodies appear as irregular, curved and ramified filaments or
TEXT-FIG. 3.

el Tar

=

\ Mig
\ , ao
~~ » CO 6
; te _
4 i
ii ges ae va
- x, j

Glandular or salivary cell of Lum bricus sp. showing a vesicular
nucleus with large nucleolus and with numerous intraproto-
plasmic mitochondrial bodies. x 2,200,

patches composed of small darkly-staimmg granules, and are
distributed throughout the protoplasm, not bemg confined to
its basophile portions. Their number and size varies in different
cells, some of which are crowded with them, while in others they
are more or less scattered.

As to the nature of the granular substance filling the clear
parts of the protoplasm of these cells, from the sections pre-
pared by an ordinary method (fixation in Bouin and staiming in
Haemalum), I had already ample evidence that it is ordinary
mucin. On the other hand, as the supposition of a secretion of
mucin by these cells was absolutely denied by Stephenson, I had
to study these glands in sections prepared by special methods

44 D. KEILIN

(Mucihaematein or Thionin), which enable one to detect the
most minute quantities of mucin. Moreover, to obtain a
definite result by these methods, it was important to apply
them simultaneously to the pharyngeal gland and to some other
elandular cells which are known to contaim mucin. The best
control tissue of this kind is undoubtedly the external integument
of the same earthworm. In sections, not only of an extracted
pharyngeal gland, but of the whole anterior portion of the
earthworm, it is always possible to make a comparison of
the staining reactions of the pharyngeal gland with those of the
mucin cells of the skin. We willnow examine the longitudinal
median sections of the anterior segments of Allolobo-
phora chlorotica stamed by the Mucihaematem method
(see p. 88 of this paper). These sections, after thirty seconds to
two minutes staining in 10 per cent. solution of Mucihaematein,
show already a very clear picture of the distribution of mucin in
the different tissues. These sections, when counterstained with
Magenta-red and Picro-Indigo-carmine, become still more
instructive ; the skin then shows clearly (Pl. 3, fig. 1), (1) the
epidermal cells with greenish-yellow protoplasm and red nuclei,
and (2) the mucin cells (mu. ¢.), m all stages of secretion of
mucin, stained deep violet ; the small nuclei of these cells are
displaced laterally or basally by the mucin (mu.), which in
some cells is seen to issue from a small pore in the cuticle (cu.).

The same sections show also the salivary secretion of the
pharyngeal gland cells (Pl. 3, figs. 2 and 4, m. gl.).

The basophile protoplasm of these cells is staimed red, while
the clear protoplasmic areas are now seen to be composed of
sranular mass (mu.) stamed, like the mucin of the cutaneous
sland, deep violet. This shows that the granular substance of
the pharyngeal gland cells, which has been already mentioned
by Stephenson, is composed of ordinary mucin. The results
obtained by the Mucihaematem method were corroborated
by the Thionin method. Seetions of the anterior portion of
Allolobophora foetida prepared by this method have
also shown the pharyngeal gland cells filled (PL. 3, fig. 9, m. gl.)
with granules of mucin (mu.) similar to those of the mucin cells

PHARYNGEAL GLAND OF THE EARTHWORM 45

of the skin (PI. 3, fig. 10, mu. ¢.). In these sections the mucin
is stained red, while the rest of the tissue stains in all shades of

blue.

(b) Conductive or musculo-vascular portion.

As one follows them continuously from the deep glandular
portion to the muscular or central region of the pharyngeal
bulb, the glandular cells gradually change their structure (Pl. 3,
fig. 5, m. gl.). They become smaller, their basophile protoplasm
becomes more and more reduced, while the clear protoplasm,
filled with granules of mucin, rapidly increases in quantity.
These granular mucinous portions of the cells fuse together
and form wide strands of mucin, the granules of which are
regularly distributed in a multitude of sinuous rows (mu.).
Nearer to the pharynx several small cells with basophile
protoplasm may still be found embedded in this mucin, but
usually one finds on the surface of these mucin ducts a
few small nuclei (Pl. 3, fig. 6, d. mu.) filled with chromatic
granules. These large mucin ducts subdivide and _ pass
gradually into smaller ducts which are interlaced with the
muscle fibres (m.) and blood-vessels (v.) This gradual passage of
the glandular salivary cells into the salivary or mucin ducts was
misinterpreted by Stephenson for a gradual transformation of
his ‘ chromophile ’ cells into fibrillar or reticular packing tissue
(‘ Fullegewebe ’). It is also evident that the connective tissue
deseribed by Stephenson is no other than the above-described
salivary ducts containing precipitated and stained mucin.
The musculo-vascular portion of the pharyngeal gland thus
contains : (1) very strongly developed muscle fibres, (2) blood-
vessels, and (3) salivary ducts filled with mucin.

To these we can now add: (4) nerve fibres, (5) nephrocytes
or excretory cells similar to the yellow cells of the alimentary
canal, and, finally, (6) cells with bacteroids or. crystals of
uric acid (PI. 3, figs. 2and 9, wr.). Concerning the nature of the
last two elements I have more to say in the supplementary notes
to this paper (p. 54).

46 D. KBEILIN

(c) Superficial or epithelial portion.

It is a matter of surprise that, in spite of the fact that he
absolutely condemns Eisen’s observations as to the existence
of ductules in the pharyngeal epithelium, Stephenson made
no special study of this particular portion of the pharynx,
although such study is all-essential for making a correct inter-
pretation of the function of the pharyngeal gland cells.

The lumen of the pharynx (Text-figs. 1, 2, and 6, A) in all
earthworms is divided by means of two longitudinal folds of the
lateral walls mto dorsal and ventral chambers. An elongated
median slit, bordered by the free margin of these folds, estab-
lishes a communication between these portions of the pharyngeal
lumen. The lateral folds meet posteriorly in the median line to
form a posterior dorsal pharyngeal pocket which communicates
with the two lateral pockets and forms the dorsal or salivary
chamber of the pharynx (Text-fig. 1, ph. d., and Text-fig. 6, A,
ph. d.), while the ventral chamber (ph. v.) is continued into the
oesophagus (0e).

Of all the pharyngeal epithelium, the dorsal portion only, to
which the pharyngeal bulb is attached, is composed of ciliated
cells. The cells of the remaining portion of the pharyngeal
epithelium are covered by a thin cuticular layer similar to that
which lines the oesophagus.

The dorsal portion of the pharyngeal epithelium of Allolo-
bophora chlorotica (Pl. 3, fig. 3) is composed of
elongated cells, the oval nuclei of which are provided each with
one or two nucleoli besides the chromatic granulation. These
cells are usually so crowded that, in sections, their nuclei
appear to he at different levels. The free border of the cells
bears the vibratile cilia (el.).

The basal ends of the cells are very narrow and covered with
a basal membrane. Near the free border of the epithelium one
often sees the darkly-stained nuclei in all stages of the karyo-
kinesis. As one follows their approach to the internal surface
of the pharyngeal epithelium, the mucin ducts (PI. 3, fig. 3,
d. mu.), which, as we have previously seen, are interlaced with

PHARYNGEAL GLAND OF THE EARTHWORM AT

the muscle fibres (m.) and blood-vessels, are seen to become
parallel to each other and perpendicular to the epithelium.
Reaching the basal membrane of the latter, these salivary
ducts give off numerous small ductules (dl. mu.) which penetrate
between the epithelial cells and terminate separately in a
multitude of small pockets (d. p.) of mucin lying immediately

TEXT-FIG. 4.

\ )
| i
Section of the ciliated pharyngeal epithelium of All. foetida
(stained with Mucihaematein only, showing the intra-cpithelial

mucin ductules = d/. mu., ending in the discharge pockets = d. p. ;
Co— Cilia) < 750:

\ XN

()

beneath the free surface at the base of the cilia. These fine
ductules, with the terminal discharge pockets, are very clearly
seen in sections stained by Mucihaematein alone (‘Text-fig. 4),
or combined with Magenta-red, Picro-Indigo-carmine, or by the
Thionin method. In the first two cases they are all stained
violet while the surrounding protoplasm is either unstained or
greenish yellow in colour (PI. 3, fig. 3), m the second case
(ex. All. foetida) these ductules are red, while the rest of the

io 6)

4) D. KEILIN

tissue is blue (PI. 8, figs. 7 and 8). Some of the sections of
All. foetida stained by the latter method showed the
actual discharge of the mucin from the terminal or dis-
charge pockets (d. p.) into the pharyngeal lumen (PI. 3,
fig. 8 d. p. and mu.). The latter in all sections is shown to
be filled with mucin (mu.), which flows partly towards the buceal
cavity and partly towards the oesophagus. It is very
important to examine now a number of observations of
certain histologists, who, treating of the minute structure
of this organ from quite a different standpoint, and
using a totally different technique, discovered nevertheless the
ductules with their discharge pockets in the pharyngeal
epithelium, but unfortunately completely misunderstood their
nature and their function. I am alluding here to the papers
dealing with the study of the peripheral nerve endings and
sensory cells of earthworms.

In 1892 Retzius discovered in the pharyngeal epithelium
special fibrils which he named clubbed fibrils— Kolbenférmige
fasern’—and which he supposed to be the gustatory sensory cells.

In 1894 Smirnow, to whom we owe the discovery of free nerve
endings in the skin and the pharyngeal epithelium of the earth-
worm, using Golgi’s method, detected in the pharyngeal
epithelium the clubbed cells of Retzius.*

Smirnow’s description of these cells closely resembles that of
Retzius ; he found in the pharyngeal epithelium an enormous
number of these cells, which in their terminal dilated portion
seem to contain nuclei. Their elongated portion he described
as somewhat tubular with the lumen filled with a granular
substance, and the whole structure of the club-shaped cells
leaves, according to Smirnow, some doubt as to their nervous
origin.

A year later (1895) Retzius confirmed Smirnow’s discovery
of the free nerve endings of the skin and the pharyngeal
epithelium of the earthworm ; and, returring to the subject
of his clubbed fibrils, he now denied the existence of nuclei in the

' It may be mentioned that, under the name of oesophagus, Smirnow was
actually dealing with the salivary portion of the pharynx,

PHARYNGEAL GLAND OF THE EARTHWORM 49

dilated terminal portion of these fibrils ; he also disagreed with
Smirnow as to their tubular structure and he described them
once more in some detail. These fibrils in traversing the
pharyngeal epithelium do not ramify and are completely
devoid of the varicose nodules so characteristic of the nerve
fibrils which are met with in the same pharyngeal epithelium.
He failed again to detect the origin of these fibrils and still con-
sidered them to be nervous elements, but he added that further
study, and especially the discovery of their central origin,
would finally solve the problem as to their nature and their
function.

The same year Langdon (1895), relying upon Smirnow’s
description, denied the nervous nature of the clubbed fibrils and
considered them to be glandular or mucous cells.

More recently, Dechant (1906) demonstrated the same fibrils
by a metallic impregnation method, and, in accordance with
Retzius, described them as nervous elements.

I myself have recognized the structures described as clubbed
fibrils by Retzius in the pharyngeal epithelium of Lumbricus
sp. fixed with Champy and stained with Iron Haematoxylin.
The fibrils, in enormous numbers, run between the pharyngeal
cells and are either straight or smuous ; they all terminate in a
very dilated portion filled with granular substance (Text-fig. 5,
A and B).

The merest glance at the structures convinced me that I was
dealing with the same mucin ductules and thei discharge
pockets. The only difference between these structures and
those previously described consists mainly in the fact that,
while previously we stained only the mucin which fills the
ductules and the pockets, now we stained the ductules and the
pockets themselves. Moreover, the figures of the clubbed fibrils
as shown in the papers of Retzius, Smirnow, and Dechant are
similar in all respects to my figures of the intra-epithelial mucin
ductules and their discharge pockets (Pl. 3, figs. 3, 7, and 8,
and Text-figs. 4and 5). On the other hand, the fact that these
authors sueceeded in detecting these salivary ductules by
metallic impregnation methods is not surprising, as these

NO. 257 B

50 D. KEILIN

methods were already advocated by Miller (1895), Zimmermann
(1898), and Retzius himself, for the detection of minute, or even
intracellular, capillary ductules of secretion.

(d) Septal glands.

The salivary gland cells in all earthworms are intimately
connected with some other cell aggregates which, being cyto-

TEXtT-FIG. 5.

A and B. Sections of the ciliated pharyngeal epithelium of Lum-
bricus sp. (fixed in Champy’s solution and stained with Iron-
haematoxylin) demonstrating that the clubbed nerve fibrillae
of Retzius are the intra-epithelial mucin ductules (dl. mu.) with
their discharge pockets (d. p.); c.= cilia; mu. = contracted mucin
in some of the discharge pockets. A x 734; B x 734.

logically similar to the salivary cells, differ from the latter m the
fact that they are completely devoid of mucin (PI. 3, fig. 2, e. gl.).

Similar glandular aggregates, devoid of mucin, are found
posteriorly in the coelomic cavity, surrounding the oesophagus.
In places [ believe that I have been able to trace a communica-
tion between these deeply-lying glandular elements (septal
glands) and the pharyngeal bulb. In other places, although I

PHARYNGEAL GLAND OF THE EARTHWORM 51

could not trace any communication between these cell aggregates
and the pharyngeal or oesophageal walls, on account of the
difficulty of following the course of these fine ductules in
sections, I nevertheless believe that such communication
exists. The function of these cells, as we shall see later, consists
probably in elaborating a digestive enzyme which is discharged
into the lumen of the pharynx or oesophagus.

4. FUNCTION OF THE PHARYNGEAL GLAND CELLS.

All the foregoing has proved, beyond doubt, that the pharyn-
geal bulb of the earthworm is a true salivary gland, which
pours its secretion (mucin) into the lumen of the dorsal or
salivary chamber of the pharynx. The mucinous salivary
secretion accumulates in the pharyngeal cavity and oesophagus,
and there it performs an important service during the operation
of feeding. In view of the unusual diet of earthworms in general,
it would be a matter of surprise to find that no special pro-
vision was made by which the relatively enormous quantities of
earthy matter, composed, in great part, of hard and insoluble
particles, could be conveniently passed through the alimentary
tract.’

In addition to the function of the formation of the food bolus,
the salivary secretion has also a digestive function. In con-
nexion with this digestive function of the pharyngeal bulb, it
is interesting to examine briefly the available information con-
cerning the digestive ferments of earthworms.

Frédéricq (1878) was the first to discover in the alimentary
canal of the earthworm the existence of two ferments : the one
amylolytic, and the other proteolytic, the latter being active in
either a slightly alkaline or a slightly acid medium.

Darwin (1881, pp. 35-48), in his classical observations on the
habits of earthworms, stated that they emit from the mouth an
alkaline secretion, containing a ferment similar to the pancreatic

1 In several earthworms, according to Vejdovsky and Eisen, the salivary
portion of the pharyngeal wall is very easily protruded or evaginated from
the buccal cavity and serves a more or less prehensile function,

BE 2

52, D. KEILIN

enzyme, which digests the leaves which are dragged into the
burrows before they are taken into the alimentary canal. This
mode of extra-stomachal digestion he compares to that of
insectivorous plants, as Drosera or Dionaea.

The amylolytic and proteolytic ferments in earthworms were
also deseribed by Willem and Minne (1899), and more recently by
Lesser and Taschenberg (1908). The last two authors found,
in addition, the following enzymes: (1) an enzyme capable of
hydrolysing glycogen, (2) Invertase, (3) Lipase, (4) Katalase,
and (5) one which very probably was an Aldehydase.

Of the work cited above, that of Willem and Minne is of
especial interest, inasmuch as they prepared extracts separately
from the individual parts of the alimentary tract, while the
other authors used extracts of the entire alimentary canal.
Thus the extract which they obtained from the isolated
pharynges of several earthworms digested fibrin in alkaline media
and produced peptone. According to these authors this pro-
teolytie ferment is derived only from the pharyngeal gland cells,
aithough they failed to establish the existence of an actual com-
munication between their ductules and the pharyngeal lumen.!

The pharyngeal bulb, with its accessory
glandular aggregates, has, then, a double funce-
tion: (1)secretionof mucin, and (2) secretionofa
proteolytic enzyme. Wehave seen, on the other hand,
that the glandular aggregates comprise two kinds of cells, the
one containing the mucin, and the other devoid of it ; it is then
very probable that the cellular aggregates devoid of mucin are
those which elaborate the proteolytic ferment. This is cor-

1 The following is a quotation from the papers of Willem and Minne
(pp. 2 and 8) relating to this question : * I] est trés pénible de suivre sur les
coupes le trajet des conduits glandulaires ; on en retrouve des troncgons
au sein de la masse des fibres musculaires, et ’épithélium cylindrique du
cul-de-sac pharyngien dorsal présente entre ses cellules des lumiéres quinous
paraissent correspondre aux extrémités de ces canaux. Les éléments dont
nous parlons sont les seuls de la masse pharyngienne dont Ja structure soit
compatible avec une fonction glandulaire, on doit leur attribuer la séerétion
du ferment peptonisant dont nous avons constaté lexistence dans les
parois de Porgane.’

PHARYNGEAL GLAND OF THE EARTHWORM 53

roborated by the fact that the extracts from the oesophageal
portion, which, as we have seen, is surrounded only by the
non-mucinous glandular cells, contaims, according to Willem
and Minne, a proteolytic ferment, although in smaller quantity
than that of the pharyngeal bulb.

Having established the glandular nature of the pharyngeal
bulb, and having shown its function, it seems to me quite
superfluous to seek further proof in a study of the development
of the pharyngeal glandular cells. As to the origin of these
cells, Stephenson’s statement that they are derived from the
peritoneal layer appears to me to be doubtful. His deserip-
tion, and especially his figures, do not give the shghtest support
to this opinion, and I consider that the question of the develop-
ment of the pharyngeal gland cells remains still open for
further investigations.

5. SUMMARY AND CONCLUSIONS.

1. The pharyngeal dorsal bulb of the earthworm is a true
salivary gland.

2. The function of the basophile cell-ageregates of this bulb
is the production of mucin and a proteolytic enzyme.

3. These products of secretion are collected in a system of
salivary ducts lymg in the conductive musculo-vascular por-
tion of the pharyngeal bulb. The salivary ducts, on reaching
the pharyngeal ciliated epithelium, divide into mnumerable
fine ductules which penetrate between the epithelial cells and
terminate near the free surface in the discharge pockets. The
salivary secretion accumulates in these pockets before it 1s
discharged into the dorsal or salivary chamber of the pharynx.

4. The club-shaped fibrillae of the pharyngeal epithelium
discovered by Retzius are not of a nervous nature, as he
supposed ; they are the ordinary salivary ductules with their
discharge pockets.

5. The question as to the development of the pharyngeal
bulb of the earthworms remains open for further investigations.

6. In addition to the glandular cells with their ducts, muscles,
nerve fibres, and blood-vessels, the pharyngeal bulb contains

54 D. KEILIN

bacteroid or uric acid cells and amoebocytes, similar to the
yellow cells of the alimentary canal.

6. SUPPLEMENTARY NOTES.

According to Cuénot (1897) and Willem and Minne (1899)
there are five different excretory organs in earthworms :
(1) nephrida, (2) chloragogenous cells, which contain guanine,
(3) cells with bacteroids or with crystals of uric acid, (4) yellow
cells of the walls of alimentary canal, (5) amoebocytes of the
blood. As the two latter elements are found in the pharyngeal
bulb, we will examine them in greater detail.

(a) Cells with bacteroids or crystals of wrie acid.

These cells are very common in earthworms, bemg found in
enormous numbers on the peritoneum, the septa, between the
muscle fibres, on the nerve ganglia, in the nephridia, &&. In
the case of Allolobophora foetida, I found them in
large numbers between the muscles of the pharyngeal bulb
(cf. p. 45 of this paper). These cells, of various shapes
and sizes, are filled with elongated crystalline bodies. In
sections, or in stained smears, these bodies so closely resemble
bacteria, that several authors have considered them to be such.
Thus, according to Cuénot, Cerfontaine (1890) described them
as bacilli; he also thinks that the tubercle bacilli, found in
such numbers by Lortet and Despeignes (1892) in the bodies of
earthworms which lived in soil mixed with the sputum of
tuberculous patients, were also the bacteroids of these excretory
cells, and, moreover, Cuénot believes that among the three
kinds of commensal bacteria, found by Lim Boon Keng (1895)
in the coelomic fluid of earthworms, there were undoubtedly
some of the bacteroids which had become accidentally freed from
the cells. The crystalline nature of these bacteroid bodies was
demonstrated by Cuénot, while their chemical composition (i.e.
that they are formed of uric acid) was proved-by Willem and
Minne."

' It isimportant to mention here that Willem and Minne (1899, pp. 16-19)
have completely misunderstood Cuénot, in ascribing to him the opinion

PHARYNGEAL GLAND OF THE EARTHWORM 55

(b) Yellow cells of the alimentary canal.

In the wall of the alimentary canal of the earthworm, between
the epithelial cells, there are often found special cells filled with
yellow spherules. These cells vary in size and shape; they
may be either spherical or oblong, or even irregular and amoe-
boid. The number of nuclei depends upon the size of the cell,
and the cells oceupy a variable position in the wall of the gut,
being either very deeply placed in the epithelium, near the
coelomic cavity, or extending themselves to the lumen of the
cut. Cuénot, to whom we owe a very good description of these
cells, considered them as belonging to the intestinal epithelium,
and ascribed to them an excretory function. According to
Willem and Minne these cells do not belong to the alimentary
canal, but are amoebocytes which originate from the haematic
system.

They make their way through the walls of the blood-vessels
and the epithelial cells of the mid-gut, which they destroy on
their way, and then, filled with the products of excretion, they
leave the organism by way of the intestine.

The distribution of these cells in different specimens is very
irregular ; in some specimens they are rare and difficult to
find, while in others they are very numerous.

Up to the present these cells have only been mentioned as
occurring in the wall of the alimentary canal between the crop
and anus. During this study I frequently found them in the
pharyngeal bulb and especially in the wall of the oesophagus,
which they traverse in the same manner as they do the wall of
the intestine. Text-figure 6, B and OC, shows these cells
lying in the wall of the oesophagus, their protoplasm being filled
with corpuscles of excretion, fat spherules, and some albuminoid
bodies. Onseveral occasions I found the cuticle of the oesopha-
gus perforated at the place of contact of the yellow cells,
thus establishing a communication (Text-fig. 6, C, 0.) between

that the bacteroid bodies are the real bacilli. Throughout his work Cuénot
criticized this opinion, and described and figured these bodies as ‘cristal-
loides’ of excretion.

56 D. KEILIN

the latter and the oesophageal lumen. It is very easy to con-
ceive that a violent contraction of the earthworm will expel
these cells, with their contents, into the lumen of the alimentary

TrxtT-FIG. 6.

A. Schematic figure representing a transverse section of the
pharynx of the earthworm : ph, d. = dorsal or salivary chamber of
pharynx; ph./f.=lateral folds of the pharyngeal wall; ph. v. = :
ventral chamber of pharynx.

Band ©. Sections of the oesophageal wall of All. foetida,
showing a yellow cell or excretory amoebocyte in the act of
traversing it. x 500. ae.=amoebocyte; cu. = cuticle of oesopha-
geal epithelium ; ¢.=oesophageal epithelium ; 0.=opening in
the oesophageal wall through which the amoebocyte will pass
into the lumen of the alimentary canal.

canal. The fact that these excretory cells are found indif-
ferently in all the portions of the alimentary canal corroborates
the supposition of Willem and Minne, that these cells do not

PHARYNGEAL GLAND OF THE EARTHWORM 57

belong to the intestinal epithelium but are amoebocytes of the
haematic system which fulfil an exeretory function.

(c) Reserve substance im Oligochaetes.

From the work of Gegenbaur, Beddard, and Cuénot it is known
that the usual nutrient reserve substance of Oligochaetes is
glycogen, which is localized in the special peritoneal cells which
surround the nephridia. These authors have also mentioned
that in some earthworms the glycogen is replaced by fat.

TEXT-FIG. 7.

oe ‘
hd

‘
a

ee?

Coelomic cells containing droplets of fat (cf. Text-fig. 1, fic., p. 40
of this paper). x 1,100.

More recently Willem and Minne (1899 a), who have made
complete analyses of earthworms, found that their reserve
substance is composed of fat and glycogen, the first being
localized in the ciliated cells of the intestinal epithelium,
while the second is found in the peritoneal cells."

1 The following is a quotation from the paper of these authors: *On
rencontre chez les lombrics, comme produits de réserve, de la graisse et du
glycogéne ; la premiére, constituée surtout par de Voléine, est localisée
dans des cellules ciliées de l’épithélium intestinal ; le glycogene s’observe
dans des cellules péritonéales et fournit, comme dérivé, de la dextrine *

(pp. 42-3).

58 D. KEILIN

In Allolophobora foetida, I found that the coelome
of segments 5, 6 and 7 is often filled with a crowded mass of
cells surrounding the glandular portion of the pharyngeal bulb
(Text-fig. 1, f. c.). These cells, in sections fixed with Carnoy or
Brazil, show a central nucleus lymg im a condensed central
portion of the protoplasm, while the remaining part of the latter
is filled with vacuoles (Text-fig. 7).

Sections of specimens fixed with Champy’s solution show
that the external or vacuolar portion of these cells contains
numerous globules stained in all shades, from dark brown to
black. These globules are undoubtedly droplets of fat, which,
in specimens fixed with Carnoy, are dissolved. It is quite pos-
sible that this accumulation of fat, not only in the cells of the
alimentary canal or peritoneal cells, but in the free coelomic
cells, is only seasonal, and is related to the period of sexual
activity of the earthworm.

7. REFERENCES.

Beddard, F. E. (1890).—** Contributions to the anatomy of earthworms,
with descriptions of some new species ’’, ‘ Quart. Journ. of Micros. Se.’,
New Series, xxx, pp. 421-79, Pls. xxix—xxx.

—— (1895).—** A Monograph of the order Oligochaeta’’, Oxford.

Cuénot, L. (1897).—‘‘ Etudes physiologiques sur les oligochétes ’’, ‘ Arch.
de Biologie ’, xv. pp. 79-124, Pls. iv—v.

Cerfontaine, (1890).—** Recherches sur le systéme cutané et le systéme
musculaire du lombric terrestre ’’, ‘ Arch. de Biologie’, x, pp. 327-428,
Pls. xi-xiv.

Darwin, Ch, (1881).—*‘The formation of vegetable mould through the
action of worms ”’, London, 1881, John Murray. See pp. 35-43.
Dechant, E. (1906).—** Beitrag zur Kenntnis des peripheren Nerven-
systems des Regenwurmes”’, ‘ Arbeit. aus dem Zool. Inst. Wien’,

Xvi, pp. 361-82, 2 Pls.

Kisen, G. (1894).—** On Californian Eudrilidae ”’, ‘Mem. of Calif. Acad. of

Se.’, ii, pp. 21-62, Pls, xii-xxix.
~ (1895).—* Pacific coast Oligochaeta, part I”, ibid., pp. 63-122,
Pls. xxx-xlv.

— — (1896).—* Pacific coast Oligochaeta, part Il’, ibid., pp. 123-200,
Pls, xlvi-lvii.

Kredéricg, L. (1878).—‘‘ La digestion des matiéres albuminoides chez

PHARYNGEAL GLAND OF THE EARTHWORM 59

quelques invertébrés”’, ‘Arch. Zool. Expér.’, vii. 391-400. See
pp. 394-6.

Hari, P. (1901).—‘‘ Modificirte Hoyer’sche Schleimfirbung mittelst
Thionin ’’, ‘ Arch. mikr. Anat.’, Iviii, pp. 678-85, Pl. xxxv.

Hesse, R. (1894).—*‘ Beitrage zur Kenntnis des Baues der Enchytraeiden ”’,
‘ Zeitschr. f. wiss. Zool.’, Wien, lvii, pp. 1-17, Pl. i.

Hoyer, H. (1890).—‘* Uber den Nachweis des Mucins in Geweben mittels
d. Farbenmethode”’, * Arch. f. mikr. Anat.’, xxxvi, pp. 310-74.

—— (1903).—** Schleimfarbung”’ in * Encyklopidie der mikroskopischen
‘Technik’, herausgegeben von P. Ehrlich, R. Krause, &c., vol. ii,
pp. 1197-1210.

Krause, R. (1895).—** Zur Histologie der Speicheldriisen *’, ‘ Arch. mikr.
Anat.’, xlv, pp. 93-133, Pls. vii—viii.

Langdon, Fanny E. (1895).—‘‘ The sense organs of Lumbricus agricola
Hoffm.”’, ‘ Journ. of Morph.’, xi, pp. 193-234, Pls. xiii-xiv. See p. 228.

Langley, J. N. (1889).—*‘ On the histology of the mucous salivary glands,
and on the behaviour of their mucous constituents’’, ‘Journ. of
Physiol.’, vol. x, pp. 433-57, Pl. xxx.

Lankester, E. Ray (1864).—“ The anatomy of the Earthworm, part I”’.
“Quart. Journ. of Micros. Se.’, N. 8., vol. iv, pp. 258-68. See p. 264.

Lesser, E. J., and Taschenberg, E. W. (1908).—‘‘ Uber Fermente des
Regenwurms ”’, ‘ Zeitsch. f. Biologie’, xxxii, pp. 445-55.

Lim Boon Keng (1895).—‘“‘ On the coelomic fluid of Lumbricus terrestris
in reference to a protective mechanism”’, ‘ Phil. Trans. Roy. Soc.,
London’, clxxxvi, p. 383.

Lortet et Despeignes (1892).—*‘ Vers de terre et tuberculose ’’, “C. R. de
1 Acad. des Sc. Paris’, exv, pp. 65-6.

Maximow, A. (1896).—“‘ Beitrage zur Histologie und Physiologie der
Speicheldriisen ”, “ Arch. mikr. Anat.’, Iviii, pp. 1-134, Pls. i-ii.
Mayer, P. (1896).—** Ueber Schleimfirbung ”’, ‘ Mitt. Zool. Stat. Neapel ’,

xii.

Michaelis, L. (1903).—** Metachromasie”’, in ‘ Encyklop.der mikr. Technik’,
vol. ii, pp. 797-803.

Miiller, E. (1895).—*‘ Ueber Sekretkapillaren ’’, ‘ Arch. f. mikr. Anat.’,
xlv, pp. 463-74, Pl. xxvii.

Retzius, G. (1895).—‘* Die Smirnow’schen freien Nervendigungen im
Epithel des Regenwurms”’, *‘ Anat. Anz.’, x, pp. 112-23.

Ribeaucourt, E. de (1900).—‘** Etude sur l’'anatomie comparée des lombri-
cides’, “ Bull. scient. de la Fr. et de la Belg.’, xxxv, pp. 211-311,
Pls. ix—xvi.

Smirnow, A. (1894).—‘‘ Ueber freie Nervendigungen im Epithel des
Regenwurms ”’, ‘ Anat. Anz.’, ix, pp. 570-8.

Stephenson, J. (1917).—‘‘ On the so-called pharyngeal gland-cells ot
Earthworms ”’, ‘ Quart. Journ. of Micros. Se.’, Lxii, pp. 253-86, PI. xix.

60 D. KEILIN

Vejdovsky, F. (1884).—** System und Morphologie der Oligochaeten ”’,
Prag. See pp. 101-6.

Vogt, C., and Jung, E. (1888).—‘* Lehrbuch der praktischen vergleichen-
den Anatomie”. Braunschweig, t. I., pp. 461-3.

Willem, V.. and Minne, A. (1899).—‘‘ Recherches sur la digestion et
l'absorption intestinale chez le lombric ”, ‘Livre jubilaire dédié a
Charles van Bambeke ’, pp. 1-22, with 1 PI.

Willem, V.. and Minne, A. (1899).—‘‘ Recherches sur l’excrétion chez
quelques annélides”’, ‘Mém. couronnés et Mém. des Savants
étrangers. Acad. R. de Belgique, Classe des Sciences ’, lvili, 72 pp.,
Pls. i-iv.

Zimmermann, K. W. (1898).—‘‘ Beitriige zur Kenntnis einiger Driisen und
Epithelien ’’, ‘ Arch. f. mikr. Anat.’, lii, pp. 552-706, Pls. xxvii-xxix.

EXPLANATION OF PLATE 38.

Illustrating Dr. Keilin’s paper: ‘On the Pharyngeal or
Salivary Gland of the Earthworm.’
Key to Lettering on Plate.

el. = cilia.

cu. = cuticle.

dl. mu. = intra-epithelial mucin ductules.

d. mu. = mucin ducts.

d. p.= mucin discharge pockets.

e. gl. = enzyme-secreting glandular cells.

m. = muscles.

m. gl. = mucin-secreting pharyngeal, or salivary cells.

mu, = mucin.

mu. c. = rmucin cells of the skin.

v. = blood-vessels.

ur. = crystals of uric acid or bacteroids,

Figs. 1 to 6 concern Allolobophora chlorotica Sav. All the
sections were stained with the Mucihaematein of P. Mayer, and Magenta-red
and Picro-Indigo-carmine (see pp. 38-9 of this paper).

Figs. 7 to 10 represent sections of Allolobophora foetida stained
by the Thionin method (see pp. 38-9 of this paper). The nuclei of the cells
are of a dark-blue colour, not purple as shown in these figures.

Fig. 1.—Section of the skin of All. chlorotica, showing mucin cells
(mu. c.) in different stages of activity. x 825,

Fig. 2.—Deep glandular portion of the pharyngeal bulb showing the
mucin-secreting salivary cells (m. gl.)and the enzyme secreting-cells (e. gl.).
x 825.

PHARYNGEAL GLAND OF THE EARTHWORM 61

Fig. 3.—Epithelial and subepithelial portion of the pharyngeal bulb,
showing the salivary or mucin ducts (d. mu.) dividing into a multitude of
fine ductules (dl. mu.), which penetrate between the cells of the pharyngeal
epithelium and terminate in the discharge pockets (d. p.) lying beneath the
cilia (c/.) of the epithelial cells. x 562.

Fig. 4. —Glandular or salivary portion of the pharyngeal bulb, showing
eranules of mucin within the cells. x 825.

Fig. 5.—Portion of the pharyngeal bulb showing the transition between
the glandular and the conductive regions. The mucin-secreting, basophile
cells are widely separated by strands of mucin. x 825.

Fig. 6.— Conductive portion of the pharyngeal bulb, showing the mucin
ducts (d.mu.), muscles (m.), and blood-vessels (v.). x 825.

Fig. 7.—Epithelial portion of the pharyngeal bulb of All. foetida
stained by the Thionin method. Section similar to that of AII.
chlorotica represented by fig. 3, but with mucin stained red. x 562.

Fig. 8.—Portion of the pharyngeal epithelium of All. foetida
showing the emission of mucin from the discharge pockets into the pharyn-
geallumen. x 825.

Fig. 9.—Section of the glandular portion of the pharyngeal bulb of
All. foetida showing the basophile cells filled with mucin. x 825.

Fig. 10. - Portion of the skin of All. foetida showing the mucin
cells. x 825.

7o7PNt@ ia OY te 10 Halen inY

he “ts 79. els 4 (lvnitioatea Was "

s Py wi : wii t lie : : Pes chest igs 4D »
«@ 1 “, i ote itvwt dil a
ie } ra voly Ht Te ;

fer;
dae are Piel

- \ighe ol
ot wef (iP aat ”

a tH ee ft a \
aalius ee : ale ee: id
‘Aare baste

Sa Eps eo bey: to

P args ferret of! ~~
) fast adig Oty TL ee Waa eal i =
- a wa 4
tah, Si, be Det peliet@anty, eee Tare fee
‘ iis LP wipe alin! 1 weet al ii be
Ate wat i-th oh 4 hi

*e

naece” - CNA
heat 1

_

=
‘ 7 is
-~ ie
«te Ga ag)
de 7 | a)
® an rs > 7 2
2 +a ' ! ete P EER A a. as
’ ian 0. ee — BR thd (ae ae
ie) oe) ee oo) Liao
ye Ue ST Shi ae Se ot

5 Fhe). 0G"
oe F¢ ee » ~~. y- eres ee
— <i Giptet Ves ws i s card

i
}
|
é
|

WAH
ay

D.Keilin del,

Huth, Jondon

Some Observations on Caudal Autotomy and
Regeneration in the Gecko (Hemidactylus
flaviviridis, Riippel), with Notes on the
Tails of Sphenodon and Pygopus.

By

W. N. F. Woodland, D.Se. (Lond.),
Indian Edueational Service,
Senior Professor of Zoology, Muir Central College, Allahabad, India.

With 6 Text-figures.

CONTENTS.
PAGE
INTRODUCTORY . ; : ‘ : i= = 64:
NAKED-EYE OBSERVATIONS ON Gre ee : oe

NAKED-EYE OBSERVATIONS ON NorMAL CAUDAL TPR aAREON . 70
THE GENERAL STRUCTURE OF THE ORIGINAL Tat OF Hemidac-
tylus flaviviridis . : : : : i Sys il
THE CaupDAL FLEXOR MUSCLES: THEIR ATTACHMENTS AND MopE
oF ACTION IN AUTOTOMY . . TUT
A BRIEF COMPARISON OF THE STRUCTURE OF THE Guard Pair WITH
THAT OF THE Tat or A Non-autotomous Lizarp, Calotes

We CL slic OG Ly .. : , : : : : 81
PLANES OR LINES OF Geaavien IN Atrcuanty : , Ee mete
THE STRUCTURE OF THE REGENERATED TAIL OF THE eens: i 3
THE HISTOGENESIS OF NORMAL CAUDAL REGENERATION : ; 87
CAUDAL REGENERATION UNDER ABNORMAL CONDITIONS : : 88
Nores oN TECHNIQUE F 94

Nores ON THE ORIGINAL AND ah: ease Tite OF Bp pene
don punctatus

R&suME , : F : é : ; :

APPENDIX. NOTE ON THE REGENERATION OF DIGITS IN AN INDIAN
TOAD

64 W. N. F. WOODLAND

INTRODUCTORY.

DuRING the rainy seasons (July to October) of 1914 and 1918
I made a large number of observations and experiments on the
facts of caudal autotomy and regeneration in the common
Indian Gecko, Hemidactylus flaviviridis* (Rippel),
a very familiar and useful ‘ snapper-up of unconsidered [insect |
trifles * found on the walls of every bungalow in the United
Provinces. In January 1915 I read a brief paper™ on the
subject at Madras on the occasion of the second meeting of
the Indian Science Congress, but until the present year I have
not had an opportunity of writmg up a complete description
of the results obtained by me.

The more conspicuous features of caudal autotomy and
regeneration in Jacertilia, such as e.g. the intravertebral
position of the cleavage or autotomy plane, the substitution
of a continuous cartilaginous tube for vertebral centra, of an
epithelial tube (an extension of the epithelium liming the
canalis centralis) for the spinal cord, the change in lepidosis,
absence of segmentation and subdivision of the muscles in the
regenerated tail, and other features, have of course been known
for many years (vide e. g. Fraisse ® in 1885, Brindley * in 1895,
and Tornier ’ in 1897) ; on the other hand, judging from recent
literature on the subject known to me, there still appears to be
a certain amount of uncertainty respecting even some of the
main facts. E.g..in Morgan’s ‘ Regeneration’ ® it is stated

1 The H. coctaeiof the * Fauna of British India’. At least two other
species or genera of Geckos are common in Allahabad, but the facts
described in this paper apply to all.

* Published in brief abstract in the official account of the Proceedings
of the Congress issued by the Asiatic Society of Bengal, and in several
Madras daily papers.

5 Fraisse, P., ‘Die Regeneration von Geweben und Organen bei den
Wirbelthieren, besonders Amphibien und Reptilien ’, Cassel, 1885.

4 Brindley, H. H., “Some Cases of Caudal Abnormality in Mabula
carinata and other Lizards”’, ‘ Jour. Bombay Nat. Hist. Soc.’, vol. xi,
1897-8, p. 680. .

> Tornier, G., “‘ Uber experimentell erzeugte dreischwanzige Eidechsen
und Doppelgliedmassen von Molchen’’, * Zool, Anzeig.’, Band xx, 1897,
p- 356.

® * Regeneration ’, T. H. Morgan, 1901.

CAUDAL AUTOTOMY AND REGENERATION IN THE GECKO 65

in a foot-note (p. 198) that ‘the attachment of the muscles may
be the cause of the break in the middle of the vertebrae, rather
than between two vertebrae’, and this statement (true to
a large extent), coupled with Powell White’s recent assertion 4
that ‘ there is no special autotomy-site as in the legs of crabs,
but apparently any vertebra may be involved’ (also true in
one sense), might very easily convey the impression that caudal
autotomy is a mere mechanical fracture of any given vertebra,
with the connected muscles and skin. The whole truth is,
as Leydig I believe first pointed out, that instead of there
being only one autotomy plane as in the crab’s claw, there are
as many autotomy planes, each as complex in form as that
of the Crustacean, as there are caudal segments. Further,
I have not yet met with satisfactory accounts of the conditions
under which autotomy occurs, of the exact modus operandi
of autotomy, or of regeneration under certain experimental
conditions, nor with any account of the mechanism by means
of which haemorrhage is stopped when autotomy occurs, and
I believe, therefore, that there is justification for describing the
facts as a whole in the case of Hemidactylus flavi-
viridis.

NAKED-FYE OBSERVATIONS ON CAuDAL AUTCTOMY,

(Statement 1) That caudal autotomy is very common among
Geckos may be concluded from the fact that over 50 per cent.
of two hundred specimens used in my experiments bore
regenerated tails, and that it is an easy process may be proved
by the simple expedient of catching a Gecko by any portion
of the tail posterior to the unsegmented base; thus I have
caught hold of the remaining end of the tail of a young Gecko
five times in almost as many seconds, and on each occasion
a portion came off in my fingers. In nature the animals
usually shed their tails when bitten by other Geckos or other
animals (e.g. out of twelve perfect Geckos placed together in
a box on one oceasion five had shed their tails within an hour).

(2) Geckos never shed their tails ‘ spontaneously ’ or from

1 Vide ‘Report of Brit. Assoc. Advancement Science’, Manchester,
September, 1915, pp. 472, 473.
NO. 257 F

Ob W. N. F. WOODLAND

mere alarm.’ This I have proved repeatedly by catching the
animals by parts of the body other than the tail. Further,
mere lateral flexion (the tail is not flexed to any extent dorso-
ventrally) of the tail is insufficient to cause autotomy, as may
be seen when, on being chloroformed, the animal lashes its tail
vigorously. In fact, an all-essential condition for caudal
autotomy is that the tail should be held a little distance
posteriorly to the plane at which autotomy is to oceur, a ful-
erum thus being provided for the action of the muscles. I have
proved this by anaesthetizing (with ether) a number of Geckos
and tying cotton thread round the tail in different positions,
the other end of the thread being fixed. On recovering from
the ether, the captive Gecko would at first try to run away
(though quite unalarmed, since I observed it from a good
distance away) and only find itself a prisoner by the cotton
becoming taut It would then, after several further attempts
at escape, suddenly stretch the cotton to its full extent and
deliberately autotomize one segment in front of the segment
held by the cotton. This autotomy was not a mere result of
the longitudinal pull on the tail (it requires a considerable
force to pull off a portion of the tail in the direction of its
length,” though the tail can be easily broken off by sharply

' Gilbert White, in a foot-note on page 64 of ‘The Natural History of
Selbourne’, states that,* the blind-worm or slow-worm does not need a blow
to induce it to cast off its tail. A sudden fright is sufficient.’ This is also
stated to be the case for the American Opheosaurus ventralis, the
‘glass snake’. If these statements be true (and the extreme brittleness
of the tail is doubtless correlated with the rigidity of the tail assumed
when the animal is alarmed, all the muscles contracting strongly), it is
proof that autotomy is a much easier process in these forms than in
the Gecko. Such forms as Anolis principalis, the American
‘Chameleon’, which can usually be captured by seizing the tail,
the animal only being able to autotomize by a great effort, and
Uromastix spinipes, which allows its tail to be pulled off rather
than release its hold on its burrow, on the other hand, lie at the other end
of the scale.

* Tn six recently killed Geckos, varying in length from 9-9 em. to 13-4
em,, and in body-weight from 2-4 grm, to 5-5 grm., with the cotton thread
suspending the weight tied midway in the length of the tail (hanging
vertically), the weights necessary to break the original tail varied between
54 grm, and 129 grm., as kindly determined for me by Mr. B. K, Bas, M.Sc.

a

ee

CAUDAL AUTOTOMY AND REGENERATION IN THE GECKO 67

bending it at one point laterally), but was a result of powerful
localized contraction of the tail muscles causing sudden flexion
at one point. These observations prove that autotomy is
a purely voluntary process, and this is confirmed by the
further fact that Geckos, caught by the tail, sometimes refuse
to autotomize when they perceive that escape is impossible
(compare the refusal even to attempt to fill the gas bladder
with more oxygen when a fish is over-weighted!). On one
occasion a Gecko, tied up by thread, remained captive for
three days, though it frequently tried to run away when
I approached, and it was only when I held the tip of the tail
that autotomy occurred—apparently the fulerum provided by
the cotton attachment was insufficient in this case.

(3) The original (non-regenerated) tail of H. flaviviridis
consists of a basal unsegmented region (the ‘ base’) covered
only with small inconspicuous scales, and about thirty autotomy
segments, each of which can be distinguished dorsally (Text-
fig. 1, A, D) by the presence on its extreme posterior edge of
six large projecting scales (three on each side of the middle line),
the outermost scale on each side being the largest ; ventrally
each segment extends lengthwise over two of the large median
transversely-elongated flat scales (Text-fig. 1, B, HE). As
experiments prove, autotomy can occur at the posterior edge
of the base of the tail or of any subsequent segment, but
cannot occur in front of the posterior border of the base.
In fifty captured specimens I have found examples of autotomy
having occurred naturally at every segment situated between
the base of the tail and the sixteenth segment: thus m seven
specimens autotomy had occurred at the posterior edge of the
base, 1.e. the whole of the segmented tail had been shed; m
ten specimens autotomy had occurred between the first and
second segments, and so on, the examples decreasing in number
the more posteriorly situated the site of autotomy. In nature
autotomy usually occurs in the anterior half of the segmented

region (Text-fig. 1, D, E), but may of course also occur pos-
teriorly to this.

1 Woodland, W. N. F., ‘ Anat. Anzeig.’, Bd. XL, 1911, p. 225.
.F 2

WOODLAND

Ni. is

W.

OS

IXT-FIG. 1.

TE

CAUDAL AUTOTOMY AND REGENERATION IN THE GECKO 69

The Gecko usually only sheds that portion of its tail necessary
for escape ; in other words, autotomy usually occurs at a seg-
ment situated not more than two segments in front of the point
of seizure of the tail. This is proved by the results of the
following experiments :

Thread tied between Tail shed behind
3rd and 4th segments ; 2nd segment ;
lith and 12th ..,, 10th. ts
Thread tied across middle of

Sth segment ; (thy
DOG AOD
Shea: .. 12th *,
ithe 2 Gs
ih Gyles ot xe iti, see
OG oi: 20th he
2 Gt hs =, 25th =

AD SS cera Ae

The Original and Regenerated Tails of the Gecko (all figures about
natural size).

A.Hemidactylus flaviviridis with original tail, dorsal vicw.

B. A Pe Original tail, ventral aspect.

y Tail regenerated from the
base, dorsal aspect.

ID} c: # Tail regenerated from the
5thsegment of the original
tail, dorsal aspect.

E. a is Tail regenerated from the 4th
segment of the original
tail, ventral aspect.

B=unsegmented base of tail; or=origina] tail; RT =regenerated
tail ; s=one autotomy segment of the original tail; v= cloacal
aperture.

F. End-on aspect of a tail segment 3 days after autotomy has
occurred. The edge of the original skin shows no sign of exten-
sion over the ‘wound’ surface. G. The same, 6 days afterwards.
The surface of the wound is now covered over with a new young
skin, formed by the histogenetic cells, hiding the transverse
processes of the vertebra. H. The same, 9 days afterwards. The
multiplication of the histogenetic cells has now produced a
slight protuberance. J. The same, 11 days afterwards. K. The
same, 13 days afterwards, dorsal and end-on aspects. The
protuberance is now well marked. L, N, O, P, Q represent stages
of growth after 15, 17, 19, 33, and 50 days respectively.

< 99 29

70 Ww. N. F. WOODLAND

When a Gecko is wounded on the tail, it usually subsequently
sheds the tail immediately anterior to the wound as the
easiest method of repairing the injury.’

(4) The regenerated tail, not being segmented in character
(see description of structure below), cannot be shed in parts
(its thin fragile extremity can, however, be easily broken or
bitten off), though it may be shed as a whole either at its
junction with the stump of the original tail or attached to
a few segments of the original tail. This has been proved by
numerous experiments which I need not record. Usually (Gin
eleven out of thirteen experiments) when a Gecko is caught by
the thick anterior portion of the regenerated tail, the whole of
the regenerated tail is shed at 1ts Junction with the stump of the
original tail ; im some cases, however (in two out of the thirteen
experiments), the regenerated tail is shed with either one or two
(rarely more) of the posterior segments of the original tail
attached ; in other words, autotomy at the junction of the
regenerated with the original tail is only a little more easy than
autotomy at any ordinary joint of the original tail.

(5) Whenever autotomy occurs, the escape of blood from
the caudal artery is practically nil. If, however, a segment
be eut through in the middle, haemorrhage is a little more
pronounced, and if the base of the tail be cut through (i.e.
anterior to the first joint or autotomy plane) bleeding is profuse.
The explanation of these facts will be found in the description
of the structure of the origimal tail given below.

NAKED-EYE OBSERVATIONS ON NoRMAL CAUDAL
REGENERATION.

(6) Regeneration of a tail only normally occurs at the
posterior surface (a) of the unsegmented base of the original
tail, or (b) of a segment of the original tail, or (c) of the end of
the regenerated tail which has had a portion broken off (not
autotomized). ‘Text-fig. 1, F-Q, shows the stages of develop-

* In these cases, apparently, the weakening of the joint caused by the
wound renders seizure of the tail posteriorly unnecessary.

EE

CAUDAL AUTOTOMY AND REGENERATION IN THE GECKO 71

ment of the regenerated tail in H.flaviviridis up to that
of seven weeks, and Text-fig. 5, J’, K’, L’, shows a second tail
bemg regenerated on the broken-off stump of a first regenerated
tail.

The exact length of time it takes nm H. flaviviridis
for a new tail regenerated from the base to attain the full length
of the original tail I do not know, but it is certainly not less
than four or five months, and is probably more.

(7) There is apparently no limit (save that imposed by the
longevity of the animal) to the number of times a tail can be
regenerated.

(8) The skin of the regenerated tail 1s not a mere extension
of that of the original tail but is a new product, as shown by
both lepidosis (‘Text-fig. 1, D) and texture. The skin of the
original tail is, hke that of the trunk, head, and limbs, very soft
and rubs off easily (the tail i consequence not being easy to
skin), whereas the skin of the full-grown regenerated tail is
relatively tough and the tail is easily skinned. After auto-
tomy the origmal skin shows no signs whatever of growing
over the raw exposed surface, and remains quite distinct from
the new skin which covers the outgrowing regenerated tail

(Text-fig. 1, H-P).

THE GENERAL STRUCTURE OF THE ORIGINAL JAIL OF
Hemidactylus flaviviridis.

In those Lacertilia in which the tail is of distinct use to the
animal for purposes other than that associated with autotomy,
e.g. for prehension (as in the Gecko Ceratolophus auri-
culatus, Bavay, and in Chameleons), for swimming (as im
aquatic Monitors, Iguanidae, Amblyrhynchus, Lophurus, and
Physignathus), for steermg (Basiliscus in water, Ptychozoon
in air), or for balancing in air (Draco), caudal autotomy naturally
does not occur, but it appears to me that the tail in the more
common Lacertilia (Lacertidae, Agamidae, &¢.) can be of but
very little use to its owners. It is in these forms not used for
swimming (as may be proved by throwimg lizards into water,

72 W. N. F. WOODLAND

when progression is seen to be effected mostly by undulations
of the trunk, the tail only waving as an appendage of the
trunk), nor for leverage (like the tails of hounds in turning
corners), nor for balancing (lizards with amputated tails
appear to be at no disadvantage in climbing or running), nor
for means of offence (I have chased large Monitors in the
jungles on the south coast of Ceylon and in southern India
and on no occasion have they attempted to strike with their
tails, though they can lash them) ; on the other hand, the tail
must often be a positive disadvantage, since it 1s easy to catch
most lizards by their tails. It is also a fact that in many
lizards the tail muscles are more or less degenerate (the white
muscles being valued as food in many cases), or at least ncapable
of exerting much force (m Central India the snake-charmers tie,
without cord, the tails of small Monitors in loops round their
necks, the bases of the tails then serving as convenient handles!).

These being the facts, it is not surprising that numerous
Lacertilia have discovered in their tails, otherwise useless and
indeed a danger, a means of self-preservation by the adoption
of caudal autotomy. As we shall see in the Gecko, the whole
structure of the tail is adapted for autotomy at every joint,
and if, after describing these adaptations, we glance at the
structure of the tails of lhzards which are non-autotomous
(e. g. Calotes), we shall appreciate the considerable simplification
of structure which must have taken place in the ancestors of
the Gecko in order to produce an autotomous tail.

If we examine longitudinal (Text-fig. 2, A, C) and transverse
(Text-fig. 2, B) sections of the original Gecko tail we shall
observe the following features. (a) The skin is divided into
cylindrical regions, each covermg one complete autotomy
segment, by lines of cleavage (described in detail later),
each of which extends round the entire circumference of the
tail, and the small seales forming the uniform covering of the
skin are arranged (‘T'ext-fig. 2, A) im correspondence with these
regions : at the anterior or posterior edge of the region bordering
the line of cleavage the small scales are arranged in a transverse
circumferential line, whereas in the space between the lines

CAUDAL AUTOTOMY AND REGENERATION IN THE GECKO 73

TEXT-FIG. 2.
As VS beGRIE AUT.

ROOAT FB

ey Cee
SOE Oe MUM eo TT LRG Cr
moth ar ge a ST TR sae een 4 js Sere mee eae aang FM

Sr ge SENS Dg a, LN a lage ER Ley
é ST WEE WN Torey Se pei pp Ca ag!
BL Rr cr cg Gr ua Wr Wr py lat EI OD pp RG
qT i tern SP eer, LL Gy Thm a —
soot AIT I A eae LT mt Ae :
' \ \ N Pe Aware
Dame co eyo NGS ee

Structure of the Original and Regenerated Tails of the Gecko,

A. Semi-diagrammatic thick sagittal section through the original tail of the
Gecko (x cir. 8). Aas=extent of autotomy segment through the vertebrae ;
B=unsegmented base of tail; cs=sphincter on caudal artery; cvc=
constriction of caudal vein anterior to autotomy plane; rB=fat band ;
FL= subcutaneous fat layer ; rm= flexor muscle ; HA=haemalarch ; mMB=
muscles of base; NC=spinal cord; Ncc=notochordal canal ; N.sp. =neural
Spine; PL.AUT. = plane of autotomy ; TP= transverse processes of vertebra ;
v=cloacal opening ; vs= extent of vertebral segment; v.sc. = transversely-
elongated ventral scales. B. Semi-diagrammatic thick transverse inter-
vertebral section through the original tail of the Gecko ( x cir. 8); FB and FL
as in A; SN=spinal nerves. The numerals indicating the flexor muscles
seen in transverse section are for identification of these muscles with those
shown in Text-fig. 3 (A, B, C, D, E, F). C. Sagittal section through the
junction of the original and the regenerated tails ( x cir. 60). Most letters
asin A. cr=cartilaginous tube of the regenerated tail ; N.c.’ = extension of
epithelium lining the canalis centralis into the regenerated tail. The general
character of the hyaline septa which mark the autotomy planes is well shown.

74 WwW. N. F. WOODLAND

of cleavage, the scales of adjacent longitudinal rows alternate
with each other. (b) Underlying the skin is a layer of fat
cells, thin dorsally, thick laterally, and extremely thin ventrally
(Text-fig. 2, B). This subcutaneous fat layer of the
tail is also divided into cylindrical segments by lines of cleavage
continuous with those of the skin; on their internal surface
the fat cells are bounded by a thin dense layer of connective
tissue. Text-fig. 3, K represents the fat layer which has been
cut through in the mid-dorsal line and flattened out. The
extreme thickness of the two laterally-situated regions is well
shown. (c) Lying internally to the subcutaneous fat layer is
a layer of muscles, the caudal flexor muscles, the
attachments of which will be described later. The laterally-
situated flexor muscles are the thickest, as might be expected.
On their external surfaces these muscles abut agaist the dense
connective tissue lining of the subcutaneous fat layer, and on
their internal surfaces they are for the most part attached to
the outer surfaces of the submuscular fat bands. (d) Lying
internally to the layer of caudal flexor muscles are the s u b-
muscular fat bands. These thick bands are four in
number, two on each side of the vertebral column, and on each
side one lying dorsally to the transverse process of the vertebra
and the other ventrally. These fat bands are, like the sub-
cutaneous fat layer, chiefly composed of fat cells, and are
segmented by lines or rather planes of cleavage continuous
with those already mentioned. The four fat bands are traversed
by eight longitudinal radiatmg connective tissue septa (one
dorsal, one ventral, two lateral, and four in between these),
which unite the dense connective tissue layer lining the mternal
surface of the fat layer with the thin layer of connective tissue
investing the vertebral column. These and minor septa
(shown in Text-fig. 3, J, im which the fat layer has been cut
through along eight lines, and the muscles and skin removed)
separate the individual muscle processes from each other and
serve to some extent for the attachment of the muscles.
(ec) Internally to the submuscular fat bands and forming the
axis of the tail is the caudal vertebral column.

ee

a a ae

a

CAUDAL AUTOTOMY AND REGENERATION IN THE GECKO 175

Each elongated saddle-shaped vertebra consists of an elongated
centrum containing a notochordal canal (full of tissue not
shown in the figures) running continuously between successive
centra but closed at the planes of cleavage (autotomy planes),
to be mentioned shortly. Successive centra are separated by
intercentral pads of cartilage (perforated by the notochord),
to which the bony haemal arches (chevron bones) are attached
below. Midway in its length, and on the anterior side of the
vertebral cleavage plane, each centrum gives off laterally on
each side a large transverse process, which extends
outwards and posteriorly to the outer surface of the sub-
muscular fat bands. On the ventral side of each intervertebral
joint and attached to the joint (not the centrum) is the ha ema]
areh which bears a median haemal spine for the attach-
ment of muscles. Dorsal to the centrum is the neural arch
which mid-vertebrally bears a conspicuous neural spine.
The well-known feature of the vertebral column in the seg-
mented region of the tail is the presence of a vertebral
cleavage plane dividing the whole vertebra (centrum
and neural arch) into two pieces in the middle of its length,
each autotomy segment thus containing the two halves of two
successive vertebrae. This vertebral cleavage plane is marked
by a hyaline septum which is continuous with the similar
septa marking the cleavage planes of the skin, subcutaneous
fat layer, muscular layer, and the submuscular fat bands, and
it is therefore obvious that, with the exceptions of the spinal
cord, spinal nerves, caudal artery, and caudal vein, and certain
longitudinal blood-vessels, the whole substance of the
tail is traversed at each intersegmental
joint by a hyaline septum marking a con-
tinuous cleavage plane.

Nor do the adaptations to autotomy im the various systems
of organs cease here. Though naturally the spinal cord and
small longitudinal nerves and blood-vessels show no signs of
cleavage planes, yet when we examine the two big blood-
vessels of the tail we find special mechanisms for stopping
haemorrhage when autotomy occurs. (f) The caudal

76 W. N. F. WOODLAND
artery, when observed in longitudinal (Text-figs. 2, A, C,
and 8, P) and serial transverse sections, is seen to possess in
its course a number of regions in which its walls are very thick
and its lumen therefore small. These thick-walled small-
lumened regions constitute sphincters for the closure
of the artery lumen, and each one of these sphincters
is found to be situated immediately anterior to an autotomy
plane (and behind the haemal arch of each vertebra) in the
region of autotomy, and there is also one in front of the first
autotomy plane (behind the last haemal arch of the unseg-
mented base of the tail), as might be anticipated. When
autotomy occurs at any segment it is the closure of the sphincter
on the eaudal artery immediately in front of this segment that
prevents haemorrhage. As far as I am aware, this is the only
instance yet described of a sphincter muscle bemg developed
on a blood-vessel. (g) The caudal vein does not possess
sphincters and this is not surprising, since the flow of blood is
towards the body and therefore away from the portion of tail
which is cast off. Nevertheless, to prevent undue loss of
blood when autotomy occurs, the vein becomes constricted im
the region of each plane of cleavage and dilates at each m-
between region (Text-fig. 2, A, C), i.e. in the region of each
haemal arch, and when the tail is shed the open lumen apparently
becomes temporarily plugged up by blood-clotting. (h) Con-
cerning the spinal cord there is nothing worth remarking,
save perhaps that it contains as usual Reissner’s fibre (I have
also observed this in the tail of Pygopus which is autotomous).
It maintains an approximately uniform diameter throughout
its course. On the ventral side of the spmal cord and in
contact with its substance is a subneural vessel; also
contained in the neural arch but lying external and ventral to
the spinal cord are usually to be seen two lateral neural
vessels, which in reality are part of a plexus of blood-
vessels.

The above-named structures are to be found in the segmented
portion of the original tail of the Gecko. There remains for
description the unsegmented base (Text-fig. 2, A) of the tail. The

————

CAUDAL AUTOTOMY AND REGENERATION IN THE GECKO 77
skin and subeutaneous fat layer in this region are unseemented.
The submuscular fat bands are absent, their position and that
of the muscles of the tail segments being occupied by large
longitudinally-disposed masses of muscle doubtless concerned
with the occasional movements of the tail base. The type of
muscles found in the segments of the tail is quite absent.
In the base, i.e. the region between the cloacal aperture and
the first autotomy plane (marking the anterior border of the
first segment), two and a half vertebrae are to be found in the
adult Gecko (1 found three and a half vertebrae in the base of
a young Gecko), and the base thus consists of the equivalents of
two and a half tailsegments. Only two haemal arches are present
in the region of the base, these being attached to the last two
intercentral cartilages, the first intercentral cartilage only pos-
sessing, like the trunk vertebrae, a small ventral nodule of bone.

THe CaupAL FuExor MuscuEs: THEIR ATTACHMENTS AND
Mops or AcTION IN AUTOTOMY.

If we catch a Gecko by its original tail and examine the
front aspect of the piece shed, we see (Text-fig. 8, A) that
lying external to and arising from the four submuscular fat
bands are eight projecting muscle processes (numbered 2’ 2”,
4’ 4”, 6’ 6”, 8’ 8” on each side of the segment), two arising from
each fat band. If we examine the hind aspect of the portion
of tail left attached to the animal (Text-fig. 8, B) we again see
the four fat bands, external to which are eight cavities which,
before autotomy, lodged the eight muscle processes just
mentioned ; there are also to be seen two pairs of small tapering
muscle extremities, one in the mid-dorsal line (labelled 1’)
and one in the mid-ventral line (labelled 10’). The transverse
processes of the vertebra are also conspicuous. If we now
remove from a single shed segment of the tail the skin and the
subcutaneous fat layer, the entire musculature of the segment
becomes visible (Text-fig. 8, C). Anteriorly the eight muscle
processes are to be seen; posteriorly each of the four dorsal
processes is seen to bifurcate, the halves of each, however,
uniting with adjacent halves, except in the case of the two

F, WOODLAND

N.

W.

3.

TEXT-FIG.

CAUDAL AUTOTOMY AND REGENERATION IN THE GECKO 79

dorsalmost halves which are separated by the vertebral neural
spine, to form altogether six posterior muscle extremities.
Each of the four ventral processes (Text-fig. 8, D) end pos-
teriorly ina similar manner. ‘Thus on the posterior side of the
segment there are altogether ten points of termination of the
muscles instead of sixteen, since twelve of these fuse together
in pairs and only those in the mid-dorsal and mid-ventral lines
persist separately. Text-fig. 3, EH, F illustrates respectively
dorsal and lateral views of the musculature of several adjacent
segments, from the latter of which it will be seen that the
lateral posterior processes, which contract most vigorously in
tail flexion or autotomy, become attached to the strong pro-

TEXT-FIG. 3.

Structure of the Original Tail of the Gecko and of the Tail of Calotes.

A. Front end-on aspect of the piece of separated-off tail after
autotomy (x cir.2). Eight large muscle processes are seen which
were, before autotomy, lodged in the eight interseptal recesses
seen in fig. B, B. Posterior end-on aspect of the stump of the tail
after autotomy (x cir. 2). Eight recesses (situated under the
subcutaneous fat layer) are visible, separated from each other
by radiating septa of connective tissue: these lodged the eight
muscle processes seen in fig. A. The transverse processes are
visible, also the extreme hind end of the haemal process. C. Dorsal
aspect of the flexor muscles of a single tail (autotomy) segment
(x cir. 2). D. Ventral aspect of the posterior flexor muscles of
asingle tail segment (x cir. 2). KE. Dorsal aspect of the arrange-
ment of the flexor muscles ( x cir. 2). F. Lateral ditto. G. Attach-
ment of the flexor muscles to the fat bands seen in longitudinal
sections (x cir. 2). H. The segmented subcutaneous fat layer
exposed after removal of the skin from three of the tail segments
(xcir. 2). J. Transverse section through the posterior half of
a tail segment showing the central septal attachments of the fat
layer. The spaces between the (cut) fat layer and the fat bands
are empty and form the eight recesses referred to in figs. A and B.
In the anterior half of a tail segment the fat layer is attached all
round to the outer surface of the flexor muscles. K. The fat layer
of three segments cut through in the mid-dorsal line and spread
out. Very few fat cells are present in the thin mid-ventral area
(x cir. 3). Lines of cleavage are visible. L. Transverse section of
the tail of Calotes (x cir. 3). The multiple subdivision of the
peripheral muscles and the absence of a fat layer and fat bands
are noteworthy. The large internally-situated muscles run
longitudinally the whole length of the tail. M. Portion of dorsal
skin of the tail of Calotes (x cir. 2). N. Dorsal aspect of muscles
of tail of Calotes after removal of skin ( x cir. 2). O. Lateral ditto.
P. Sphincter on caudal artery of Gecko seen in longitudinal
section (magnification unrecorded, but about 70 diameters).

SO W. N. F. WOODLAND

jecting transverse processes of the vertebra. I have labelled
each of the anterior muscle processes and their posterior
extremities in order that the muscle masses shown in the
fizure of the transverse section of a segment (Text-fig. 2, B)
inay be compared with those of Text-fig. 3, A, C, D, E, F.
In short, dissection and serial sections show that all the pos-
terior continuations of the eight anterior muscle processes are
firmly attached posteriorly to the vertebral
axis, directly dorsally and ventrally to the neural and haemal
spines respectively, and laterally to the transverse processes,
and indirectly by connexion with the eight radiating septa of
connective tissue which join the connective tissue internal
lining of the fat layer with the connective tissue external
investment of the vertebrae—these traversing the area of the
fat bands. The muscles are also firmly attached on their
inner surfaces to the fat bands (Text-fig. 3, G), which them-
selves are firmly connected with the connective tissue invest-
ment of the vertebrae. The eight anterior muscle processes,
on the other hand, are only feebly attached to the septa
separating successive muscle segments. Usually the tail of
a Gecko merely depends from the body, but when the animal
is excited (as when pursuing a fly) the tail can be slowly flexed
from side to side. During these lateral flexions of the tail the
muscles of many segments on one side of the tail contract and
the strains on the slender anterior attachments of the muscles
are relatively slight, however violent the flexion (as when the
animal is being chloroformed), because the muscles of many
segments are involved, i.e. the effect is distributed between
them and the tail is freely movable. On the other hand, when
the tail is seized by another Gecko, the part of the tail seized
is relatively fixed, and since the body is also fixed in position,
and the muscular contraction involved in autotomy is limited
to one segment (see Statement 2) and is therefore proportion-
ately violent, the contraction of the muscle, in trying to flex
relatively inflexible segments, i.e. in trying to cause to approach
each other the sides of two adjacent segments which, under the
conditions, can only approach to a very small extent, ig then

CAUDAL AUTOTOMY AND REGENERATION IN THE GECKO Sl

and only then able to effect the disruption of the feeble anterior
attachments of the muscles. Disruption of the muscles having
occurred on one side (and with it disruption of the skin, fat
layer, fat bands, and vertebrae along their cleavage planes),
the muscles of the other side of the segment contract violently
in their turn and so complete the process of autotomy. This
interpretation of the action of the muscles in autotomy explains
why it is that the Gecko cannot shed its tail unless it is held,
i.e. relatively fixed, a fact which I have already remarked upon.

A Brizr CoMPARISON OF THE STRUCTURE OF THE GECKO TAIL
WITH THAT OF THE J'alt OF A NoON-AuUTOTOMOUsS LizARD—
Calotes versicolor.

If we examine the tail of a typical non-autotomous lizard,
such as Calotes versicolor,’ we find conspicuous dif-
ferences from the Gecko tail. In Calotes the tail is covered
with equal-sized scales arranged in longitudinal rows, all the
scales of adjacent longitudinal rows alternating with each other
in position (Text-fig. 3, M); thus the arrangement of the
scales shows no signs of segmentation, and lines of cleavage
are of course absent. The annular arrangement of the scales at
the ends of the autotomy segments of the Gecko tail must
therefore have arisen secondarily in relation to autotomy.
Internally in the Calotes tail, fat layer and fat bands are both
absent, the entire space between the skin and the vertebral
column being occupied by muscles. The general arrangement
of these muscles, which can be seen when the tail is skinned
(Text-fig. 8, N, O) and from transverse sections (‘Text-fig. 8, L),
is much more complicated than in the Gecko tail. In Calotes
all the superficial muscles are arranged in a zigzag myotome
fashion, but those internally situated are continuous (not
myomeric) and run longitudinally the greater part or the whole
of the length of the tail. In Varanus a similar arrangement
of the muscles obtains. From these facts it will appear

4 The cut tails of two Calotes showed no signs of regeneration after one
anda half months of captivity, and I have never met with a regenerated
tail in this animal in nature, nor in Varanus.

NO. 257 G

S

32 W. N. F. WOODLAND

(

i

probable that in the Gecko tail the four sub-
muscularfat bands must represent centrally-
situated longitudinal unsegmented muscles
which have degenerated into fat and become
secondarily segmented for autotomy. It is also
certain that the superficial muscles of the Gecko tail have
become secondarily simplified and segmented in relation to
autotomy.

PLANES OR LINES OF CLEAVAGE IN AUTOTOMY.

The annular lines of cleavage in the skin are indicated
(1) by the arrangement of the scales in the skin, a regular
straight transverse row of scales bordering each side of the line
of cleavage (Text-fig. 2, A), and (2) by the presence of a line
of very thin transparent substance, devoid of pigment and
other cells, separating the two straight lmes of scales of adjacent
segments. Apparently in this line of tissue the epidermis and
dermis of the integument have become extremely attenuated
and practically reduced to a layer of non-cellular hyaline
matrix, only occasionally traversed by capillaries and nerves
passing from one segment to another. In the subcutaneous
fat layer (Text-fig. 4, E) the lmes of cleavage are denoted by
similar lines, alone composed of this non-cellular hyaline
matrix and bordered by several rows of connective tissue cells,
outside which lie the cells of the fat layer. Similar sheets of
matrix separate the muscle segments of the tail and the seg-
mented parts of the longitudinal fat bands (Text-fig. 2, C).
With reference to the plane of cleavage dividing the middle of
each centrum and neural arch, Gadow + (p. 494) describes this

as a ‘ cartilagmous septum .. . which coincides exactly with
the line of transverse division of the vertebra . .. where the

tail breaks off and whence it is removed’. This is a mistake ;
the vertebral plane of cleavage simply consists (Text-fig. 2, C),
like the planes and lines of cleavage already mentioned, of
a sheet of non-cellular hyaline substance which is continuous

* The Cambridge Natural History. Volume on Amphibia and Reptiles,
H. Gadow, 1909.

CAUDAL AUTOTOMY AND REGENERATION IN THE GECKO 83

with those separating the other tissue of adjacent segments ;
also the plane of cleavage lics immediately behind the trans-
verse process of the centrum, which is therefore not affected
by autotomy and remains projecting from the posterior surface
of the portion of tail retained by the animal (Text-fig. 1, TF).
I have verified these statements in numerous longitudinal and
transverse microtome sections, also in hand-cut sccetions, these
latter proving, in virtue of their thickness, more useful on the
whole than the former.

I may add here that there is apparently great general
similarity between these cleavage planes in the Gecko tail and
the ‘ breaking plane’ which Paul? has recently described in
detail in Decapod Crustacea. In fact the only conspicuous
difference between the two is as regards number—in the
Crustacean there is only one plane for each limb, whereas in
the Gecko (as in the Ophivroid arm) there are as many planes
as there are jomts. And just as there is a sphincter on the
Gecko caudal artery to stop haemorrhage, so in the Crustacean
there is a diaphragm developed for the same purpose. In all
cases muscular action affects autotomy of the shed part along
the cleavage plane.

THE STRUCTURE OF THE REGENERATED TAIL OF THE GECKO.

The most conspicuous difference between the regenerated
tail and the original tail is the total absence of any signs of
segmentation in the former, either on the surface or in internal
structure. On the dorsal surface of the tail the skin bears
a uniform covering of the usual small scales (‘Text-fig. 1, C, D),
i.e. the small scales are arranged in the same somewhat irregular
manner throughout the length of the tail, and no larger scales
are present. On the sides of the tail the scales are larger, and
on the median ventral surface there is a longitudinal series
of large laterally-elongated scales (Text-fig. 1, E). The sub-
cutaneous fat layer is present (‘Text-fig. 4, A), very thin dorsally

1 ** Regeneration of the Legs of Decapod Crustacea from the Preformed
Breaking Plane”’, J. H. Paul, ‘ Proc. Royal Soc., Edinburgh’, vol. xxxv,

1914-15, p. 78.
G2

W. N. F. WOODLAND

TEXT-FIG. 4.

Structure of the Regenerated Tail of the Gecko.

A. Semi-diagrammatic transverse section of the regenerated tail of
the Gecko ( « cir. 8). The multiplication of the flexor muscles (FM)
seen in transverse section (and the resulting largenumber of radi-
ating septa traversing the fat bands—rs) is noteworthy, also their
lack of connexion with the cartilaginous tube (cr), FrL= fat layer,
The caudal artery and vein are seen underneath the cartilaginous
tube. B. Dissection of the fat bands and flexor muscles, showing
the longitudinal course of the latter (nat. size), C. Transverse
section of the cartilaginous tube (x cir. 150). Bv = blood-vessel ;

DN eee ee LC LL

CAUDAL AUTOTOMY AND REGENERATION IN THE GECKO 8)

and ventrally and thick laterally, and as usual lined internally
by a thin dense layer of connective tissue ; if shows no signs of
lines of cleavage, being continuous the whole length of the tail.
Internally to the fat layer is the muscle layer, consisting of
from twenty to thirty (in different specimens) slender muscle
bands, separated from each other by a corresponding number of
radiating connective tissue septa (continuations of the dense
connective tissue lining of the fat layer which extend inwards
through the fat band to the similar, and here thick, connective
tissue investment surrounding the axial cartilaginous tube
enclosing the regenerated spinal cord) and running in a straight
line the entire length of the regenerated tail (Text-fig. 4, B).
The fibres of these muscle bands appear to run obliquely from
the central fat bands outwards to the subcutaneous fat layer
and have no special connexions in their course, except that
anteriorly all the bands are attached to the connective tissue
septa bounding the hind ends of the muscles of the base or
other portion of original tail. In autotomy the separation of
the regenerated tail from the part in front of it must be solely
effected by the contraction of these longitudinal muscle bands
away from their connective tissue junction with the last
intermuscular septum, this forcible separation causing the
simultaneous separation of the slender junctions of the other
organs. In other words, the tail being seized and held, these
muscles contract, and since the whole body cannot be dragged
back, the inevitable result is the separation of the tail.
Between the layer of muscle bands and the axial tube
enclosing the regenerated spinal cord lies the substance of the
submuscular fat bands already mentioned; these are con-
tinuous from end to end of the tail (cleavage planes being

cc =calcified cartilage at periphery (the tube is also calcified on
the inner edge); cr=cartilaginous tube; Nc=extension of
spinal cord; oc=unealcified cartilage; Prco=pigment cell.
D. Longitudinal section of the spinal cord extension in the
regenerated tail ( x cir. 580). cc=canalis centralis ; RF = Reiss-
ner’s fibre. E. Section through autotomy plane in the region of
the fat bands( x cir. 250). The hyaline septum is shown, bordered
by connective tissue cells, outside which lie the fat cells. Similar
hyaline septa extend through the vertebrae, muscles, and skin.

S6 W. N. F. WOODLAND

absent) and are radially subdivided by the numerous con-
nective tissue septa above described. Forming the central
axis of the regenerated tail is a thick-walled cartilaginous
tube (Text-fig. 4, C). The cartilage of this tube is calcified *
on its outer surface (next the fat bands) and on its mner
surface (next the spinal canal), the space between these two
concentric cylinders of calcified cartilage consisting of ordinary
unealcified cartilage. Anteriorly this cartilaginous tube joins
on to the ring of bony tissue formed by the centrum and
neural arch of the last vertebra (Text-fig. 2, C) and so secures
a continuation of the spinal canal. The cartilaginous tube is
quite continuous—no planes of cleavage being present—and
it bears no processes of any kind, neural spines and haemal
arches both bemg absent. The contents of the cartilagimous
tube are (a) a very attenuated extension of the spinal cord
(about a quarter or less of the diameter of the original) which
practically consists of a continuation of the cellular linmg of
the canalis centralis, with little or none of the external nerve-
fibre substance ; (b) a network of capillaries which lies for the
most part ventrally to the spinal cord extension ; and (¢) an
arachnoid meshwork containing pigment cells. In view of the
fact that no nerves are given off from this slender extension of
the spinal cord into the regenerated tail, it is evidently quite
a useless structure so far as muscular innervation is concerned ;
it, however, contains a well-developed Reissner’s fibre (Text-
fig. 4, D). It may here be mentioned that the nerves supplying
the slender muscle bands are all derived from the last two or
three pairs (I have not determined the exact number) of nerve
roots in the stump of the original tail (according to Powell
White, the nerves are, in Lacerta vivipara, derived
from the last three pairs) and, as stated, have no connexion
with the regenerated spinal cord. In the abstract of Powell
White’s paper it is stated that m Lacerta viridis the
cartilaginous tube enclosing the spinal cord is ‘ unsegmented

! This calcification of the cartilage is apparent in thick unstained hand-

cut sections of aceto-bichromate-fixed material ; in ordinary microtome
sections it is not easily seen,

CAUDAL AUTOTOMY AND REGENERATION IN THE GECKO 87

and continuous except for some perforations through which
blood-vessels pass to the interior’. In the fully-regenerated
tail of the Gecko no perforations at all exist in the length of
the tube, not even for blood-vessels, though perforations (for
vessels) are fairly numerous in the young growing
cartilaginous tube, and I suspect that this is also the case in
Lacerta. It certainly is so in Pygopus, sections of which
Professor J. P. Hill has kindly shown to me. At the extreme
posterior end, however, of the cartilaginous tube in one series
of sections of a fully-developed regenerated tail I have found
one median ventral terminal opening and two lateral sub-
terminal openings through which blood-vessels pass, but these
are the only openings I have discovered. In another series of
sections of a young regenerated tail (6 mm. in length) I found
that the spinal cord continuation actually bifurcated at its
posterior extremity, one branch piercing the cartilaginous tube
through a mid-ventral subterminal opening, the other branch
continuing to the end of the tube, but I suspect this to be
a freak.

The caudal artery extends back into the regenerated tail
lying underneath the cartilagmous tube, and only differs from
that of the original tail in not being enclosed in a haemal canal
and in being devoid of sphincters; it gives off branches at
intervals. The caudal vein extends posteriorly under the
caudal artery and is uniform in diameter.

Tor HistoGENESIS oF NoRMAL CAUDAL REGENERATION.

Under this heading I can only confirm and correct previous
accounts. As Powell White says, ‘ The wound after autotomy
is quickly covered with new skin [not derived from the old
skin covering the stump of the original tail], beneath which is
a mass of spindle cells [| quasi-embryonic tissue | which apparently
originates in the connective tissue. This cellular mass acts as
a growing-point to the new tail, and from it the various struc-
tures are developed. The cartilage, fat, and blood-vessels
arise by differentiation from the spindle cells. The muscle
fibres arise segmentally in groups, the groups nearest the tip

SS W. N. F. WOODLAND

being the least differentiated. The muscles in the stump play
no part in the process.’ It is also possible that the continuation
of the lining epithelium of the canalis centralis of the spinal
cord is produced by these histogenetie cells. On the other
hand, it appears that the nerve trunks of the regenerated tail
are produced by the growth into the regenerating tail of the
torn ends of the trunks in the original tail, the posterior root
ganglia of which ‘are increased in size or number owing to
increase in size of the nerve bundles *. The preceding account,
which I can confirm in full as regards the origin of the skin,
muscles, fat layer, fat bundles, and cartilaginous tube, is thus
in distinct opposition to the views of Fraisse, who believed that
the skin, connective tissue, cartilagmous tube, and muscles of
the regenerated tail are all derived ultimately from the corre-
sponding tissues of the original tail—that new tissues can only
be reproduced from tissues like themselves. This belief is, in
the main, not only disproved by actual observation, but is also
contradicted by some of the results obtained from caudal
regeneration under abnormal conditions now to be described.

CauDAL REGENERATION UNDER ABNORMAL CONDITIONS.

Intervertebral Regeneration. Though Fraisse
rightly came to the conclusion that the remnants of the old
notochord (even if these be exposed by the mjury) take no
part in the formation of new skeletal tissue, yet simece a more
recent writer like Gadow (p. 494) is of opinion that ‘repro-
duction of centra {in the regencrated tail] is precluded by the
previous normal reduction of the chorda, around which alone
proper bony centra could be formed’ (though Fraisse has
shown that in the regenerated tail of Urodeles new vertebrae can
be produced in the total absence of a notochord), it may be as
well to quote, first of all, the results of my experiments on
caudal regeneration from the posterior surfaces of caudal seg-
ments which were cut in half, i.e. cut transversely between
anytwo autotomy planes, i.e.intervertebrally.
These experiments were successful on four occasions (Text-fig.
5, B, B’) and in each ease, though the notochord was well

CAUDAL AUTOTOMY AND REGENERATION IN THE GECKO 8&9

exposed by the cut, the endoskeleton of the regenerated tail
was of the normal cartilagmous tube type. These experiments
also prove that the tissues bordermg the autotomy plane are
not indispensable for the regeneration of the tail—histogenetic
cells are distributed throughout the tail tissues. I may add
that in one of these four experiments I held the animal by the
regenerated tail (of 66 days growth) but I could not induce
autotomy either from the junction of the regenerated tail with
the original stump or from a true autotomy plane anterior to it.

Regeneration from the Cut Base of the Tail.
I have stated that the first autotomy plane in the Gecko tail
is situated between the posterior surface of the base of the tail
and the anterior surface of the first autotomy segment. Since
we now know that caudal regeneration can occur at the surface
of any autotomy segment cut imtervertebrally, i.e. between
two successive autotomy planes, it is of interest to inquire
whether regeneration can occur from the posterior surface of
the base of the tail if this becut through anterior to the
first autotomy plane. ‘The answer to this question
is also of special interest when we reflect that the structure
of the base of the tail is different in several respects from that
of the segmented tail proper—in the absence of segmentation,
in the absence of fat bands, and in the arrangement of the
muscles—and that it has been contended that (in the Gecko
and the other Lacertilia which it resembles in this respect) the
regenerated tail differs in type from the original tail solely
because in development the former is shut off from the con-
trolling influence of the main organism by the hyaline septa
of the autotomy planes, whereas the original tail is developed
before autotomy planes (which are only produced after the
original tail is formed) are present.

I performed this experiment of cutting through the base of
the tail four times in 1914 but in no instance did regeneration
occur, though the Geckos were kept for two months. In 1918,
however, when I repeated the experiment, five of the Geckos
regenerated tails of the normal regenerate type
(Text-fig. 5, A’, A”), as shown by the cartilaginous tube, fatty

90 W. N. F. WOODLAND

TEXT-FIG. 5.

CAUDAL AUTOTOMY AND REGENERATION IN THE GECKO 91

tissue, nerve cord, and other features seen in sections. This
result is of importance because it proves (1) that regenerative
cells are present in a part of the body which, under normal
conditions, never reproduces a tail, and (2) that the peculiar
characters of the regenerated tail are not due to mere lack of
continuity with the rest of the organism. ‘The real reason for
the regenerate tail differimg in character from the original tail
appears to be that the organism as a whole ‘ knows’ that the

TEXT-FIG. 5.

Experimental Regeneration of the Gecko Tail (all figures nat. size).

A’. Tail regenerated from cut base, after 20 days. A”. Ditto,
after 43 days. B. Diagram showing direction of cut through the
middle of an autotomy segment. B’. Tail regenerated from cut-
through autotomy segment (cut B), after 15 days. C. Diagram
showing direction of oblique lateral cut through one autotomy
segment. C’. Tail regenerated from the cut C, after 64 days.
D. Diagram showing direction of oblique lateral cut through two
autotomy segments. D’. Tail regenerated from the cut D, after
52 days. E. Diagram showing direction of oblique dorso-ventral
cut through one autotomy segment. E’. Tail regenerated from
the cut E, after 67 days. EK”. Tail regenerated from the cut E,
after 87 days (the tail was shed when held at the point shown).
KE”. Diagram showing absence of endoskeleton in the lower
division of the bifid tail of EK”. F. Diagram showing direction of
oblique dorso-ventral cut through two autotomy segments.
F’, Tail regenerated from the cut F, after 76 days (the tail was
shed when held at the point shown). G. Diagram showing direc-
tion of two oblique lateral cuts through an autotomy segment.
x, Straight tail regenerated from the cut G, after 82 days.
H. Diagram showing direction of two oblique dorso-ventral cuts
through one autotomy segment. H’. Tail regenerated from the
cut H, after 80 days (the tail was shed when held at the point
shown). J. Diagram showing direction of oblique lateral cut
through regenerated tail (cf. cut C). J’. Tail regenerated from the
cut J, after 73 days (the tail was shed when held at the point
shown). K. Diagram showing direction of oblique dorso-ventral
cut through regenerated tail. K’. Tail regenerated from the
cut K, after 73 days. L. Diagram showing two oblique dorso-
ventral cuts through regenerated tail. L’. Tail regenerated from
the cut L, after 73 days (the tail was shed when held at the point
shown). M. Diagram showing position and extent of ventral
wound made on original tail. M’. Two tails regenerated from the
wound M, after 80 days. M”. Diagram showing absence of endo-
skeleton in the lower division of the bifid tail of M’.. N. Diagram
showing absence of endoskeleton in the lower division of a ventral
accessory tail produced from a wound similar to M, after 80 days.

B= base of tail; pc=direction of cut; oT=original tail; RT =
regenerated tail.

99 W. N. F. WOODLAND

=

reproduced tail is only reproduced for the purpose of being
shed, and in consequence the regenerated tail is grown on
cheap ‘ jerry-built ’ lines sufficient for the end im view. That
this is the explanation will be clear, on the one hand, when we
call to mind the regenerated tails and limbs of Urodeles, arms
of Starfishes and Ophiuroids, and limbs of Crabs, Centipedes,
and Plasmids (walking-stick imsects), all of which, when
regenerated, are required for use as integral parts of
the organism and are therefore of normal type’; on the
other hand, the fact that the organism can actively mould
an autotomous appendage so as to adapt it for functions not
connected with its own individuality is shown in such cases
as those of the hectocotylized arm of Dibranchiate Cephalopods
and the heteronereis segments of Polychaetes. According to
this explanation then, the aberrant scaling of the regenerated
Gecko tail is to be regarded as that form of scaling most easy
to be produced under the circumstances, just as the simple
longitudinal muscles (devoid of connexion with the endo-
skeleton) and regenerated nerve cord (devoid of white matter,
sanclion cells, and nerves) are to be regarded as similar products
of a ‘ jerry-building’ policy, and not due to a mere reversion-
to-type tendency, as supposed by Boulenger.* The type of
scaling of the regenerated tail may happen to be of an ancestral
type simply because this latter chances to be a ‘ cheaper’ or
‘to-hand * form of lepidosis, but it is quite evident that since
the ‘reversion to an ancestral type’ explanation does not
apply to the internal structure of the regenerated tail, it also
cannot be held to be sufficient to account for the scaling.
[may mention that previous to preserving the tail (of 45 days’
srowth) of one of these five Geckos, I held it with my fingers
1 The well-known examples of an antenna being generated on the eye-
stalk of Palinurus, of a mandible being substituted for a first antenna in
Asellus, and a wing replacing the hind leg of the moth Zygaena (vide
Bateson, ‘ Material for the Study of Variation’, 1894), and other similar
examples are of the same category, the ‘controlling’ influence of the
organism as a whole, however, being at fault, the reproduced part being

out of position.
* Boulenger, G. A., ‘ Proc. Zool. Soc.’, Lond., 1888, p. 351.

CAUDAL AUTOTOMY AND REGENERATION IN THE GECKO 93

and the animal shed it ‘not very easily’. The stump bled to
some extent, but not profusely.

Regeneration from obliquely-cut Ends of
the Original Tail. When one or two segments of the
original tail are cut through obliquely from left to right (‘Text-
fig. 5, C, C’) or from right to left (Text-fig. 5, D, D’), the axis
of the regenerated tail is usually bent out of the straight line
in order to place itself at right angles to the plane of the cut
(six experiments).

When one segment of the original tail is cut obliquely
ventro-dorsally and postero-anteriorly (Text-fig. 5, E, E’, E”)
the axis of the regenerated tail is usually bent downwards in
order to place itself at right angles to the plane of the cut
(three experiments).

When one or two segments of the original tail are cut obliquely
dorso-ventrally and antero-posteriorly (Text-fig. 5, F', F’) the
axis of the regenerated tail is usually bent upwards, the more
so if the number of cut segments be two (four experiments).

In four experiments in which one segment of the origimal
tail was cut to a point by left and right lateral cuts (Text-fig.
5, G, G’) the axis of the regenerated tail remained in the
straight line.

In three experiments in which similar cuts were made
dorsally and ventrally (Text-fig. 5, H, H’) the same result was
obtained.

Regeneration from the Regenerated Tail. A
transverse cut through a regenerated tail merely leads to
a second regenerated tail being produced (two experiments).

When the regenerated tail is cut obliquely (‘Text-fig. 5,
J, J’, K, K’, L, L’) the second regenerated tail behaves in the
manner already described for regeneration from the original
tail (at least six experiments).

Accessory Tails. Inall the 1918 experiments chronicled
above (which are only a selection of the experiments I
actually performed), and in a number of similar experiments
which I conducted in 1914, I only obtained four examples m
which accessory tails were produced. Text-fig. 5, M, M’ shows

4 W. N. F. WOODLAND

the result I obtained after making a wound on the ventral
surface of an original tail. In this case the tail evidently
autotomized at the autotomy plane separating the two segments
involved in the wound, and the surface thereby exposed
produced two tails. The upper tail was a normal regenerated
tail in every respect, but the larger lower accessory tail differed
in the essential respect that it was entirely devoid of a carti-
laginous tube (Text-fig. 5, M”).

‘Text-fig. 5, N shows another small accessory tail produced
as the result of a wound on the ventral surface of a regenerated
tail. An endoskeleton was also absent in this case, as also in
another similar case which I have not recorded.

In Text-fig.5, E’’’ is shown a small accessory tail produced
as the result of the oblique dorso-ventral cut already described
(Text-fig. 5, E). The lower lobe of the bifid tail was devoid
of a cartilaginous tube.

I have described these four examples of accessory tails
because, to judge from the paper by Tornier, the reader might
imagine that an accessory tail without a cartilagmous tube is
an impossibility. This is by no means the case, as these four
examples and the examples in Anolis grahami, deseribed
by Brindley in 1898, prove. Assuming the statements of
Tornier to be correct, it would appear that the injury must
reach the vertebral column in order that the accessory tail
produced may contain a cartilaginous tube.

Notes ON TECHNIQUE.

All Geckos were kept in large flower-pots, covered over with
mosquito-netting, and were fed on house-flies. ‘Tails preserved
for section-cutting were fixed for 24 hours or longer in a saturated
solution of potassium bichromate (100 parts), to which 5 parts
of acetic acid had been added ; they were afterwards washed
in running water for the same length of time, and then kept m
70 per cent. alcohol until required for use. For the study of
the gross structure of the tail, nothing is better than thick
hand-sections (longitudinal and transverse) of the spirit-
preserved material, dehydrated and mounted unstained im

i ee ee

OO 2 -$G=— €-6@-Baa——:« 46 4

7+

CAUDAL AUTOTOMY AND REGENERATION IN THE GECKO 995

balsam, the bichromate fixative acting as a stain for many of
the tissues. For histology, the spirit-preserved material was
first decalcified by leaving it in alcohol plus nitric acid for
several weeks, and subsequently dehydrated, embedded, cut,
and stained with haemotoxylin. For dissection of the muscles
preliminary maceration of the tails in weak alcohol or water
plus nitric acid gave good results.

NovTES ON THE ORIGINAL AND REGENERATED TAILS OF
Sphenodon punctatus.

I have examined the origimal and regenerated tails of
Spbenodon punctatus kindly given to me by Professor
Arthur Dendy. The scales are arranged in the original tail in
accordance with the planes of autotomy, each autotomy
segment bearing dorsally one of the large mid-dorsal scales,
and ventrally two transverse rows of the large hexagonal
scales. The muscles, after removal of the skin, have a super-
ficial arrangement closely resembling that of the Gecko shown
in Text-fig. 3, E, F, only the muscles are more numerous.
In lateral aspect, e. g., there appear to be four muscle layers
(and processes) instead of two as in the Gecko (Text-fig. 3, F).
In transverse section the muscles are also seen to be more
numerous than in the Gecko, and they extend inwards from the
skin to the vertebral column, fat bands being entirely absent.
The muscles are separated from each other by thin radiating
septa of dense connective tissue. I dissected out a piece of
the caudai artery about 9 cm. in length and cleared if im
creosote, when it was evident that sphincters were not present.
The regenerated tail is of course not segmented and the sealing
(irregular small scales) is quite irregular. A cartilaginous
tube is present, the cartilage of which is calcified in the middle
of the thickness of the ring, not on its inner and outer edges as
in the Gecko. The muscles are very numerous in transverse
section (about fifty bands cut across), and these are separated
from the eartilagmous tube not by fat bands but by dense
conncetive tissue, which is continuous with the subcutancous

96 W. N. F. WOODLAND

connective tissue by means of the radiating septa separating
the muscle bands. The tails of Sphenodon, therefore, appear
to be less specialized for autotomy than the tails of the Gecko,
though the presence of definite autotomy planes, the evident
simplification of the muscles, and the presence of the carti-
laginous tube indicate that considerable progress has been made
in that direction. :

ResuME.

1. The Gecko original tailis made up of numerous (about thirty)
autotomy segments, separated from each other by as many
hyaline septa marking autotomy or cleavage planes. Autotomy
ean occur voluntarily at any plane provided that the tail be
held a short distance posteriorly to the poimt of separation.
Autotomy in the Gecko is never ‘spontaneous’ or the result
of mere alarm.

2. The structure of the origmal Gecko tail is described.
The caudal artery develops a sphincter muscle in its walls
immediately anterior to each autotomy plane as a means of
avoiding haemorrhage after autotomy. JI am not aware that
a sphincter muscle has previously been described in connexion
with a blood-vessel. The caudal vein is similarly constricted
in front of each autotomy plane. The base of the tail differs
from the segmented portion in the absence of fat bands and
in the arrangement of the muscles. The flexor muscles of each
tail segment are firmly attached posteriorly to the vertebra
and the outer surface of the fat bands; anteriorly, however,
they are only attached to the connective tissue of the hyaline
matrix in the autotomy plane and are threfore easily separated.
Autotomy is effected by the strong localized contraction of
these muscles separating their weak anterior attachment.

3. Comparison of the Gecko tail with the non-autotomous
tail of Calotes shows that in order to effect autotomy the former
has become greatly sunplified. The seales have become
rearranged at the extremities of each autotomy segment, the
superficial muscles have also become rearranged on a more
simple plan, and the internal longitudinal continuous niuscle

CAUDAL AUTOTOMY AND REGENERATION IN THE GECKO 97

bands have degenerated into the fat bands and become
secondarily segmented.

4. The autotomy planes are marked by simple septa of a
hyaline matrix bordered by connective tissue, which traverse
and separate into segments the entire substance of the tail.
The spinal cord, nerves, and blood-vessels are, however, con-
tinuous.

5. The structure of the regenerated tail is described. Reiss-
ner’s fibre is present in the regenerated spinal cord, as in the
cord of the original tail. Boulenger’s explanation of the
changed character of the lepidosis of the regenerated tail when
compared with that of the original tail, viz. that it is a reversion
to an ancestral type, does not apply to the internal anatomical
features which distinguish the regenerated from the original
tail. A more probable explanation of the differences between
the regenerated and original tails is that the former, being
merely produced for autotomy purposes, is ‘jerry-built ’—an
appropriate description of a tail in which the muscles have no
direct connexion with the endoskeleton and the spinal cord is
devoid of nerves, ganglion cells, and fibres.

6. Tails of the normal regenerated type can be produced
from cut surfaces situated between the autotomy planes and
anterior to the first autotomy plane in the base of the tail.
This is proof (a) that the histogenetic cells oceur throughout
the tail substance and quite apart from the hyaline septa,
(b) that the peculiar features of the regenerated tail are not
due to a lack of organic connexion with the rest of the body
caused by the interposition of the autotomy plane septa.

7. The axis of the regenerated tail usually tends to be
placed at right angles to the plane of the cut on the tail stump.
In four of my experiments accessory tails were produced, none
of which contained a cartilaginous tube endoskeleton.

8. The tails of Pygopus and Lacerta viridis are ap-
parently almost identical in structure with those of the Gecko,
and in Sphenodon punctatus the tails only differ
essentially in the absence of the fat bands and the absence of
sphincters on the caudal artery.

NO. 257 H

QS

Ww. N. F. WOODLAND

TEXT-FIG. 6.

CAUDAL AUTOTOMY AND REGENERATION IN THE GECKO 99

In conclusion, I wish to express my thanks to Professor
Arthur Dendy, F.R.5., for his kind gift of two tails (one
regenerated) of Sphenodon punctatus, to Professor
J. P. Hill, F.R.S., for the loan of three slides of the tail of
Pygopus sp., to my pupil Mr. B. K. Das, M.Se., University
of Allahabad Research Scholar in Zoology, for much assistance
in the practical work connected with caudal regeneration under
abnormal conditions, and to Professor D. R. Bhattacharya,
M.Se., for some aid in 1914.

AppENDIxX. NoTE ON THE REGENERATION OF DIGITS IN AN
INDIAN TOAD.

Since, so far as I am aware, only one instance! has yet been
deseribed of a very limited regeneration of amputated digits
having occurred in adult Anura, I reproduce here drawings
(Text-fig. 6) made by my former pupil, Mr. N. K. Patwardhan,
M.Sc., of regenerated digits in the Indian toad; Bufo melan-
ostictus. These digits had been removed (in all cases they
were cut off with scissors to a little below the level of the
bases of the adjoming digits) for purposes of identification.
All the figures represent the amount of regeneration which had
occurred within 73 days of amputation, excepting figs. C, B’,

1 T refer to Gadow’s statement (Cambridge Natural History, vol. on
Amphibia and Reptilia, p. 67) that in two specimens of Rana tempo-
raria in which the hand was amputated from the wrist, “within a year
this changed into a four-cornered stump and two of the protuberances
developed a little further, reaching a length of about 4mm. These
specimens lived for four years without further changes.’

TEXT-FIG. 6.

Regenerated Digits of the Indian Toad, Bufo melanostictus,
from the dorsal aspect (all figures x cir. 3).

The arrows indicate the regenerated digits. B”,C,and D’ (all males)
represent 94 days growth; all the others (all females) 73 days
growth. It is noteworthy that in A and A’ the first digit has
grown more rapidly than any of the other digits, though these
animals were females, and the digit therefore was not used for
the nuptial embrace.

H 2

100 W. N. F. WOODLAND

and D’, in which the period was 94 days. The latter maximum
. period of thirteen weeks, three days was therefore considerably
less than the year referred to by Gadow, and in this connexion
I may mention that in another toad (a toad labelled J, celebrated
in its way since it was the only animal in which the renal
afferent veins, each cut in two, became regenerated), in which
Iamputated the 5th toe on both hind legs, that on the left leg
became completely regenerated within fifteen
months, though that on the right leg was not re-formed to
any considerable extent. Unfortunately I neglected to make
a drawing of this before I left India. Figures A and A’ repre-
sent the regenerated Ist digits on the left and right arms
respectively, and it is noticeable that though the period of
regeneration was only 73 days, and though both specimens
were females (the digits therefore not being used for the
nuptial embrace), yet they are better developed than any of
the other digits. The other figures show the partial regenera-
tion of the 2nd, 3rd, and 4th fingers. |

On the Bionomics and Development of Lygo-
cerus testaceimanus, Kieffer, and
Lygocerus cameroni, Kieffer (Procto-
trypoidea-Ceraphronidae), parasites of A phi-
dius (Braconidae).

By

Maud D. Haviland,

Fellow of Newnham College, Cambridge.

With 18 Text-figures.

INTRODUCTION.

Tue Proctotrypoidea have been less studied than most other
groups of the Hymenoptera Parasitica. Ganin (1869) was the first
to study the embryology of certam members of the group (8).
In 1884, Ayers (2) described the development of the Scelionid,
Teleas. In 1898, Kulagin (14) resumed the study of Plat y-
gaster; and in 1906, Marchal (18) published the results of
his elaborate researches into the embryology and development
of that family. In recent years much work has been done on
this group from the systematic standpoimt, notably in the
monographs of Ashmead (1) in America, and of Kieffer (18) in
Europe, but the life-histories of most of the families are com-
paratively little known.

The following is an account of the bionomics and _ post-
embryonic development of two species of the genus Ly gocerus,
of the sub-family Ceraphroninae. These forms are parasites of
the larvae and pupae of certain Braconidae, of the family
Aphidiidae, which are themseives internal parasites of various
plant-lice.*

I would here express my sincere thanks to Professor Stanley

' A preliminary note on these observations by the writer appeared in

the ‘ Proceedings of the Cambridge Philosophical Society’, 1920, vol. xix,
Ptrvr

102 MAUD D. HAVILAND

Gardiner, who gave me facilities to carry out the work in the
Zoological Laboratory at Cambridge ; and my obligations to
Professor J. J. Kieffer, and to Mr. G. T. Lyle, who kindly
determined the specimens of Proctotrypoidea and Braconidae
submitted to them respectively.

BIOLOGICAL STATUS.

The genus Lygocerus was founded by Forster, and is
included in the sub-family Ceraphroninae. Ashmead (1, p. 103)
and Kieffer (18) state that the Ceraphronmae are almost
exclusively parasitic upon Homoptera (Aphidae) and Diptera
(Cecidomiidae, &c.). Riley is said to have reared a Ly go-
cerus from a tortricid larva (Lepidoptera), but Ashmead
considers the observation to be of doubtful accuracy. ‘The genus
contains a number of species obtaimed from aphides, but their
bionomics have hitherto been in doubt, authorities disagreeing
as to whether they are parasites or hyperparasites.

Curtis believed correctly that they were hyperparasites, and
Buckton (4) agreed with him; but later writers have reverted
to the view that these Proctotrypids are directly parasitic upon
the aphides from which they are reared. Thus Ashmead
(1, p. 21), who says that the larvae all feed upon the host
internally, contmues: ‘Lygocerus and allied genera
living in the Aphidae, gnaw a hole through the ventral surface
of the aphis, and after securely fastening the aphid by a silk-
like seeretion to the leaf or twig upon which it has been feeding,
pupate within the body of their host, which, in leu of a cocoon,
affords ample protection to the larvae to undergo their trans-
formations.’ Gatenby (9) says, ‘I am inclined to support the
view that the Proctotrypid is a parasite and not a hyper-
parasite ’.

The subjects of this paper, Lygocerus testacelimanus,
Kieff., and L. cameroni, WKieff., are both hyperparasites.
The eggs are laid and the larva stages are passed outside the
body of the host. The Aphidius larva, in the course of its
development, devours the internal organs of the aphis im which

ay Ve nt are

a eee ee ee

ay > i ee ee ee

BIONOMICS AND DEVELOPMENT OF LYGOCERUS SP. 103

it is reared; and when it is full-fed, it lines the empty skin
with silk, and pupates within it. At this time, it is itself liable
to parasitisation by the Proctotrypids (fig. 1).
Lygocerus does not confine itself to Aphidiidae. Twice
I have observed its larvae upon newly-transformed and dead
pupae of its own species. The aphidivorous Braconidae are
known to be parasitised by certain Chalcidae and Cynipidae,
some of which were reared from material collected in the
field in the course of this work. Lygocerus cameroni

Text-Fic. l.

Skin of Macrosiphum urticae cut open to show the full-grown
larva of its parasite, Aphidius ervi, which has in turn been
attacked by Lygocerus cameroni. Anegg, and third stage
larva of the hyperparasite are represented.

occurred occasionally upon the adult larvae of a Chalcid, prob-
ably Asaphes vulgaris, and also upon a second species,
not yet determined, which is possibly a Cynipid (Allotria
sp.). Apart from the two cases mentioned above, where the
larva had been hyperparasitised by its own species, Ly go-
cerus was never found to be attacked by another hymen-
opteron.

One remarkable instance of hyperparasitisation came under
notice. An aphis (Macrosiphum urticae) had been
parasitised by Aphidius ervi. The latter had been
hyperparasitised by an undetermined species of Chalcid.
This form, after metamorphosis, had been devoured except for
the frass, the head, and part of the thorax, by a second hyper-
parasite, whose life-history 1s not yet worked out. his larva
was full-grown when the cocoon was opened, but it had itself

104 MAUD D. HAVILAND

been recently hyperparasitised by Ly gocerus cameron.
Hence, within certain limits, this species seems to be poly-
phagous.

MATERIAL.

The material used was obtained in Cambridge m the summer
of 1919. At the end of June,a variety of L. testaceimanus,
Kieff., was reared from the larvae of Aphidius salicis,
Hal., parasitic in the sexuales of Aphis saliceti, Kalt.,on
the willow ; and as the host material became scarce, I subse-
quently induced it to oviposit on larvae of Aphidius ervi,
Hal.,in Macrosiphum urticae, Kalt., onthe nettle. In
July, I reared a number of L. cameroni, Kieff., from the
latter material collected round Newnham ; and as the host was
plentiful, and, owing to its larger size, easier of dissection than
the parasites from the willow, I worked with it exclusively in
July and August. The followimg account therefore applies
especially to L. cameroni, though the life-history of
L.testaceimanus is essentially the same.

Aphides parasitised by A. ervi were collected in the field,
but a proportion of these were found to be already hyperpara-
sitised by certain Chalcidae and Cynipidae. ‘To ensure a ‘ pure
culture’ of Lygocerus, nettles infested with Macrosi-
phum urticae were placed in water under bell-jars in the
open air insectary, and exposed to Aphidius ervi. The
aphides were kept under cover during the development of the
parasite, and when the latter were about to transform, the
leaf was cut off, and placed in a glass tube with a fertilized
female of Lygocerus. Thus the possibility of an mfection
by another hyperparasite was virtually eliminated.

I tried many times to cut open a flap on the dorsal side of the
aphis skin, hoping by this means to follow the complete develop-
ment of the hyperparasite from day to day, but the attempt
always failed through the death of both the Aphidius and
the Proctotrypid within a few hours.

BIONOMICS AND DEVELOPMENT OF LYGOCERUS sP. 105

PAIRING.

No parthenogenetic ovipositions were observed, and about
40 per cent. of the imagos reared were males. Pairing took
place a few hours after emergence. It was noticed that the
males paired only once. Thus Lygocerus differs from its
Braconid host, in which a single male will fertilise two or three
females successively.

OVIPOSITION.

The female Ly gocerus, when about to oviposit, runs in
an agitated manner over the leaves infested with plant-lice.
Living aphides, whether parasitised or not, are ignored, and
I have never seen the Lygocerus make the mistake of
ovipositing on an A phidius which had not begun to spin silk.
The necessity is obvious, for until just before metamorphosis,
the host is still bathed in the juices of the aphis, in which the
egg of the hyperparasite could hardly develop. Sometimes
a pupa is chosen instead of a full-grown larva ; but these
are never attacked in the later stages when the chitin is
hardening.

When a suitable host is found, the Ly gocerus runs round
and over it with much excitement, tapping it repeatedly with
her antennae. The act of oviposition usually takes from
30-60 seconds. The Proctotrypid stands either on the thorax
of the aphis skin, facing the head, or on the leaf behind it with
the tip of her abdomen against its posterior part. Hither
way, the result is to bring the ovipositor, when exserted, into
the curve formed by the body of the Aphidius as it lies,
bent head to tail, in the cocoon. The ovipositor seems to
penetrate the aphis skin with little effort. Sometimes it is
partly withdrawn and inserted again, but only one egg is
deposited on the host. Occasionally two females may be seen
to oviposit simultaneously on the same Aphidius; and,
later, it is not uncommon to find two or three young larvae,
but only one of the latter reaches maturity, and two imagos
were never reared from the same cocoon.

106 MAUD D. HAVILAND

The number of eggs laid by a single Ly go cerus is uncertain,
but from observations made on females in captivity, and from
dissections of mature ovaries, it does not appear to be more
than fifteen or twenty, at most twenty-five. Calculation by the
latter method is difficult, as the eggs do not all mature at the
same time; and if the hosts be removed from the cage of a
captive female, and restored two or three days later, she will
recommence and complete oviposition.

THE Hee.

The egg of the hyperparasite, when newly laid, is elliptical,
and measures ‘25 x:10 mm. It is white and semi-translucent,
with a minute protuberance at one end. Under the bigh power

TEXtT-FIG. 2.

The egg immediately after oviposition. x 100.

of the microscope, the chorion shows numerous longitudinal
striae. ‘Treatment with Aman’s lacto-phenol and cotton-blue
reveals the presence of bodies resembling the symbiotes of the
‘ pseudo-vitellus ’ of aphides. The egg is laid upon the upper
surface of the host’s body, and hatches in abovt twenty hours.
As the development of the embryo proceeds, the egg becomes
more spherical, and the jaws, gut, &c., of the future larva are
visible through the chorion.

First Stace Larva.
Dimensions -45 x -22 mm.

The larva of the first instar is white and transparent, with
a distinct head and thirteen body segments. The form is
cylindrical, the greatest diameter being through the thorax, and
the segments diminish regularly to the last which bears the anus.
It removed to a slide, the larva can progress fairly actively by

.

BIONOMICS AND DEVELOPMENT OF LYGOCERUS sp. 107

a kind of peristaltic movement of the body, but under normal
conditions it probably does not need to move from where it
was hatched, provided that the host be a larva. If the latter
be a pupa, the hyperparasite is generally found feeding on the
posterior part of the abdomen, where the integument is still
soft. As the egg, as previously described, is always deposited
on the third or fourth segment of the Aphidius, the hyper-
parasite must needs seek the new situation for itself after
hatching.

TEXT-FIG. 3.

The larva, newly hatched, showing tracheal and nervous systems.
x 200.

The internal anatomy, with the exception of the tracheal
system, does not change essentially during development, so
that an account of it is left to the description of the fourth
instar. The mouth, which is very small and transversely oval,
is furnished with two slerder mandibles, set behind the hood-
like labrum, and the labium (fig. 5). The head is furnished
with two tactile papillae. he mid-gut, which at this stage,
as with the other parasitic Hymenoptera, does not communi-
cate with the proctodaeum, is large and globose, and its
contents tinge the otherwise transparent larva pale yellow.

The tracheal system consists of a pair of lateral trunks, united
by an anterior commissure passing above the oesophagus im

108 MAUD D. HAVILAND

front, and a posterior commissure passing beneath the gut, in
the eleventh segment, behind. Simple dorso-lateral, and
ventro-lateral, branches are given off in segments 1, 3-8. When
newly hatched there are only two pairs of open spiracles, the
first between the first and second segments, and the second on
the anterior part of the fourth, but the spiracles of the third
and fifth segments open shortly afterwards. (See ‘ Moults ’.)

Seurat (26, p. 100) states that the young larva of the Chaleid,
Torymus propinguis, has likewise four open spiracles,
but situated on the first, fourth, fifth, and sixth segments.

This stage lasts from twenty to twenty-four hours.

SECOND StTaGE LARVA.

Dimensions -70 x °35 mm.

The second stage larva differs from the first chiefly in the
tracheal system, and in the greater development of the anterior

TEXT-FIG. 4.

The larva of the second instar, showing tracheal system. x 200.

part of the body in proportion to the head, so that the latter
appears constricted off from the thorax, and the body resembles

BIONOMICS AND DEVELOPMENT OF LYGOCERUS sp. 109

a cone with the head projecting from the blunt end. The
tracheal system is more complex: the ramifications of its
branches are more numerous, and those of the second segment
appear at this stage. The stigmatic trunks of segments six,

TExt-Fic. 5.

Mouthparts of second stage larva. x 400. Ventral view. 0.sal.=
aperture of salivary duct. sk.=endoskeleton of head. lab.=
labium. m. lab. = muscles of labium. Jbr.=labrum. md. =man-
dibles. sal. d.=salivary duct.

seven, and eight are visible at the junction of the dorso-
lateral branches with the main stem of the tracheae, but
the corresponding spiracles are still closed. This stage lasts
about thirty-six hours, and during this time the host dies and
becomes black and shrunken. ‘The hyperparasite seems to
feed by suction, and the skin of the Aphidius, otherwise
uninjured, is gradually emptied of its contents. As the fluid
from the decomposing tissues passes into the mesenteron of
the Proctotrypid, the latter changes in colour from yellow to
brown.

THIRD STAGE LARVA.

Dimensions 1-00 x -75 mm.

TEXT-FIG. 6.

Larva of the third instar, showing tracheal system. x 49.

In the third stage the body becomes globose, owing to the
increased proportionate development of the first seven or eight
segments to accommodate the distended mesenteron. The result

110 MAUD D. HAVILAND

of this distension is to bend the head round ventrally to form an
acute angle with the long axis of the body. The papillae on the
head disappear. The branching of the tracheal system is more
elaborate, and the spiracles of segments six, seven, and eight
open in the order named, while the stigmatic trunk of the
second segment appears. This stage is longer than the two
preceding, and lasts about forty hours. The parasite is bathed
in the fluid that oozes from the decomposing body of the host.

FourtH STAGE LARVA.

The larva in the fourth instar differs considerably from that of
the preceding stages im size and form. Immediately after
ecdysis, the dimensions are not much greater than those of the
third instar, and the body is transparent; but as the larva
ingests the remainder of its host, it grows rapidly, and when
fully fed, measures 1:67 x -88 mm. At the same time it be-
comes creamy white and opaque.

The first four body segments are greatly developed. The small
head is bent completely round to the ventral side, and is almost
hidden by the large prothorax. The abdominal segments
diminish in diameter posteriorly, and the last bears dorsally a

conical caudal appendage. The function of this is unknown,

unless it is used as a lever by the larva which is able to turn
round freely in the cocoon. Seurat (26, p. 99) has described
a somewhat similar appendage in a Chaleid, Eneyrtus sp.,
and supposes that its purpose is locomotion (fig. 9, cd.).

Both the caudal appendage and body bear short chitmous
papillae or spines. The head is without larval antennae or palpi.
The mouth, which is very small and transversely oval,is bounded
anteriorly by a large horseshoe-shaped labrum, and posteriorly
by a smaller square labium. Between these, and deeply set
within the buccal cavity, are two stout little mandibles (fig. 8).
The salivary glands extend from the dorsal part of the fourth
segment forwards on either side of the mid-gut as two straight
tubes with a considerable lumen. ‘They are formed of poly-
hedral cells with large nuclei and granular cytoplasm, which
stains deeply with haematoxylin. Each gland runs obliquely

_————ia—LK =

BIONOMICS AND DEVELOPMENT OF LYGOCERUS sp. 111

TEXT-FIG. 7.

Larva of the fourth instar, showing tracheal system. x 49. 7. st.=
rudimentary stigmatic trunks of segments 9 and 10.

TEXT-FIG. 8.

The mandibles of the full-grown larva, x 400.

TEXT-FIG. 9.

Diagram of the general structure of the fourth stage larva, a.=anus.
ed.=cauda. g.=gonad. dc. o0.=imaginal disk of ovipositor.
1. m.=longitudinal muscles, M. t.=Malpighian tube. mes.=
mesenteron. m.—mouth. ». c.=nerve cord. oes,—oesophagus.
gl. s.=salivary gland. sp. gn.=supra-oesophageal ganglion.

112 MAUD D. HAVILAND

forwards and downwards, and between the first and second
segments enters a duct lined with epithelial cells, very similar
to those of the oesophagus (fig. 10). The two ducts unite
behind the head to form the common salivary duct, which
opens just inside, on the floor of the mouth. Under high power,
the ducts have the trachea-like structure found in most insects.
On either side of the salivary aperture is inserted a small muscle,
which runs outwards and backwards to the endoskeleton of the
head. When these contract, the labium, and consequently

TExtT-Fic. 10.

Longitudinal section through the salivary gland and duct of a larva
of the fourth instar. x 300,

the opening of the salivary duct, is slightly everted from the
mouth (fig. 5).

Two pairs of buccal muscles are connected with the labrum,
and by their contraction enlarge the buccal cavity. The
anterior, and more lateral, pair arise from the exoskeleton of
the front of the head, just above the labrum, on either side of the
median line, and running directly downwards (or, having regard
to the position of the head, backwards) are inserted on the roof
of the mouth. The posterior and median pair arise together
behind the last, and, running forwards obliquely between them,
are inserted on the distal half of the labrum (fig. 18). The

ae ee ee ee Be ee

BIONOMICS AND DEVELOPMENT OF LYGOCERUS sp. 113

short oesophagus opens into the mid-gut, which fills the greater
part of the body cavity, and is lined with glandular cells, rather
wider than deep, with well-marked nuclei. It contains a mass
of fluid food material, which is churned to and fro by incessant
muscular contractions of the body, but until just before meta-
morphosis there is no communication with the hind-gut. Two

Trext-ric. 11.

Longitudinal section through the Malpighian tube of a larva of
the fourth instar, showing lumen. x 300,

large Malpighian tubes extend from the fourth segment,
ventral to the salivary glands, and run back on either side of the
mesenteron. They are somewhat dilated at their anterior
extremities, and in sections show a considerable lumen, sur-
rounded by large flattened cells with great nuclei, resembling
those of the salivary glands (fig. 11). In the posterior half of
the tubes the lumen is very small and the cells are rounded.
The tubes open into the ampulla of the proctodacum, that is,
the cup-like anterior end of the hind-gut, which abuts on the
mid-gut in the eleventh segment (fig. 14).

The muscular system is well developed, especially the dorsal

NO, 257 I

114 MAUD D. HAVILAND

longitudinal, and lateral muscles of the posterior segments
(fig. 9).

The circulatory system calls for no particular comment.

In the tracheal system of the fourth instar larva there are still
seven pairs of open spiracles, for the eighth (mesothoracic) does
not become functional until metamorphosis. The first spiracle
is situated between the first and second segments, and the
second on the posterior side of the third segment, while the
remainder are on the five following segments. In addition,
two rudimentary stigmatic trunks can sometimes be seen on
the ninth and tenth segments, and the anterior one is occasion-
ally visible durmg the third instar. It appears that these
trunks are never functional, and they were not always apparent
in the larvae examined. Imms (11) has described vestigial
stigmatic trunks on the eleventh segment of the full-grown
larva of Aphycus melanostomatus, which has nine
pairs of functional spiracles. These do not appear in the
Lygocerus larva, in which the spiracles have evidently been
reduced in number from behind forwards. The aborted
trunks of segments nine and ten are probably vestiges mherited
from an ancestral form with ten open spiracles. The rest of the
tracheal system differs from that of the preceding stage only
in the greater calibre and more elaborate ramifications of the
tubes. It should, however, be remarked that there is no ana-
stomosis of the tracheal branches of the two sides of the body,
such as Seurat (26) describes in certain Ichneumonidae and
Braconidae (fig. 7).

The nervous system consists of two supra-oesophageal
sangha, united by a broad commissure, and connected with the
sub-oesophageal ganglion by two short, thick cireum-oesophageal
commissures. ‘The ventral nerve cord contains eleven ganglia.
The four anterior are well marked ; the five following are less
distinct, and appear as a wide, slightly-segmented band. The
cord terminates in a bulbous swelling, composed of two
ganglia, that of the eleventh segment being fused with that of
the tenth (fig. 12).

The genital organs lie above the mid-gut on either side as

eS

BIONOMICS AND DEVELOPMENT OF LYGOCERUS sp. 115

Trxt-Fig. 12.

Sp gr

Nervous system of larva (partly diagrammatic). sb.=sub-oeso-
phageal ganglion. sp.—supra-oesophageal ganglion. Th. 1-3 =
Thoracic ganglia. 4—11.=abdominal ganglia.

‘Rexcr=Kie. Vs:

Vertical section through the head of a larva of the fourth instar.
(The muscles of the labrum are shown somewhat diagrammatic-
ally.) x 200. cl.=cuticle. dc. op.=imaginal disk of eye. fb.=
fat body. hp.=hypoderm. J. m. /br.=lateral muscles of labrum,
m.m. lbr.=median muscles of labrum = md.=mandible. o. sal.=
aperture of salivary duct. sb. gn.=sub-oesophageal ganglion.
8p. gn. =Supra-oesophageal ganglion.

12

116 MAUD D. HAVILAND

Text-ria. 14.

Vertical section through the posterior region of the body of a larva
of the fourth instar. x 350. a.=anus. ct.=cuticle. de. st.=
imaginal disk of stylets. dc. s. v.=imaginal disk of sheath and
valves. fb.=fat body. gn. 10.=ganglion of segment 10. hp.=
hypoderm. m.ep.=wall of mesenteron. mes.=mesenteron.
pr.am.=ampulla of proctodaeum. pr.=proctodaeum,

two oval bodies, the testis bemg more elongated than the ovary
(fig. 9). The complete development of the accessory genital
apparatus was not observed, but in the fourth imstar the
female armature exists as two imaginal disks on the eleventh
and twelfth segments. In Lygocerus the relationship
of the parts is somewhat obscured, owing to the curvature of
the body and crowding together of the segments in the posterior
ventral region, but my observations on the origin of the

pr:

BIONOMICS AND DEVELOPMENT OF LYGOCERUS SP. 117

TExT-FIG. 15.

Q.
de.v.
SVs = y eS

VR eee NE de. st.

de. shz AY LOS Sa LS EN

2 \\

Wy) \

gl.p. A

= ——
= :

W ‘ gE
& Ss

Vertical section through the developing genital armature of a female
larva of the fourth instar. x 350. a.=anus. ct.=cuticle.
dc. sh.=imaginal disk of sheath, de. st.=imaginal disk of stylet.
de. v.=imaginal disk of valve. gl. p.=‘ poison gland’.

ovipositor, as far as they go, are substantially in agreement
with those of Seurat on Doryctes gallicus. The stylets
arise from the posterior ventral wall of the eleventh segment,
and the sheath and valves are derived from the reduplication
of the imaginal disks of the twelfth segment. A tubular
glandular structure is formed by constriction from the hypo-
dermal cells at the base of the latter. In its origin and position
it corresponds with that described by Seurat as ‘la glande a
venin’. Whether this organ is actually a poison gland in the
Ceraphronidae I am unable to say. Saunders, quoted by
Woodward (Ashmead, 1), records that he was stung by a female

118 MAUD D. HAVILAND

Trxt-Fia. 16.

PAN

DET

6
FENG

Ware ee

The same as fig. 15, more advanced. x 350. a.=anus. ct.=
cuticle. de. sh.=imaginal disk of sheath. de. st.=imaginal disk
of stylet. de v.=imaginal disk of valve. gl. p.=‘ poison gland ’.
hp.=hypoderm. ss. 11.=sternite of segment 11.

of Scleroderma linearis; and of other parasitic
Hymenoptera, the female Ichneumonid. Ophion, will
sometimes pierce with the ovipositor when handled. The pain
is more severe and persistent than a mere mechanical stab
would produce, so that presumably some secretion enters the
wound. Bordas and others have described structures in various
Terebrantia which appear to be homologous morphologically
with the poison glands of the Aculeata, but their function is

es ha

BIONOMICS AND DEVELOPMENT OF LYGOCERUS SP. 119

TExt-FiG. 17.

; hp.
The same as in figs. 15 and 16, shortly before metamorphosis.
x 350. a.=anus. ct.=cuticle. dc. sh.=imaginal disk of sheath.
dc. st.=imaginal disk of stylet. de.v.=imaginal disk of valve.
gl. p.=‘ poison gland’. hp.=hypoderm. s. 11.=sternite of seg-
ment 11.
still uncertain. The tubular gland, ‘ glande tubuleuse’, that
Seurat describes in Doryctes, I have not traced in Ly go-
cerus (figs. 14, 15, 16, 17).

Owing to lack of suitable material, the whole ontogeny of
the male genital armature was not followed, but it appears to
arise, as described by Seurat, from the imaginal disks of the
twelfth segment only. In the fourth instar two terminal
lateral processes appear at the end of the disks, and probably
represent the future stipites (forcipes) (fig. 18).

120 MAUD D. HAVILAND

The fourth instar lasts between two and three days. Ly go-
eerus does not spin silk, but pupates within the cocoon pre-
viously woven by the Aphidius. Just before metamorphosis,
the mid-gut opens into the hind-gut, and the contents are
voided. The larva is active, and by its movements the frass,
together with the now empty skin of the host, are welded into

TErxt-rie. 18.

Vertical section through the developing genital armature of a male
larva of the fourth instar. x 350. de. f.=imaginal disk of forcipes.
ct.=cuticle hp.=hypoderm.

a compact, moist pellet at the ventral side of the body. The
frass of the Proctotrypid, Chaleid, and Cynipid parasites of
Aphidius ean readily be distinguished from one another,
for that of Lygocerus is invariably a single black mass,
whereas that of the Chalecidae and Cynipidae consists of several
pieces of a different form and colour.

Moutts.

The determination given of the number of moults, and the
duration of the instars, is based on the examination of many in-
dividuals of different ages, and may be somewhat arbitrary ;
but it was the only practicable method to employ, since it

BIONOMICS AND DEVELOPMENT OF LYGOCERUS sp. 121

proved impossible to keep one larva alive for observation from
day to day. The reasons for determining the different instars
thus are as follows :

The newly hatched (first instar) larva of Lygocerus
possesses only two pairs of open spiracles, but examples twelve
hours old have four. At one time I believed that these forms were
separated by a moult, though it was never observed. On the
other hand, I noticed a larva twenty-four hours old which had
the cast skin attached to the hind part of the body. The
exuviae were too much torn to show the spiracles, but the larva
itself had four (fig. 4). For purposes of convenience, there-
fore, I have referred all stages up to that represented in that
figure to first mstar, and assumed that the spiracles of the third
and fifth segments opened as the stadium proceeded ; but it
may well be that there is a moult between the forms with two
and those with four spiracles. We should then have five
larval stages, separated by four moults.

Similarly, the actual ecdysis between instars two and three, as
here described, has never been observed, but the differences in
the external form and respiratory system seem sufficient to
place them in separate instars.

The difference in size and form between instars three and four
is so great that, if a large number of larvae had not been
examined, there would have been doubt in referring them to the
same species. The fourth instar, immediately after the moult, is
transparent, and half the size of that represented in fig. 7. But
the caudal appendage and tracheal system are unmistakable, so
that although the actual ecdysis has not been seen, this form has
been described as the fourth instar.

PUPATION AND EMERGENCE.

The period of pupation is from fourteen to sixteen days. I
disturbed, the pupa jerks its abdomen vigorously from side to
side. It is possible that this habit, which is marked im both
the larva and the pupa, and in which they differ from the
A phidius itself, and from its Chaleid and Cynipid parasites

122 MAUD D. HAVILAND

may in some degree protect them from ovipositions by the
females of their own and other families.

When ready to emerge, the imago gnaws a hole somewhere on
the dorsal side of the cocoon and creeps out. As Gatenby (9)
has remarked, this hole differs from that made by Aphidius
in having irregular edges, and is not necessarily placed in the
dorso-posterior region of the aphid’s skin.

The number of broods occurring in one year is not known, and
probably depends on the number of species of Aphidius
upon which the hyperparasites can live. Two broods were
reared from Aphidius ervi im 1919; but the host did
not appear in any numbers before July, and it is possible that
earlier broods may have occurred with a different host. All the
imagos of Lygocerus had emerged by the end of August,
and there is no evidence to show whether the species over-winters
as larva or pupa.

In captivity the imagos generally live five or six days, but
sometimes as long as ten. They were observed to feed on
sugar and water, on honey-dew from the aphides, and on sap
oozing from cut leaves, but they seemed to live as long, and to
remain as vigorous, when no food was supplied.

COMPARISON OF LARVAL CHARACTERS WITH THOSE
OF OTHER SuUB-FAMILIES.

The most complete comparative account of the larvae of
entomophagous Hymenoptera is that of Seuiat (26), who
studied certain Ichneumonidae, Braconidae, and Chalcidae.
Unfortunately he did not include the Proctotrypidae, and our
knowledge of the larval morphology of this family, as already
remarked, is very scanty. Seurat emphasized the importance
of the tracheal system in determining the larvae of the different
groups, but, as Lichtenstein and Picard have recently pointed
out (15), increased knowledge has somewhat modified this view.

Some authorities have considered that the Proctotrypoidea
are allied to the Chalcidoidea, but Ashmead (1) disputes this, and
thinks them in every respect more nearly related to the Hymen-

EE eee

Sey 6 ~<a +

ae ee ee ee

BIONOMICS AND DEVELOPMENT OF LYGOCERUS SP. 123

optera Aculeata, and among Terebrantia, to the parasitic
Cynipidae. The discussion of the affinities of the group is
outside the scope of this paper, but it should be pointed out
that the larval form of this particular genus of Ceraphroninae
differs from the Chaleid larvae described by Seurat (26), Imms
(11), Embleton (6), &c., in several respects. As regards the
tracheal system, the late opening of the spiracle of the
second segment is common to many larvae of the entomopha-
gous Hymenoptera. On the other hand, the larva of Ly go-
cerus is remarkable for the reduced number of abdominal
spiracles, and the rudimentary nature of the stigmatic trunks
of segments nine and ten, and differs from the Ichneumonidae
and Braconidae studied by Seurat in the absence of anastomosis
of the tracheal vessels of either side ; though as Lichtenstein
and Picard (15) have shown for the Braconid, Sycosoter
lavagnel, this is not an invariable character of the external
feeding Braconidae.

The reduction in the number of spiracles is earried still
further in Platygaster. Marchal (18) figures four spiracles
in Platygaster ornatus, the first between the first
and second segments, and those succeeding on the third, fourth,
and fifth. The spiracle of the fourth segment (the propodaeum
of the imago) differs from the others in its larger size, and the
greater proliferation of the hypoderm cells surrounding it.
“Tl est pareil a une sorte d’histoblaste aux dépens duquel devra
se former plus tard le grand stigmate du segment médiaire de
ladulte. Further, in Platygaster, the main tracheal
trunks are not joimed posteriorly by acommissure. In Ly go-
cerus a posterior commissure exists, and the spiracle of the
fourth segment is indistinguishable from the rest.

Likewise M‘Colloch (20) describes ‘ four or five pairs of well-
developed spiracles’ im the larva of the Scelionid, Eumi-
erosoma benefica; but Ganin (8) states that there are
nine spiracles in the third stage larva of the form of Platy-
gaster that he studied, and that spiracles are lacking only on
the first, second, and three last segments.

Kulagin (14) for Platygaster, and Ayers (2) for

124 MAUD D. HAVILAND

Teleas, do not describe the later stages of the larvae, and say
nothing about the tracheal system. Keilin and Thompson (12)
describe nine pairs of spiracles ina Dryinid larva, parasitic in
Typhlocyba (Homoptera). The relative positions are not
determined, but from the figure it seems as if the meso-, or
possibly the metathorax, bears no spiracles.

I can find no other account of the tracheal system of the
Proctotrypoidea, and until we have more knowledge of the
hymenopterous larvae which live upon their hosts as external
parasites, we cannot tell how far the characters observed
indicate true phylogenetic relationships, or are merely
secondary adaptations. Moreover, it is unwise to compare a
highly modified internal parasite, such as Platygaster,
with the more generalized external forms; though in this
connexion it may be significant that the third stage larvae
of Platygaster and Humicrosoma have a certain
resemblance to the early larva of Ly gocerus.

The differences are not confined to the tracheal system.
Marchal describes ten ganglia in the nerve cord, and three
Malpighian tubes, in Synopeas rhanis. Keilin and
Thompson observed thirteen ganglia, and no Malpighian tubes,
in the Dryiid that they studied. This diversity of structure
indicates either that little reliance can be placed on larval
characters, which are often adaptive, or that the Proctotry-
poidea as at present understood are, m some respects, an
arbitrary group.

Economic Stratus.

From an economic standpomt Lygocsrus must be
regarded as an injurious insect. Parasitisation by Braconidae
is an important natural check upon the increase of plant-lice ;
and this Proetotrypid, like the hyperparasitic Chalcidae and
Cynipidae, is an enemy of the beneficial Aphidius. Unless,
as seems improbable, it confines its attacks to a single species,
it must destroy considerable numbers of Aphidiidae.’ A p hi-

1 Kieffer records that L. testaceimanus has been reared from
arose aphis (? Macrosiphum rosae) (18, p. 51).

BIONOMICS AND DEVELOPMENT OF LYGOCERUS sp. 125

dius ervi, and the nearly related species A. avenae, are
parasites of such pests as Macrosiphum granarium,
the grain aphis, and according to Marvhall (19) are polyphagous,
preying indiscriminately on various species of aphides. If
their parasites follow them to other hosts, their efficiency
as controls of plant-lice must be seriously impaired. For
instance, two collections of A. ervi from M. urticae, made
from different places round Cambridge im August, gave the
following results :

Number Parasitised by Parasitised by Total % we ery i
examined. other families. Lygocerus. parasitised. Lv; teen ee
ygocerus.
I 50 12 16 56 32
It 38 6 L7 60 44

Other collections, of which exact records were not kept,
likewise showed a high percentage of hyperparasitisation by
these Proctotrypids.

Aphidius is at least twice as prolific as its parasite,
and each female destroyed by the latter means the loss of
thirty or forty ovipositions, which would lull, or at least impair
the fertility of, the same number of aphides. If this high
rate of hyperparasitisation should occur in a grain crop infested
by Macrosiphum granarium, attacked by Aphi-
dius, the efficiency of this natural control might be lowered
by 50 per cent.

SUMMARY.

1. Lygocerus testaceimanus, Kieff. is a hyper-
parasite of Aphis saliceti, Kalt., through the primary
parasite, Aphidius salicis, Hal.; and L. cameroni,
Kieff. is similarly a hyperparasite of Macrosiphum
urticae, Kalt., through the primary parasite, A phidius
ervi, Hal.

2. The Aphidius is attacked immediately before or
after metamorphosis, when lying within the empty skin of the
aphis within which it is reared.

3. The egg is laid, and post-embryonic development takes
place, outside the body of the host.

126 MAUD D. HAVILAND

4. The evidence points to the conclusion that there are four
larval instars and three moults.

5. The larvae differ in several particulars from those of the
families of Proctotrypoidea previously described, and there
is considerable difference in form between the early and later
instars. ;

6. During development, which lasts about six days, the
larva devours its host, and then pupates within the skin of the
aphis for a further period of two weeks.

7. Two, and possibly more, broods are reared in the season ;
and it is probable that the hyperparasite is a considerable check
on the Aphidius in its control of plant-lice infestation.

8. Ly gocerus, though occasionally attacked by its own
species, was never found to be parasitised by another hymen-
opteron. This immunity is probably due to the active move-
ments with which the larva and pupa in the cocoon respond
to external stimuli.

BIBLIOGRAPHY.

1. Ashmead, W. H. (1893).—‘* A Monograph of the North American
Proctotrypidae ”’, * Bull. U.S. Nat. Mus.’, vols. 44-6, pp. 1-472,
Pls. i-xviii.

2. Ayers, H. (1884).—‘ On the development of Oecanthus niveus
and its parasite Teleas”, ‘Mem. Bost. Soc. Nat. Hist.’, vol. iii,
no. 8, pp. 261-81, Pls. xxiii-xxv.

3, Berlese, A. (1909).—“ Gli Insetti ’’, Milan.

4. Buckton, G. (1879).—‘ A Monograph of British Aphides”, Ray
Society's Publications, vol. ii.

5. Devitz, H. (1874).—“ Ueber Bau und Entwickelung des Stachels und
der Legescheide einiger Hymenoptera ”’, ‘ Zeit. wiss. Zool.’, vol. xxv,
pp. 174-200, Taf. xii—xiii.

Embleton, Alice (1904).—‘‘ On the Anatomy and Development of
Comys infelix”, ‘Trans. Linn, Soc.’, vol. ix, pt. 5, pp. 231-
54, Pls, 11-12.

Forster, A. (1856).—‘* Hymenopterologische Studien ”, ii.

8. Ganin, M. (1869).—‘ Beitrige zur Kenntniss der Entwickelungs-

geschichte bei den Insecten ’’, ‘ Zeit. wiss. Zool.’, lxix, pp. 381-448,
Taf, xxx—xxxiii.
9. Gatenby, J. Bronté (1919).—*‘ Notes on the Bionomics, Embryology,

Sa

=

10.
i,

12.

13

* 14.

15.

16

17.

18.

19

20.

21.

22.

23.

24.

25.
26.

27

BIONOMICS AND DEVELOPMENT OF LYGOCERUS sP. 127

and Anatomy of certain Hymenoptera Parasitica’”’, ‘ Journ. Linn,
Soc.’, vol. xxxiii, pp. 387-416.

Henneguy, L. F. (1904).—‘*‘ Les Insectes ”’, Paris.

Imms, A. D. (1919).—*‘ Insect Parasites of some Coccidae ”’, * Quart.
Journ. Micros. Sci.’, vol. lxiii, pp. 293-374, figs. 1-34.

Keilin, D., and W. R. Thompson (1915).—‘*‘ Le cycle évolutif des
Dryinide’’, ‘C. R. Soc. Biol. Paris’, no. 78, pp. 83-7, Pls. x-xi.
Kieffer, J. J. (1907).—Proctotrypoidea : in André’s “Species des
Hyménoptéres d’Europe et d’Algérie’, Fasc. 97-100, pp. 1-288,

Pls. i-x.

Kulagin, N. (1898).—** Beitrage zur Kenntniss der Entwickelungs -
geschichte von Platygaster’”’, ‘Zeit. wiss. Zool.’, |xiii, 195-235,
Taf. x-xi.

Lichtenstein, J. L.; and F, Picard (1918).—‘‘ Etude morphologique
et biologiquedu Sycosoter lavagnei’’, “Travaux del Institut
de Zoologie de Montpellier’, mém. 29, 2° série, pp. 440-74,
i-xxxill.

Marchal, Paul (1897).—-‘‘ Les Cécidomyies des céréales et leurs para-
sites’, ‘ Ann. Soc. Entom. France’, vol. xvi, pp. 1-105, Pls. 1-8.

(1900).—‘‘ Notes biologiques sur les Chalcidiens et Proctotry-
pides ’, ‘ Ann. Soc. Entom. France’, vol. lxix, pp. 102-12.

—— (1906).—“‘Les Platygasters”, ‘Arch. Zool. Expér.’, 4°
série, t. iv, pp. 485-640, Pls. xvii-xxiv.

Marshall, T. A. (1899).—‘‘ British Braconidae”’ (Flexiliventres),
‘Trans. Ent. Soc. London’, pt. i, pp. 1-79, Pl. 1.

M‘Colloch, J. W. (1915).—* Further data on the Economy of the
Chinch Bug Egg Parasite, Eumicrosoma benefica, Gahan”,
* Journ. Econ. Entom. ’, vol. viii, pp. 248-60.

Packard, A. G. (1898).—* Text-book of Entomology ’, London.

Perkins, R. C. L. (1906).—** Leaf-hoppers and their Enemies ”’, ‘ Sugar
Plant. Ass. Hawaii, Entom. Bull. Honolulu Exp. Station’, no. 1.

Radoszdowski, O. (1884).—** Revision des armures copulatrices des
males du genre Bom bus’”’, ‘ Bull. Soc. Imp. Moscow’, t. Ix, pp. 51-
92, Pls. i-iv.

Saunders, E. (1884).—‘*‘ Further notes on the terminal segments of
Aculeate Hymenoptera ”’, ‘ Trans. Ent. Soc. London’, pp. 251-67,
Pl ica.

—— (1896).—‘ British Hymenoptera Aculeata ’, London.

Seurat, L. G. (1899).—‘“‘ Contributions 4 étude des Hyménoptéres
entomophages ”, “Ann. Sci. Nat.’, 8° série, t. 10, pp. 1-159,
Pls. i-v.

Sharp, D. (1899).—* Camb. Nat. Hist.’, “* Insects ”’, pt. 1, pp, 520 et seq.

On the Terrestrial Planarians from the Islands
of Mauritius and Rodrigues; with a Note
upon the Canal connecting the Female Genital
Organ with the Intestine.

By
Tokio Kaburaki,

Zoological Laboratory, The Museums, Cambridge.

With Plate 4 and 6 Text-figures.

Few terrestrial planarians are as yet known to occur in the
islands remote from continental land in the Indian Ocean.
Such are Pelmatoplana mahéensis von Graff and P.
braueri von Graff (12, 20) from the Seychelles, Placo -
cephalus isabellinus Geba (11) from Mauritius, Geo -
plana whartoni Gulliver (1) from Rodrigues, and
Rhynchodemus ceylonicus von Graff (17) from Male
Atoll.

The material serving as a basis of the present report was
collected by Mr. H. P. Thomasset in October, 1918, in the island
of Mauritius, and by Mr. H. J. Snell in November and December
of the same year in the island of Rodrigues. The specimens
were sent to Professor J. S. Gardiner, who kindly turned them
over to me for examination.

In this communication it may not be out of place to add
a brief account of G. whartoni Gulliver, known as occurring
in the Rodrigues Island. I am much indebted to the Director
of the British Museum of Natural History for the privilege of
studying this species.

The following is a list of the species dealt with in the present
paper :

Geoplana whartoni Gulliver ;
Placocephalus isabellinus Geba:
Rhynchodemus ceylonicus von Graff -
Amblyplana trifuscolineata, n.sp.

NO, 257 K

130 TOKIO KABURAKI

The planarian fauna of the islands mentioned above is
regarded by von Graff to be derived from that of the Aethiopian
region, whilst the species referrable to Rh. ceylonicus has
been clearly brought from Ceylon through the agency of man.

Before proceeding further, it gives me great pleasure to
express my deep indebtedness to Professor J. 5. Gardiner for his
suggestions and kind assistance throughcut this work in his
laboratory. I deem it my duty to mention my indebtedness to
Dr. Sir A. E. Shipley for his kind help in many respects. My
best thanks are also due to Dr. H. A. Baylis for providing
me with opportunities and accommodation for the examination
of the Museum material.

Geoplana whartoni Gulliver.
(Text-fig. 1.)

Geoplana whartoni Gulliver (1), pp. 561, 562, Pl. lv, fig. 1.—
von Graff (12), p. 347, Pl. iv, figs. 12-14, Pl. xxvi, fig. 4.

This species, according to Gulliver’s statement, occurs in
situations similar to those in which the nemertean, Tetra-
stemma rodericanus, lives, and, indeed, is often found
together with it. He collected some specimens on rotten wood.

Kxternal Characters.—tThe body is elongate, slender,
and for the most part nearly uniformly broad, though it tapers
off considerably in front. The sole on the mid-ventral surface
is slender, and corresponds to about one-ninth the width of the
body. Well-grown specimens in the preserved state measure
15-20 mm. long by about 2 mm. broad.

The ground colour of the dorsal surface is cream, with three
dark-brown stripes which run almost throughout the whole
length of the body, and anteriorly merge into the general colour
of the head-end, without revealing a dark tip. The ventral
surface is a somewhat paler shade of the same colour as the
dorsal, without any markings.

The numerous eye-spots are arranged in a single row round

the anterior tip, and continue sparsely for some distance down
the sides.

PLANARIANS FROM MAURITIUS AND RODRIGUES 131

The mouth-opening, which leads into the peripharyngeal
chamber, is placed somewhat behind the centre of the body, in
the mid-ventral line.

The common genital opening les nearer to the posterior end
of the body than to the mouth-opening.

Epidermis and Body-glands.—the following ac-
count is based on a single specimen received from the British
Museum. The epidermis consists of a layer of columnar cells,
which are about equally high on the dorsal and ventral surface,
and possess cilia, which, however, are confined to the latter sur-
face. It contains spindle-shaped rhabdites on the dorsal surface
only, where they are found in enormous quantities, evidently
situated between the epidermal cells. Immediately beneath the
superficial muscular system there occur such rhabdites as are still
contained in their mother-cells. These are scattered in sparse
numbers in the parenchyma. There are enormous quantities
of slime glands, deeply situated in the parenchyma, opening not
only to the exterior all over the surface of the sole, but in
a narrow zone of the ventral surface along and just within the
margin of the body.

Muscular System.—tThe musculature of the body pre-
sents no noteworthy features, consisting, as it does, of two
systems, superficial and deep, which are rather more strongly
developed on the ventral than on the dorsal side, doubtless in
relation to the movements. Dorso-ventral fibres occur also in
the usual manner.

Digestive System.—The mouth-opening is situated
somewhat behind the middle of the body and at nearly the
centre of the peripharyngeal cavity, with the pharynx horizon-
tally disposed. The pharynx is a cylindrical tube, terminating
conically at the free end. Embedded in the parenchyma
in front of the pharynx-insertion are numerous salivary
slands, which continue their way to the free end of the
pharynx.

All the three main trunks of the intestine give off numerous
lateral branches, which are sometimes bifurcated and sometimes
‘multifureated ’. The epithelium consists, as usual, of a single

K 2

132 TOKIO KABURAKI

layer of high cylindrical cells. So far as I have observed, special
glands are altogether absent in the lining epithelium.

Nervous System.—the exact arrangement of the nervous
system could not be ascertaimed, but it seemed to be quite
similar to that previously observed in several forms of this genus.
Each half of the bilobed braim-mass is continuous posteriorly
with one of the longitudinal nerve cords, which proceed, running
nearly parallel to each other, to the hind end of the body, and
are connected together by transverse commissures. Lateral
nerves are given off from the cords towards the nerve plexus,
which lies directly beneath the superficial muscular system.

The eye consists simply of a small pigment cup, partly filled
with a peculiar cellular substance, whose true nature could not
be ascertained from any of the sections available.

Reproductive Organs.—tThe genital organs are in
accordance with those described by von Graff. The common
genital opening leads directly into the penis-sheath, which
receives from behind the openings of the seminal receptacle
(uterus) and the glandular canal. The cavity is lined with a
single epithelium resting upon a fine basement membrane,
beneath which are found circular and longitudinal muscular
layers.

Male Organs.—The numerous testes occur close together
in the ventral parts of the body, arranged in two longitudinal
lateral zones which extend from behind the ovaries to nearly the
region of the copulatory organs. Each testis is, as usual, made
up of sperm-mother-cells and spermatozoa in all stages of
development, surrounded by the tunica propria. Probably
they are all connected by testicular ductules, but these could
not be definitely made out. Not far in front of the penis the
vasa deferentia rise obliquely upwards to enter the penis-bulb
separately at the upper lateral sides and finally open into the
lumen of the penis or the seminal vesicle. The vas deferens,
which is filled with spermatozoa, has a wall consisting of an
epithelium and an outer layer of circular muscular fibres.

In the penis there can be distinguished the conical intromit-
tent part, lying nearly horizontally in the penis-sheath, and the

aed

PLANARIANS FROM MAURITIUS AND RODRIGUES loo

bulbous part of muscular nature, which contains a cavity of
somewhat irregular contour, the seminal vesicle. The vesicle
gives rise to the moderately wide ejaculatory duct which opens
at the tip of the penis. The muscular fibres of which the penis
is composed are arranged in two principal sets, circular and
longitudinal, the fibres of the two sets occurring intermingled
with one another. Embedded in the parenchyma of the penis
are numerous glands, the ducts of which open into its lumen
lined by a layer of small columnar cells.

TExt-Fic. I.

Diagrammatic representation figure of the sexual organs of
Geoplana whartoni Gulliver. ed.=ejaculatory duct. gce.=
glandular canal. go.=genital opening. od.=oviduct. ps.=
penis-sheath. sr.=seminal receptacle. sv.=seminal vesicle. vd.=
vas deferens.

Female Organs.—The paired ovary occupies a ventral
position somewhat behind the brain. It is a nearly oval body
made up of egg-cells in several stages of development. From
the lateral aspect of the ovary the oviduct starts as an ampul-
laceous passage, which soon takes the character of a narrow
duct and proceeds backwards just outside the longitudinal
nerve cords, receiving the vitellme glands at numerous points.
The vitelline glands axe represented by branching cellular masses,
which are extensively distributed in the interstices between the
cut diverticulae. The mode of the connexion of the glands with
the oviduct is effected by means of the short branches of the

134 TOKIO KABURAKI

latter. Far behind the genital opening the oviduct rises obliquely
upwards, to unite with its fellow of the opposite side into
a single common duct, the glandular canal, which opens into
the penis-sheath from behind, after receiving numerous glands.
The duct exhibits a distinct lumen throughout the entire
length. Its direct wall is lined by a ciliated epithelium, outside
which is a layer of circular muscular fibres.

At a short distance below the opening of the glandular canal
the penis-sheath gives rise to a narrow passage, which pursues
a somewhat tortuous course obliquely backwards and upwards,
becoming gradually wider at the same time. Beyond the junc-
tion point of the oviducts it extends further backwards. This
organ, which doubtless represents the seminal receptacle, has
a wall consisting of a non-ciliated epithelium and a fine muscular
coating; in the cavity are found enormous quantities of
spermatozoa.

Placocephalus isabellinus Geba.
(Pl. 4, figs. 1, 2.—Text-fig. 2.)
Placocephalus isabellinus Geba (11), pp. 385, 386.

Three specimens of the species, which I identify with
Placocephalus isabellinus described by Geba from
the Mauritius Island, were collected by Mr. Thomasset under
half-rotten logs and rocks in damp places in the same island.

The head in the preserved state is of a semilunar shape and
not wider than the trunk, from which it is distinctly marked
off by a constriction. The trunk is dorso-ventrally depressed,
elongate, and nearly uniformly broad for the most part of its
length, though it tapers in the hind parts down to the bluntly
pointed end. The sole, scarcely raised above the general level,
extends from the neck to the posterior extremity, its width
being about a quarter that of the body. The large specimen
was 120 mm. long by 4 mm. broad, while the small was 50 mm.
long by 3 mm. broad.

As mentioned by Geba, the ground colour of the dorsal surface
is an umber brown with five longitudinal black stripes, a median

PLANARIANS FROM MAURITIUS AND RODRIGUES 135

and two pairs of laterals. The median stripe is very fine,
extending from the neck to the posterior extremity, and widen-
ing slightly above the pharyngeal region. ‘The inner pair are
much the strongest of all, and the outer pair at the edge of the
body become indistinct as they approach the hind end; on
either side both coalesce at the neck into a black patch. The
head is marked with a crescentic black pattern. Ventrally, the
worm is similar in coloration to the dorsal surface, with a
darker shade at the outer edge and also next to the surface of
the sole ; this latter is very pale.

TEXxT-FIG. 2.

Eyes of Placocephalus isabellinus Geba.

The numerous eye-spots are distributed all round the head,
and are continued sparsely for a considerable distance along the
sides of the body. At the sides of the neck they extend some-
what to the ventral surface and form a patch, as seen in Text-
fig. 2.

The mouth-opening, which leads into the peripharyngeal
chamber, is placed at some distance in front of the centre of the
body. In the specimens examined the pharynx was protruded
through the mouth-opening as a creamy frill.

The genital organs were unfortunately yet undeveloped in
the individuals examined. Like some other forms, this species
may to some extent reproduce asexually by transverse fission,
as stated by von Graff. On two occasions the severed hind end
presented a concave edge, apparently forming the new tail-end.

136 TOKIO KABURAKI

Rhynchodemus ceylonicus von Graff.

(Pl. 4, figs. 3, 6-8.—Text-figs. 3, 4.)
Rhynchodemus ceylonicus von Graff (12), pp. 499, Pl. xv,
figs. 35-38.—Laidlaw (17), p. 579.

The material was collected by Mr. Snell in the island of
Rodrigues. At a glance it appeared to be identical with
Geoplana whartoni described above, as dealt with by the
collector, but a closer examination has revealed the fact that
this is not so. After some hesitation I have referred it to von
Graff's Rh. ceylonicus, which has‘not been adequately
deseribed, as Laidlaw referred a worm from Male Atoll to this
species, but with some doubt.

This species appears to be fairly common in this island, as it
has been procured in enormous quantities at Grande Montagne
and also at Mount Malartic. According to Mr. Snell’s statement,
it is found under decaying logs, on the bark, under the bark, or
in the wood; the nemertean appeared to exist in far greater
quantities than the terrestrial planarians, but these often live
together in the same place.

External Characters (PI. 4, fig. 3)—_The body in the
preserved state is nearly oval in transverse section, elongate,
slender, and for the greater part of a uniform width, though it
gradually tapers off towards the anterior and posterior ends,
which are bluntly pointed. The ventral surface is made up of
the median somewhat raised sole, on which the animal creeps.
It extends over almost the whole length of the body and is
rather less than one-fourth the width of the body. This species
is wholly devoid of any trace of a sensory pit at the anterior tip.
In length the animals range from 22mm. to 45mm.; the
difference in length depending upon the state of contraction.
The 45 mm. specimen was not less than 3 mm. across.

Von Graff is speaking of the coloration of the body as a whole
when he states in his deseription : ‘ Die Grundfarbe ist lebhaft
gelb (sulphureo-citrinus) und der Ricken mit drei sehr kraftigen
schwarzbraunen Streifen versehen, von welchen aber die beiden

PLANARIANS FROM MAURITIUS AND RODRIGUES 137

lateralen mehr als doppelt so breit sind als der mediane. Hinten
convergiren die femer werdenden Liingsstreifen, ohne aber
zusammenzufliessen, vorne verschwimmen sie inder graubraunen
Pigmentirung des nur an der dussersten Spitze farblosen
Vorderendes. Eine gleiche Triibung findet sich auch auf der
Bauchseite des Vorderkérpers. Sie verschwindet erst gegen die
Mitte der Korperlinge und erstreckt sich vom Aussenrande der
Seitenstreifen des Riickens bis an die Kriechleiste, in deren
Umegebung sie am dunkelsten wird.’

In the specimen I have examined, the dorsal surface is of a
uniform orange colour with a slight touch of grey and marked
with three fine black longitudinal stripes, comprised of one

TExtT-Fic. 3.

ana

f \

fawn ib

NZ

Eyes of Rhynechodemus ceylonicus von Graff.
median and two lateral, these latter converging towards the
extremities of the body and meeting the median one. At the
anterior end the lines thicken and then coalesce, revealing a dark
tip unlike von Graff’s form, in which the anterior tip is light. In
most instances the lateral mes are much thicker than the
median. Sometimes the former get slightly lighter and are less
strongly marked than the latter. The ventral surface is much
paler than the dorsal, except on the sole, where the colour is
nearly white.

The eyes, which are only two in number, occur on either side
near the anterior tip of the body.

The mouth-opening which leads into the peripharyngeal
chamber lies nearly in the middle of the body, differing from
von Graff’s form, in which it is situated at the commencement
of the posterior fifth of the body. The pharynx in the normal
condition is usually completely retracted and hidden within the

138 TOKIO KABURAKI

peripharyngeal chamber. In some preserved specimens, it was
protruded through the mouth-opening as a cylindrical organ of
a creamy or white colour.

The common genital aperture is situated about half-way be-
tween the mouth-opening and the posterior extremity of the
body.

EK pidermis.—tThe specimens had not been preserved in
a condition satisfactory for the purpose of minute examination.
The epidermis is not of the same thickness all over the body,
bemg thickest on the dorsal surface, gradually becoming thinner
as it passes round to the mid-ventral surface. The cilia, though
stated by some investigators to exist over the entire surface of
the body, in this species are present on the surface of the sole
only. Dorsally and laterally the epidermis, as is well known, 1s
made up of closely packed, elongated, columnar cells resting
upon a basement membrane, each with an oval nucleus at its
base. Apparently wedged in between these cells, except those
that are on the head-surface, are found spindle-shaped bodies,
the rhabdites, which originate from their mother-cells, scattered
in fair abundance in the parenchyma beneath the dermal
musculature. In some cases the rhabdites are seen to be im
connexion with their mother-cells. Also there are some unicellu-
lar glands which open to the exterior here and there. Between
the epidermal cells are found some * gland cells ’ with granular
contents. These, though having been regarded by Dendy (9)
as masses of hardened mucus originating from the rhabdite-
forming cells, appear to me to be masses of mucus derived from
the glandular cells. Except on the surface of the sole the
epidermis on the ventral surface is constructed in the same
manner as that on the dorsal. Embedded in the parenchyma
are unicellular glands, which are much more abundant on the
ventral than on the dorsal surface, and these make their way
to the surface generally, instead of opening on the ventral
surface, more especially submarginally, as they do in some
other terrestrial forms as well as in all the freshwater and
marine Triclads. The epidermis on the surface of the sole, as
has been already indicated, is composed of closely packed,

PLANARIANS FROM MAURITIUS AND RODRIGUES 139

short, columnar cells, each bearing a large number of short cilia
onits outer surface. Inno cases have I been able to demonstrate
rod-like bodies, wedged in between the cells. Deeply situated
in the parenchyma there are enormous quantities of slime glands,
which open to the exterior all over the surface of the sole.

Basement Membrane.—tThe basement membrane,
which is in connexion with the epidermis, is distinetly visible as
a very thin, structureless, homogeneous layer. It is perforated
at various points by the passages of the rhabdite-forming cells
and the glands which lhe deep down in the parenchyma.

Muscular System.—tThe musculature of the body, as is
well known, is differentiated into two systems, superficial and
deep.

The superficial muscular system consists, as usual, of circular,
transverse, and longitudinal fibres. Immediately beneath the
basement membrane is a thin muscular layer made up of closely
apposed circular fibres. The transverse fibres, crossing those
of the other set obliquely, are just inside the circular layer. The
longitudinal fibres form a thick layer, the external longitudinal
layer, which is more strongly developed on the ventral surface
than on the dorsal. The muscles appear separated into a series
of bands, each made up of a few fibres. Through the intervals
between the bands the rhabdites and the glands make their way
to the surface.

The deep muscular system, separated from the superficial by
a zone of tissue, forms a layer thicker than the latter, and
consists principally of two distinct sets of fibres, longitudinal
and circular, which occur intermingled in the same mass,
without bemg arranged in definite layers. The longitudinal
fibres are more strongly developed than the circular. In addition
to these dorso-ventral muscles are found, which run between
the branches of the intestine.

Parenchy ma .—tThe tissue filling all the interspaces be-
tween the various organs and structures assumes, as usual, the
appearance of an irregular network, in the ground substance of
which is found a number of nucleated cells of a more or less stel-
late shape. Embedded in the superficial parts of the dorsal

140 TOKIO KABURAKI

parenchyma are the fme pigment granules in enormous quanti-
ties, which are of an irregular outline and of a dirty olive-lke
colour. The pigments, though rather few, occur on the ventral
side also.

Body-glands.—situated in the intervening zone between
the superficial and deep muscular systems are two distinct
kinds of glands, the mother-cells of the rhabdites and the
unicellular glands, as already mentioned. On some occasions
the mother-cells of the rhabdites have a very stout, horny-
looking cell-wall with a greatly elongated narrow tube tapering
off into a long process, each of which makes its way between
the epidermal cells at various points. Due to the action of
reagents, the cells vary in appearance. In some cases there
oceur such rhabdites as are still contained in the mother-cells.

The rhabdites vary in form and appearance. Some present
a slender spindle-like shape, while others are nearly oval in
shape. In no eases have I been able to demonstrate the vermi-
form bodies which were described by Dendy and others. Some-
times the rhabdites appear almost homogeneous, and sometimes
finely granular, but I have no doubt that they are all one and
the same thing. In some sections the dorsal surface of the
worm, outside the epidermic cells, is seen to be partly covered
with a layer of hardened mucus which reveals a character quite
similar to the rhabdites. They may possibly, by making the
animal extremely unpalatable, serve as a protection for its own
body, and also help to hold its prey more securely.

Scattered in sparse numbers in the parenchyma are unicellular
slands, which have the finely granular contents and open to the
exterior at various points of the body-surface, as mentioned
above.

Besides those glands there are slime glands which occur
deeply embedded in the parenchyma along the median plane of
the body and open out on the surface of the sole. They occur
in enormous quantities, and are distinguished from the glands
opening out over the whole surface of the body by a closer
affinity for borax carmine. In the terrestrial planarians the
movements are effected by the action of cilia in mucus which is

PLANARIANS FROM MAURITIUS AND RODRIGUES 141

constantly being secreted in greater or less quantities, and gives
rise to a thin layer between the ventral surface of the body and
the substratum. In this case rhythmical wavy motions of the
muscles stand, of course, in intimate relation to the movements.

Digestive System.—The mouth-opening, which lies
nearly in the centre of the body, leads, as usual, into the wide
peripharyngeal cavity with the pharynx horizontally disposed.
The cavity is limed with a single layer of epithelial cells made
up of pear-shaped cells of a glandular nature, as has been stated
by Dendy in G.spenceri. The epithelium rests upon a fine
basement membrane, beneath which are two layers of circular
and longitudinal muscular fibres. Situated in the parenchyma
around the cavity are unicellular glands which open into the
cavity

The pharynx is a short, tubular body of a cylindrical shape,
which arises from the dorso-anterior wall of the peripharyngeal
cavity, with its free end posteriorly directed. The outermost
layer of the wall is represented by a very thin, richly ciliated
epithelium, immediately beneath which come, as usual, two
thin layers of external longitudinal and internal circular
muscles. The circular layer is followed, after an interval in
which glandular and nervous tissues exist, by a very thick layer
of longitudinal muscular fibres. Just external to this layer
comes a layer of circular fibres, immediately surrounding the
lumen of the pharynx, which is lined by a single layer of non-
ciated cells. Besides the muscles mentioned above, there are
found a number of radial fibres, running from the inner circular
layer towards the outside.

The lumen of the pharynx leads anteriorly into the intestinal
canal, which is of the triclad type. The anterior trunk extends
to a point above the brain and usually gives off on each side
numerous lateral branches, which are sometimes bifurcated and
sometimes trifurcated. The posterior trunks proceed back-
wards nearly to the hind end of the body, one on each side of
the middle line, and are provided with numerous outwardly
directed, subdivided branches. The wall of the intestine is
a single epithelium made up of high cylindrical cells, which are

142 TOKIO KABURAKI

placed very closely together and rest on the surrounding tissue.
The cells, each with an oval nucleus in its basal portion, contain
a great number of coarse, highly refractive granules in the finely
granular protoplasm. In some cases the cells were observed to
be vacuolated in the distal portion of the cell. So far as I have
observed, any special glandular cells are altogether absent in
the epithehum.

Nervous System.—the brainis a bilobed organ, situated
at the anterior end of the body between the ventral wall and the
anterior termination of the intestinal canal. From the brain-
mass arise numerous nerves which are distributed over the
various parts of the anterior end of the body. But their
arrangements were not clearly made out. Each half of the
organ is formed of a very finely granular ground-substance, in
which small nerve cells occur much more abundantly towards
the periphery than in the central part. At various points the
mass is perforated by fine muscular fibres in the dorso-ventral
direction.

Each half of the brain-mass is continuous posteriorly with
one of the longitudinal nerve cords, which proceed straight
backwards, until finally they join together at the posterior end
of the body. The cords themselves are very thick and usually
present, in cross-section, the characteristic spongy or finely
reticulate appearance. Small nerve cells are scattered in sparse
numbers in the substance of the cords. Throughout their
entire course the longitudinal nerve cords are connected by
very numerous transverse commissures. Laterally they give
off numerous branches towards the nerve plexus, which lies
beneath the outer longitudinal muscles of the body and extends
completely round the body. The plexus consists of a close
network of fine fibres.

Eyes (PI. 4, fig. 6).—The only special sense-organs which
I have seen in the present species are the eyes. Each consists,
as usual, of a pigment cup and of numerous visual rods. The
pigment cup is of a bell-lke shape with its opening directed
outwards and upwards, and is as usual formed of very minute,
closely packed, spherical granules, of a dark-brown colour.

PLANARIANS FROM MAURITIUS AND RODRIGUES 1438

Enclosed in the cup is a mass of visual rods, the outer extremity
of which projects for a short distance beyond the margin of the
pigment cup. Between the pigment cup and the mass of
the visual rods, and also just in front of the outer surface of the
rods, small spaces are visible, doubtless caused by shrinkage of
the tissues. Hach rod is an elongated, faintly staming, very
finely granular body, which at the periphery shows a closer
affinity for borax carmine than in the central part. In front of
the opening is a collection of nervous matter, viz. granular
substance and fibres surrounded by numerous cells, apparently
belonging to nerve cells. The fibres pass over into the cavity of
the pigment cup, but how the nerves stand in connexion with the
visual rods I was unable to determine.

Reproductive Organs (PI. 4, figs. 7, 8).—The common
genital aperture, lying nearly mid-way between the mouth-
opening and the posterior extremity of the body, leads into
the wide, annularly outbulged vestibulum, which receives the
opening of the penis-sheath from above. Both the vestibulum
and the penis-sheath are lined with a single epithelium resting
upon a fine basement membrane, beneath which are found
circular and longitudinal muscular layers. Especially around
the penis-sheath the muscular layer presents a thick, compact
mass, which chiefly consists of circular fibres and is continuous
with that of the penis. In the diaphragmatic part between
both the cavities just mentioned the radial muscular fibres are
present in a strongly developed condition.

Male Organs.—Numerous follicular testes are placed close
together in the ventral parts of the body, arranged in a single
row on either side of the anterior main gut trunk, just on the
dorso-lateral side of the longitudinal nerve cord. The row
begins on each side slightly behind the ovary, and extends
backwards nearly to the insertion of the pharynx. Each testis,
of an oval shape, is made up of sperm-mother-cells and sperma-
tozoa in all stages of development, surrounded by the tunica
propria. In contact with the epithelium accumulations of the
mother-cells occur, which contain very large, deeply staining,
highly granular nuclei. In the cavity of the testis, and separated

144 TOKIO KABURAKI

from the accumulations of the mother-cells le some compact
masses of metamorphosing spermatozoa. The spermatoblast
in a further stage of development presents an elongated, pear-
shaped protoplasmic body, in the broad end of which the
nucleus is visible as a distinct, deeply staining spot. It is then
changed into a spermatozoon, the nucleus forming the head and

TEXT-FIG. 4,

od gY Gg? vd

Diagrammatic representation of the genital organs of Rh. ceylonicus
von Graff. gv.=genital vestibulum. Other letters as in Text-fig. 1.

the protoplasm having greatly stretched out and elongated
itself into a thin thread to form the tail of the spermatozoon.
Each testis gives rise, on its lower side, to a short canal which
communicates soon with the vas deferens. The vasa deferentia,
proceeding backwards close along the dorsal sides of the longi-
tudinal nerve cords, rise obliquely upwards to enter, each
separately, the bulbous part of the penis at the upper lateral
sides, and finally open into the lumen of the penis or the seminal
vesicle, The vas deferens, which is filled with spermatozoa, is

PLANARIANS FROM MAURITIUS AND RODRIGUES 145

led by a thin, flattened epithelium of nucleated cells resting
upon a basement membrane.

The penis consists of two parts, viz. the free, conical intro-
mittent part lying subvertically in the penis-sheath, and the
bulbous basal part of muscular nature. Enclosed in the latter
part is a wide cavity of somewhat irregular contour, the seminal
vesicle, into the anterior extremity of which open the vasa
deferentia ; posteriorly this is continuous with the ejaculatory
duct which opens into the penis-sheath at the tip of the penis.
The cavity is lined by a layer of columnar glandular cells,
beneath which is a circular muscular layer. Embedded in the
parenchyma of the penis are numerous glands which open into
the seminal vesicle and the ejaculatory duct. Externally the
penis is covered with a thin epithelium which becomes thicker
towards the proximal portion, and at the same time is provided
with cilia. The epithelium surrounds a muscular layer
consisting of external, thick, circular, and internal, thin,
longitudinal fibres. On some occasions the penis at the proximal
parts gives rise to special processes which are covered with an
epithelium made up of ciliated, columnar cells.

Female Organs.—the paired ovary is situated far behind
the brain, one on either side close to the dorso-lateral side of
the longitudinal nerve cord. Each ovary is nearly oval in
shape, and its cavity is lined with a thin epithelium, composed
apparently of a single layer of flattened cells. In the interior of
the ovary, ova in various stages of development are met with.
Occupying the periphery of the ovary occur numerous young
ova, each with an oval, large, and highly granular nucleus. In
the successive stages of development the ovum assumes a nearly
spindle-like shape, as has been mentioned by Dendy. The large
nucleus sometimes shows a very distinct chromatin network.
Situated in the central and lower regions of the ovary are the
ripe ova, which present a round shape and enclose a very large
nucleus, revealing a transparent, vesicular aspect.

The vitelline glands are represented by irregularly ramified
masses of cells, which are extensively distributed in the inter-
stices between the diverticulae of the intestinal trunk and stand

NO, 257 L

146 TOKIO KABURAKI

at many points in connexion with the oviduct. The vitelline
zlands consist of large round cells closely packed, each of which
contains a highly granular nucleus and highly refractive proto-
plasmic bodies. Probably, at the time when the ova are
passing down, the cells break down and make their way into the
oviduct. They are considered to take part in connexion with the
nutrition of the ova and also with the formation of the cocoon
capsule.

The oviduct arises from the mid-ventral aspect of the ovary
as a wide passage ; this soon assumes the character of a narrow
canal, which proceeds straight backwards, just along the out-
side of the nerve cord. In the region of the genital opening
the oviduct nears the median line, rismg upwards at the same
time, and finally unites with its fellow of the opposite side, at
a point behind the penis, to form the rather wide glandular
canal. The oviduct shows a distinct lumen along its entire length.
Its actual wall is made up of a layer of distinctly nucleated
columnar cells, with well-developed cilia projecting into the
lumen of the oviduct. Immediately external to the layer
mentioned comes a layer of circular muscular fibres.

As already indicated, the oviduct receives the vitelline glands
at several points of its course. The mode of connexion seems
nearly similar to that described by Moseley (22), Dendy, von
Graff, and others, im several forms. The glands stand in com-
munication with the oviducts by means of the short branches
of the latter, which are situated at tolerably regular intervals.

The glandular canal, mentioned above, runs anteriorly and
obliquely downwards to open from behind into the atrial
passage, between the penis-sheath and the vestibulum. The
canal is constructed in the same manner as the oviduct, and is
lined with an epithelium made up of ciliated columnar cells
resting upon a fine basement membrane, beneath which exists
a muscular layer composed of circular and longitudinal fibres.
Numerous glands are found all round the canal, into which they
Open.

‘The present species is wholly devoid of any trace of the organ
representing the seminal receptacle. As already indicated, the

rgd

PLANARIANS FROM MAURITIUS AND RODRIGUES 147

vestibulum is supphed with an annular outbulging, which
extends more deeply backwards than forwards. ‘To me, this
outbulging appears to serve as a seminal receptacle during
copulation.

Amblyplana trifuscolineata, n. sp.

(Pl. 4, figs. 4, 5.—Text-figs. 5, 6.)

This new species is represented by a single specimen which
was taken by Mr. Thomasset under a half-rotten log in the
island of Mauritius.

Hxternal Characters (PI. 4, figs. 4, 5)—The body,
which is nearly circular in cross-section, 1s rounded at the
posterior end, and has the lateral margins even and nearly
parallel for a large part of its length, but tapering in front to
the bluntly pointed extremity. The sole corresponds nearly to
one-third the width of the body, extending to both extremities.
It measures 25 mm. long by about 8 mm. across in the broadest
part.

Tn coloration this species nearly resembles Geba’s Amb ly -
plana tristriata, described by that author from the
Comoro Island. The dorsal surface is of a dark colour with a
touch of olive-like brown, and marked with three longitudinal
black stripes, a median and a pair of laterals, the latter con-
verging towards the extremities of the body, without coalescing.
Ventrally, the colour is similar to that of the dorsal side, except
for the creepmg surface which is pale, while each side of it has
a diffused brownish black tinge.

Near the anterior tip of the body he the eyes, one on each
side, as shown in 'l'ext-fig. 5.

The mouth-opening, which leads into the peripharyngeal
chamber, is situated at a short distance behind the centre of the
body. I could make out its position by a slight protrusion of
the pharynx.

The common genital opening les at the hind end of the first
third of the distance from the mouth-opening to the posterior
extremity of the body.

L 2

148 TOKIO KABURAKI

Epidermis and Body-glands.—tThe epidermis con-
sists, as usual, of a layer of columnar cells, which are of a greater
height on the dorsal than on the ventral side. Wedged in
between these cells, except on the ventral surface, are spindle-
like rhabdites which appear almost homogeneous. In some
sections they are seen to be discharged on to the exterior,
revealing a layer of hardened mucus over the epidermis. The
rhabdites enclosed in the subcutaneous cells occur widely

TEXT-FIG. 5.

4 r

Eyes of Amblyplana trifuscolineata, n. sp.

distributed on the dorsal side of the body. In addition to the
glands deeply situated in the middle of the body and opening
to the exterior on the surface of the sole, there are some glands
which open in scattered distribution all over the ventral surface.

Muscular System.—Immediately beneath the fairly
well-developed basement membrane is the superficial muscular
system composed of the outer circular and the mner longitudinal
layers. The deep muscular system, which chiefly consists of
longitudinal fibres, is well developed all round in the paren-
chyma as a thick and continuous sheet surrounding the intestine
and the nerve cords.

Digestive System.—The mouth-opening is placed at
about the centre of the peripharyngeal chamber, in which is
disposed the pharynx of a cylindrical shape. It is conically
pointed at the free end. The gut trunks are provided with
numerous subdivided branches, the epithelium of which presents
no noteworthy features, consisting, ‘as it does, of high columnar
cells.

Reproductive Organs.—tThe genital apparatus is
nearly sunilar in appearance to that of Am. tristriata Geba.
The genital opening leads into the vestibulum, which forms an

PLANARIANS FROM MAURITIUS AND RODRIGUES 149

oblique upwardly directed, annular outbulging, and which
receives the penis-sheath from above. The vestibulum has
a wall consisting of a single epithelium and a muscular layer,
while the penis-sheath is lned with a ciliated epithelium,
outside which is a thick muscular coating, chiefly composed of
circular fibres.

Male Organs.—tThe numerous testes, containing sperma-
tozoa in several stages of development, are arranged in a row on
each side of the body close to the upper side of the longitudinal
nerve cords, extending from behind the ovary to the insertion
of the pharynx. The vasa deferentia run backwards, just along
the inside of the nerve cords. Shortly in front of the penis they
gradually bend inwards and upwards, finally to open as a rule
separately into a moderately wide seminal vesicle. The vas
deferens shows a definite wall consisting of a thin epithelium
and a feeble muscular layer of circular fibres.

The penis is a conical body, hanging from above subvertically
in the pear-shaped penis-sheath, and encloses a cavity, the
seminal vesicle, which gives rise to the ejaculatory duct, opening
into the sheath at the tip of the penis. The vesicle is coated
internally with a thick glandular epithelium, which projects
into the lumen of the organ in folds. Embedded in the body-
parenchyma around the penis-bulb are numerous glands, the
ducts of which enter the penis at the base and open into the
penis-sheath over the surface of it.

Female Organs.—I am unable to give an account of the
ovary, as I have been reluctant to sacrifice the anterior half of
the body to the microtome. Probably the paired ovary occurs
in the usual manner. The vitelline glands, which are composed
of large cells closely packed, extensively fill up the mterstices
between the gut diverticulae. They are in connexion with the
oviduct at numerous points by means of a short cylindrical
duct.

The oviducts lie close to the dorso-lateral side of the nerve
cords, one on each side, in which position they proceed straight
backwards, receiving the contents of many vitelline glands.
Behind the genital opening they near the median line, slightly

150 TOKIO KABURAKI

rising at the same time, and finally jom ito a single median
duct, the glandular canal. The oviduct is characterized by the
possession of ciliated epithelial cells, beneath which comes a thin
muscular coating, and between which open numerous glands
for some little distance before forming a common duct.

The glandular canal pursues a course obliquely forwards and
downwards, and finally opens into the vestibulum at a point
on the right side, after receiving in its course a short duct from

TEXT-FIG. 6.

|
|
|
!
|
I
|
|

va

Genital organs of Am. trifuscolineata in sagittal section,
diagrammatically shown. gi.=genito-intestinal canal, a=
intestine. Other letters as in Text-figs. 1 and 4.

above, which stands in communication with one of the intestinal
coeca, so that there is, as in the Heterocotylean Trematodes,
a genito-intestinal canal. This is similar to that deseribed by
Geba (11)in Am. tristriata andAm. mediostriata.
The canals are constructed in the same manner as the oviduct,
and are lined with an epithelium composed of ciliated columnar
cells ; outside this is a thin muscular layer.

PLANARIANS FROM MAURITIUS AND RODRIGUES 151

The intestinal coecum is coated internally with an epithelium
made up, as usual, of high columnar cells, which near the junction
point of the canal exhibit a close affinity for borax carmine ; in
the cavity are contained spermatozoa enveloped in a coagulum
of the secretion. ‘This organ seems to me to serve as a seminal
receptacle.

As stated above, the present species closely resembles
Am. tristriata described by Geba. But it differs from this
in the arrangement of the parts of the genital organ.

NotTE UPON THE CANAL CONNECTING THE FEMALE
GENITAL ORGAN WITH THE INTESTINE.

The peculiar canal connecting the female genital organ with
the intestine is of somewhat frequent occurrence in other
terrestrial planarians, as is the case with Rhynchodemus
terrestris Mill, Rh. attemsi Bendl, Pelmatoplana
mahéensis von Graff, and P. braueri von Graff. In
Rh. terrestris, according to von Graff (12), the two ducts,
one on each side, spring from the anterior parts of the seminal
receptacle and take a course obliquely upwards and backwards,
finally opening into the posterior trunk of either side. But
these connexions appear to be inconstant in occurrence and
arrangement, for on sore occasions there exists, according to
Bend] (2, 8, 5), a right connexion only, which is well developed.
He has also placed on record a case of Rh. at temsi, in which
the receptacle is in direct connexion with the left posterior
trunk of the intestine, without passing by any distinct duct.
According to Mell (20), the vagina in both P. mahéensis and
P. braueri is continuous with a canal which communicates
with the right posterior trunk of the intestine.

An arrangement of this kind is also known to occur in other
Turbellarian groups. Such are Oersted’s Phaenocora uni-
punctata, an Acoela (4, 5),and Haswell’s Enterogonia
pigrans, a Polyclad (18,14). In the former the receptacle
communicates with the intestine by a short median duct, while
in the latter the dorsal passage of the vagina, after receiving on

152 TOKIO KABURAKI

its ventral side the common duct formed by union of the lateral
uterine ducts, proceeds backwards as a narrow tube, which
opens into the median posterior branch of the intestine. To
me, such frequent occurrence of the genito-intestinal connexion
appears in favour of the view that this is certainly not abnormal.
The discovery of the canal in question helps to connect more
definitely the seminal receptacle of some Polyelads and Triclads
with parts that occur in other Platodes. It cannot well be
doubted, it seems to me, that this canal corresponds to the
similarly named canal in the Heterocotylean Trematodes.
In this group the duct passes from the oviduct, opposite the
opening of the yolk-duct, to the right limb of the intestine.
Now let us proceed to review the arrangement of the terminal
part of the female genital organ, which is of interest from the
morphological point of view. The vaginal canal, after almost
invariably receiving the unpaired common uterime duct, either
ends blindly, as in Stylochus and some others, or proceeds
backwards to join the seminal receptacle, as in some Triclads,
which is unpaired in most, but paired in some, genera (Disco -
celis, Woodworthia, Shelfordia, and Diploso-
lenia). This agrees closely with the condition of the duct
found in the Aspidocotylean Trematodes, which are provided
with a duct, arising from the oviduct, near or opposite the
opening of the yolk-duct and leading to the vitelline receptacle.
On some occasions the dorsal passage of the vagina, instead of
swelling into a receptacle and opening into one of the intestinal
coeca, pursues a course backwards, finally to open to the
exterior at a certain point of the surface of the body. In
Cryptophallus and Bergendalia it proceeds back-
wards and downwards, describing an arched course, and finally
opens into the female atrium closely behind the vaginal aperture
and just inside the external female aperture. In the case of
Trigonoporus, Copidoplana, and Tripylocelis
the duct terminates behind on the ventral surface of the body by
the second female aperture. In Polyporus the second female
opening lies near the hind end of the body, whilein Laidlawia
it occurs, occupying a position on the dorsal, but not on the

PLANARIANS FROM MAURITIUS AND RODRIGUES 153

ventral, surface. Such an opening dorsally situated is also
known to occur in Acoelean forms, suchas Cylindrostoma
quadrioculatum Jens, and C. klostermanni Jens.

The discovery of Laidlawia (15) mentioned above may
be regarded as of some importance, as it may constitute an
additional link in the chain of evidence against the homology
of the part of the duct, as has been suggested by Lang (18). He,
in his monograph, has the following passage : ‘ In morphologi-
scher Beziehung erinnert der Canal, insofern er eine Verbindung
zwischen der Kinmiindungsstelle des Uterus in den Eiergang
einerseits und der Aussenwelt anderseits darstellt, emigermassen
an den Laurerschen Canal der Trematoden und Cestoden.’
A comparison with the Laurer’s canal of the Malacocotylean
Trematodes, which passes up from the oviduct, in the neigh-
bourhood of the ootype, and opens by a minute pore on the
dorsal surface, obviously suggests itself.

Great interest is attached to the existence of some Polyclads
having the dorsal passage of the vagina, which opens either to
the exterior on the surface of the body, or into one of the intes-
tinal coeca, as stated above. The homology between the genito-
intestinal canal of the Heterocotylea, the Laurer’s canal of
the Malacocotylea, and the duct leading to the receptacle in
the Aspidocotylea, though it may be open to question, seems
to have the balance of evidence in its favour. Haswell (14) has
put forward the view that there can be regarded as representing
Laurer’s canal in the Polyclads not only the genito-intestinal
canal of Enterogonia, but the seminal receptacle of the
Acotylea in general and the posterior female passage, which
opens to the exterior, as has been observed in some forms.
IT am inclined not only to concur with him, but further to
develop to a certain extent this view even to the Triclads. In
this communication, however, I have intentionally abstained
from making any such attempt, leaving the problem to future
consideration.

2. Bendl, W. E., 1908. “‘ Beitriige zur Kenntnis des Genus Rhyncho-
demus ”’, ‘ Zeitschr. f. wiss. Zool.’, Bd. Ixxxix.

3. —— 1909. ‘‘ Europiiische Rhynchodemiden, i ”’, ibid., Bd. xcii.

4. —— 1909. ‘‘ Rhabdocédle Turbellarien aus Innerasien ”’, ‘ Mitth. d.
Naturwiss. Ver. f. Steiermark ’, Bd. xlv.

5. ——1909. “Der Ductus genito-intestinalis der Platheminthen”’,
‘Zool. Anz.’, Bd. xxxiv.

6. Bock, Sixten, 1913. ‘“‘ Studien iiber Polycladen ”’, ‘ Zoologiska Bidrag
f. Upsala ’, Bd. ii.

7. Bohmig, L., 1890. ‘‘ Untersuchungen iiber rhabdocéle Turbellarien, ii,
Plagiostomia und Cylindrostomia v. Graff”, ‘Zeitschr. f. wiss.
Zool.’, Bd. li.

8. Busson, B., 1903. ‘‘ Ueber einige Landplanarien”’, ‘ Sitzungsber.
Akad. Wien ’, vol. exii, pp. 375-429.

9. Dendy, A., 1889. “‘The Anatomy of an Austrarian Landplanarian ”’,
‘Trans. Roy. Soc. Victoria ’, vol. i, part 2.

10. Fuhrmann, O., 1912. ‘‘ Voyage d’exploration scientifique en Colombie.
Planaires terrestres de Colombie’’, ‘Mém. Soc. neuchateloise se.
nat.’, vol. v, pp. 748-92.

11. Geba, J., 1909. ‘‘ Landplanarien von Madagaskar und den Comoren ”’,
‘ Voeltzkow, Reise in Ostafrika in den Jahren 1903-1905. Wissen-
schaftliche Ergebnisse’, Bd. ii.

12. Graff, L. von, 1899. ‘‘ Monographie der Turbellarien. ii. Triclada
Terricola (Landplanarien).”

13. Haswell, W. A., 1907. “ A genito-intestinal canal in Polyclads ”’, ‘ Zool.
Anz.’, Bd. xxxi.

14. —— 1907. ‘‘ Observation on Australian Polyclads ”, ‘Trans. Linn.
Soc., London ’, 2nd ser., vol. ix.

15. Herzig, E. M., 1905. ‘* Laidlawia trigonopora, n. gen. n. sp.”’, * Zool.
Anz.’, Bd. xxix.

16. Ikeda, I., 1911. ‘‘ Note on a new Land Planarian from Ceylon ”’,
‘Spolia Zeylonica ’, vol. vii, part xxvii.

17. Laidlaw, F. F., 1903. ‘‘ On a Land Planarian from Hurule, Male Atoll,
with a note on Leptoplana pardalis Laidlaw ”’, ‘ Fauna and Geogr.
Maldive Laceadive Archip.’, vol. ii, part i, p. 579.

18. Lang, A., 1884. ‘“‘ Die Polycladen”’, ‘ Fauna u. Flora des Golfes von’

TOKIO KABURAKI

REFERENCES.

. Balfour, J. B., and others, 1878. “ An Account of the Petrological,

Botanical, and Zoological Collections made in Rodrigues during the
Transit of Venus Expeditions in 1874-5 ”’, ‘ Phil. Trans. Roy. Soc.’,
vol. elxviii (extra volume), pp. 561, 562.

Neapel’. xi. Monographie.

it, 1 Py a Py 2 Up

PLANARIANS FROM MAURITIUS AND RODRIGUES 155

19. Meixner, A., 1906. ‘“‘ Zwei neue Landplanarien ”’, ‘ Zool. Anz.’, Bd. xxix,
p. 665.
Mell, C., 1903. ‘‘ Landplanarien der Madagassischen Subregion ”’,
‘ Abhandl. d. Senkenb. naturf. Ges. Frankfurt ’, Bd. xxvii.
21. —— 1904. ‘‘ Die von Oscar Neumann in Nordost-Afrika gesam-
melten Land-Planarien ”’, ‘ Zool. Jahrb., Abt. Syst.’, Bd. xx.
Moseley, H. N., 1875. ‘‘ On the Anatomy and Histology of the Land-
planarians of Ceylon, &c.”’, ‘ Phil. Trans. Roy. Soc.’, vol. clxiv.
Scharff, R. F., 1900. “‘ Rhynchodemus Howesi: a new European
Species of Terrestrial Planarian Worm”, ‘Journ. Linn. Soc.’,
vol. xxviii.
24. Snell, H. J., and Tams, W. H. J., 1920. ‘‘ The Natural History of the
Island of Rodrigues”, ‘Proc. Cambridge Phil. Soc.’, vol. xix,
part 6, p. 287.
25. Whitehouse, R. H., 1914. ‘“‘ Land Planarians ”’, ‘ Ree. Indian Mus.’,
vol. viii, part 6.
26. —— 1919. ‘Indian Land Planarians ”’, ibid., vol. xvi, part 1.

20

22

23

EXPLANATION OF PLATE 4.

Fig. 1.—Placocephalus isabellinus Geba in the preserved state,
seen from the dorsal side. About natural size.

Fig. 2.—Ditto. Ventral view.

Fig. 3—Rhynchodemus ceylonicus von Graff in the preserved
state, seen from the dorsal side. About 1:5 x.

Fig. 4—Amblyplana trifuscolineata, n. sp. in the preserved
state, seen from the dorsal side. About 2 x.

Fig. 5.—Ditto. Ventral aspect.

Fig. 6.—Rh. ceylonicus. Longitudinal section of an eye.

Fig. 7.—Ditto. Transverse section through the ovarian region.

Fig. 8.—Ditto. Median sagittal section through the region of the
copulatory organs.

ABBREVIATIONS USED IN THE EXPLANATION OF PLATE.

ed. ejaculatory duct. gc.=glandular canal. gv.=genital vestibulum.
i.=intestine. %In.—longitudinal nerve cord. n.=nerve. od.=oviduct.
ov.=ovary. p.=pigment. ps.=penis-sheath. s.=sole. sv.=seminal
vesicle. vd.=vas deferens. vr.=visual rod.

5 whe, f
a 7 z 7 ; well ps 7" 7 ?

co

ye 1 eehee
tenw) oe
eee an
contol ionge a
7 ik ‘he, anrcal nal aal

Nel bQ = ‘dhe
saline Leow el
pre Lele
s 5 oles
rs ee
i a

Quart. Journ. Micr. Sct. Vol. 65, N.S., Pl. 4.

|

op if r i

se sv va rey

TATEN) SOQLUASMAUITD

-

Kaburaki del.

a

Gonospora minchinii, n. sp., a Gregarine
inhabiting the egg of Arenicola.

By
Edwin 8S. Goodrich, F.R.S., and H. L. M. Pixell Goodrich, D.Se.

With Plates 5 and 6.

WHEN examining the contents of the coelom of an Areni-
cola ecaudata Johnston, at the Marie Biological Labora-
tory in Plymouth last winter, we discovered a new Gregarine
of considerable interest, since it appears to be the first instance
on record of such a parasite inhabiting the ovum of its host.t

This gregarine belongs to the genus Gonospora. It does not
seem to occur at all in the male worm, and of the females
examined only about 30 per cent. were infected. However,
since the parasite was not found in any but female worms whose
ovaries were fairly ripe and had begun to shed their products
into the coelom, it is probable that it often inhabits less mature
hosts, but in some situation not yet determined. We have
looked for it without success in the immature ovary. Frequently
it occurs simultaneously with the larger and well-known coelomic
sregarine Gonospora (Kalpidorhynehus) areni-
colae Cunningham.

The immature ovary of Arenicola ecaudata isalobu-
lated organ with finger-shaped processes (see Gamble and Ash-
worth, 1). Inside it the germ-cells multiply, accumulating in its
lumen, and later bursting through its wall. The ova thus escape
into the coelom at various stages in development ; some quite
small and oval, others larger, more rounded, loaded with yolk,
and surrounded by a thick covering. This shell is formed of two

1 Since this was written we have learnt from Sir Ray Lankester that
many years ago he discovered a somewhat similar parasite in the eggs of
Thalassema. From an inspection of some unpublished drawings of the

trophozoite, which he kindly sent to us, we conclude that it is not the
Gregarine described in this paper.

158 EDWIN 8S. GOODRICH

distinct layers : an outer thin refringent membrane, the original
vitelline membrane ; and an inner much thicker and probably
less dense perivitelline layer (fig. 10). A full-grown ovum with
its covering is about 120 to 130 microns in longest diameter.

Young Trophozoites.—The youngest stages of the
parasite observed were small rounded trophozoites embedded
in the ege close to its nucleus. Fig. 1 shows such a stage
where the gregarie is 12 in diameter; far smaller than the
nucleus of the immature egg it has invaded, and indeed only
about twice the diameter of its nucleolus. It will be noticed
that even at this early stage the nucleus of the parasite is
distinguished from that of the ovum by the possession of two
karyosomes, while the latter is almost invariably provided with
only one nucleolus. The trophozoite continues to grow at the
expense of the egg, enlarging and becoming stored with granules
of paraglycogen (figs. 2, 3). As it acquires the shape and size of
the adult (fig. 7) the egg and its nucleus become more and more
compressed against the surrounding membranes.

Penetration into the egg.—lt has been stated above
that the ovum of Arenicola is protected not only by a vitelline
membrane, but also when full-grown by a thick perivitelline
layer. How does the parasite penetrate to the egg? is a
question which at once suggests itself. Now it is probable that
fully-developed eggs are safe from invasion, since infected eggs
are rarely, if ever, found with the perivitelline layer fully formed.
By far the greater number of eggs infected are provided with
a vitelline membrane only (figs. 4, 7), or with but a thin peri-
vitelline layer as well (figs. 6, 11). The parasite enters the egg
by boring a round hole through these membranes, and usually
the margin of the hole is found turned inwards (figs. 2, 4). The
aperture so formed may remain open ; but sometimes it seems
to close up almost entirely (fig. 6), presumably when the egg is
invaded at a very young stage.

Position and growth of trophozoite in eggs.
—It is often very difficult to decide whether the parasite, having
picrced the egg-membranes, really enters the egg-cell or merely
bulges into it. Except perhaps in the very earliest stages it

id ee” ee ee Ee en eee, ee!

GONOSPORA MINCHINII 159

certainly lies as a rule outside the egg-cell, between it and the
membranes (figs. 2, 4). It compresses the ege more and more
as it grows and is separated from it by a space, except at that
one region opposite the point of entrance where the epimerite
of the parasite adheres closely to the egg-cytoplasm near the
germinal vesicle (fig. 9). Here are developed, in that part of the
eregarine which is fixed to its host, small club-shaped bodies
staining deeply in haematoxylin or fuchsin. They appear to be
hollow, with long narrow necks reaching to the surface (fig. 9).
These strange structures somewhat resemble the ‘ lamelles
mucoides ’ described by Léger and Duboseq in Nina (2); but
their function would appear to be connected with the absorption
of nutriment from the egg, or possibly merely with fixation.
Meanwhile, as the parasite grows it enlarges the deep depres-
sion it causes in the egg; the margin of this hollow is at first
smooth (fig. 2), it soon becomes notched, and finally drawn out
into delicate protoplasmic processes converging towards the
point of entrance (figs. 3, 4, 5).

Effect of parasite on host egg.—tThe very young
ovum has little or no yolk ; but with advancing age the yolk
granules imcrease in humber until the fully-developed egg
becomes so heavily loaded that it looks quite opaque. In
parasitized eggs, however, the yolk is absorbed by the gregarine
almost as fast as it 1s laid down, so that in late stages the com-
pressed ovum is relatively clear, while the parasite on the
contrary is densely granular (fig. 4). The nucleus of the egg
is also influenced, for its nucleolus, instead of undergoing the
orderly series of changes seen to occur in normal eggs, lags
behind in differentiation, remaining in fact apparently at that
stage of development it had reached when the egg was invaded.
Thus the nucleolus in most parasitized ova resembles that of the
quite young ovum when it is still small and has but little yolk
(figs. 2, 7).

Another peculiar and somewhat similar effect is seen on the
egg-envelopes. There is no reason to think that the peri-
vitelline layer when once formed can be reabsorbed, and since
it is, as a rule, almost or quite absent from parasitized eggs,

160 EDWIN 8. GOODRICH

even when these have reached full size, there can be little doubt
that the presence of the gregarin> checks its deposition. Never
have we observed full-sized eggs without parasites in which this
layer was not present.

Emergence of parasite from egg.—When the tro-
phozoite has completed its growth 1t emerges from the egg-shell
by around hole, whichis probably the enlarged original opening
through which it entered, or at least formed afresh in the same
place (figs. 5, 8). The gregarine first pushes out its pointed
‘ tail’ end, the rest of the body following after.

Fate of parasitized egg.—aAssoonas the parasite has
thus abandoned the egg, leaving a large space partially sur-
rounded by the emaciated host-cell and communicating with the
exterior by an aperture of considerable size, leucocytes from
the coelomic fluid make their way in (figs. 8, 11, 12). They
gather in large numbers in the cavity, and proceed to attack the
already depleted ovum, the cytoplasm of which becomes
vacuolated. Strange thread-hke structures, which stain im
acid-fuchsin, are now visible round the edge of the egg (th., fig. 12)
before its final breaking up.

The free trophozoite.—the full-grown trophozoite
free in the coelomic fluid is usually pear-shaped, the epimerite
being at the blunt end. Asa rule the nucleus is provided with
two conspicuous karyosomes, but additional small granules
may be present. Often the gregarines hang together in groups,
sometimes in masses of ten or twelve individuals.

Association and spore-formation.—tThe associa-
tion of two trophozoites is terminal (fig. 13), the ‘ head’ end of
one penetrating deeply into that of the other mm the manner so
characteristic of the genus Gonospora (8, 4). At the extremity
of the embedded epimerite may be seen in sections a cap of
dense substance tipped with a deeply-staining granule, possibly
of nuclear origin (fig. 14). At this stage, before the formation of
a cyst, the two associates can still be separated by pressure.
As soon as the cyst wall is secreted round the pair their opposed
faces flatten out. Gamete formation and syngamy then take
place as usual in these gregarines.

|
|
|
|
|

GONOSPORA MINCHINIL 161

A spore with its eight sporozoites is shown in fig. 15 ; it is from
8 to 10 » in length. The sporocyst is thin, one pole being
rounded and the other provided with a slight thickening, but
there is no well-developed funnel such as occurs in Gono-
spora glycerae (8).

For this new gregarine we propose the name Gonospora
mine hini1,

Summary .—The new species of gregarine described above,
and to which we have given the name Gonospora min-
ehinii, occurs in the coelomic fluid of the female Arenicola
ecaudata. The adult trophozoite is pear-shaped, and the ripe
spore has a thin cyst without distinct funnel. The young tropho-
zoite lives in the egg floating in the coelomic fluid of the Areni-
cola, where it grows at the expense of the food-material stored in
the ovum. To reach the ovum it pierces the viteline membrane
and perivitelline layer. The growing trophozoite occupies a deep
depression it causes in the egg, to which it adheres by its
epimerite. The margin of this depression becomes drawn out
into delicate protoplasmic processes. The cytoplasm and
nucleus of the host-cell, and also the development of the peri-
vitelline layer, are affected by the presence of the parasite.
When full-grown the trophozoite escapes from the egg by a hole
pierced in its envelopes, and leucocytes then enter the space
so left to complete the destruction of the ovum.

REFERENCES.

1. Gamble, F. W., and Ashworth, J. H.—** The Anatomy and Classifica-
tion of the Arenicolidae ’’, ‘ Quart. Journ. Micros Sci.’, vol. xliii,
1900.

2. Léger, L., et Duboseq, O.—“* Etudes sur la sexualité chez les Gré-
garines ”’, ‘ Arch. f. Protistenk.’, Bd. xvii, 1909.

3. Pixell Goodrich, H. L. M.—‘‘ The Gregarines of Glycera siphonostoma ”’,
‘Quart. Journ. Micros. Sci.’, vol. lxi, 1916.

4. Trégouboff, G—‘‘ Etude monographique de Gonospora testiculi
Trég.”’, ‘ Arch. Zool. Expér.’, vol. lvii, 1918.

NO. 257 M

162 EDWIN S. GOODRICH

EXPLANATION OF PLATES 5 AND 6.

Fig. 1.—Young egg of Arenicola with smal] trophozoite inside it. Whole
preparation ; Formol-iodine, Paracarmine. x 500.

Fig. 2.—Later stage showing opening in vitelline membrane, and depres-
sion in egg in which lies the parasite. Whole preparation; Formol-
corrosive, Paracarmine. x 500.

Fig. 3.—Nearly full-grown parasite in egg ; from the living. x 500.

Fig. 4.—Semi-diagrammatic optical section of egg with contained para-
site. x 500.

Fig. 5.—Trophozoite emerging from egg ; from the living x 500.

Fig. 6.—Portion of a section of an infected egg showing the’ young
trophozoite. Bouin, Iron-Haematoxylin.. x 1.100.

Fig. 7.—Optical section of whole preparation of egg with full-grown
trophozoite. Formol-corrosive-acetic, Paracarmine. x 500.

Fig. 8.—Infected egg from which the parasite has escaped. Leucocytes
are making their way into the cavity. From the living. x 500.

Fig. 9.—Portion of a section of full-grown trophozoite which is fixed to
host-cell near flattened nucleus, and showing deeply-staining bodies, a.
Chrom-osmic ; Iron-haemat., Light-green. x 1,100.

Fig. 10.—Part of section of uninfected egg, showing normal development
of vitelline and subvitelline membranes. Bouin, Iron-haemat. x 1,100.

Fig. 11.—Section of an infected egg from which parasite has escaped.
Chrom-osmic ; Iron-haemat. x 500. ;

Fig. 12.—Similar egg at later stage showing its destruction by invading
leucocytes. Chrom-osmic, Iron-haemat., Light-green.

Fig. 13.—Two Gonospora minchinii in association. Whole pre-
paration. x 120.

Fig. 14.—Section through dovetailing epimerites of associates. Chrom-
osmic; Iron-haemat., Light-green. x 1,100.

Fig. 15.—Spore with eight sporozoites. x 3,000.

REFERENCE LETTERS.

a.=deeply-staining bodies at edge of trophozoite fixed to ovum. c¢.=
cytoplasm of ovum. cp.=cytoplasm of parasite. /.=leucocyte in cavity
vacated by parasite. //.=limit between associated trophozoites. mp.=
minute pore, probably contracted pore of entrance. m».—=nucleus of egg.
ne.=nucleolus. mnp.=nucleus of parasite. o.=ovum. op.=opening,.
p.=parasite. p' and p*.=associates. pr.—protoplasmic process. sp.=
space left by parasite. sv.—perivitelline layer. ¢.=‘ tail’ end of tropho-
zoite th.=threadlike structures in outer zone of egg. v.=vitelline
membrane

Ae
5 Pa

Aa
Za!

Py Od eee:

Quart. fourn.Micm Sei. VABS,NS. P5.

Huth, Londen

BR.

Quart. ourn. Mier Sci. Vol. 65.NS. 4.6.

E.S.Goodrich del. Huth, London.

The Eye of Peripatus.
By
William J. Dakin, D.Se., F.L.S., F.Z.S.,

Derby Professor of Zoology, University of Liverpool ;
late Professor of Biology in the University of Western Australia.

With Plate 7 and 3 Text-figures.

THE first description of the minute structure of the Eye
of Peripatus was given by Balfour (1) in his memorable
paper on the anatomy of Peripatus capensis. So far
as I am aware nothing has been added to our knowledge of
the structure since that date, despite the advances in micro-
scopical technique, and the rather thorough investigation
of invertebrate visual organs. Other arthropod eyes have
received considerable attention, and this seems strange at
first because a comparison of the Peripatus eye with that
of other arthropods should be highly interesting by reason of
the phyletic position occupied by the Onychophora.

The development of the eye was followed by Sedgwick (4),
but nothing was added to the previous knowledge of the
structure of the adult eye, although the origin of the different
parts was very clearly shown.

In Balfour’s illustration, the structure of the eye of P e ri-
patus capensis is shown in longitudinal section through
the head. This figure has been often recopied, and it will
be well to take note of the details brought out (see Text-fig. 1,
which is a copy of that after Balfour in this Journal,
vol. 23). The general cuticle of the body wall is continued
as a thin layer over the eye. Below this is the cornea—a layer
of epithelial cells, which are continuous with the epidermis.
Between the cornea and the lens there is another cell layer
which appears to terminate peripherally against the region
marked pigment. There is no evidence to show that the
structures masked by the pigment were ever brought to light.

NO. 258 N

164 WILLIAM J. DAKIN

From-the illustration it would appear as if the pigment formed
a separate layer which acted as a kind of capsule enclosing the
retina and bounding the eye internally. This impression is
strengthened by the fact that the cells below the pigment are
marked ‘ optic ganglion ’.

The space within the structures enumerated above is
oceupied by the lens, and by a layer termed the rods.

TeExt-Fia. |.

Longitudinal section of the Eye of Peripatus capensis after
Balfour, ‘Quart. Jour. Mier. Sc.’, vol. 23, plate 18, fig. 24. cor. =
cornea; /.=lens; op.=optic ganglion; op.n.=optic nerve;
pi.r. =pigment; Re, =rods; s.p.=secondary papilla.

Now let us turn to the results of the present investigation.
The species utilized was Peripatoides occidentalis from
Western Australia. A large number of preparations had to be
made, including sections and maceration preparations. No
single method ean be singled out, the usual series of fixatives
and stains must be adopted, one method giving a little informa-
tion, another a little more (see Dakin, “ Eye of Pecten”’,
* Quart. Journ. Micros. Scei.’, 1909).

The Eve of Peripatus is not stalked although the dista]

t } oD

THE EYE OF PERIPATUS 165

surface forms a dome-shaped protuberance on the skin. The
whole of this bulge appears to be occupied by the lens. In
sections which have not been depigmented (see left side of
fig. 1) the eye appears to be made up of three regions—the
lens, the region previously known as the retina (or rod region),
and the so-called optic ganglion. Now it will clear matters
up at once if we state that the rod layer does not consist
of cells but only of parts of cells—i.e. the distal halves
of cells whose nuclei lie internally to the pigment. In
other words, the so-called optic ganglion plus the
rod layer together make up the retina. The units
of these layers are not separated by a layer of pigment ; the
pigment is actually enclosed within the cells (see fig. 2).

The Cuticle overlying the eye (fig. 1, Cut.) differs
from that of the surrounding regions in being free from the
small projections so characteristic elsewhere. Not only are
the minute spines absent, but the dermal papillae which are
present over the entire body wall are missing here.

The Epidermis is continued over the eye to from the
Cornea (fig. 1, Cor.). Most of the cells of the general
epidermis are somewhat cubical or pyramidal in form, with
large nuclei. The corneal cells are very different, being quite
flat. The nuclei are decidedly compressed and the protoplasm
is reduced in amount.

The Subcorneal layer of cells may be said to form
a capsule which encloses the lens. It is seen as a well-marked
layer where it covers the lens and extends down over the rod
layer (fig. 1, Sub. Cor.). There is nothing of importance to
add further regarding it except that im the development of the
eye it formed the outer portion of a complete vesicle, the
proximal cells of which have given rise to the retina (see
fig. 5).

The Lens is non-cellular and forms a homogeneous mass
which stains readily with eosin. The face towards the retina
appears almost flat in well-preserved sections, whilst the distal
surface is highly convex, so that the entire structure is practi-
cally a dome. In all the well-preserved sections the proximal]

N 2

166 WILLIAM J. DAKIN

surface of the lens was in contact with the face of the retina.
A delicate non-nucleated sheath appears to bound the lens,
but it is in all probability only the outermost layer of the lens
substance.

Tur STRUCTURE OF THE RETINA.

Very little trouble will suffice to show quite clearly the
structure of the dioptrical part of the eye described above.
The elucidation of the structure of the retina is a much more
difficult task, and it is quite natural that this essential part
of the eye has remained misunderstood.

As we have already seen, the pigment band does not enclose
the retina, but is made up of pigment granules lying within the
retinal elements. We shall keep the term Rods for the real
constituents of the rod layer, the part marked Re. in Balfour’s
figure. This rod layer in poor, or even in moderately
good sections, appears to be made up of rather long ‘ rods ’
separated by clear spaces. The ‘rods’ also have a pecuhar
broken-up appearance even when not cut obliquely, as appears
most frequently to have been the case. Now as a matter of
fact these dark-staining bodies are not the rods. Macera-
tion preparations, but still more certain, transverse sections
in the plane of the retina, show quite clearly that the rod layer
is not exactly what it seems. It comes as a surprise, in fact,
to discover that the dark-stainng part of the rod layer
appears in transverse sections as a grating or net (see fig. 3).
It now requires the study of depigmented longitudinal sections
and maceration preparations to explain the above. Really
the explanation is simple. The retina is built up of one kind
of unit only, and there are no supporting cells or other non-
visual elements. Each visual unit consists of a rod-cell bearing
a rod.

The Rod-cells and Rods. A rod-cell (see fig. 2, and
fig. 1, Rod-cell) consists of a columnar portion containing finely-
granular protoplasm and crowded with pigment granules, and
a proximal constricted and unpigmented part swollen out by
the nucleus. As the rod-cells are numerous and the nuclei

THE EYE OF PERIPATUS 167

rather large, the latter are arranged at different levels in the
cells. It is the nuclei of the rod-cells which collectively have
been mistaken for an optic ganglion.

Proximally the rod-cells are continued as nerve fibres,
which form the very short optic nerve. The distal portions
of the rod-cells are hexagonal in section, so that all fit together
closely to form a mosaic (fig. 4).

The rods are projections from the rod-cells, but the main
part, the axis, of the rod is composed of a rather non-staining
material. Thus in longitudinal sections the axes of the rods
lie between the stained column-like bodies, whilst in trans-
verse sections the rods would be the meshes of the grating
(see fig. 3). The next question is, naturally, what is the
‘grating’ itself, the part so easily mistaken for the rods in
longitudinal section. It would appear as if this staining
substance was simply the peripheral portions of the rods.

Each rod can be seen in maceration preparations to bear
peripheral ‘ Stiftchen "—short processes very characteristic of
invertebrate visual cells. These ‘ Stiftchen’ clothe each rod
completely, and it is the ‘Stiftchen’, or the ‘Stiftchen ’-
borders, of the rods which stain up so readily and actually
appear to be the rods in longitudinal sections. ‘This explains
why they show up as a kind of grating when cut transversely,
for the * Stiftchen ’-borders of adjacent rods touch each other
(see figs. 1 and 3).

Underlying the layer of rod-cells is a collecting region of
nerve fibres—the prolongations of the sensory cells. These
collect to form a short optic nerve (fig. 1, Op. N.) which enters
the brain. The optic tract is traceable for some distance within
the ‘ Punktsubstanz’. A delicate layer of connective tissue
forms a capsule bounding both retina and optic nerve.

CoMPARISON OF THE EYE oF PERIPATUS WITH THAT OF OTHER
ARTHROPODS, AND WITH THE POLYCHAETE EYE.

The Eye of Peripatus is in reality a very simple struc-

ture compared with some insect ocelli. It is developed,

as was discovered by Sedgwick (4), as a simple vesicular

168 WILLIAM J. DAKIN

invagination of the ectoderm. ‘The vesicle cut off gives rise
to the subcorneal layer and the retina (see fig. 5). The lens is
secreted within this vesicle and is non-cellular. It has no
connexion directly with the cuticle of the body-wall, nor is
the latter thickened as it passes over the cornea.

The description already given shows clearly that we can
exclude the complicated compound eyes of the Insects and
Crustacea so far as our comparison is concerned. No informa-
tion regarding the origin of the compound eye of the arthropoda
is likely to be obtamed by the study of the Eye of Peripatus.
Comparison must be made, then, with the lower and more

TEXT-FIG. 2.

Insect ocellus (Helophilus) after Hesse, somewhat modified.
C. =cuticle; C.L. = cutic. lens; Conn. = connective tissue; hy. =
hypodermis; #.c.=rod-cells of retina. Note difference in
character of lens from that of Peripatus. The formation of lens
by thickening of cuticle over eye is very characteristic in Insecta.

simple arthropod visual organs, the simple eyes. We shall
also exclude the Arachnoid eyes, the structure of which (see
Lankester (6), and Watase) is again different in type. We
are left with the Myriapod eyes and the larval eyes and ocelli
of insects.

A marked difference is easily recognized between the Eye
of Peripatus and the above. In the ocelli of insects (Helo-
philus, Ceratopsyllus, &c., see Text-fig. 2) and in the larval
eyes, we usually find that the ectoderm is invaginated
to form the retina (see literature 2 and 8). We do not
find a complete vesicle. ‘The ectoderm does not give

THE EYE OF PERIPATUS 169

rise to a completely separated vesicle, part of which becomes
a subcorneal layer. On the other hand the retinal layer can
be traced into the ectoderm.

With this marked difference we must also note that the lens
in the Insecta and the Myriapoda is directly continuous with the
cuticle and is indeed a local thickening of the same, whilst in
Peripatus it is secreted within the vesicle.

The modern work contirms, therefore, the statements of
Lankester (5), when in his article on the structure and
classification of the Arthropoda he adds, ‘... the Chaetopod
eye, which is found only in the Onychophora where the true
Arthropod eye is absent. The essential difference between
these two kinds of eye appears to be that the Chaetopod eye
(in its higher developments) is a vesicle enclosing the lens,
whereas the Arthropod eye is a pit or series of pits into which
the heavy chitinous cuticle dips and enlarges knobwise as
a lens’.

Thus whilst we can homologize the cuticle, cornea, sub-
corneal layer, &c., of Peripatus with parts of the simple eyes
of the Myriapoda and Insecta, the Peripatus eye is not primitive
so far as the dioptrical parts are concerned, but has developed
along its own lines and resembles that of the highly-developed
Chaetopoda. The Eye of Peripatus has, however, not evolved
very far, and its retina is quite simple and indeed not at all
unlike that of the median ocelli of Helophilus (one of the
Diptera) or of the eye of Scolopendra. In both these examples
we have retinas consisting solely of visual cells. These cells
bear rods which are remarkably lke those of Peripatus and
have the same marginal (lateral) ‘Stiftchensaum’. Indeed,
the rods of the Scolopendra retina stain very like those of
Peripatus.

Hesse speaks of the retinal elements of these eyes as being
of a very original type. It is particularly interesting, therefore,
to find the agreement with Peripatus.

The histology of the Polychaete eye has been investigated
in some detail by R. Hesse (8). We can find material for
comparison in his papers.

170 WILLIAM J. DAKIN

Eyes are to be found of very varying form and complexity
of development. In a great many cases an open cup-shaped
retina is to be seen (resulting from ectodermal invagination),
but there is no lens, cuticular or otherwise. The retina in nearly
all cases consists of rod-cells bearing rods which are directed
distally. In a large number of the eyes, the histology of which
has been investigated, the details are not very similar to the
Kye of Peripatus. Hesse’s figure of the eye of Sipho-
nostomum diplochaetos is, however, curiously like that
of the early illustrations of the Peripatus eye so far as the retina

TEXT-FIG. 3.

Diagram of lens and corneal layers of eye of Polychaete (Vanadis
formosa), modified after Hesse. Note similarity of arrange-
ment of layers to that found in Peripatus. C.o. = outer cornea ;

C.i, = inner cornea; Cu. = cuticle; Hy. = hypodermis; L. = lens;
R.=retina (structure not shown).
is concerned. Both the vertical sections and those taken im
the plane of the retina indicate this, and no doubt the structure
is almost exactly the same as that of the Eye of Peripatus.
A detailed re-examination with up-to-date methods would be
necessary to make it certain.

The remaining features (dioptrical) of this Polychaete eye
are quite unlike those of Peripatus. The eye is not nearly so
well developed as that of the latter.

One of the best-developed Polychaete eyes is found in the
group Alciopidae. We have here a vesicular eye (see Text-
fig. 3) with enclosed and well-developed lens. There are
many resemblances to the Eye of Peripatus. The cuticle, for

THE EYE OF PERIPATUS 7A

example, is continued over the eye without thickening. Below
this, and between it and the lens, there are two cellular layers—
an outer cornea and an inner cornea. ‘These correspond
exactly to the corneal and subcorneal layers in Peripatus.
The lens is non-cellular.

We need not carry our comparisons further ; they may be
summed up as follows: (1) The retina of the Eye of Peripatus
is of a simple and primitive type, and is found again in the ocelli
of certain Diptera and in the eyes of some Myriapoda. It 1s
also not unlike that of some Polychaeta. (2) The dioptrical
parts of the Eye of Peripatus (lens and corneal layers) are well
developed and, as pointed out by Lankester, are arranged
in a manner quite unlike that met with in the Diptera, Myria-
poda, or Crustacea. These parts, on the other hand, resemble
very closely the similar structures of the Polychaete Vanadis.
(3) The Eye of Peripatus possesses some features of a simple
type met with in other Arthropod groups and in the Polychaeta,
but so far as the Arthropoda are concerned it has followed its
own line of evolution and remains quite distinct.

LITERATURE CITED IN TEXT.

_

. Balfour, F. M.—** The Anatomy and Development of P. capensis”,
“Quart. Journ. Micro. Sci.’, Vol. 23. 1883.

2. Carriére, J—‘ Die Sehorgane der Thiere’. Miinchen u. Leipzig. 1885.

3. Hesse, R.—‘* Untersuch. ii. die Organe der Lichtempfindung: Poly-
chaeta ”’, ‘ Zeit. f. wiss. Zoologie ’, Bd. 65. 1898-9. “ Arthropoda”’,
loc. cit., Bd. 70. 1901.

4, Sedgwick, A.—‘* Monog. of the Development of Peripatus
capensis ”, ‘ Quart. Journ. Micro. Sci.’ 1885-8.

5. Lankester, E. Ray.— On the Structure and Classification of the
Arthropoda ”, *‘ Quart. Journ. Micro. Sci.’, Vol. 47. 1904.

6. ——- “ On the Structure and Classification of the Arachnida ’’, * Quart.

Journ. Micro. Sci.’, Vol. 48. 1904.

172 WILLIAM J. DAKIN

EXPLANATION OF PLATE 7.

Illustrating Prof. W. J. Dakin’s paper on ‘ The Hye of
Peripatus ’.

~

Vig. 1—The Eye of Peripatoides occidentalis in vertical
section (longitudinal through the eye). The right half of the retina is
represented in the depigmented condition, the left side in the natural
state. x 740. Cor.=cornea; Cut. = cuticle; Sub. Cor. = subcorneal
layer; Op. N.=optic nerve; Epid.=epidermis; Mus. = muscle-cells ;
L. = lens.

Fig. 2.—Complete rod-cell with rod isolated from the retina. Macera-
tion preparation. 1,500. Pig. = pigment; Nuc. = nucleus of rod-cell.

Fig. 3.—Transverse section through retina in plane of the rods (stained
haematoxylin, Ehrlich). x 1,500.

Fig. 4.—Transverse section through retina, in plane of rod-cells in
the region where pigment is present. (Depigmented section.) x 1,500.

Fig. 5.—Diagrams illustrating the development of the Eye of Peripatus.

(a) Invagination of ectoderm.

(6) Invagination of ectoderm complete.

(c) Ectodermal vesicle cut off.

(d) Proximal cells give rise to retina, the distal becomes the sub-
corneal layer.

(e) Retina developed, lens secreted by cells of vesicle.

Quart Pourn. Mer Scé.Vot.C8,NS.H1.7.
fod

id

esa

35 x
=¢ <=
= a

5. Huth, London.
PERIPATUS.

7 ine
he
7 “ 7

eee oe say

« ®
#)* Thess
1 7 tae z .
ar iy =" oer
a a A
AEC. :
er
: + ;
_ ‘
RS
; abla,
- i?
A it: °
ie
ay : +
Ly sa,
t#.
ar
9 wer
*
‘>
‘
=
‘
€
4
»
*
‘
om ’
'
’
\
. a
- ’
bi
ac
1 ;
Lae
" -
j - i" :
i?
*) { .

On the Development of Cucumaria
echinata v. Marenzeller.

By

Hiroshi Ohshima.

With Plates 8 and 9 and 11 Text-figures.

CONTENTS.

PAGE
1. INTRODUCTORY REMARKS ; ‘ ; 3 ; q eee lifes
2. PREVIOUS STUDIES ON ALLIED FoRMS . A : ; 5 LT
3. MrerHops OF INVESTIGATION . F : ‘ - : Se by)
4. BREEDING SEASON > : : i ; : : 5 tsi!
5. GENITAL ORGANS . : é : : : : , . 182
6. SPAWNING . : : , ; ; 3 4 - SG
7. NEWLY-SHED Eacs ; : : : ; : ; . 188
8. CLEAVAGE . : : ‘ : P ; ; ; S190
9. BLASTULA «. : A ‘ : : - : 3 = 192
10. GASTRULA . ; : , : 3 : : ‘ . 195
11. DreLevRULA F 5 A 3 ; : j ; . 199
12. DoLIoLARIA ‘ : ‘ : : : 4 ; . 204
13. METADOLIOLARIA : : : ; 2 5 : . eS
14, PENTACTULA 3 , ‘ ; ; : j , e221
15. YounG 224
16. SUMMARY . ; ; : : : : F ; 5 BE
17. BIBLIOGRAPHY 240

1. InrRopucTORY REMARKS.

Tue following passages, which are to be found in the diary of
the Marine Biological Station of Misaki, concern the spawning of
Cucumaria echinata, and are written chiefly by the late
Professor Dr. Kakichi Mitsukuri.

174 HIROSHI OHSHIMA

‘ June 18, 1899. Mitsukuri arrived at the Station to-day,
being informed of the fact that in these days Messrs. Aoki
and T'suchida had observed the spawning of Cucumaria
echinata.’

‘ June 20. If some freshly caught specimens of C.echinata
are kept in a glass vessel, it 1s almost certain that they will spawn
in the evening.’

‘July 21. C. echinata spawned !’

‘July 29. C. echinata caught in the morning began to
spawn at 5 p.m.’

‘August 11, 1902. Kuma Aoki dredged several scores of
C. echinata at the mouth of the inlet of Koajiro. They
began to spawn at 7.30 p.m. and went on till about 10.30 p.m. ;
a very large number of eggs were spawned. Mitsukuri
engages in the study of them.’

In his memoir on pedate holothurians (81, 1912, p. 242) the
late professor has recorded with reference to the above facts
as follows: ‘In the breeding season (the summer) the ripe
individuals throw out reproductive elements. The males shoot
forth the spermatic fluid, after which the females begin to
shed eggs, which easily undergo development under observation.’

So far as the results of his study are concerned, unfortunately
no report was made, except his two short addresses delivered
at the monthly meetings of the Toky6 Zoological Society. The
contents of those addresses can now be recovered only from
unpublished notes made by a member who attended the meetings.

The first address was given on December 16, 1899, under the
title ‘General Account of the Embryology of Holothurians ’.
The notes may be translated as follows :

‘To obtain eggs of C. echinata it is necessary to tease
the animal and not to change the water. Animals captured
in the morning will lay eggs in the evening. They are in an
extended posture while laying eggs; the genital products are
emitted from the interradially situated genital pore in the form
of a streak. The amount of the products of each emission is
very remarkable. No peculiar points are noticeable in the
segmentation of the eggs. In the five-tentacled stage a pre-oral

DEVELOPMENT OF CUCUMARIA ECHINATA 5

hood is formed containing a mass of food-yolk of a green colour.
The first pair of pedicels appear on the ventral side, and at this
stage large calcareous plates, which differ in shape from those
found in the adult, appear in the interradii. Next to the five
primary tentacles (Text-fig. 1, I) three more (IL, II,, II,) arise,
of which in most cases the right-hand one (II,) appears first.
The larva attains about 2 mm. in length without having developed
the two remaining tentacles (IIT). The manner of the branching
of the tentacles seems to follow a certain rule, just as in the

TEXxtT-FIq, 1.

V

Diagram showing the order of appearance of the tentacles. Anterior
view. Larger circles (I) represent the primary tentacles; medium-
sized (II), secondary ones appearing next ; and smallest ones (ITI),
those last to appear. mes. =position of the dorsal mesentery.
(After Mitsukuri.)

phyllotaxis among plants, and the second pinnule is the largest
of all (Text-fig. 2).

‘The first pair of pedicels are followed by the third (Text-fig.
3, 3) appearing on the left side of the midventral radius, then
the fourth (4) appears on the ventral side of the right ventral
radius. A further increase of pedicels may be seen in the diagram.
To some extent the pedicels increase forwards from the height
of the first pair, while later some appear behind it.’

The second report was contained in his address on the change

176 HIROSHI OHSHIMA

of calcareous deposits in Holothuria vagabunda, read
on February 21, 1903. Here he spoke again on the sequence of
the appearance of the secondary tentacles. Judging from the
figure he indicated a difference from his former statement, since
the dorsal tentacle (Text-fig. 1, IT,) was shown as the first to
appear among the secondaries.

Having been engaged in the study of holothurians since the
lamentable death of Professor Mitsukuri on September 17,
1909, I have made a new and careful examination of the subject.
On August 12, 1910, my first attempt ended in failure because
the spawning season was already over in that year. During

TEXxtT-FIGS, 2 & 3.

| f\
"6
=
5} 2
eo 2
Me ee
| 2 Hag
ay
1} 4
A 2 B

Diagram showing the manner of Diagram showing the order of

branching. in tentacles of the appearance of pedicels in the
young, A=NSide view of a part young. Ventral view. Numbers
of the stem to show the spiral 1-8 indicate the order of appear-
arrangement of * pinnules’ 1, 2, ance, (After Mitsukuri.)

3. B=Profile of the stem to show
positions of pinnules. (After
Mitsukuri.)

the next few years I was unable to visit Misaki at the right
season, but at last, on July 25 and August 1, 1916, I was fortunate
enough to observe the animal spawn and to rear the larvae.
sy the kindness of Professor Dr. I. [jima I was permitted to
examine the valuable material left by the late professor, on
which, and on my own collection, my studies have been based ;
and incomplete though it may be, owing to several gaps in the
developmental stages in the materials available for examination,

DEVELOPMENT OF CUCUMARIA ECHINATA 177

I believe I have been able to throw some light upon the embryo-
logy of the group.

The present work was started while in the Zoological Institute
of the Tdéky6 Imperial University, but the greater part of it
was carried on in the biological laboratory of the Fifth High
School in Kumamoto, and it was completed in November of
1918. Owing, however, to the pressure of various affairs 1t could
not have been published before I had left Kumamoto the next
summer. The manuscript was then brought with me to London,
and has been subjected to Professor EK. W. MacBride’s
kind and careful revision. It is my pleasant duty here to express
my hearty thanks to Professor I. 1]1ma and Professor 8. Goto
for their kindly supervision, and to various others for their help.
Further, I extend my gratitude to Mr. K. Yoshioka, Director
of the Fifth High School, by whose favour I have been able to
enjoy many facilities to assist me in my studies. Lastly, to
Professor MacBride, to whose courtesy the appearance of
this paper is entirely due, I beg to tender my deepest indebtedness.

9. Previous StuDIES ON ALLIED Forms.

So far as the embryology of apodous holothurians and our
knowledge of typical auriculariae are concerned, the famous
works of J. Miller, Metschnikoff, Semon, Bury,
Mortensen, and Clark are unrivalled; but our knowledge
of the pedate forms is meagre and, for the most part, fragmentary.
This meagreness is due, I think, partly to the fact that artificial
fertilization is very difficult,t and partly to the shortness of the
larval stage, the auricularia stage being usually omitted,? so
that the larva easily escapes the eyes of investigators of the

1 Clark (7, 1898, p. 58), Mitsukuri (80, 1903, p. 11), and
Edwards (12, 1909, p. 212) never succeeded, while Selenka’s
experiment (45, 1876, p. 157) resulted in the malformation. Only one
case of successful artificial fertilization is recorded by Mortensen in
Holothuria nigra, though in this case only a small percentage
of the eggs developed (35, 1913, pp. 17-18).

* Holothuria tubulosa and H. nigra pass through a typical
auricularia stage (Selenka, 45; Mortensen, 35).

178 HIROSHI OHSHIMA

plankton. Among pedate holothurians, the forms which offer
the materials for the study of embryology belong to two families
only, i.e. Holothuriidae and Cucumariidae, both of which are
chiefly dwellers in shallow water. Ostergren (42, 1912,
p. 388) attributed the shortness of larval hfe in Cucumariidae
to the fact that they live near the coast. Any longer period
of larval life, he thought, would expose the larvae to greater
danger of being swept away by violent currents to destruction.

Again, our knowledge of their development is fragmentary,
because, first, many of the observations have been made on those
forms which have brooding habits in which it is not easy to
secure a complete series of developmental stages, and, secondly,
there are some difficulties in the technique of investigation, the
egg being yolky and the larvae mostly opaque.

The embryology of pedate holothurians was first studied by
Danielssen and Koren (11, 1856). Their material was
identified as Holothuria tremula, but Ludwig (21,
1889-92, pp. 249, 251) and Mortensen (88, 1898, p. 24)
alike denied this and suggested that it was dendrochirote. Later,
Louis des Arts (2, 1910, pp. 9-10) and Ostergren (42,
1912, p. 388) have both identified it as Cucumaria fron-
dosa. Kowalewsky (1%, 1867) was the second who had
the good fortune to observe the spawning of C. planei
(=Pentacta doliolum) and C. kirchsbergii (=Pso-
linus brevis); he managed to rear larvae for over ten
days, which attained the pentactula with a pair of primary
pedicels.

Selenka’s work (45, 1876) on Holothuria tubulosa
and Guecumaria planci (=C. doliolum) entered into
a more detailed account than those of his forerunners, since
he employed the paraffin-section method. Ludwig (22,
1891) most excellently explained the processes of organ forma-
tion, and elucidated many points upon the origin of various
organs in C. planci. He escaped the failure which befell
Selenka, since he was successful in obtaining well-orientated
serial sections, whereas Selenka adopted the method of

‘ Masseneinbettung ’, i.e, embedding large numbers of embryos

DEVELOPMENT OF CUCUMARIA ECHINATA 179

together and trusting to chance to obtain some orientated in
the right direction. But unfortunately his materials were fixed
only once every day, so that there were gaps between the stages
he obtaimed. He was able to publish only preliminary notes
without figures, and no final report. In his ‘ Holothurien’ of
Bronn’s ‘Klassen und Ordnungen des Thierreichs’ (21,
1889-92) he summarized the facts known up to that time chiefly
in Labidoplax digitata, Holothuria tubulosa,
and Cucumaria planel.

Mortensen (82, 1894) described in detail the young of
a brooding form, Cucumaria glacialis. Another brood-
ing form, Phyllophorus urna, was then studied by
Ludwig (24, 1898). Lo Bianco’s skill managed to keep
the young of that species alive for two months outside the
mother’s body in a small aquarium.

In EKdwards’s study (12, 1909) on Holothuria
floridana much attention was paid to the development
of the ambulacral appendages, some important details in the
early changes of other organs being left unnoticed. Des
Arts (2, 1910) succeeded in rearing larvae of Cucumaria
frondosa and observed some early changes and pathological
accounts caused by change of temperature. The most promising
work has been carried on by Newth on C. saxicola and
C. norman, yet we know of his results only through a pre-
liminary note (86, 1916).

Thus far the important works so often referred to in the
present paper. Other works left unnoticed in the above enumera-
tion will be cited later on as occasion may require.

3. Mretruops or INVESTIGATION.

The newly-shed eggs were transferred to a larger glass vessel
hy a pipette. The vessel was cylindrical in shape, 28 em. in
diameter and 18 em. in height, and was filled with clean filtered
sea-water. ‘The water was changed the next morning, and was
afterwards left unchanged. The vessel was tightly covered with
a glass plate, and was soaked in cool well-water which was
changed once or twice a day.

NO, 258 oO

180 HIROSHI OHSHIMA

To kill and fix the larvae I used a chromo-acetic mixture
containing a shght amount of osmie acid. Later examination
showed that this fixative proved very satisfactory for the study
of tissues, but for embryological purposes it is bad, since the
free cells contained in the hydrocoele and enterocoele are fixed
in a very expanded state, almost filling up the lumen of these
vesicles. Newth used Bouin’s picro-formol-acetic and
Flemming’s strong solution, and seems to have experienced
a similar difficulty, judging from his following remarks: ‘ Even
after the tentacles are well established, and can be protruded
and retracted, their lumen is obliterated in some places by
the vacuolated inner ends of their cells ’, and ‘ there is a complete
suppression of the typical curved hydrocoele crescent owing to
the large size and close crowding together of its lobes and to the
thickness of their walls’ (86, p. 636).

The late Professor Mitsukuri seems to have used acetic
sublimate, and his materials proved very good for the study
of those internal cavities, which remained very wide and distinct.
Some larvae of C. planci which he obtained at Naples on
March 25-7, 1898, labelled as fixed with acetic sublimate, are
in a precisely similar condition to his material of C. echinata,
Ludwig (22, p. 604) doubted the wisdom of Selenka’s
employment of chrom-osmic mixture, and recommended a careful
alcohol method. For observations on calcareous deposits in some-
what advanced larvae the latter were simply killed in alcohol.

For the orientation of the material to be sectioned, the
double embedding in celloidin-paraffin gave good results. First,
the material was put in celloidin-clove-oil mixture, and then
hardened with chloroform. The hardened block was then
clarified with carbol-xylol and transferred into melted paraffin.
Even by employing the celloidin-paraffin method of sectioning,
shrinkage of the material by about at least 15 per cent. diameter
is unavoidable. That is why the Text-fig. 5, which is drawn
from a whole mount, is larger than other figures obtaimed from
sections. Sections were cut of a thickness of 5p, except in the
case of the egg and quite advanced young, which were cut
into sections 8 p thick,

DEVELOPMENT OF CUCUMARIA ECHINATA 181

The sections were stained with Heidenhain’s haema-
toxylin and orange-G, and very satisfactory results were obtained
even from the material which had been preserved for nearly
twenty years, enabling pictures to be taken by photomicrography.

For the reconstruction of sections, the graphical method with
a glass-plate with parallel lines proved very simple and quite
satisfactory.

4. BREEDING SEASON.

As mentioned above the late Professor Mitsukuri
observed the spawning of Cucumaria echinataon July 21
and 29, 1899, and on August 11, 1902, and Messrs. Aoki
and Tsuchida seem to have observed it previous to June 18.
IT myself observed the same phenomenon on July 25 and
August 1, 1916, but as stated above, on August 12, 1910, I was
too late to secure the spawning individuals,

Though the animal occurs in April and May ‘in such abun-
dance that many boats dredge for them day after day, and by
evening each one is loaded down with them’ (Mitsukuri, 81,
p. 241), no spawning takes place in the evening, and an examina-
tion of the gonads reveals the fact that they are still immature.
It is strange that in July and August the animal does not occur
in such abundance as in Apriland May. Further, from the latter
part of July to the early part of August we meet with minute
young provided with five to ten tentacles and pedicels of varying
numbers, mingled with sand and broken shells, in which adult
animals are found embedded. These young measured only
1-545 mm. in length on August 12-17, 1910, while on July 20-5,
1916, they were remarkably large and in a far advanced state
of development (see Table V).

From these facts we may conclude that the breeding season
begins about the middle of June and comes to an end in the
early part of August, though it varies to some extent according
to different years. The larger young found in the latter part
of July may have developed from eggs spawned quite early in
the season.

Sex-Ratio._Selenka (45, p. 157) noticed that in

02

182 HIROSHI OHSHIMA

Holothuria tubulosa both sexes occur in approximately
equal numbers. Lo Bianco (19, 1899, p. 476) found that in
a certain locality males of Phyllophorus urna oceur
much more abundantly than females. I could pay no special
attention to the sex-ratio of C. echinata during the breeding
season. But of the specimens collected on April 9, 1914, an
examination of over 8,000 individuals showed that there were
1.627 males to 1,596 females, or that the ratio of males per
1,000 females is 1,019. Both sexes thus occur in almost equal
numbers.

5. GENITAL ORGANS.

In addition to the statements of systematists, such as that
given by v. Marenzeller (28, 1881, p. 128), I may deseribe
some points about the genital organs.

Genital Tubes.—Both in males and females the genital
tube consists of an external epithelium, a connective tissue
layer, and an internal germinal epithelium. The external
epithelium, which is continuous with the peritoneum, is very
thick in an inactive stage of the gonads (Pl. 8, fig. 2, ep), and
the connective tissue layer 1s very thin. I have no positive
evidence to prove the presence in C. echinata of the muscle
layer which was found in C. glacialis (Mortensen,
$2, p. 715) and CO. laevigata (Ackermann, 1, 1901,
p. 781). In Pseudocucumis africanus, however,
1 found in the specimens fixed on July 28, 1915, when the
breeding season was over, rather scattered circular and longitu-
dinal muscle fibres, a connective tissue layer, and a very high
external epithelium. Though having the same habit of carrymg
the brood inside the body-cavity, the structure is totally different
from the peculiar feature found by Clark (7, 1898, pp. 58-9)
in Synaptula hydriformis. Here he observed a very
thin external epithelium and scanty connective tissue, which,
according to that observer, probably break so as to allow the ripe
egos to escape into the body-cavity.

Female Gonad.—In a still unripe condition, as was met
with in the specimens collected on March 27, 1914, the female
gonad is light purplish grey in colour, with a very thick wall

DEVELOPMENT OF CUCUMARIA ECHINATA 183

owing to the high external epitheliam as stated above. It
contains eggs of various sizes. The egg is slightly flattened and
is attached to the wall of the genital tube by its broad surface
(Pl. 8, fig. 2), not by a slender stalk formed of the follicular
epithelium, as described by Semper (47, 1867-8, p. 144),
Jourdan (16, 1883, p. 52), and others in Holothuria.!
The germinal vesicle (n) is large and spherical, lying eccentric-
ally near the free end of the egg. The centre of the free surface
is indicated by a minute conical process of the cytoplasm (ma).

The gonad in the breeding season (Pl. 8, fig. 3) has become
thin-walled and yellowish in colour, containing still very small
eges as well as large ones, which latter are ready for spawning.
The germinal vesicle (n) has now approached much nearer the
free end than in the foregoing stage, leaving a thin layer of
yolky cytoplasm between it and the egg-membrane. Germinal
spots (gs) are more mimute and fewer compared with those in
the early stage.

The centre of the free surface 1s now very peculiar in structure.
Here we see a compact cytoplasmic mass, of a somewhat fibrous
nature, forming a short rod-like body protruding through the
ega-membrane, with its free enlarged knob-like end attached
to the folhcular epithelium (Pl. 8, fig. 3, ma). Usually the
proximal end reaches the germinal vesicle and becomes con-
tinuous with its membrane, but in rare cases it ends apart
from the germinal vesicle, where the latter 1s not very near to
this pole of the egg. The space between the egg-membrane
and the follicular epithelium (7) was perhaps occupied by
a gelatinous layer. This remarkable structure, which I may call
a micropyle appendage,” is only conspicuous in full-grown eggs,

! According to Hamann (15, 1884, p. 89) H. tubulosa is excep-
tional ; here a fibrous bundle of connective tissue serves to fasten the egg.

2 As early as in 1851, J. Miller remarked that the ovarian egg of
Pentacta doliolum is flat, and that at the centre of a flattened
surface there is formed a yolky process which passes through the thick
jelly layer investing the egg. He further observed a similar structure
in some other Holothurians—T hyone fusus, Holothuria tubu-
losa, &c. (‘Monatsber. Akad. Berlin’, April, 1851; *Phys. Abhandl.
kén. Akad. Wiss. Berlin’, 1852 (1850), p. 77, Taf. LX, fig.8,9; Miller’s
“Arch. f. Anat. u. Physiol.’, 1854, p. 60).—Jan. 28, 1921.

184 HIROSHI OHSHIMA

while in the very early stages no such structure appears (fig. 1).
In a rather small one I only once made out a slight indication
of this structure.

A similar feature was observed by Semper (47, p. 144).
He found that n Holothuria impatiens and others the
jelly-canal (‘ Mikropylkanal ’) of all the egg is directed without
exception to the internal lumen of the ovarial tube, and in his
figs. 6 and 7 of Pl. xxxviit is shown that the ‘ Stiel des Kernes ’
is penetrating the canal. On the formation of the egg and the
significance of the micropyle he writes as follows :

‘Eine der Zellen des einfachen glatten Hpithels vergrossert
sich und hebt dabei die anliegenden Zellen etwas mit in die
Hohe. In diesem Stadium scheint das Ei lediglich aus einem
Keimblaschen mit sehr geringer Dottermasse zu bestehen. Wie
schon vorher die einzelnen Zellen des Epithels miteinander
zusammenhingen, so bleibt auch jetzt noch die Eizelle in inniger
Verbindung mit den niichstliegenden Epithelzellen, welche sie
bei stetem Wachsthum mehr und mehr mit sich in die Hohe
zieht. . . . die Mikropyle ist allerdings ein Stigma, namlich die
bis zur volligen Reife bestehende Verbindungsstelle mit den zur
Hihaut umgewandelten Epithelzellen, und der Stiel, an welchem
die Kier hangen, erklirt sich auf die einfachste Weise durch
allmiliges Auswachsen und Abtrennen von der inneren Fol-
likelhaut ’ (pp. 144-5).

Tcould not find the youngest stage of egas showing the relation-
ship with the follicular epithelium, as figured by Semper
(fig. 10, a, b, c). But, as stated above, this peculiar structure
develops quite late, apparently simultaneously with the appear-
ance of the gelatinous coating. I feel justified in thinking that
this has something to do with future changes of the egg, above
all with the maturation, and is not a mere vestige of the attached
part of the egg. Hamann (15, pp. 88-9, fig. 3) observed a
similar structure in H. tubulosa. He noticed a cytoplasmic
cord, which breaks through the transparent albuminous layer,
ending in a round nucleus-like body outside the follicular
epithelium. In my case it was different in that the rounded end
always lay inside the epithelium. I could find no trace in any of
iny specimens of the peculiar feature observed by Mortensen
(32, p. 715, Pl. xxxi, fig. 22) in C. glacialis, viz. that the

=e

DEVELOPMENT OF CUCUMARIA ECHINATA 185

chromatic substance gathered on one side of the egg in the form
of a dish.

Male Gonad.—I am unable to give any detailed account
of the male gonad. In the specimens collected during March
I found spermatogonia and spermatocytes near the germinal
epithelium, while the central part of the tube was filled with
unripe spermatozoa or probably spermatids. In the breeding
season the internal space of the tube is filled with ripe sperma-
tozoa, only a thin layer of larger cells—probably spermatocytes
being found near the germinal epithelium. Radiating bunches
of spermatozoa, termed by Mortensen Spermatogemmae,
were found in C. glacialis (82, p. 715, Pl. xxxi, fig. 23) and
were also observed in C. ijimat.

Genital Papilla.—The genital papilla is situated im-
mediately behind the tentacular crown along the mid-dorsal
line. A very singular case is noticed by Ludwig (20, 1887,
p. 1233), who observed that in C. crocea the papilla is located
far backwards, In extreme cases 8 mm. back and far from the
tentacular crown in an Individual, which measured 40 mm. in
length, and 8-5 mm. in another individual, 42 mm. long; that
is one-fifth or more of the body length.

But a more noteworthy fact about the genital papilla was
discovered by Edwards (18, 1910, pp. 338-9, PI. xii, figs. 2-5).
He found in C. frondosa that the male genital papilla 1s
subdivided into from four to thirty or more parts with a general
average of ten, each branch ending in a terminal pore. The
papilla in females is usually simple, ending in a single pore, but
sometimes, though rarely, two or five pores may be met with.
Further, the same author (14, 1910, pp. 599-608, PI. xix, fig. 1)
proved the presence of a similar feature in such allied species
as C. californica, C. miniata, and C. fallax.

CG. echinata offers another example of the same thing. In
a male specimen which I examined the genital pores were at
least fifty in number, while in females there were from five
to about twenty-five pores. Previous to branching, the end
of the genital duct is dilated into a wide cavity just below the
cluster of minute papillae. This cavity, as well as the branched
canals, is lined with a ciliated epithelium followed by a thin

186 HIROSHI OHSHIMA

layer of connective tissue and a layer of loose irregularly arranged
muscular fibres.

6. SPAWNING.

The late Professor Mitsukuri observed on one occasion
that the spawning had begun at 5 p.m., and im another that
it had continued from 7.30 to 10.80 p.m. My experience is as
follows :

At 5 p.m. on July 25, 1916, several specimens were brought
to the Station, and soon after that I found two imdividuals
emitting spermatic fluid. Half an hour afterwards, at 5.30,
ancther individual began to lay eggs. In the morning of August 1
of the same year some few specimens were got and brought into
the Station at 1.15 p.m. Soon afterwards two individuals
were found emitting sperm, but im this case no laying of eggs
was seen to follow. At about 2.30 p.m. of the same day plenty
of large specimens were brought in. Here the emission of
sperm by a male was found to begin at 5.25 p.m., and a female
which was lying about 5 em. distant from the former began to
shed eggs at 5.40, which continued till about 6.30 p.m. In
the specimens of the same lot kept in another jar the emission
of sperm by one began at 4.50 p.m. followed by several others,
but no shedding of eggs took place here.

These specimens were all quite big, and later examination
showed that all of them were sexually ripe and contained sperma-
tozoa or eggs in abundance. In the individuals which missed
ihe chance of shedding genital elements, I neticed that 1t never
took place the next evening or at any later time. Even exposing
the animals in a warm sunny place with very little sea-water,
or putting them in the dark, could not cause them to spawn.
According to Mitsukuri (ante, p. 174) unclean water makes
the animal lay eggs.

The male, while emitting spermatic fluid, stretched out its
tentacles half-way and kept them very quiet. The spermatic
duct could be seen through the body-wall as a white streak
which appeared to perform a peristaltic movement. In conse-
quence of the subdivision of the spermatic duct at the genital
papilla, that white streak could be seen divided into five or six

ee ign em’ &

ee ee ee ee es

DEVELOPMENT OF CUCUMARIA ECHINATA 187

branches, and a white thread-like spermatic fluid flowed out from
each. Being heavier than sea-water, the spermatic fluid sank
down as a milky white cloud on to the bottom of the vessel.

In females the shedding of eggs seemed to be accompanied by
no waving of tentacles, but eggs were thrown out intermittently
on the’ hinder aspect. The egg is much heavier than sea-water
and very soon sinks to the bottom.

According to Kowalewsky (17, p. 1), who put about
fifty freshly-caught individuals of C. kirchsbergii ito a
large vessel through which fresh sea-water flowed, the emission
of sperm by males occurred within two hours. The spermatic
fluid formed a white thread streaming out of a pore situated
between tentacles, and was then stirred up by the waving move-
ment of the latter. The emission of sperm lasted for about an
hour, and in the next hour a female lying near began to shed
egos. He seems to have believed that the egg was fertilized
inside the mother’s body and was expelled through the pore
in the body-wall, and that im a viviparous form, Phyllo-
phorus urna, this pore serves as a birth-pore. Selenka
(45, p. 166) observed no waving of tentacles in the male
C. planci while emitting sperm, and according to him the
females which began to shed eggs as soon as the males emitted
sperm moved their tentacles very actively. Des Arts
(2, 1910, p. 3) put a great number of C. frondosa into an
aquarium, and the first night saw the laying of eggs which
were soon fertilized. According to Newth (86, p. 653) the
spawning of C. norman takes place in the night, and generally
near midnight. On one occasion males and females lying
together in the same tank began to spawn within a few minutes
of one another. Isolated individuals of both sexes are said
to have spawned too, but he was never successful in fertilizing
the egg so shed by adding sperm suspension, and I cannot
help doubting whether he was careful to ascertain that the
females really spawned without being stimulated by spermatic
fluid.

Of other pedate holothurians records are given by Selenka
(45, p. 157) on Holothuria tubulosa, and by Edwards
(12, p. 212) on H. floridana. Among some dozens of big

[88 HIROSHI OHSHIMA

specimens of H. tubulosa kept in a large box, the males
emitted sperm in the form of long white threads at intervals
of two to ten minutes. After some hours fertilized eggs were
found on the bottom of the box. Edwards adopted
Slenka’s live-box method and obtained fertilized eggs within
four to ten hours.

7. NEWLY-SHED Eaes.

The egg is slightly flattened, especially so on the side of the
animal pole (Pl. 8, fig. 4), as is known to be the case in C. nor-
manifrom Newth’s observation (86, p. 633). Along the axis
through the poles it measures about 300-35 p, and the greatest
diameter as measured along the equatorial plane is about 340-
400 », most commonly 4004. Externally the egg is covered with
a radially striated gelatinous layer which is 50-72, thick.
At the centre of the more flattened surface, the animal pole,
the jelly canal can be distinctly seen. The egg is heavier than
sea-water.

According to Kowalewsky (17, pp. 2, 6) the egg of
C. kirehsbergil is opaque with a greenish yolk, and is
heavier than sea-water. The egg of C. plane is said to be
four to five times larger than that of the former species, and,
according to Selenka (45, p. 167), it is ighter than sea-water
and floats immediately below the surface of the water. The
ege of C. frondosa is, as observed by Des Arts (2, p. 3),
intransparent and of a red colour, with a distinct micropyle.
Newth (86, p. 633) observed that the egg of C. normani
tends to float with its animal pole directed upwards, and though
no definite micropyle could be found the ‘ umbilicus’ of the
follicle seemed to act instead.

Remarkable records of large eggs are known among deep-
sea forms, e.g. Enypniastes eximia has an ovarian egg
of 8-0-8-5 mm. diameter, and in both Benthodytes goto1
and Kuphronides depressa the ovarian egg measures
2-5 mm. in diameter (Ohshima, 38, 1915, p. 214). Besides
these cases, large eggs are met with in Cucumaridae, especially
in those forms which are accustomed to care for their brood.

*

DEVELOPMENT OF CUCUMARIA ECHINATA 189

The following table shows the sizes of eggs observed in the
family by various writers :
TABLE I.

Diameter of egg (mm.).
Lag newly shed

Ovarian or found in
Species. Bq. brood-pouch. Observer.
Cucumaria parva 0-2 — Ludwig (28, 1898)
C. echinata — 0-44 Ohshima (40, 1918)
C. frondosa i 0-46 Des Arts (2, 1910)
Psolus granulosus — 0-5 Vaney (48, 1906)
Cucumaria ijimai 0:5-0-55 — Ohshima (88, 1915)
C. crocea 0-6-0-65 0:7 Ludwig (20, 1887;
23, 1898)
C. lamperti 0-8 a= Ohshima (88, 1915)
C. glacialis —- 1-0 Mortensen (82, 1894)
C. lateralis — 1-0 Vaney (48, 1906)
C. curata _- 1-0 Cowles (10, 1907)
C. laevigata 1-0 Lampert (18, 1889)

> 1-34-1-5 Ackermann (1, 1901)
Thyone imbricata 1:2 — Ohshima (88, 1915)

Among the twelve species in the table, there are only two which
have no brooding habit and lay eggs freely in water, namely
C. echinata and C. frondosa; all the others have the
brooding habit.

Maturation.—Examination of sections of the egg fixed
immediately after being shed show that the egg has just extruded
the first polar body (PI. 8, fig. 4, pb), and that the second
maturation spindle (ps) can be seen orientated either obliquely
or vertically with reference to the circumference of the egg,
while the sperm head (sp) has in most cases just entered. When
the spindle is perpendicular to the surface a conical cytoplasmic
process is formed projecting into the canal through the jelly,
through which the first polar body may have been expelled.
According to Boveri (4, 1901, p. 147) the canal through the
jelly of the egg of a sea-urchin, Strongylocentrotus
lividus, is widened at the maturation period and serves
as the way through which the polar bodies are given out.
Selenka (45, p. 167) noticed for the first time a pol: r body
in the egg of C. plancei and described it as ‘ der oth des
Hies’. The egg of CO. normani when taken from among

190 HIROSHI OHSHIMA

the tentacles of the mother is, according to Newth (86,
p. 633), undergoing. or has just completed the second matura-
tion division. In the egg of Holothuria floridana
Kdwards (12) saw three polar bodies, one of which was
remarkably larger than the others.

Fertilization.—The sperm head is still minute and
stains intensely ; it measures about 2°5 in diameter It
is found situated rather peripherally and quite distant from the
animal pole, and rather near the equator (Pl. 8, fig. 4, sp),
so that it is doubtful whether it is at the animal pole that the
spermatozoon penetrates into the egg. It is highly probable
that the spermatozoon enters the egg after, or at the same
time as, the protrusion of the first polar body. In some sections
the sperm nucleus is seen approaching the centre of the egg
and becoming somewhat vesicular. In an egg fixed at fifty
iniputes after being shed the sperm nucleus is found lying
close to the egg nucleus ; it is of the same size as the latter and
encircled with astral rays.

S. CLEAVAGE.

Among the eges fixed fifty minutes after bemg shed were
found some showing the first cleavage spindle lying horizontally
at the centre of the egg. Thus the first cleavage seems to begin
in about an hour. The amount of the material which I was
fortunate to rear was so limited that, from fear of destroying
the whole culture or in any case of losing much before any further
development could be observed, I was unable to examine closely
the living embryos during cleavage, &c. The following state-
ments are given from preserved materials.

The eggs fixed within two and a half to three hours after beng
shed show various stages between four-cell and thirty-two-cell
Stages.

Four-Cell stage.—The blastomeres are equal in size,
clongated along the egg-axis, and flattened or even slightly
concave on the axial surface. Usually the blastomeres inter-

' 'The sperm head before entering the egg measures about 2p in diameter.

DEVELOPMENT OF CUCUMARIA ECHINATA 191

lock, i.e. a pair situated diagonally do not lie parallel to each
other but their ends approach ai one pole, while the other pair
approach at the other pole. In an extreme case, these two
pairs come to lie in different planes, one pair being high above
the other.

Kight-Cell stage.—In consequence of the mterlocking
of the blastomeres 'n the preceding stage, the two tiers of blasto-
meres in the eight-cell stage tend to lie shifting 45° above the
other. Much irregularity in regard to the size and position of
blastomeres is often met with.

Sixteen-Cell stage.—Here eight blastomeres in a tier
lie above the other set consisting of eight. Very frequently.
however, each tier shows a zigzag arrangement, thus the alter-
nating four are slightly above the remaining four, and in an
extreme case they form a tier of themselves at each pole.

Thirty-two-Cell stage.—Now the embryo is globular
in shape ieaving a spacious blastocoele inside it. In the most
regular cases there are four cells in a tier on each pole, and
between these there are three tiers of eight cells arranged in
zigzag rows. No remarkable difference in size among the
blastomeres can usually be discerned.

Above Sixty-four-Cell stage.—Now it is hardly
possible to recognize any special arrangement of blastomeres.
Hereafter those at and near the vegetative pole are found to
be a little larger than those of the opposite pole. No coagulable
matter is as yet found in the blastocoele. The blastula is found
still wrapped within the egg-membrane.

According to Kowalewsky (17),Selenka (45), Edwards
(12),and Des Arts (2), the cleavage of the eggs of C. kirchs-
bergil, C.planci,C. frondosa, Holothuria tubu-
losa,and H. floridana is total and equal or approximately
equal. Selenka noticed that in C. planci the difference of
size among blastomeres is evident only after the thirty-two-
cell stage, and that the cleavage ends at the beginning of the
second day. Edwards stated that in H. floridana the
four-cell stage is reached within three hours and the sixteen-
cell stage within four hours, Des Arts observed in the ege

192 HIROSHI OHSHIMA

of C. frondosa various regular stages of cleavage on the
second day. InC.normaniand C. saxicola, as examined
by Newth (86), the cleavage is not absolutely regular, in that
the four blastomeres may rearrange themselves diagonally, and no
orderly scheme could be detected in the cleavage later than the
sixteen-cell stage. Only in a few individuals of the latter-
named species perfect symmetry up to the sixteen-cell stage
was met with. A very curious feature is seen in 0. glacialis
as reported by Mortensen (82, pp. 722-3). As a remark-
able exception among echinoderms, the cleavage here is said to
be superficial in that the divided nuclei migrate towards the
periphery, increasing in size, and at last there is formed an
epithelium, each nucleus being separated by a cell wall.

9. BLASTULA.

The blastula when free from egg-membrane floats at about
the middle layer of the water, rotating actively by means of
cilia. Its diameter as measured in life is about 83854. Though
I was unable to observe its emergence from the egg-membrane,
the presence of a wrinkled stage inside the membrane is hardly
conceivable in view of the fact that no remarkable increase in
size of the free-swimming blastula as compared with the em-
bryonic blastula is to be found. The wall consists of a layer of
very high slender cells, the vegetative pole being indicated by
a thicker wall. In the blastocoele a coagulable fluid now appears,
known as blastocoele jelly or ‘ Gallertkern ’ (PI. 8, figs. 5, 6, by),
which increases in density with the growth of the embryo.

The blastula of C. kirechsbergii is said to be still
covered with egg-membrane (Kowalewsky, 17). In C.
planei the blastula is formed at the end of the first day
(Ludwig, 22, p. 605) or in ten hours (Kowalewsky,
17, p. 3), and the cleavage ends early on the second day
(Selenka, 45, p. 168). According to Selenka cilia arise
here and there at the end of the first day, and when the cleavage
is ended every cell is beset with a cilium; the embryo then
gets out of the egg-membrane, and swims usually near the
surface of the water. During the course of twelve hours the

DEVELOPMENT OF CUCUMARIA ECHINATA 198

blastula diminishes in size by one-fifth of its diameter, and
the internal cavity becomes filled with blastocoele jelly.
Ludwig (22, p. 605) denied the diminution of size in the
blastula. For my part, I should think an increase of size would
seem more probable in such a form where the blastula is wrinkled
while inside the egg-membrane. Des Arts (2, p. 5) observed
that nm C. frondosa the blastula is formed on the third day,
and that on the fourth day the cells are so multiplied that
many folds appear on the surface and an irregularly formed
internal cavity makes its appearance. As late as on the sixth
day it is still covered with egg-membrane, but it then acquires
cilia and rotates actively inside the membrane. On the seventh
day it emerges from the membrane and is then 405 » in diameter,
and on the next day a thickening at the vegetative pole occurs.
The same author gives further the results of the influence of
the temperature upon the embryo. Besides the syncytium-
formation which usually results by its bemg put in a warm
place, the blastula, being accelerated by warmth, begins to
rotate on the fifth day, and on the next day it casts off the
membrane and the vegetative pole thickens. The discrepancy
found between my culture and those of Mitsukuri with
regard to the growth-rate, as will be stated later on, seems
to be due largely, if not exclusively, to the influence of tem-
perature. I cannot therefore lay much stress upon the time-
record.

Similarly wrinkled blastulae are reported by Newth (86,
p. 633) inC. normani and C. saxicola. In these species
the morula is solid, and the blastocoele first appears during the
formation of a wrinkled blastula. At the latter stage cilia
appear and the embryo soon emerges from the egg-membrane
and begins to rotate slowly at the bottom. The rotation in
C. normani is counter-clockwise in direction as seen from
above, while in C. saxicola it is clockwise. The wrinkled
surface smoothes out before invagination occurs. According to
Selenka (45, p. 160) the blastula of H. tubulosa acquires
cilia near the end of cleavage, and at the twentieth hour it
comes out of the membrane. The blastula of H. floridana

194 HIROSHI OHSHIMA

is, as observed by Edwards (12, p. 213), reached at the
fourteenth hour.

Before invagination begins mesenchyme cells are formed by
the active proliferation of the cells at the vegetative pole
(Pl. 8, fig. 5). Having become free from the wall, these cells
wander into the blastocoele, some lying attached to the wall near
the animal pole (PI. 8, fig. 6, me).

The mesenchyme-formation begins, according to the species,
either before or after the invagination, or sometimes at the same
time as the latter. In C. frondosa and C. echinata
the mesenchyme-formation precedes the invagination. The
same is true for C. planci in normal cases (Selenka, 45)
but it may occur afterwards (Ludwig, 22). In H. tubu-
losa and H. floridana both the processes occur at the same
time, while in Synaptids the mesenchyme cells are formed from
the tip of the already formed archenteron. Ludwig (21,
1889-92, p. 258) noticed this fact and concluded that these
differences are proportional to the rapidity of development.
Thus in a form whose development is rapid mesenchyme is
formed later, and vice versa. I may poimt out further that in
those forms where the mesenchyme-formation takes place
early the cells are generally very numerous and they readily
fill up the blastocoele, while in those where invagmation precedes
the mesenchyme-formation the cells are generally few.

As to the origin of the mesenchyme Ludwig (21, p. 258)
surmised that some mesenchyme cells may arise from the
blastoderm in other places than that where the future endoderm
is situated, and from his study in C. planei (22, p. 605) he
claimed to have proved this statement. His view could not
be confirmed by Newth (86, p. 635), while Clark (7, p. 61),
in his observations on Synaptula hydriformis, felt
‘no hesitation in affirming that the mesenchyme arises
exclusively from the endodermal cells’. It is highly
probable that Ludwig saw those cells attached to the
future ectoderm, as I have mentioned above. I could not,
however, find any positive evidence to support his view, and
in contrast to the vegetative part where many mitotic figures

DEVELOPMENT OF CUCUMARIA ECHINATA 195

are to be met with, no such thing is found in the ectoderm.
Newth observed some enucleated cytoplasmic droplets
attached to the ectoderm. All I have seen were nucleated
cells showing no notable difference from other mesenchyme cells
suspended in the jelly. However, I cannot deny the role that
the ectoderm plays in mesenchyme-formation in a later stage
of the gastrula, followed by the appearance of stomodaeum, as
will be described below.

Selenka (45, pp. 160-1, 168) observed some peculiar cells
consisting partly of those detached from the blastoderm and
partly of those which arose from subdivisions of the former,
and called them ‘ Mesodermkeim ’. Every subsequent observer,
however, denies their presence.

10. GASTRULA.

Invagination begins early in the morning of the next day,
i.e. about at the fifteenth hour. The larva gradually mereases
in length in accordance with the growth of the inmvaginated
archenteron and the multiplication of mesenchyme cells. It
swims with the apical end forwards, at the same time rotating
around its longitudinal axis. Cilia usually beat towards the
oral pole. According to Ludwig (22, p. 605) the gastrula of
C. planci is complete at the end of the second day, while
Des Arts (2, p. 8) records that in C. frondosa the gastrula
is formed as late as on the tenth day. Newth (86, p. 633)
noticed in C. saxicola that the direction of rotation mostly
changes, and at the gastrulation is the reverse of that seen in the
blastula.

The invaginated pit is beset with especially long marked cilia
(Pl. 8, figs. 6, 7, ¢), which remain forming a bundle attached to
the end of the archenteron for some period, still being visible
even when a slight twisting has occurred in the archenteron
(Pl. 8, fig. 8, c). The cells of the archenteron increase very
actively, which fact is shown by many mitotic figures lying
always near the surface and parallel to it (Pl. 8, fig. 7). The
top of the archenteron shows no definite cell boundaries
on the side towards the blastocoele; it here assumes the

NO, 258 P

196 HIROSHI OHSHIMA

appearance of a syncytium. Mitotic figures are found here
and, as a result of rapid proliferation, the cells of the distal
part detach themselves and move into the blastocoele. These
detached cells continue to multiply after bemg free in the
blastocoele. While the free mesenchyme cells are amoeboid
in shape, the dividing ones are readily distinguishable by their
rounded shape (PI. 8, fig. 10).

While rapidly increasing in body length, no mitotic figures
are found in the ectoderm. The cells here seem simply to
decrease in height and to extend in surface. It is thickest at
the hind end near the blastopore, gradually thinning out as it
approaches the apical end. The nuclei le near the internal
surface in the hinder half, while in the apical half they are nearer
to the outer surface. Only in abnormal embryos, which grow to
an enormous size without developing beyond the gastrula, were
many mitotic figures found in the ectoderm.

When the gastrula reaches its full length the archenteron
almost exceeds half the length of the whole body continuing
active cell-division. Selenka (45, p. 164) noticed that the
archenteron lies in H. tubulosa near the future ventral side,
and Ludwig (22, p. 605) also found in C. planei that it
bends slightly ventrad. I have noticed no such feature im the
ease of C. echinata. At about this stage the archenteron
begins to flatten in the anterior portion, and then an unequal
crowth of the wall occurs, resulting in the characteristic twisting
of the hind part.

The above-mentioned bundle of long cilia at the bottom of the
archenteron remains until about this stage. It then seems to
disappear, and in the same place the archenteric wall now
begins to bud off cells into the lumen of the archenteron (PI. 8,
fiz. 9, bl). The cells thus liberated into the archenteron, some-
times called ‘ blood corpuscles ’, vary in amount according to
different individuals, in some being tolerably numerous while
there are none at all in others.

Archenteron.—The archenteron of the fully developed
gastrula is very characteristically twisted in the sinistrorse
direction, and may be described in three parts : the flat expanded

DEVELOPMENT OF CUCUMARIA ECHINATA 197

free end, the transverse ring-shaped middle part, and the longi-
tudinal tubular end opening at the blastopore (PI. 8, figs. 8, 11,
a2 yar 5).

The first part les perpendicularly to the frontal plane (PI. 8,
fig. 11 a, ar,), and its posterior end approaches the left side
(Pl. 8, fig. 12.4, ar). The frontal plane is determined from
the position of the stomodaeum, which soon afterwards makes
its appearance. This part is round in outlme, with thickened
walls near the centre, thinning out towards the periphery, making
the internal lamen appear usually as a slender dumb-bell shape
in transverse section. This resembles the feature found by
Newth in a younger gastrula of C. normani (86, p. 685).
Tt is continuous at its postero-ventral end with the second part.

The second part runs transversely round the dorsal side,
across the mid-dorsal line, and bends ventrad on the right side
of the body, slightly turning anteriorly (Pl. 8, figs. 11 B, 12 B,
ary). The wall is not very thick, the lumen being somewhat
compressed, with the greater diameter along the body-axis.

The third part is directly continuous with the second at the
ventral end of the right limb of the latter, It runs posteriorly, and
is slightly oblique to the left (PI. 8, fig. 12 a, ary). In transverse
section it is round, containing a narrow, often almost obliterated,
internal lumen (PI. 8, fig. 11 c). The posterior end is continuous
with the blastopore.

The internal surface of the archenteron is probably lined with
cilia all over, though I could not demonstrate their presence in
sections.

Selenka (45, p.170) observed in C. planci that when the
tip of the archenteron reaches the centre of the blastocoele it
begins to bifureate. The dorsal branch increases in size very
rapidly, lying obliquely forward and ventrad, while the other
ventrally situated branch remains short. This stage is. said
to have been met with on the fourth day. According to Newth
(386, p. 685; Pl. 8, fig. 8) the circular flat archenteron of the
sastrula of ©. normani turns to bend in an S-shape at right
angles to its plane of flattening, and the anterior flattened sac
hes obliquely to the body-axis.

P 2

198 HIROSHI OHSHIMA

My own observation on the specimens of C. planei brought
back from Naples by the late Professor Mitsukuri shows
clearly that the archenteron is twisted exactly in the same
manner as in C.echinata. It seems to me highly probable
that Selenka’s figure (Pl. xl, fig. 21) was obtained from
a thick section, as he was apparently unable to get a good
series of well-orientated sections. His figure is said to represent
a sagittal section, but. really it is a frontal one. From Newth’s
figure of a longitudinal section of a forty-fourth-hour gastrula of
CG. normani (PI. 8, fig. 8) it is obvious that the archenteron
is not simply folded in an S-shape, but is twisted in a spiral.
The figure, too, is a frontal section, I believe, not a sagittal one
as he supposed, /

It was found by Edwards (12, p. 213) in Holothuria
floridana that by the twenty-second hour a plug of cells
crows out from the blind end of the archenteron towards the
blastopore, and that by this plug the archenteron is divided
into the dorsal and ventral branches. No such changes were
observed by Selenka in H. tubulosa.

In C. planei the position of the blastopore changes, accord-
to Selenka, slightly towards the future dorsal side, but
according to Ludwig it is said to shift ventrad. I could
not decide which of the two holds true in my ease. In most
cases the blastopore opens at the hind end.

Stomodaeum. The stomodaeum makes its first appear-
ance in the quite old gastrula, where the archenteron begins to
divide into hydro-enterocoele and gut (PI. 8, figs. 18, 14 4, s#).
It is preceded by a thickening of ectoderm on the ventral side
at, about the middle of the body (Pl. 8, fig. 11 4, sy). This
is partly due to a sinking down of the ectodermal cells and
partly to an accumulation of the multiplymg mesenchyme
cells. Here a syncytium is formed, the internal surface of the
ectoderm not being a definite one, touching the ventral edge
of the flattened part of the archenteron. The surface of the
latter is still clearly cut, no mitotic figures being found on this side,

The stomodaeal depression then comes in sight a little on the
left side of the plane in which the flattened part of the archen-

DEVELOPMENT OF CUCUMARIA ECHINATA 199

teron has been lying (PI. 8, fig. 14 4, st). These changes very
much resemble the mesenchyme-formation and the invagination
process occurring in the late blastula. Only in this place does
Ludwig’s opinion, that the ectoderm shares in the mesen-
chyme-formation, seem to be true.

This author observed in the third-day larva of C. planeci
that the stomodaeum appeared on the ventral side immediately
behind the pre-oral hood (22, p. 606). According to Newth
(86, p. 634), the stomodaeum appears in C. saxicola and
C. normani in forty-eight hours, i.e. on the middle of the
third day, as a crescentic invagination at the junction of the
opaque and less opaque regions. The horns of the crescent extend
backwards and ultimately fuse up and the enclosed area sinks in.
It lies very obviously to the left of the mid-ventral line as deter-
mined by the pedicels. Similarly a crescentic depression appears
in the second-day embryo of H. floridana, according to
Kdwards (12, p. 213), but it gradually deepens and straightens,
growing out to either side until it extends entirely across the
ventral surface. The plane in which the groove lies is at an
augle of 50° with the sagittal plane of the adult holothurian.

11. DieLeuRULA.

Under the term ‘ dipleurula > I mean the stage which connects
the gastrula with the barrel-shaped larva or doliolaria. This
stage is characterized by remarkable changes occurring in the
archenteron accompanied by a rapid increase of the mesenchyme
cells. As seen from the exterior the larva has become slightly
shorter than in the foregoing stage, the stomodaeum has
appeared, and it differs from the next stage in having no ciliary
bands. This stage is passed during the thirtieth to fortieth
hours, i.e. from the end of the second day till early in the
morning of the third.

Ludwig (21, pp. 274-5) suggested that there might be
a stage, reminding us of the auricularia, during the changes
which take place between the gastrula and the barrel-shaped
stage. In his study of C. planci (22, p. 606) he pomted out.
the fact that the buccal cavity has at the beginning a garland-

2.00 HIROSHI OHSHIMA

shaped thickening on its edge, which he believed to be homo-
logus with the ciliary band of the auricularia. I was unable
to verify Ludwig’s opinion, but the stage which I call
dipleurula is homologous with the auricularia in respect to the
arrangement of the internal vesicles.

The arrangement of the cilary bands as well as the degrees
of development of the alimentary tract cannot, I believe, help
us in discussmg homologies among different forms, because
they vary in degree according to different modes of living. In
the free-living auricularia of Labidoplax digitata the
alimentary canal is well differentiated imto fore-, mid-, and
hind-gut, and the ciliary band is typically developed, as is well
known from the records of many observers. The embryo of
Synaptula hydriformis developing inside the mother’s
body-cavity retains an elliptical shape of body, showing no trace
of any ciliary band, the gut being quite rudimentary (Clark,
7, p. 62). Another viviparous form, Chiridota rotifera,
shows an intermediate feature between the above two (Clark,
9, p. 501).

The division of the twisted archenteron in the old gastrula
occurs first at the postero-ventral end of the second part, where
a solid cell-mass with obliterated lumen connects the divided
portions for a while (PI. 8, fig. 13). The larger vesicular portion,
consisting of the first and second parts of the archenteron is
now to be called hydro-enterocoele or vaso-peritoneal vesicle,
while the smaller one which was the third part is the future
gut. The latter has a very narrow lumen, in most cases bemg
still continuous with the exterior through the blastopore.

The next change occurring in the larva is the displacement,
change of shape, and division of the hydro-enterocoele, and the
enormous multiplication of the mesenchyme cells which fill up
the blastocoele, so that no external examination of the internal
structure on clarified material is now possible.

The anterior part of the hydro-enterocoele, which in the late
gastrula was concave on the right side (Pl. 8, fig. 144, ary),
now moves round to the right across the dorsal side and becomes
narrow in breadth (PI. 8, fig. 164, hy). The posterior part

ve & tee

DEVELOPMENT OF CUCUMARIA ECHINATA 201

of the same vesicle, on the contrary, moves to the left through
the dorsal side (fig. 16 c, en), and, as the two parts thus move
in opposite directions, they gradually begin to be cut off from
each other a little on the left side (PI. 8, fig. 15).

The anterior part, which will give rise to hydrocoele (hy),
gives out from the postero-dorsal margin obliquely backwards
a conical process which finally unites with the dorsal ectoderm
(pe). This is the rudiment of the pore-canal. The posterior
part, which is the future enterocoele (en), is a little smaller
than the anterior part and lies on the left side, extending round
the body-axis and stretching from the antero-ventral to the
postero-dorsal side.

The walls of the hydrocoele and enterocoele consist of a single
layer of cells, clearly distinct from the free mesenchyme cells,
and the latter do not yet attach themselves to the surface of
the former.

The first observer who traced the fate of the archenteron was
Selenka (45, p. 170). He noticed that in C. planci the
archenteron bifurcated at the top, and that the dorsal branch
increased rapidly in size, bending obliquely antero-ventrad,
and at last becomes separated as a vaso-peritoneal vesicle from
the other branch, which latter was stunted and later gave rise
to the gut (‘ Kérperdarm’). In my opinion, his two branches
are a complete vaso-peritoneal vesicle, and he seems to have
overlooked the separation of the gut from that vesicle. He
further stated that after the separation of the two vesicles the
vaso-peritoneal vesicle shifted to the left side of the gut, while
the latter rapidly grew forwards and at last united with a ventral
invagination (‘Munddarm’). He is right in saying that the
first part then lies on the left side, but that the gut breaks
through to the stomodaeum is improbable at such an early
stage, and moreover, the part which he called ‘ KGrperdarm ’
is, I think, to be identified with the enterocoele, which should
never have any communication with the stomodaeum at all.
A careful comparison of his figs. 21, Pl. xi, and 22s, Pl. xu,
leads one to conclude that the part he denoted P (enterocoele)
in the fig. 22 B is derived from that part denoted Bin the fig. 21.

202 HIROSHI OHSHIMA

His fig. 228 resembles very much what I observed in
C. echinata in a corresponding stage.

Ludwig (22, p. 605) observed in the third-day larva of
C. planci that the hydro-enterocoele had separated from the
rest of the archenteron, and again divided into the hydrocoele
and two enterocoele vesicles. Some larvae were somewhat
younger, and in them the hydro-enterocoele was still in con-
nexion with the archenteron. In these statements he seems to
have been unable to give the time and sequence of the separations
of those vesicles. Selenka was of the opinion that the hydro-
enterocoele, at first as long as it is connected with the gut, hes
on the dorsal side of the latter, but shifts to the left side about
the time when the separation sets in. Ludwig observed,
contrary to Selenka, that it had been lying on this side from
the beginning. I agree with Ludwig on this point.

Newth (86, p. 635) could not be sure about the breaking
off of the archenteron in C. saxicola and C. normani,
bemg only able to say that the water-vascular system, the
posterior (perivisceral) coelom, and the gut are separated from
successive regions of the archenteron in the order named, begin-
ning at the anterior end. In C. normani the separation of
the enterocoele was complete by the middle of the third day,
though in some individuals the hydro-enterocoele connexion
was not then broken.

The hydrocoele then increases in breadth again, stretching
round the right side of the body, and its free anterior margin
begins to divide into lobes, which are indistinct at first but
rapidly become distinct processes. These changes as well as
those of the enterocoele to be mentioned below vary very
much according to different individuals. The following state-
ments seem, however, to represent what is most frequently met
with.

Hydrocoele.—In the beginning three lobes are formed
(Pl. 9, fig. 17, hy). The first is narrow, situated at the left
anterior corner, the second is broad, formed of the greater part
of the anterior margin of the hydrocoele, and the third is again
narrow, situated on the right ventral edge of the hydrocoele

DEVELOPMENT OF CUCUMARIA ECHINATA 203

directed transversely. The second broad part then divides into
three lobes almost equal in breadth, while another lobe arises
on the left edge, behind the first lobe (PI. 9, fig. 19). Thus the
hydrocoele is now fan-shaped, with a narrow conical process
directed postero-dorsally, which is the future pore-canal, and
an expanded anterior margin, wavy with six lobes as just
described. From that transverse lobe, at first numbered third,
is formed the mid-ventral radial canal (mv), while the other
five lobes are rudiments of the five primary tentacles (t). Except
the mid-ventral one no other radial canals appear at this stage.
The pore-canal opens to the exterior through the dorsal pore
about the end of these changes (dp).

Kowalewsky (1%, p. 4) and Selenka (45, p. 171) are
of the same opinion that,inC. kirchsbergiiandC. planei
respectively, there are first formed only three tentacles situated
dorsally, and the remaining two appear after the hydrococle
ring has closed. Ludwig (22, p. 608) found, on the contrary,
that in C. planci the five primary tentacles appear simul-
taneously as outgrowths of the radial canals. This divergence
in view from other observers results probably from the fact that
he originally dealt with eighth-day larvae without examining
any earlier stages. I agree with Kowalewsky and Selenka
in assuming the appearance of the tentacles to be earlier than
that of the radial canals, but those three lobes which first
appeared are, im my opinion, not the dorsal three of the primary
tentacles. Lud wig’s account of the early features of the hydro-
coele is very incomplete owing to the lack of any intermediate
stages. Accordmg to him the hydrocoele is of an irregular
horseshoe shape, whose arched part lies dorsally, the right limb
is short, stretching obliquely antero-ventrad (‘nach vorn und
unten ’), and the left limb is longer, stretching postero-ventrad
(‘nach unten und hinten’). I never observed such a condition,
and am convinced that he was in error in these statements,
from which he drew an incorrect conclusion that the hydrocoele
ring probably closed on the right side.

Enterocoele.—psoon after being separated from the hydro-
coele the enterocoele divides into two vesicles, one larger and

204 HIROSHI OHSHIMA

antero-ventral in position, the other smaller and on the left
dorsal side stretching posteriorly (Pl. 9, figs. 17, 18, le, re).
The former corresponds with the left enterocoele of other echino-
derms, while the latter, although situated at first on the left
dorsal side, is the right enterocoele. Selenka (45, p. 171)
noticed in C. planci that the peritoneal vesicle (enterocoele)
divides, immediately after being separated from the vascular
vesicle (hydrocoele), into two ellipsoid vesicles lying on the right
and left sides of the gut respectively. Likewise in H. tubu-
losa the enterocoele which stretches behind and below the gut
divides into two vesicles which he symmetrically on each side
of the gut.

Stomodaeum.—The stomodaeum is formed by an en-
circling of the slit-like depression and a sinking down of the
included area. It contains a thin lumen, extending parallel to
and below the external surface, which opens through a narrow
orifice to the exterior (Pl. 8, fig. 164; Pl. 9, fig. 18 a, st).
The syneytium (sy) extending over the stomodaeum grows
between the hydrocoele and enterocoeles to form a solid cell-
mass running backwards to join with the gut. The gut is of
a single layer of cells but very thick, leaving a narrow lumen
inside (Pl. 9, fig. 18 B, 9).

12. Doxiouarta.

The doliolaria, or barrel-shaped stage, 1s reached about at
the fortieth to the fiftieth hour, i.e. on the third day. This is
characterized by the acquisition of three transverse ciliary
bands on the posterior half of the body, the appearance of
rudiments of the pedicels, and the further development of the
five primary tentacles and radial canals. This stage lasts until
the fourth day or even the eighth day or more.

The larva measures above 500» in length, and swims usually
immediately beneath the surface of the water, being either vertical
or oblique in position. Cilia beat usually towards the posterior

! My own culture showed no evidence of changing into the pentactula-

stage even on the eighth day, when I had to leave Misaki and could not
follow any further changes,

de tae aia a a AD

DEVELOPMENT OF CUCUMARIA ECHINATA 205

end but often reverse, which latter movement makes the larva
sink to the deeper part of the water. Besides these two kinds
of locomotion, rotation around the body-axis is observed at
the same time. In no ease is the pre-oral hood directed down-
wards.

Although no marked change is visible externally, the latter
half of the stage had better be treated under a distinct heading,
Metadoliolaria, owing to its internal changes. Here in the
present chapter I will confine myself to the earlier part, dololaria
in the narrow sense.

In the corresponding stage of C. frondosa, Danielssen
and K oren (11) found that rudiments of the tentacles appear on
the tenth day and a pair of the primary pedicels on the twentieth.
Des Arts (2, p. 9) observed in the same species that the larva
measures on the fifteenth day 510» by 3875p, and that the
tentacles are visible in section on the twenty-first day, but are
observable externally so late as on the twenty-fourth day, and
the pedicels make their first appearance on the thirty-seventh
day. The internal structure of the doliolaria of C. kirchs-
bergii was described and figured by Kowalewsky (17%,
fig. 12). The same author gave an external view of the larva
of ©. planci (figs. 16, 17), while Selenka (45) and Ludwig
(22) made much closer observations. From the observations of
Newth (86), we gather that the corresponding stage in
C. saxicola and ©. normani is not distinguishable
externally from lack of the ciliary bands which are so charac-
teristic of the stage in other species.

Ciliation of the Ectoderm.—the presence of three,
very rarely four, transversely-running bands of cilia is a very
marked character of doliolaria. They seem to appear simul-
taneously. The most anterior band lies about on the middle
of the body (Pl. 9, fig. 25, ¢,), the second and third run parallel
to the former and in such a way that they divide the posterior
half of the body into three equal divisions, or, as is often the case,
the hindermost third is a little broader than the other two
(cy-3). In preserved specimens the cilia are extremely difficult
to make out, but they can easily be found in the living state.

206 HIROSHI OHSHIMA

The length, as roughly measured, is about 254. Much weaker
cilia are found uniformly covering both the parts anterior to
the first ciliary band and that posterior to the third. The areas
lying between the bands seem to be devoid of them.

In C. plancei the ciliary bands present are four (K o wale w-
sky) or very rarely five (Selenka) in number, besides the
uniform ciliation all over the pre-oral hood and anal field. After
the appearance of pedicels and tentacles these uniform weaker
cilia disappear (Selenka, 45, p. 172). Mortensen (88,
pp. 23-4; Pl. i, fig. 8, a, b, c) could not ascertain the presence
of ciliary bands in preserved specimens of doliolaria which are
about 1 mm. long, of a light reddish colour, and which were
found in the Southern Kattegat. He referred them to Psolus
phantapus and suspected the presence of three cilary
bands. Iam much inclined to believe that there are four bands
running along the circular spaces free from calcareous bodies
(fig. 8, c). In C. frondosa we have no record of ciliary
bands (Danielssen and Koren, 11; Des Arts, 2),
and in C. kirchsbergii the bands seem to be really absent
(Kowalewsky, 1%). Doliolariae of C. saxicola and
GC. normaniare ciliated uniformly all over as in other stages,
no segregation of cilia into bands bemg found (Newth, 86).
The larva of Phyllophorus urna, too, shows no zonary
distribution of cilia while actively swimming inside the mother’s
body-cavity (Kowalewsky, 1%, p. 7).

When examined in section, the cila are very obscure and
markedly short, due to shrinkage. The ectoderm is thickened
at the band, bemg about twice as thick as other parts, and of
a lens shape in transverse section (PI. 9, fig. 25, ¢,-3). The nuclei
are situated near the base of the cells. As to these cells I could
find no distinction between * Wimperzellen’ and ‘ Reserve-
zellen’ as Reimers (48, 1912, p. 270) did in his observations on
the larva of Labidoplax digitata. Further, I could not
clearly make out either ‘ Binnenfaser’ or ‘ Basalstiibchen ’ as
clearly figured by him (PI. ii, figs. 5-9).

Hydrocoele.—the lobe no. 6 of the hydrocoele, as
numbered from the dorsal one towards the ventrum, stretches

DEVELOPMENT OF CUCUMARIA ECHINATA 207

out ventrad, bringing together with it the lobes nos. 4 and 5,
and on reaching the mid-ventral line it turns posteriorly to give
rise to the mid-ventral radial canal (PI. 9, fig. 21.4, mv). The
other lobes except no. 1 are now differentiated into cylindrical
tubes—primary tentacles (f,-;)—connected at the base by
a rather narrow canal, which forms a horseshoe-shaped rudiment
of the rig canal. The lobe no. 1 eventually gives rise both
to the remaining one of the primary tentacles and to the free
end of the dorsal limb of the open hydrocoele ring. It remains
for a while as an inconspicuous outgrowth.

Radial Canals.—The free end of the rudimentary mid-
ventral radial canal then dilates laterally to form a rhombic
vesicle in ventral view, and then takes on a cross shape (Pl. 9,
fig. 23.4, mv). The transverse branches thus formed are the
primary pedicel canals (rpe, Ipc), and in correspondence with each
of them a rudiment of the primary pair of pedicels (rp, lp) is formed.

The four radial canals, other than the mid-ventral one, are
formed comparatively late, especially the ventral pair (rd, Id,
rv, lw). They arise at first as small knobs on the anterior margin
of the rmg canal, one in each interval of the primary tentacles.
The knobs then bend outwards and immediately turn pos-
teriorly, and they are remarkably thin as compared with the
mid-ventral one.

Kowalewsky (1%, p. 4) first. described in C. kirchs-
bergii the first appearance of the mid-ventral radial canal,
which soon divided into two, pushing the body-wall outwards
to form a pair of pedicels. In C. planci Selenka (45,
p. 171) aseribed the development of the mid-ventral radial canal
to too late a period, stating that it made its first appearance
soon after the closure of the hydrocoele ring, and that four
other radial canals and the Polian vesicle followed it. As to the
fact that the four radial canals, other than the mid-ventral one,
appear later than the latter, all observers are unanimous.
Ludwig (22, p: 181) further noticed that among those four the
ventral pair are shorter and narrower than the dorsal pair, the
difference being observable throughout the life of the young ;
they grow to be equal much later on.

208 HIROSHI OHSHIMA

Primary Pedicels.—A pair of the primary pedicels are
first indicated by cireular pits formed on the ectoderm corre-
sponding to the pedicel canals branching from the mid-ventral
radial canal (Pl. 9, fig. 22D, p; fig. 24 Fr, lp, rp). These I may
call pedal pits. The formation of the pedal pits a good deal
resembles that of the stomodaeum, the rudiment of the
pedicel being formed by the syncytium helow the pit and arising
from the bottom of the pit covered by the ectoderm. Finally
it projects from the pit, the latter being soon flattened out.
These changes strikingly resemble those found in the primary
tentacles. The pits are situated between the second and third
ciliary bands and at an angle of about 40° on each side of the
sagittal plane. Specimens are often found in which only one
of the pair has just appeared. Of six cases of such specimens
I observed all had only the left pedicel (PI. 9, figs. 21, 22, p).
Out of seventeen cases where both the pedicels had appeared,
eight cases showed that the left pedicel hes more or less anteriorly
to the right one, while seven cases were the reverse, and in
the remaining two cases the two lay on the same level. Thus
we can find no constant feature as to the relative position of the
two primary pedicels.

According to Ludwig (22, pp. 185, 607), in C. planei
the pedal pits appear in most cases near the end of the fourth
day, and the pair of pedicel. canals appear from the mid-ventral
radial canal either at the end of that day or early on the next.
Of the pair of pedal pits the right one always lies a little anterior
to the left one, and the same holds true in Phyllophorus
urna as observed by the same author (24, p. 97). Newth
(36, p. 637) found in the third-day larvae of C. saxicola
and GC. normani that the posterior end of the mid-ventral
radial canal formed a rhombic dilation representing the rudiments
of pedicel canals. Here, he says, the left pedicel hes further forward
than the right, just contrary to the feature seen in C. planei
and Ph. urna.

From Edwards’s observation (12, pp. 222-3) we learn that
in H. floridana the first pedicel is unpaired and appears
at the posterior end of the mid-ventral radial canal. It is said

‘DEVELOPMENT OF CUCUMARIA ECHINATA 209

that J. Miller had observed in the young developed from
* Auricularia mit Kugeln’ that the first pedicel was put forth
on the right from the mid-ventral radial canal (Ludwig,
21, p. 291).

Primary Tentacles.—Now the five rudiments of
primary tentacles can be referred to the respective interradius
(Pl. 9, figs. 2834,B; 24a-p; #,). No. 2 (t,) lies in the mid-
dorsal interradius, no. 3 (t;) in the right dorsal, no. 4 (t,) in the
right ventral, no. 5 (t;) in the left ventral, while the remaining
one, no. 1 (t,), which should arise from the dorsalmost lobe,
and is the last to appear, is in the left dorsal interradius.

In their early stages the five primary tentacles arise directly
from the ring canal, or more properly speaking, the hydrocoele
differentiates into the tentacle-rudiments and the ring canal.
In the course of growth a peculiar grouping of tentacles begins
to appear in relation to the radial canals (Text-fig. 4, a—c).
It seems very probable that the rudiments of the radial canals
develop at the expense of the adjoming parts of the hydrocoele
ring, attracting towards them the bases of the tentacle-rudiments,
so that the latter seemingly become branches given out from
the radial canals, totally independent of the ring canal. The
srouping of tentacles is precisely the same as that found by
Ludwig (22, pp. 188, 608) in C. planei, i.e. the ventral
pair gather at the mid-ventral radius, that of the right dorsal
interradius moves towards the right dorsal radius, and the
remaining two meet with the left dorsal radial canal (Text-
fic. 9a, I, I’). Ludwig insisted upon the opinion that all
the five primary tentacles appeared simultaneously as branches
of the radial canals, not directly from the rig canal. One must
keep in mind that Ludwig’s first observations were made
on the eighth-day larvae, which are quite well established dolio-
lariae, and that in his second report he tried to trace the fact
as far back as to the fourth-day larvae. Newth (86, p. 637-8),
confirming Selenka’s view, claims that the primary tentacles
are originally interradial in position, arising from the ring canal
directly and alternating with the radial canals. He found that
some individuals had shown on the third day that curious grouping

210 HIROSHI OHSHIMA

of the three dorsal tentacles, while still on the fourth day the
ventral two retain their original interradial position, their
bases, however, beginning to approach the mid-ventral radial
eanal. According to Edwards (12, p. 216) the feature is quite
different in H. floridana. Here on the fourth day each

TEXxtT-FIG. 4.

Cross-sections cut at the height of the ring canal, to show the
gradual displacement of primary tentacles. <A. Doliolaria,
reconstructed from several sections. B. Metadoliolaria, (, Still
advanced metadoliolaria. x 200. ep =epineural canal; g =
gut ; le = left enterocoele ; Jv = left ventral radial canal; mv =
mid-ventral radial canal ; re=ring canal; re= right entero-
coele ; rv = right ventral radial canal ; ¢,, = primary tentacles ;
v= ventilating apparatus.

of the mid-ventral and the left dorsal radial canals produces
a tentacle on the right side, and each of the paired ventral

radial canals sends out one on the dorsal side. Of these four,
the former two seem to be the first to develop. The fifth appears

eer 4H eh SG OL RE

= or. aa

DEVELOPMENT OF CUCUMARIA ECHINATA DA la |

last on the left side of the mid-ventral radial canal (‘Text-fig. 9B,
iF 1’); this attaims a size equal to the other four as late as
the seventh day. Becher (8, 1908, p. 4) attributes this differ-
ence between Cucumaria and Holothuria to the higher
tentacle number of the latter in the adult. This must not be
rashly concluded since the feature in question has not yet been
recorded in other many-tentacled forms such as Phyllo-
phorus, Pseudocucumis, &e. Apart from the differ-
ence in the sequence of their appearance, the distribution of
the primary tentacles with regard to the radii does not differ
very much in Cucumaria and Holothuria. Indeed,
the two agree in the circumstance that the tentacles alternate
with the radial canals as pointed out by Newth (p. 689).
The only notable difference is that in Cucumaria each of
the pair lying in the dorso-lateral interradii is supplied by
a tentacular canal sent ventrad from the dorso-lateral
radial canal of each side, while in Holothuria the corre-
sponding tentacles are supplied by branches sent dorsad
from the ventro-lateral radial canals. Now, if we are right
to admit that the primary tentacles all originate directly from
the ring canal, but not from radial canals, such difference as
found between these two genera does not seem to me so great
and fundamental. As will be shown later,in Cucumaria the
dorsal pair of radial canals grow faster than the ventral pair,
so that the bases of the tentacles in question shift dorsally and
at last become branches of the dorsal pair. In Holothuria,
on the contrary, the ventral pair of radial canals, bemg more
vigorous in growth than the dorsal pair, might have conquered
in pulling together the bases of those tentacles. As a matter of
fact, in Holothuria the ventro-lateral radial canals are
in adult state generally more strongly developed than the dorsal
pair, while in Cucumaria these two pairs differ very little.
Another common feature in both cases is that two of the
primary tentacles belong to the mid-ventral radial canal. It
is a noteworthy fact that even in such a form as C. echinata,
whose ventral pair of tentacles are markedly smaller than the
remaining eight in the adult stage, the former are represented
NO, 258 Q

212 HIROSHI OHSHIMA

in the five primary tentacles (compare Becher, 38, p. 4).
It is an open question how the mid-ventral radial canal behaves
in the early stages of Sphaerothuria bitentaculata,
which possesses only eight tentacles when adult, the mid-
ventral radial canal supplying no tentacles at all.

Polian Vesicle.—tThe Polian vesicle appears at the free
end of the ventral limb of the open hydrocoele ring, directed
posteriorly and lying inside the enterocoele vesicle (Pl. 9,
figs. 23.4,B; 24D; pv). It may often appear after the closure
of the ring.

Stone-Canal.—About at the same time as the appearance
of the Polian vesicle the pore-canal swells out dorsally at. its
middle part, and as a result of it the canal slightly bends at this
point (Pl. 9, fig. 25, as). In transverse section this swelled
part shows a very characteristic feature in that the wall of the
axial side is of a very high epithelium, while along the dorsal
side the wall is very thin. Bury (5, 1889, p. 427; Pl. xxxix,
fig. 26) first noticed this structure in a Cucumaria, and
considered it as a vestigial anterior enterocoele, the presence
of which he had proved in auricularia. Ludwig (22, p. 609)
observed it in the fifth-day larva of C. planei and ealled
it the madreporie vesicle, with an assumption that it is only
a secondary outgrowth. In his later paper Bury (6, 1895,
pp. 53-4) insisted upon his former view, and suggested that
future and closer examinations would reveal changes similar
to those in auricularia, proving its origin from the enterocoele.
According to Newth (86, p. 637) this enlargement occurs
by an up-pushing of the antero-dorsal wall of the canal on the
third day in C. saxicola andC. normani. On the next
day the cells of the antero-dorsal wall of this vesicle become
large and clear. This swelling up of the cells seemed to him
to be a preliminary stage in the thinning out of the part as seen
by Bury, Ludwig, &e. I was unable to find either the
change which Bury suggested to be present or the swelling
up of the cells in the early stages of this structure. The same
structure has further been proved to be present in Phyllo-
phorus urna by Ludwig (24 p. 98) and by Russo

DEVELOPMENT OF CUCUMARIA ECHINATA 213

(44, p. 45; Pl. ui, fig. 52), n Holothuria floridana
by Edwards (12, p. 214), n Cucumaria crocea and
another antarctic Cueumaria by Mac Bride (25, pp. 7, 8;
27, p. 4). The last-named author called this vesicle the axial
sinus.

The distal portion of the canal, which should now properly
be called the pore-canal, runs through the dorsal body-wall
and opens to the exterior. The opening, or dorsal pore, is situated
between the second and third ciliary bands, and is in most
cases slightly on the right of the mid-dorsal line (PI. 9, fig. 22 B,
dp; fig. 245, pe). Ludwig (22, p. 186) also found that the
pore opens on the right.

Closure of the Ring Canal.—From the fact that
the rudiment of the left ventral radial canal appears on the
ventral limb of the open hydrocoele ring, while that of the left
dorsal radial canal belongs to the dorsal limb of the same, it is
clear that the closure of the ring occurs on the left dorsal
interradius. The Polian vesicle lies at first very near to the left
ventral radius, but later it moves towards the middle of the
dorsal interradius, which is its normal position as found in
adult individuals.

As to the time and position of the closure of the ring no
entirely satisfactory observations have been given. Kowalew-
sky (17, p. 4) and Selenka (45, p. 171) were in agreement in
the opinion that the ring closed after the formation of three
dorsal tentacles, while the remaining two developed from the
closed rng. Ludwig (22, p. 607) observed in C. planei
that the rmg was complete at the end of the fourth day, and
the closure seemed to have taken place on the right side of the
body. From this incorrect view he concluded that the Polian
vesicle which lay on the left dorsal interradius could not be
an indication of the point of closure. Newth (87, p. 637)
is quite right in concluding that in C. saxicola the ring
closed in the left dorsal interradius on the third day, when
the radu can be identified. He could not determine to which
limb of the free ends of the unclosed ring the rudiment of
the Polian vesicle belonged, being only able to say that it

Q 2

914 HIROSHI OHSHIMA

was found as a small blunt outgrowth produced at the point of
closure.

From the table on p. 215 the following features may be
summarized :

1. Of the five primary tentacles the one situated in the left
dorsal interradius appears last (nos. 1-4).

2. To the four tentacles and one mid-ventral radial canal the
right dorsal radial canal is first added (nos. 2-4).

3. The left dorsal radial canal appears at about the same time
as the appearance of the fifth tentacle, after which the right
ventral radial canal follows immediately (nos. 5-7).

4. The appearance of the left ventral radial canal is still
later (no. 10).

5. The appearance of the Polian vesicle in some cases precedes
the closure of the ring canal (nos. 8, 9, and 12), and in others
it is later (nos. 6, 11, 18, and 15).

6. The formation of the axial sinus also in some cases
precedes the closure of the ring (nos. 8, 9) and in others it is
later (nos. 6, 11).

7. The closure of the ring takes place in most cases after five
radial canals have all appeared.

Stomodaeum.—Now the position of the stomodaeum
can be determined by the establishment of the mid-ventral
radial canal. It lies in front of the first ciliary band, and at about
30° to the left of the sagittal plane. The ectoderm covering the
interior of the atrial cavity is pushed up by the growing tentacles,
forming an epidermal covering for the latter (Pl. 9, fig. 25, at).
The orifice is often found plugged up by the left ventral tentacle
which lies nearest to the stomodaeum (‘Text-fig. 5, t,). Newth
(36, p. 634) noticed the asymmetrical position of the stomodaeum,
while Ludwig (22, p. 610) observed the same fact but inter-
preted it erroneously. He was of the opinion that the larval
symmetry plane is not coincident with that of the adult, and
thought that the left ventral tentacle stands nearest to the
mid-ventral line.

Simultaneously with the growth of the primary tentacles
and the diminution of the pre-oral hood, the stomodaeum

tll 9c he ah oy el Io on pI WW

> eeeepasg gaye Pm

Tt BS RL

OF CUCUMARIA ECHINATA

DEVELOPMENT

- ‘$z ‘Zz ‘s3y “G ‘Iq Ul UMOYsS st UOMMDAdS STITT, 5 ‘ez ‘Sy ‘G ‘Iq Ul UMoYsS st uoMMTOAdS sIYT, ,
‘ZZ ‘IZ “SB ‘G ‘Tq Ul UMOYs st uOMIIOEdS SILT, g ‘uado ood

[estop oy} eAvy suouTIOeds quonbasqns ]je pue Z “ou 4ynq ‘7 ‘ou ur AT[estop uedo 4aX you st euvo-orod oy], 1
x x d “pd d *d “d x x x x x ) € 91
x _ d “pd d | ‘d x x x x x 9 g GI
x x D “pd 1 “d ‘d x x x x x H) G FI
x _ » “pd 1 ‘d “d x Xx x x x i) G €1
~ % 1 “pd 1 ‘d 1 x x x x x 0 ra rol
= = D pd 1 ‘d 1 x x x x x °0 j Il
x x q “pd 1 ‘ad d ~ x x x x 4 G OL
x x _ “pd 1 ‘d ‘d x x x x x ‘0 z 6
oe x _ “pd 1 ‘d “d x x x 4 x x0) V6 283
== = — a D “| ff] x x x x x ‘oO 4 L
— — — of] — 4] 1) x x x x x oH) v2 9
— — — “Y ” ‘D q x x x x x ‘0 Zz G
“ = = ‘y — "1 - x x x x = 0 z ij
_ — ~ 4 -- D - x x x x _ 0 ra ¢
— — _ “of — ‘D - x x x x - 0 GY] ae
_ — _ “y — — _ x x x x = ‘0 é I

venue "yon = AT “AW ‘AN ‘GH ‘GT (mp) (a4) (ps) (“pm) = (“py) = youno = shpp = ‘on

ixp woyod G P € j I buy uy aby

*s]DUD wIpHYy "saqonquay Havre]

‘(pd) spewed joorpad ‘soyouraq petozxy] Jo aed v poUrtoy sey AO “(4) epIsea oIquIoYyA v oFUT popurdxea
AYSTS st 10 (4) qouy v Jo wWaAOF oy} UL ATduMIs spUe [VU [LIP [VIPUOA-prUt oY} OYJ M + (d)
Ajxorteysod 10 ‘(7) spavaqno ATpet9ye] Io ‘(y) ATLOLIayWR poyoortp Io ‘(q) spnq ATUO o.rv speUBO TeIpeT Jo
S}USMAIpNA oy} LayyoyM { (9) posop 10 (0) uado [[ys st euRo Sua oYy AoyoyA + (—) yok you AO (x)
pout} 9IV SNUIS [VIXB pUR ‘opISEA ULT]Od ‘SepPvyUE} of} ToYJoOYM pozVOTpUL ote 9TQv} OY} UT
‘[eURO SUIA OY} JO ANSTO Jo oUTT] ayy puv sadepuadde spooooapAy oy} Jo yuouudoparop oy} MOUS O],

TUTE Share

216 HIROSHI OHSHIMA

gradually widens and at last flattens out, so that the tentacles
freely protrude above the body-surface.

Alimentary Canal.—tThe rudiment of the gut has been
growing both in length and diameter by rapid cell-division
and by increase of the cells in height. Its anterior end extends
beyond the ring canal by which it is encircled (PI. 9, fig. 25, q),
and while both ends remain solid its middle portion has a dis-
tinetly discernible flat lumen lying parallel to the frontal plane
(Pl. 9, figs. 24u, F; Q).

Enterocoeles.—Early in the doliolaria stage, where
four of the primary tentacles have become apparent, the right
and left enterocoeles come into contact with each other at their
free margins. Both the enterocoeles have been rapidly growing
in size, extending across the median line and encireling the gut.
The fusion of their ends takes place on the right side, beginning
either at the anterior part or at the posterior part of the lme
of contact, leaving for a while an oblique incision at either end
of the line (PI. 9, figs. 21, 23 ; re, le).

The other ends of the two vesicles approach each other but
are separated by a narrow interval. This intervening part gives
rise to the dorsal mesentery in the end, and hes at first obliquely
on the left side, beginning anteriorly near the mid-dorsal line to
end near the mid-ventral line. It, however, gradually bends into
an S-shape, indicating the three sections as found in the future
mesentery—the first, mid-dorsal and descending section; the
second, oblique and ascending section on the left; and the
third, descending section running along the mid-ventral
line.

[ failed to find any ‘ finger-like process’ as seen by Bury
(6, p. 48) in Synaptids and verified by others. Though very
often there appears a process on the antero-dorsal end of the
enterocoele, stretching beyond the primary stone-canal to the
left, I could not follow its fate, and am uncertain whether
the peripharyngeal sinus originates from it or not.

The behaviour of the enterocoeles in C. planei was first
observed by Selenka (45, p. 171), according to whom the
union of the right and left vesicles takes place on the ventral

DEVELOPMENT OF CUCUMARIA ECHINATA 217

side so soon after the separation into two from the original
single vesicle that he at first overlooked this separation.
Ludwig (22, pp. 609, 611) observed in the fourth-day larva
that the right and left enterocoeles extended around the gut
so as to meet and break through on the ventral side, while they
remain separate on the dorsal side. At the end of the sixth
day the rudimentary mesentery begins to bend, the last section
lying on the right ventral side of the body. From Oster-
gren’s comparative study in the Dendrochirotae (41, 1898)
it has been shown that the last section of the mesentery does not
lie on the right side of the mid-ventral radius, but with the
exception of the Psolinae always on the left side. Ludwig
was wrong in this respect.

The above feature is essentially the same in the Synaptids.
I may only point out that the pomted ends of the two entero-
coeles unite on the right side of the mid-ventral line
(Reimers, 48, p. 280).

Blastocoele Jelly and Mesenchyme.—By the
time that the dipleurula is reached the pre-oral hood is filled up
with blastocoele jelly. It consists of a structureless gelatinous
substance and a few sparsely arranged mesenchyme cells sus-
pended in it. The former stains with plasma-dyes and often
shows a netted appearance in some fixatives. This substance
is seen most developed in the doliolaria stage, while near the end
of the late doliolaria it gradually diminishes, probably being
absorbed as nourishment.

Most of the other mesenchyme cells gather thickly around the
hydrocoele, enterocoele, and gut, without, however, forming a
definite cell-layer of any kind. Others lying below the ectoderm
form a loose connective tissue of cutis. Metschnikoff
(29, p. 4) showed in a Synaptid the origin of the cutis from
mesenchyme. Selenka (45, p. 169) opposed this view, claim-
ing that mesenchyme gives rise to musculature only. Later, he
(46, p. 57) corrected his former view admitting that mesenchyme
gives rise to connective tissues, and on the other hand that the
musculature of some parts origmates from other sources than
the mesenchyme.

218 HIROSHI OHSHIMA

18. METADOLIOLARIA.

Near the end of the doliolaria stage many important changes
occur internally, though but few changes are seen from the
outside. When seen externally, the pre-oral hood gradually
diminishes in size, and in consequence the tentacular crown
shifts anteriorly, calcareous deposits appear while the ciliary
bands degenerate, and the tentacles and pedicels become
prominent and visible from the outside. The internal changes
are: the further development of the hydrocoele appendages
into the adult water-vascular system, the differentiation of
musculature and nervous tissue, the widening of the enterocoele,
&e. This I may call the metadoliolaria stage. The larva now
very often sinks to the bottom from its increased specific gravity
and degenerated ciliary function.

Water-vascular System.—aAs was stated by Lud-
wig and Newth, thering canal lies a little obliquely in such
a direction that its dorsal half approaches the anterior end of the
body rather more than the ventral, but I could not find any
lateral inclination such as was observed by Ludwig, who
stated that the left half is slightly more posterior than the
right (22, p. 181).

In my culture the tentacles begin to protrude a very little
above the surface of the body at the end of the fourth day,
and early in the morning of the next day a slow movement
was observed, obviously owing to the differentiation of muscle
fibres in their wall. The tip is found covered with minute hyaline
papillae as known to Selenka, Ludwig, and _ others
(‘Text-fig. 5, p). The ramification occurs on the seventh day.
The primary pedicels now protrude as short cylindrical promi-
nences, as Clearly seen in the sixth-day larva.

Musceuli » fibres are now to
be found below the hydrocoele epithelium in the tentacles and
pedicels, and along the radial canals. They appear in the ten-
tacular wall first along the internal (axial) side and then spread
around the cavity. Those of the radial canals lie along the

d
|
q

DEVELOPMENT OF CUCUMARIA ECHINATA 219

internal (axial) side between the hydrocoele and_ peritoneal
epithelia.

In C. planci Ludwig (22) observed the first appearance
of muscle fibres in tentacles on the seventh day (p. 612), in
pedicels on the tenth day (p. 185), and along the mid-ventral
radial canal on the thirteenth day (p. 182). According to him,
all these are derived from the hydrocoele epithelium.

Nervous System.—tThe nervous tissue is well marked
in this stage. Immediately below the atrial cavity the ring
nerve 1s formed, encircling the still closed anterior end of the
gut. Anteriorly a branch, the tentacular nerve, is put forth in
each interradius to run along the oral side of the tentacle.
Posteriorly the five radial nerves appear, of which the mid ventral
is the strongest. The latter gives out a pair of branches to
the primary pedicels.

Along the oral side of each tentacle, a part of the atrial cavity
extends backwards as a thin flat canal. On reaching the ring
nerve, these canals unite with one another to form a circular
canal above the former. This is the epineural ring. From
this the epmeural canal is sent out along each radial nerve
(Text-fig. 4,B,c,ep; Pl. 9, figs. 24, a—n, enc).

I can give no further account, and will refer to Ludwig’s
detailed descriptions on the origin, differentiation, and develop-
ment of the nervous system given for C. planci (22). Accord-
ing to him, the rudiments of the nervous system first
appear on the fourth day (p. 608), the epineural ring and
canals are formed on the fifth day (p. 609), differentiation of
fibrous structure takes place on the sixth day (p. 611), the
tentacular nerves are formed on the ninth day, and the pedal
nerves are given off on the seventeenth day (p. 188). In C.
echinata I could make out all these features even in the
fifth-day larva.

Calcareous Deposits.—In my culture calcareous
deposits made their first appearance on the sixth day. They
occur at three places, i.e. in the wall of the axial sinus (‘lext-
fig. 5, mvp), at the bases of the tentacles (cr), and in the integu-
ment of the pdsterior part (cd).

2.2.0 HIROSHI OHSHIMA

The deposits formed in the wall of the axial sinus consist of
a loose basket-work, which forms a short tube opening on both
ends surrounding the stone-canal. There is another opening
which is directed dorsad, corresponding to the thin-walled part
of the vesicle. Ludwig (28, p. 27) noticed a similar structure
in the pentactula of Cucumaria parva, with a wide open-
ing directed anteriorly.

TExtT-FIG. 5.

Ht Steet eel Vey

Seventh-day metadoliolaria, Right-side view to show calcareous
deposits. x 100. cd = deposit of integument; cr = rudiment
of calcareous ring ; en = enterocoele ; g = gut ; mp = axial sinus ;

= papilla on the tip of tentacle; pr= pre-oral hood; ¢; =
tentacle.

Those which appear at the bases of the tentacles are a delicate
netted rig, giving off a pair of anteriorly-directed pointed
processes at each radius. These represent the rudiments of
the radial segments of the calcareous ring. It has been shown
by Ludwig (22, p. 611; 23, p. 27) and Clark (7%, p. 67)
that the calcareous ring is first represented by five radial seg-
ments. In C. echinata it does not consist of five separate
pieces but of a continuous ring, as stated above. »

DEVELOPMENT OF CUCUMARIA ECHINATA 221

Those which appear in the integument increase and develop
rapidly and soon cover the body on its posterior half. Their
shape is not quite regular, but is commonly a delicate lattice
plate formed of successive dichotomous branchings of the
original primary cross. ‘They he parallel to the surface em-
bedded in the dermal connective tissue formed below the ecto-
derm.

According to Kowalewsky (1%, p. 6), m C. kirchs-
bergii the calcareous body first appears in the wall of the
stone-canal. Ludwig (22, p. 610) found in C. planci that
deposits appear on the sixth day at three different places,
i.e. the stone-canal, ring canal, and pedicel canal. I failed to
notice the last-mentioned part in C. echinata, in which
the deposits in the integument are most marked among the
three kinds. Mortensen’s figure (88, Pl. 1, fig. 8,c) of
the larva of Psolus phantapus represents a similar feature,
where delicate lattice plates in the integument and the rudiment
of the calcareous ring are shown.

14. PENTACTULA.

In this stage the ciliary bands have disappeared, the
tentacular crown has assumed its terminal position from the
diminution of the pre-oral hood, and at the centre of the
tentacular crown the mouth is opened while the anus has
appeared posterodorsally. This stage is reached as early as on
the seventh day, as found among the Mitsukuri material.
The larva now begins to creep on the bottom and to feed
itself.

The internal changes taking place at this stage may be deseribed
as the further completion of all the systems and organs which
were roughly established in the preceding stage. A very con-
spicuous feature of this stage as found in sections is the large
space which the body-cavity occupies and the thinning out of
every epithelium lining the water-vascular system and body-
cavity.

222, HIROSHI OHSHIMA

Water-vascular System.—No marked change is
found in the water-vascular system. The tentacles have some

TeExt-FiGcs. 6-8.

TEXT-FIG. 6.
Pentactula viewed from ventral side. x 60.

TEXtT-FIG. 7.

Same, but still advanced, being beset with branched tentacles and the
third pedicel. x 60.

TEXT-FIG. 8.

Nine-tentacled young, 1:3 mm. long. Ventral view to show the order
of appearance of tentacles and pedicels (no. 2 represented in
Table III in the text), x30. g=gut; Up= lateral pedicel ;
mv = mid-ventral radial canal; P, p,,,= primary pedicels ;
Ps>4= Secondary pedicels; pv= Polian vesicle; rt = rudiment
of respiratory tree ; 7’, t,_, = primary tentacles ; ¢,_, = secondary
tentacles,

simple branches and stand at the anteriormost end of the body
(‘Text-fig. 6, t, ;), while the primary pair of pedicels are growing

DEVELOPMENT OF CUCUMARIA ECHINATA 228

longer and have removed near to the posterior end (p,,5). ‘Thus,
as compared with the doliolaria, the ventral surface has very
much extended. The mid-ventral radial canal is still the largest
of the five radial canals ; the other four do not as yet reach the
posterior end of the body. Muscular layers of the ring, radial,
tentacular, and pedicel canals have much developed and are
well distinguishable, but no fibres are as yet visible in the
Polhian vesicle. The pore-canal still opens to the exterior through
the body-wall.

Almost at the end of the stage, on the tenth day, the third
pedicel appears on the left side of the mid-ventral radial canal at
about the middle of the body (Text-fig. 7, ps). It is much smaller
than the primary pair, and, like the subsequent members,
develops directly above the body-surface without forming at
first any sort of pedal pit as met with in the primary pair.
Ludwig (22, p. 186) found a similar condition in the forty-
fifth-day young of C. planei, and described a rudiment of the
ampulla projecting into the body-cavity. I could not make out
any ampulla in the early stage.

Alimentary Canal.—tThe gut has now become an open
canal beginning at the mouth to end in the anus. The pharynx
seems to originate from the endoderm, the atrial wall forming
only a very beginning part of the canal. The wall has become
quite thin, and the internal lumen widened remarkably. Circular
muscle fibres are found only at the pharyngeal part, the other
part forming no such structure as yet. The intestine now
shows a characteristic coil in accordance with the peculiar
arrangement of the mesentery.

The corresponding stage was observed by Danielssen
and Koren in C. frondosa, and by Kowalewsky in
Phyllophorus urna. The larvae in both forms had
five tentacles and a pair of the primary pedicels. Ludwig
(23, p. 26) observed the pentactula of C. parva found in the
brood-pouches, measuring 0-5-0-6 mm. by 0-28-0:31 mm.
The five tentacles showing no trace of ramification, a pair of
the primary pedicels, gut, stone-canal, calcareous ring, and cal-
careous deposits of integument are described. A very interesting

294 HIROSHI OHSHIMA

case was reported by Clark (8, 1901, pp. 168-70) in another
brooding form, Psolidium nutriens. The young had the
five primary tentacles just indicated and a pair of pedicels,
which latter were very remarkable in size and apparently served
to attach them to the inner skin of the mother’s back. It is
interesting to note that in such a form characterized by the
degenerated state of the mid-ventral radial canal and its appen-
dages in contrast to a comparatively stronger development of
the lateral ventral ones, the first appearing pedicels still belong
to the former and attain such a remarkable degree of develop-
ment.

15. Youna.

In the post-larval stage which I call young, five more tentacles
are added to the primary five, the pedicels increase by degrees,
and, moreover, retractor muscles, respiratory trees, genital
organs, &¢., appear, so that a mimature adult Cucumaria
is now formed.

This stage has been known in many cases. Danielssen
and Koren (11) first deseribed and figured the young of
C. frondosa. Among others the following mstances may
be enumerated: C. glacialis by Mortensen (82),
C. crocea by Ludwig (23), MacBride and Simpson
(27), Thyone rubra by Clark (8), C. saxicola by
MacBride (25, 1912, Pl. i, fig, 41; 26, Text-fig. 402),
C.ijimai, C. lamperti and Thyone imbricata by
the present writer (88, 1915). Besides these, young referable to
Cucumaria were reported from the Antarctic Seas by
MacBride (25, pp. 3-7; Pl. i, fig. 3; Pl. ii, figs. 5-8) and
Mortensen (84, 1913, p. 87; Pl. xi, figs. 6, 7).

From want of materials in consecutive series, I am compelled
to leave untouched many important problems in connexion with
the origin of several organs. I give here only some points
of my observations.

Stone-Canal.—tThe pore-canal which has in the preceding
stage been distinetly seen lying in the dorsal body-wall has

' Identified doubtfully with C. lactea.

DEVELOPMENT OF CUCUMARIA ECHINATA 925

now utterly disappeared. The axial sinus has given rise to the
internal madreporite shaped like a folded leaf. A very peculiar
feature is seen inthe young of C. ijimai. The ten-tentacled
young of this species found in the mother’s brood-pouch measure
about 5 mm. in length. No well-marked madreporic body can
here be found, but the distal end of the stone-canal is
dilated at reaching the dorsal body-wall into a flat cavity.
The cavity extends posteriorly and ramifies like a root, each
of these branches opening to the exterior. Delicate calcareous
deposits are found at the junction of the canal and the flattened
cavity, as well as in the wall of the canal. In the young of
C. ecrocea MacBride and Simpson were able to find
the opening of the pore-canal. According to Ludwig (22,
p. 186) the pore-canal of C. plancei loses its opening on the
eighteenth to twenty-fourth days, and until the ninety-eighth
day the axial sinus opens to the body-cavity through its thin-
walled side. The canal of Phyllophorus urna remains
longer than in C. planci (Ludwig, 24, p. 98; see also
Russo, 44, p. 42).

Secondary Tentacles. Ludwig (22, p. 184) found
two more tentacles added to the primary five by the one hundred
and sixteenth day. These were sent out dorsad from each of
the lateral ventral radial canals. He was, however, unable
to observe actually the successive appearance of the remaining
three. He only assumed that the eighth should appear dorsad
from the right dorsal radial canal, the ninth and tenth ven-
trad from each of the lateral ventral radial canals. According
to Mitsukuri(ante, p. 175) the first to appear among the
secondaries in C. echinata is that given out from the right
ventral radial canal.

From observations on some specimens at my disposal I can
corroborate Ludwig’s view. In some specimens, as is seen in
no. 10 of Table ITI, there are only eight tentacles where the sixth
and seventh have attaimed a size equal to the primary five, but
the eighth, which appears dorsad from the right dorsal radial canal,
is distinctly smaller. Thus Mitsukuri’s second statement,
contradicting his first one, is obviously a mistake. Among

296 HIROSHI OHSHIMA

ten-tentacled specimens some are often found having a pair
of small and bud-like tentacles given out ventrad from each
of the lateral radial canals, as seen in nos. 8 and 8 in Table ITI.
These are the ninth and tenth. As to which of these two should
appear first, a specimen represented in Text-fig. 8 and no. 2
in Table IIT gives an indication. Here the right one of them
only is present, and thus the young is nine-tentacled (Text-

TEXT-FIG. 9.

Nae F

Diagrams showing the sequence of appearing of tentacles in
Cucumaria echinata(A)and Holothuria floridana
(B). Viewed from behind anteriorly. I= primary tentacles ;
I’ = same appearing last (dotted lines indicate the later position
of tentacular canals); 6-10 = secondary tentacles numbered
according to the order of appearing; pv = Polian vesicle ;
st = stone-canal,

fic. 9, a). There seems to be a considerable period before the
appearance of the last two, as noticed by Mitsukuri.

It is very interesting to find that this order of appearance
of the secondary tentacles in Cucumaria coincides precisely
with that observed by Edwards in Holothuria flori-
dana (12, pp. 217-20; Diagram I). As stated above, the
five primary tentacles of H. floridana arise in a manner quite
different from those of Cucumaria. But the sixth arises
dorsad from the right ventral radial canal, and the seventh, in
opposition to it, dorsad from the left ventral radial canal

DEVELOPMENT OF CUCUMARIA ECHINATA 227

The eighth is given out dorsad from the right dorsal radial
canal, and the ninth and tenth arise ventrad from either the
right or left ventral radial canal. Thus the ten-tentacled
young of Holothuria has two tentacles on each inter-
radius, but the dorsal paired radii have each only one, while
the ventral paired radii have each three (Text-fig. 9, B).

Pseudocucumis africanus, which is a_ twenty-
tentacled form, remains while young in the ten-tentacled stage
for a considerable period (Ohshima, 89, 1916). Here in this
stage each radial canal sends out a tentacular canal on each side,
just as in Cuecumaria and different from Holothuria.
According to Lud wig (24, p.97),in Phyllophorus urna,
another twenty-tentacled form, the sixth and seventh tentacles
appear between the dorsal and ventral pairs of the primary
five, just as was known in C. planci. In the ten-tentacled
stage of Ps. africanus of about 6-5 mm. in length, the
relative sizes of the tentacles indicate, to a certain extent, their
order of appearance, presumably agreeing with C. plane
and C. echinata.

Manner of Branching of the Tentacles. In one
of my former papers (87, 1914) I described the manner of
branching seen in the adult Cucumaria. Some passages
may here be translated.

‘Living specimens of ©. echinata measure, in their fully
extended state, up to 10 em. in length and 2 em. in diameter,
and the tentacles attain about 4 em. in length. The pair of
tentacles belonging to the mid-ventral radius are markedly
smaller than the others.

‘Hach of the eight tentacles, other than the ventral pair,
gives out twenty-five to thirty side branches (first order),
arranged in a dextrorse spiral, or turning “with the sun”’, with
an angular divergence of one-quarter or 90°. The first branch
(no. 1) stands at about 5 mm. above the base of the stem and
on the right of the outside (as seen from outside). The second
branch (no. 2) is the largest, standing on the left of the outside,
No. 3 is markedly smaller, standing on the left of the inner side,
and no. 4, also small, on the right of the inner side. No. 5,

NO. 258 R

228 HIROSHI OHSHIMA

again, is larger and stands just above no. 1. The same relation-
ships are to be seen in the corresponding parts in the following.
The angular divergence may often vary as much as_ two-
sevenths (ca. 102° 51% 25”), but rarely to three-elevenths (ca.
98° 19’ 5”). In the former case the branch no. 8 comes above
no. 1 with two spiral turns between them, while in the latter
no. 12 comes above no. 1 after three turns.

‘No. 1 of the first order gives out smaller branches about
fifteen in number, arranged in a dextrorse spiral, with an angular
divergence of one-quarter or 90°, or rarely one-third or 120°.
These I may eall branches of the second order. Among them
no. 1 is the largest. Hach of these branches of the second order
again gives out smaller branches, the third order, in a sinistrorse
spiral or turning “‘ against the sun ’’, with an angular divergence
of one-quarter or one-third. These of the third order produce
still smaller branches, the fourth order, in a dextrorse spiral,
and these latter once more give out the smallest branches, the
fifth order, in a sinistrorse spiral.

‘No. 2 and subsequent branches of the first order give out
a series of smaller branches in a manner quite contrary to that
found in no. 1. Here the branches of the second and fourth
orders are arranged in the sinistrorse direction, those of the
third and fifth orders in the dextrorse direction.

‘ The two ventral tentacles differ in appearance from the other
eight. But a closer examination reveals the fact that they are
only modified in the relative sizes of branches. Here no. 2
of the first order! is of a length almost equal to the main stem,
siving the tentacle the appearance of being bifurcated. Further,
no. 1 of the second order given out from no. 1 of the first order
is relatively large. Just as in the other eight tentacles the
arrangement of the smaller branches of no. 1 of the first order
is the reverse of that found in no. 2 and_ subsequent
branches.

‘Thus the tentacles of C. echinata branch according to
a definite plan like the phyllotaxis among plants. The angular

‘ In the preliminary paper (40, 1918, p. 387) I was in error in stating
that this was the first branch,

DEVELOPMENT OF CUCUMARIA ECHINATA 229

divergences, one-third, one-quarter, two-sevenths, three-elevenths,
&e., show a gradual approximation to the angle of about 99° 30’.
The angles seem to undergo no variation from different degrees
of contraction, for only longitudinal muscle fibres are present
in the wall of the tentacle.

* Hand in hand with the regular spiral arrangement of branches,

TExt-Fia. 10.

A. Tentacle of young, viewed from external side to show the manner
of branching. B. One of the ventral pair. x40. P,_,,=
branches of the first order; S$, ,,=same of the second order ;
T,_, =same of the third order,

supporting calcareous bodies lie in spiral distribution, always
on the side where a branch is given out.’

In the young, whose length exclusive of tentacles measures
2-5-4-5 mm., the regular manner of branching as referred to
above is plainly visible. In the eight tentacles, other than the
ventral pair (Text-fig. 10, a), there are about a dozen branches

R2

230 HIROSHI OHSHIMA

of the first order (P,_,3) arranged in a dextrorse spiral, and with
an angular divergence of one-quarter or two-sevenths. Of
these, no. 2 (P,) is the largest, bemg beset with eight to ten
branches of the second order (Sy). The latter are arranged
im a sinistrorse spiral except on no. 1 of the first order (P)),
where the arrangement is dextrorse. In some comparatively
larger ones of the second order, one can distinguish two to three
branches of the third order (7, 3).

In the two ventral tentacles the features are quite different
(Text-fig. 10,8). These keep for a considerable period a very
simple appearance, in that the tip is branched twice dichoto-
mously. This may probably be an adaptive change. The left
branch undoubtedly gives rise to no. 2 of the first order, which
grows as large as the main stem rising from the right branch.
They later give out branches along their whole length as seen
in the adult state. No. 1 of the first order appears later on the
outer side immediately below the bifurcated point. In none
of the other Cucumarids does such a peculiar feature seem to
have been noticed.

Mitsukuri (ante, p. 175) first noticed the regularity of
branching of the tentacles in that the ‘ pmnules’ stand in a
spiral arrangement (sinistrorse as judged from his figure), with an
angular divergence of one-quarter, and that the second pmnule
is the largest. But as regards the direction of the spiral his
statement does not agree with my observations. Ludwig
(22, p. 185; 24, p. 97) stated that both nm C. planci and
Phyllophorus urna the five primary tentacles first bifur-
cate at the tip, and then each branch produces side branches.
In C. echinata I observed no such terminal bifureation except
in the ventral pair (Text-fig. 7). Kowalewsky (17%, p. 6)
was of the opinion that the branching of the tentacles im
C. kirehsbergii occurs, not simply from terminal bifurea-
tion, but from producing a bud near the apex of the tentacle.

Inthe ten-tentacled stage of PSeudocucumis africanus
of about 6-5 mm. in length, no such differentiation of the ventral
pair is found, all being beset with several side branches.

Increase of Pedicels.—The order of the appearance

a

DEVELOPMENT OF GUCUMARIA ECHINATA Ost

of the pedicels may deserve a special notice. In the late pentac-
tula stage we have met with the third pedicel appearing on the
left side of the mid-ventral radial canal im front of the first
primary pair. Now the fourth makes its appearance on the
right side of the same radius but still in front of the third (‘Text-
fig. 8, pa). According to Mitsukuri (ante, p. 175), previous
to this, a pedicel appears on the ventral side of each lateral
ventral radii, between the height of the primary pair and the
third (Jp). Further on from this condition the appearance of
new pedicels takes place, as will be seen in the following table.

TaBLe III.

To show the number of pedicels with reference to the radii in
young of different stages.

Length
of body Date of EDS TVG Vig. aR:

No. inmm.} collection. dad.) Veale ts ae) id. ved, Total:
1 1-1 = =July 20, 1916 Lr 2h l 6
2 1:3 Latter part of 1 Ta, Ze L 1 8

July, 1897
3 1-4 x a 1 2 2 2 1 9
4 1-1 ‘ a 1 Pe Va tay 9
5 1:5 + :. 1 22 2 2 9
6 1:3 on me 1 22 2 2 1 10
il 2-2 August 1, 1916 2 oy Jy 2) 2 2 13
8 1:4 Latter part of 1 Ze oe 1 10
July, 1897
9 1-4 “i 95 1 2°2 3 2 | 1]

10 1:3 a 3 1 Zee US TZ 1 11

11 1:3. July 20, 1916 2 22) 3. 2 2 13

12 15 Latter part of 1 32) a 3 1 13

July, 1897

13 1:5 % ‘9 2 a@ 3 3 2 15

14 1-5 6 5 1 423 4 1 15

15 15 July 20, 1916 3 6 2 3 4 3 21

16 1-3 Latter part of 1 ie 3) 1 12

July, 1897

17 1-9 3 e 2 outs 93, 3 2 16

18 1-5 - 5 2 3.3 3.4 2 17

19 1:6 3 es 2 a) 6 FB} a! 2 17

20 1-6 ” Ar 3 Sear se! 2 18

21 7 “A 33 3 4.3 3 5 3 21

22 24 July 25, 1916 5 5 3: +395 5 26

! The length of body refers to the preserved state and is measured
exclusively of tentacles.

No.
23

Length
of body

in mm?
2-0

bt ee
Cust Ot GS OO

melhor
10

Fe eS las lige lig
AwWOHSCnHOwharId)d

HIROSHI OHSHIMA

TaBLE IIL (continued).

Date of ED: EV.

collection. Gd. Vid. Ue

Latter part of 3 {
July, 1897

” ” 1 4

os 1 3

” ” 2 3

” ” 3 4

” ” 3 4

July 20, 1916 4 4

Latter part of 2 3
July, 1897

” ” 3 5

August 1, 1916 4 6

Latter part of 3 5)
July, 1897

” ” 3 4

op) ” 2 3

” ” 2 4

2 ” 2 4

” 3 5)

August 1, 1916 3 5

”” 33 4 6

» ” 5) 7

” ” 8 8

39 9? 3 5)

99 33 3 7

93 33 6 8

9° «5 10 9

oOnrwW WRAP

CU OU OUST OU OU OT Ot HE HB Ot

~
Ss

=
x
-
aX
=]

Ct
w =
aS
a |
3S
ws
Ld

Co Ol He He Co He Co

bow He He Co Co Co Co Go Co an

He > OL

MINI AISP www lok Ww aw WRENN ee
to
i=)

SOOTHE WW WHR EE bo
HOD IO WH OVS Or Or > HE HH Ot

From nos. 1-15 given in the above table we get the number of
pedicels belonging to each radius as follows :

LD. Ly MY.
Total 19 37 68
Average 1-3 2-5 4-5
Percentage 11-0 21-4 39:3

TaBLE IV.

RV. RD.

32 17
2:1 1-1
18-5 9-8

1 The length of body refers to the preserved state and is measured
exclusively of tentacles,

DEVELOPMENT OF CUCUMARIA ECHINATA 233
Let us further examine the more advanced individuals :

TABLE V.

Length
of body Date of
No. inmm. collection. EDL Ne, FEV SED Total.
1 3:3 August 1, 1916 5 7 8 6 6 32
2 3:7 os - 4 6 9 7 4 30
3 3:2 ” ”” 3. 7 9 6 6 33
4 3-0 a mt 5 8 9 6 6 34
5 3:7 i ; 5 8 10 7 5 35
6 2-5 * x 9 8 10 7 a 41
7 3-9 July 20, 1916 6 $i 8 7 40
8 3°5 July 25, 1916 9 ll 11 8 8 47
9 3-9 August 1, 1916 10 12 13 14 9 58
10 4-8 is o iG 4. 14 13—=té<‘«‘i 60
11 4:8 5 i 12 16 15 16 12 71
12 4-8 an - 10 13 16 13 10 62

The summarized result of these twelve specimens is as follows :
Tasue VI.

LD. LY. MV. RV. RD.

Total : +190 118 135 110 90
Average . ES 9-8 11:3 9-2 7:5
Percentage . 16-6 21:7 24-9 20:3 16-6

Of adult individuals of different sizes the number of pedicels
with reference to the radii is as follows :

TasLe VII.
Length of

No. body in mm. LD. LY. MY. ae RD. Total.
1 9-0 29 38 42 35 33 177
2 8-0 36 40 43 40 39 198
3 9-0 41 53 50 49 47 245
4 10-0 53 60 62 62 55 292
5 18-0 64 75 ay) 80 63 361
6 16-0 72 76 80 78 73 379
i) 16-0 66 76 83 74 70 369
8 29-0 108 120 132 126 105 591
9 30-0 112 133 140 129 111 625

10 32-0 121 140 148 136 123 668

From these ten specimens the followmg summary can be
derived :
Tasie VIII.

LD. LY. MY. RV. RD.
otal ee . 7102 811 864 809 719
Average . 702 81-1 86-4 80-9 71-9
Percentage : 18:0 20-8 22:1 20-7 18-4

2354 HIROSHI OHSHIMA

From comparison of the Tables IV, VI, and VIII we may
draw the following conclusions :

1. The numbers of pedicels in each pair of lateral radu are
approximately equal, showing no asymmetrical features.

2. The pedicels of the mid-ventral radius develop early,
whereas those of the dorsal paired radii increase later. ‘Those
of the lateral ventral radii remain almost constant throughout
in regard to the ratio to the total number of pedicels.

Order of appearance of Mid-ventral Pedicels.
Of special interest 1s the examination of the order of the appear-
ance of pedicels from the mid-ventral radial canal.

As mentioned above, the fourth pedicel develops on the right
side of the radius in front of the third (Text-fig. 11, 4). This condi-
tion is seen in the specimens nos. 1-7 of Table III. The fifth (5)
appears again on the right side and in front of the primary pair.
This is observed in the specimens nos. 8-15. The sixth (6)
appears on the left side behind the primary pair, as seen in the
specimens nos. 19-22. The seventh (7) appears far forwards,
on the left side and in front of the fourth, as seen in the specimens
nos. 24-9. The eighth (8) appears again on the left side, imme-
diately in front of the primary pair, as seen in the specimens
nos. 89-41.

Among some specimens variations are found in the order and
position of newly-appearing pedicels. The specimen no, 23 has
the sixth on the right side instead of on the left, while the speci-
mens nos. 17 and 18 have the sixth in front of the fourth on the
left side. The specimen no. 16 has the fourth on the left side
instead of on the right, and the sixth on the nght in front of the
fourth. Nos. 30 and 31 have the seventh on the right instead of
on the left. Nos. 33 and 34 have the fifth on the left instead of
on the right. In no. 32 the seventh appeared on the left, imme-
diately in front of the primary pair. In nos. 36-8 the eighth
stands on the right side assuming the anteriormost position.

Increase in numbers above the nine pedicels is represented
by a few specimens. In no, 42 the ninth (9) appeared on the right
side between the fourth and fifth. tn nos. 43 and 44 the tenth
(10) appeared again on the right side between the ninth and

DEVELOPMENT OF CUCUMARIA ECHINATA 935

fifth. No, 45 has added the eleventh (11) on the right side
behind the primary pair. In no. 46 the twelfth (12) is seen on
the right side in front of the fourth, and the two behind the
primary pair stand in the reverse order to the preceding speci-
men, in that the right side one stands far behind the left.

Trxt-Fic. ll.

ae} OF
we |
|} Oa
|
ee
i
ri
u O), R
|| OP
0 ||
1 OY, | ©) 2
! oy
Gehl
6 |

Diagram showing the position and order of appearing of pedicels
belonging to the mid-ventral radius.

The stage figured by Mitsukuri (ante, Text-fig. 3) corre-
sponds with my specimens nos. 4-6 of Table HI. According
to him, of the two pairs produced from the lateral ventral radu
the right always precedes the left. In contradiction to his
statements, in my specimens nos. 8 and 8, a new pedicel is
formed only on the left ventral radius in front of the old one.
In C. frondosa Danielssen and Koren (11) described
simply that on the thirty-fourth day a pair of pedicels are added
to, and in front of, the primary pair, and on the fifty-sixth
day the third pair are added still anteriorly. In the latter
stage papillae appeared here and there on the dorsal side. A
similar stage was figured by Mac Bride (26, Pl. i, fig. 4; 26,

236 HIROSHI OHSHIMA

Text-tig. 402) for C. saxicola having developed the second pair
in front of the primaries. According to Ludwig (22, p. 186),
inC. planei the third pedicel is distinctly seen on the forty-fifth
day, constantly on the left side of the mid-ventral radius and in
front of the primary pair. The fourth makes its appearance on
the eighty-fourth day, on the right side of the radius and further
anteriorly to the third. Thus far the order and position agree
with my observations. But he differs from me in that the fifth
appears ventrad from the left dorsal radial canal near the
anterior end of the body. The same author (28, pp. 21-2)
traced the order in the young of C. crocea. The youngest
stage he examined had eight pedicels corresponding to that
figured by Mitsukuri (loe. cit.).. The ninth and tenth appear
from the mid-ventral radial canal, intervening between the
anterior and posterior pairs. Subsequently new pedicels increase
very rapidly on the ventral side of both the lateral ventral radu.
Up to the stage where the body length attains ca. 8 mm. the
dorsal paired radii are free from pedicels, while ten or more
have appeared in each of the ventral radu. These facts differ
very much from the case of C. echinata, where the dorsal
radii share in the pedicel-formation quite early when each of
the ventral radi has only one (specimen no. 2). MacBride
and Simpson’s statement (27, p. 8) referrmg to C. crocea
differs from Ludwig’s in that there are four pedicels arising
from each radial canal. Probably the observers overlooked
some others in the ventral radii from their ‘ not having reached
the surface ’.

Edwards’s laborious task of elucidating the order of the
appearance of the pedicels in Holothuria floridana
(12, pp. 222-6) shows that that species is totally different in
this respect from that seen in Cucumaria. Here in
Holothuria an unpaired pedicel first appears at the posterior
end of the mid-ventral radial canal on the fourth day. The
second appears to the left of the same canal on the seventh
day. The third and fourth follow equally on the left side.
As late as the fortieth day a pedicel appears for the first time
to the right of the radius. It seems to me highly probable that

DEVELOPMENT OF CUCUMARIA ECHINATA 2937

a similar feature occurs in Stichopus japonicus also,
judging from the figure given by Mitsukuri (80, 19038, p. 12,
fig. 3). Here the posteriormost unpaired one seems to be the
first to appear, and besides it the mid-ventral canal seems to be
provided with three pedicels to the left and one to the right,
whereas each of the ventral radii has three pedicels.

16. SUMMARY.

1. The breeding season of Cucumaria echinata seems
to begin in the middle of June and to last until the early part
of August. During that season the wall of the genital tubes
is thin, but in an inactive period it is very thick. No muscle
layer could be made out in the wall. The genital papilla is
subdivided, the branches bemg more numerous in males than in
females. Both sexes occur in almost equal numbers.

2. The ovarian egg is attached to the wall of the genital tube
by its broad vegetative half. At the animal pole which is
directed towards the internal lumen of the tube a short rod-like
cytoplasmic process is found. This structure develops near the
end of the growth of the egg, and probably has some significance
in relation to future changes of the egg.

3. Freshly captured mature animals spawn in the evening.
At first the males shed out spermatic fluid,and after some minutes
the females begin to lay eggs. During these acts no special
movements of tentacles are observed in either sexes.

4, The newly-shed egg is slightly flattened and measures about
300-400 » in diameter. It is covered with a gelatinous layer,
through which a canal opens at the animal pole. The egg is
heavier than sea-water.

5. The first polar body has been formed by the time it is shed,
when the second maturation spindle is to be seen. The spermato-
zoon enters the egg before the second maturation division, and
probably at the point near to, but not precisely identical with,
the animal pole.

6. The first cleavage spindle is formed within an hour. The
cleavage is total and equal, proceeding quite regularly up to
about the thirty-two-cell stage. Very often an interlocking

238 HIROSHI OHSHIMA

of blastomeres oceurs. Inequality im size of the blastomeres
is met with above the thirty-two-cell stage, and the embryo is
wrapped up within the egg-membrane until the blastula stage
has been attaimed.

7. The blastula is spherical but not wrinkled, and is now free
from egg-membrane. It swims about by means of cilia. The
mesenchyme-formation precedes invagination, occurring exclu-
sively at the vegetative pole. The invagination begins the next
morning.

8. In a fully-formed gastrula the archenteron shows a peculiar
twisting, enabling one to distinguish in it three parts. The
most anterior flat part is the future hydrocoele, the second
transverse part is the future enterocoele, and the hindermost
tubular part is the future gut.

9. Very late in the gastrula stage the stomodaeum makes its
first appearance, being preceded by a thickening of the ectoderm
at about the middle of the ventral side. Some mesenchyme cells
seem to be formed here by the proliferation of ectodermal cells.
The position of the stomodaeum is, as can be shown in later
stages, a little on the left of the median line.

10. The dipleurula stage begins late on the second day. In
this stage the hydro-enterocoele first becomes separated from
the gut. The former then divides into the hydrocoele and
enterocoele. The hydrocoele produces the rudiment of the
pore-canal directed postero-dorsad, and six lobes on the anterior
expanded margin. These latter are rudiments of the five primary
tentacles and of the mid-ventral radial canal. The enterocoele
divides into right and left vesicles, situated on the left dorsal
and antero-ventral sides respectively.

11. On the third day doliolaria is formed, which is charac-
terized by the possession of three ciliary bands around the
posterior half of the body besides the weaker uniform ciliation
over the pre-oral hood and on the anal field. From the hydro-
coele are first differentiated the mid-ventral radial canal and
four of the primary tentacles.

12. The primary pair of pedicels make their appearance as
ectodermal depressions (pedal pits) situated between the second

DEVELOPMENT OF CUCUMARIA ECHINATA 939

and third ciliary bands. The left pedicel is a little earlier in
appearing than the right, while neither of the two can be said
definitely to be anterior to the other in position.

13. The original position of the primary tentacles is decidedly
interradial, but their bases gradually shift towards the respective
radial canal according to a definite asymmetrical feature. The
one in the left dorsal interradius appears last.

14. The Polian vesicle appears at the free end of the ventral
limb of the hydrocoele ring, while about the same ‘ime the
axial sinus is formed as a secondary dilatation of the middle part
of the pore-canal. The dorsal pore has now opened between
the second and third ciliary bands.

15. The hydrocoele ring closes in the left dorsal interradius.
This is clearly shown by the position of the rudiments of the
dorsal and ventral radial canals of the left side, appearing usually
before the closure of the rmg. Of the four paired radial canals
the right dorsal appears first, while the left ventral is the last to
appear.

16. Fusion of the right and left enterocoeles occurs on the right
side, while on the other side the two vesicles lie close but
separated, This intervening portion gives rise to the mesentery,
which at last bends in an S-shape in agreement with the coil of
the gut in the future. The gut is almost solid, leaving but very
narrow lumen. Blastocoele jelly is most massive in the doliolaria
stage, and mesenchyme cells thickly cover all the internal
vesicles, without, however, forming any definite cell-layer.

17. The latter half of the doliolaria stage may be distinguished
by callmg it metadoliolaria. Here degeneration of the pre-oral
hood and ciliary bands sets in, while muscles and nerves are
differentiated, besides the further completion of hydrocoele and
enterocoele. Calcareous deposits, too, make their first appear-
ance In this stage. They appear in three places : the wall of the
axial sinus, the bases of the tentacles, and the integument of
the posterior part of the body.

18. In the course of a week or more the larva changes into
a creeping stage, pentactula. The five tentacles have now
a few branches and the third pedicel appears at last. The gut

240 HIROSHI OHSHIMA

is now open throughout, both at the mouth and anus, the lumen
becoming quite spacious.

19. During the transformation of the pentactula into the ten-
tentacled young, the pore-canal becomes obliterated. Of the
secondary tentacles those given out dorsad from the paired
ventral radial canals appear first, while those given out ventrad
from the same canals are completed very late. Among the
respective pair the right one appears slightly earlier than the left.

20. In the young, the branches of the tentacle can be classified
in three orders, and are sent out either in dextrorse or sinistrorse
spiral according to a definite arrangement. The angular diver-
gence of branches is about one-quarter or two-sevenths. The
ventral pair remain for a long while in a twice dichotomously
branched condition, and further branching usually takes place
very late.

21. The increase of pedicels takes place faster in the mid-
ventral radius than in the others, while those of the dorsal
radii increase slowly. In none of the stages is any asymmetrical
feature found as concerns the numbers of pedicels between
right and left.

22. Along the mid-ventral radius I could ascertain that the
pedicels up to the twelfth appear according to an almost definite
order. But pedicels above the fourth may undergo some varia-
tions with respect to the order of appearance or the position
on the right and left.

IMPERIAL COLLEGE OF SCIENCE AND TECHNOLOGY, LONDON.
February 11, 1920.

17. BIBLIOGRAPHY.

1. Ackermann, August (1902)—‘‘ Uber die Anatomie und Zwittrigkeit
der Cucumaria laevigata”, ‘Z. w. Z.’, Bd. Ixxii, 1902.

2. Arts, Louis des (1910).—‘ Uber die ersten Entwicklungsstadien von
Cucumaria frondosa unter Beriicksichtigung anormaler
Verhialtnisse ”’, ‘ Bergens Mus. Aarb.’, 1910, no. 13.

3. Becher, Siegfried (1908).—** Die Stammesgeschichte der Seewalzen ”’,
‘Ergebn. u. Fortschr. Zool.’, Bd. i, Heft 3, 1908.

SS Re Te 2 es A Ea A | ape Biter TE Pm

© Se @r = aes © mes ~

as

DEVELOPMENT OF CUCUMARIA ECHINATA 241

4. Boveri, Th. (1901).—* Uber die Polaritit des Seeigel-Kies ”’, ‘ Verh.
phys.-med. Ges. Wiirzburg ’, N.F., Bd. xxxiv, 1901.

5. Bury, H. (1889).—‘* Studies in the Embryology of Echinoderms ”’,
‘Quart. Journ. Micro. Sci.’, vol. xxix, 1889.

(1895).—‘‘ The Metamorphosis of Echinoderms”, _ ibid.,
vol, xxxvili, 1895.

7. Clark, Hubert Lyman (1898).—‘“Synapta vivipara: a Contribu-
tion to the Morphology of Echinoderms”, “Mem. Boston Soc.
Nat. Hist.’, vol. v, no. 3, 1898.

6.

8. (1901).—‘‘ The Holothurians of the Pacific Coast of North
America ”’, ‘ Zool. Anz.’, Bd. xxiv, 1901.

9. —— (1910).—‘**The Development of an Apodous Holothurian
(Chiridota rotifera)”, ‘Journ. Exper. Zool.’, vol. ix, no. 3,
1910.

10. Cowles, R. P. (1907).—“ Cucumaria curata, sp. noy.”, ‘ Johns
Hopkins Univ. Cire.’, ser, ii, 1907, no. 3.

11.1 Danielssen, D. C., et J. Koren (1856).—** Observations sur le Développe-
ment des Holothuries”’, * Fauna Littoralis Norvegicae’, vol. ii,
1856.

12. Edwards, Charles Lincoln (1909).— ‘The Development of Holo-
thuria floridana (Pourtalés), with especial Reference to the
Ambulacral Appendages ”’, * Journ. Morph.’, vol. xx, no. 2, 1909.

13. (1910).—‘‘ Revision of the Holothurioidea. 1. Cucumaria
frondosa (Gunner) 1767”, ‘Zool. Jahrb., system. Abt.’,
Bd. xxix, 1910.

14. (1910).—** Four Species of Pacific Ocean Holothurians allied

to Cucumaria frondosa (Gunner) ”, ibid.

15. Hamann, Otto (1884).—** Beitriige zur Histologie der Echinodermen ”’,
Heft i, ‘ Die Holothurien’. Jena, 1884.

16. Jourdan, Et. (1883).—‘‘ Recherches sur l’histologie des Holothuries”’,
‘Ann. Mus. Hist. Nat. Marseille, Zool.’, tom. i, mém. no. 6, 1883.

17. Kowalewsky, A. (1867).—“ Beitriige zur Entwickelungsgeschichte der
Holothurien ”’, ‘Mém. Acad. Impér. Sci. St.-Pétersbourg ’, ser. vii,
tom. xi, no. 6, 1867.

18. Lampert, Kurt (1889).—‘‘ Die wihrend der Expedition S.M.S.
‘Gazelle’ 1874-1876 von Prof. Dr. Th. Studer gesammelten
Holothurien ”, * Zool. Jahrb., system. Abt.’, Bd. iv, 1889.

19. Lo Bianco, Salvatore (1899).—‘‘ Notizie biologiche riguardanti
specialmente il periodo di maturitaé sessuale degli animali del
Golfo di Napoli”’, * Mitt. zool. Stat. Neapel,’, Bd. xiii, 1899.

' As the original paper was inaccessible to me, I followed A. Agassiz’s
*“ Embryological Monograph of Echinodermata ”’,, “Mem. Mus. Comp,
Zool. Harvard Coll,’, vol, ix, no. 2, 1883,

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

2 HIROSHI OHSHIMA

. Ludwig, Hubert (1887).—‘* Drei Mittheilungen iiber alte und neue

5

Holothurienarten ”’,

1887.

(1889-92).— “* Dr. H. G. Bronn’s Klassen und Ordnungen
des Thier-Reichs.”’, Bd. ii, Abt. 3.‘ Echinodermen. I. Buch. Die
Seewalzen.’

—— (1891).—* Zur Entwickelungsgeschichte der Holothurien”’,
‘ Sitzungsber. k. Akad. Wiss. Berlin’, no, x, 1891 (1. und 2. Mit-
theilung).

——— (1898).—‘ Holothurien”’, ‘Hamburger Magalhaensische Sam-
melreise ’, 1898.

(1898).—* Brutpflege und Entwicklung von Phyllophorus
urna Grube”, Zool. Anz.’, no. 551, 1898.

MacBride, E. W. (1912).—* National Antarctic Expedition. Natural
History’, vol. vi.“ Echinoderma, III. On a Collection of young
Holothurioids ”’.

——— (1914).— Text-book of Embryology’, vol. i. “ Invertebrata ”,
London, 1914.

and J. C. Simpson (1908).—* National Antarctic Expedition.
Natural History’, vol. iv. ‘* Echinoderma. Echinoderm
Larvae ”’.

Marenzeller, E. v. (1881)—‘*‘ Neue Holothurien von Japan und
China’, ‘ Verh. zool.-bot. Ges. Wien’, 1881.

Metschnikoff, Elias (1869).—** Studien iiber die Entwickelung der
Echinodermen und Nemertinen”’, ‘Mém. Acad. Impér. Sei. St.-
Pétersbourg ’, ser. vii, tom. xiv, no. 8, 1869.

Mitsukuri, Kakichi (1903).—* Notes on the Habits and Life-History
of Stichopus japonicus Selenka ”,* Annot. Zool. Japon.’,
vol. v, part 1, 1903.

——— (1912).—“* Studies on Actinopodous Holothurioidea ”, * Jour.
Coll. Sci. Tokyo Imp. Univ.’, vol. xxix, art. 2, 1912.

Mortensen, Theodor (1894).—* Zur Anatomie und Entwicklung der
Cucumaria glacialis (Ljungman)”,*Z. w. Z, Bd. lvii,
1894.

—— (1898).—* Die Echinodermenlarven der Plankton-Expedition
nebst einer systematischen Revision der bisher bekannten Echino-
dermenlarven”’, ‘ Ergebn. Plankton-Expedit., Humboldt-Stiftung ’,
Bd. ii, J. 1898.

(1913).—** Die Echinodermenlarven der deutschen Siidpolar-
Expedition 1901-1903”, * Deutsche Siidpolar-Expedition 1901-
1903 °, Bd. xiv, Zoologie, vi.

— (1913).—** On the Development of some British Echinoderms ”’,
‘ Journ. Mar. Biol. Ass.’, N.S., vol. x, no. 1, 1913.

* Sitzungsber, k. Akad. Wiss. Berlin’, no. liv,

DEVELOPMENT OF CUCUMARIA ECHINATA 2438

36. Newth, H. G. (1916).—** The Early Development of Cucu maria:
Preliminary Account’, * Proc. Zool. Soc.’, 1916.

37. Ohshima, Hiroshi (1914).—** On the Manner of Branching of Tentacles
in Cucumaria echinata” (Japanese), ‘ Zool. Mag.’, no. 307,
1914.

38. —— (1915).—‘* Report on the Holothurians collected by the United
States Fisheries Steamer * Albatross’ in the North-western Pacific
during the summer of 1906”, ‘ Proc. U.S. Nat. Mus.’, vol. xlviii.

39. (1916).—‘** A New Case of Brood-Caring in Holothurians ”’,
* Annot. Zool. Japon.’, vol. ix, no. 2, 1916.
40. (1918).—‘* Notes on the Development of Cucumaria

echinata ”, ibid., vol. ix, no. 4, 1918.

41. Ostergren, Hjalmar (1898).—‘* Zur Anatomie der Dendrochiroten,
nebst Beschreibungen neuer Arten”’, * Zool. Anz.’, Bd. xxi, 1898.

(1912) —‘ Uber die Brutpflege der Echinodermen in den siid-
polaren Kiistengebieten ’, ‘Z. w. Z.’, Bd. ci, 1912. (Festschrift
fir Ludwig.)

43. Reimers, Karl (1912).—** Zur Histologie der Synapta digitata”,
‘Jena. Zeitschr.’, Bd. xlviii, Heft 2, 1912.

44, Russo, Achille (1902).—‘‘ Studi sugli Echinodermi’’, * Atti d. Acc.
Gioenia ’, Ann. Ixxix, ser. 4, vol. xv, mem. 7.

45. Selenka, Emil (1876).—** Zur Entwicklung der Holothurien (Holo -
thuria tubulosa und Cucumaria doliolum), Ein Bei-
trag zur Keimblattertheorie ’, * Z. w. Z.’, Bd. xxvii, 1876.

(1883).—‘ Studien iiber Entwickelungsgeschichte der Thiere ”’,
2. Heft. ‘ Die Keimblatter der Echinodermen’. Wiesbaden, 1883.

47. Semper, Carl (1867-8).—* Reisen im Archipel der Philippinen’,
IL. Theil, wiss. Result. I. Bd. Holothurien.

48. Vaney, Clément (1906).—** Deux nouvelles Holothuries incubatrices ”’,
“Compt. rend. Assoc, Frangaise Avancem. Sci.’, 1906.

EXPLANATION OF PLATES 8 AND 9.

List OF ABBREVIATIONS.

an= anus. ar=archenteron. ar, = the anteriormost part of archen-
teron, the future hydrocoele. ar, = the middle part of the same, the future
enterocoele. ar, = the last part of the same, the future gut. as = axial
sinus, or ‘madreporic vesicle’. at= atrial cavity. be = blastocoele.
bj = blastocoele jelly. 6/= free cell in the archenteron, hydrocoele, or
enterocoele, so-called ‘blood corpuscle’. bp = blastopore. c= cilia,
C,_, = ectodermal thickenings at ciliary bands. dp= dorsal pore. en =
enterocoele. enc=epineural canal. ep=ovarial wall. /f = follicular
epithelium. g=gut. gs= germinal spot. hy=hydrocoele. j = the

NO, 258 5

944 HIROSHI OHSHIMA

space probably occupied by a jelly layer. Jd = left dorsal radial canal.
le = left enterocoele. Up = left pedal pit. Upe = left pedicel canal. lw =
left ventral radial canal. ma = micropyle appendage. me = mesenchyme,
mv = mid-ventral radial canal. = germinal vesicle. p= pedal pit.
pb = first polar body. pe = pore-canal. ps = second maturation spindle.
pv = Polian vesicle. re=ring canal. rd=right dorsal radial canal.
re= right enterocoele. rp= right pedal pit. rpe = right pedicel canal.
rv = right ventral radial canal. sp = sperm nucleus. sf = stomodaeum.
sy —syneytium. {= primary tentacle. ¢,—primary tentacle in left
dorsal interradius. ¢, = same in mid-dorsal interradius, ¢, = same in right
dorsal interradius, ¢, = same in right ventral interradius. ¢, = same in left
ventral interradius,

PLATE 8.

Fig. 1.—Very young ovarian egg, fixed on August 1, 1916. x 500.

Fig. 2.—Immature ovarian egg cut meridionally, fixed on March 27,
1914. x 200.

Fig. 3.—Same as seen in the breeding season, fixed on August 1, 1916.
x 200.

Fig. 4.—Freshly laid egg in meridional section, showing the first polar
body and sperm nucleus. x 200.

Pig. 5.—Longitudinal section of blastula in which mesenchyme-forma-
tion has begun. x 150.

Fig. 6.—Same in which invagination has begun. x 150.

Fig. 7.—Gastrula with still straight archenteron. Longitudinal section.
x 200.

Fig. 8.—Gastrula, whose archenteron has begun to bend. Longitudinal
section. 200.

Fig. 9.—Tip of the archenteron to show the origin of mesenchyme
cells and free cells in the archenteron. © x 500.

Fig, 10.—Mesenchyme cells in division. x 1,000.

Fig. 114.—Fully-formed gastrula, whose archenteron is typically
twisted. Cross-section to show ectodermal thickening towards the
ventral edge of the flattened archenteron. x 200.

Fig. 118.—The ninth section below the former in the same series,
to show the second transverse part of archenteron, x 200.

Fig. 11c.—The fifth section below the former in the same series, to show
the third tubular part of archenteron. x 200.

Fig. 124,—Gastrula of the same age as the former. Dorsal view of the
frontal section, to show the first flat and the last tubular parts of archen-
teron, x 200.

Fig. 128.—The seventh section dorsad from the former, to show the
second transverse part of archenteron., x 200.

a ee ee

CTR ee MY a), 2.25 48

DEVELOPMENT OF CUCUMARIA ECHINATA 945

Fig. 13.—Very old gastrula to show the internal feature. Viewed from
the right side. The archenteron has divided into hydro-enterocoele and
gut, and the stomodaeum has appeared. — x 200.

Fig. 144.—Posterior view of the cross-section cut along the plane 1 in
fig. 13 to show the stomodaeum. x 200,

Fig. 148.—Fifteenth section below the former in the same series, cut
along the plane 2 in fig. 13. To show the posterior part of hydro-entero-
coele and the gut separated from it. 200.

Fig. 15.—Early dipleurula viewed from the left side. The internal
cavity as a solid body ; gut not represented, x 200.

Fig. 164.—Posterior view of the cross-section cut along the plane 1 in
fig. 15. x 200.

Fig. 168.—Eleventh section below the former in the same series, cut
along the plane 2 in fig. 15. x 200.

Fig. 16c.—Fourth section below tlie former, cut along the plane 3 in
fig. 15. x 200.

Fig. 16p.—Seventh section below the former, cut along the plane 4 in
fig. 15. +200.

PLATE 9.

Fig. 17.—Dipleurula viewed from the left side. The internal cavities
shown as solid bodies ; gut not represented. x 200.

Fig. 184.—Posterior view of the cross-section cut along the plane | in
fig. 17. x 200.

Fig. 188.—NSixteenth section below the former in the same series, cut
along the plane 2 in fig: 17. x 200.

Fig. 19.—Late dipleurula viewed from the left side. The internal
cavities shown as solid bodies ; gut not represented. x 200.

Fig. 20a.—Posterior view of the cross-section cut along the plane 1
in fig. 19. x 200.

Fig. 208.—Twelfth section below the former in the same series, cut along
the plane 2 in fig. 19. x 200.

Fig. 214.—Early doliolaria (no, 2 represented in Table II in the text)
viewed from the ventral side. The internal cavities shown as solid bodies ;
gut not represented. x 200.

Fig. 21B.—Same viewed from the left side. x 200.

Fig. 224.—Posterior view of the cross-section cut along the plane 1 in
fig. 21. x 200.

Fig. 228.—Fourth section below the former in the same series, cut along
the plane 2 in fig. 21. x 200.

Fig. 220c,—Third section below the former, cut along the plane 8 in
fig. 21. x 200.

Fig. 22p, Sixth section below the former, cut along the plane 4 in
fig. 21. x 200.

246 HIROSHI OHSHIMA

Fig. 234.—Ventral view of doliolaria in which the ring canal is not yet
closed (no. 12 represented in Table II in the text). The internal cavities
shown as solid bodies ; gut not represented. x 200.

Fig. 238.—Left-side view of the same. x 200.

Fig. 244,.—Cross-section cut along the plane 1 in fig. 23. Seen from
behind anteriorly. x 200.

Fig. 248.—Section immediately next to the former. x 200.

Fig. 24c.—Section immediately next to the former, cut along the
plane 2 in fig. 23. x 200.

Fig. 24p.—Section immediately next to the former. x 200.

Fig. 24e.—Third section below the former, cut along the plane 3 in
fig. 23. x 200.

Fig. 24r.—Fifth section below the former, cut along the plane 4 in
fig. 23. 200.

Fig. 25.—Sagittal section of doliolaria cut through the pore-canal
(no. 8 represented in Table IT in the text). x 200.

H Ohshima del

Quart. Journ. Micr. Sci. Vol. 65, N.S, Pl. 8.

@--77Pb

H. Ohshima del.

Quart.

Journ. Micr. Sci.

Vol. 65, N.S., Pl. 9.

Observations on the Protozoa parasitic in

Archotermopsis wroughtoni Desn.

Part III. Pseudotrichonympha pristina.

By

D. Ward Cutler, M.A., Cantab.

With Plate 10 and 8 Text-tigures.

CONTENTS.
PAGE PAGE
1. INTRODUCTION . : . 247 5. MorpHoocy (contd.)
2. Mrruops : : . 248 (c) Striations and Granules 253
3. SYSTEMATIC POSITION . 249 (d) Nucleus . : . 2504
4, Livinc ConDITION . . 250 6. DIvIsIon : 3 RBIS
5. MoRPHOLOGY . ‘ . 252 7. GENERAL CONSIDERATIONS 258
(a) Cell Inclusions . 252 8. REFERENCES . : - 262
(b) Centroblepharoplast 253 9. EXPLANATION OF PLATES . 263

INTRODUCTION.

In previous papers are described species of Protozoa resident
in the hind gut of the Indian termite Archotermopsis
wroughtoni Desn. It is my purpose here to give an
account of a fourth species of these unicellular organisms,
already described by Imms (11) under the term Tricho-
nympha (Holomastigotoides) pristina.

The true name for the animal is undoubtedly Pseudo-
trichonympha pristina, but owing to an unfortunate
mistake made by Grassi in his earlier papers a good deal of

248 D. WARD CUTLER

confusion has arisen around the nomenclature of these forms.
In 1910 Hartmann (9) gave an account of a flagellate which
he named Trichonympha hertwigi, and described
male and female forms from which gametes were produced.
Conjugation between these gametes was supposed to oecur,
and the resulting young forms were figured. Hartmann’s
observations, however, did not bear out these assumptions,
and it is certain that they have no foundation in fact. His
conclusions were attacked in 1911 by Grassi (6), who pointed
out that ‘Trichonympha hertwigi’ was in reality
a mixture of two or more genera, the male form belonging to
the genus Holomastigotoides, the female form to the
genus Pseudotrichonympha, and the ‘young form’
was referred to Pyrsonympha. The ‘ gametes’ were
undoubtedly minute oval flagellates abundant in the intes-
tines of many termites. The confusion arose round the ‘ male ’
and ‘female’ forms of Hartmann, for Grassi’s description of
the genera, to which he referred them, did not appear to agree
with Hartmann’s account, as Franca pointed out in 1916.
In a later paper Grassi (7) rectified his error, referrmg the
‘male’ form to the genus Pseudotrichonympha and
the ‘female’ form to Holomastigotoides, thus revers-
ing his earlier statement. Unfortunately, however, the
mistake received a wide acceptance, and even in Doflein’s
latest edition of his text-book (8) it is still perpetrated. Kofoid
and Swezy (18) also in their recent paper on Trichonympha
campanula adhere to Grassi’s first classification.

The organism described in the present paper 1s undoubtedly
closely related to the ‘male’ form of T. hertwigi, and
should therefore be named Pseudotrichonympha
pristina and not Trichonympha (Holomastigo-
toides) pristina, as Imms has called it.

Merrnops.
The methods used for the study of P. pristina are
those already described in my previous papers (2), to which
I would refer those interested.

SE ————————

THE PROTOZOA PARASITIC IN ARCHOTERMOPSIS WROUGHTONI 249

General Considerations.
Systematic Position.

That P. pristina is a flagellate belonging to the order
Hy permastigina (Grassi) is indubitable. The Tricho-
nymphidae have suffered much at the hands of sys-
tematists. Stein (26) in 1878 correctly placed them among
the flagellates, though Leidy (21, 22) himself considered them
as intermediate between the gregarines and ciliates. Kent (14)
in 1882 founded the family Trichonymphidae and
placed it among the holotrichous ciliates, a view supported
by Butschli m 1889. Senn (25) in 1900 added these forms,
as an appendix to the Flagellata ; while Hickson (10) allocated
them to an appendix of the Ciliata.

In the 1911 edition of Doflein’s text-book the classification
of Senn was followed ; but in the last edition of 1916 Grassi’s
correct classification is given.

Finally, in 1913 Poche (28) added his quota to the existing
confusion by creating the new order Trichonympha,
which was placed among the Euflagellata. Kofoid and
Swezy (17, 18) have recently published papers dealing with
the flagellate affinities of these organisms, to which those
interested are referred. One point which appears to have
escaped notice is the complete absence of a micronucleus in
any of the Hypermastigina, a fact which in itself is
suggestive of their flagellate affinities, for with a few doubtful
exceptions the ciliates are all heterokaryote, as Hickson pointed
out in 1903.

P. pristina so differs from Hartmann’s male form of
T. hertwigi that the two forms cannot be regarded as
one species. Grassi distinguishes four species of Pseudo-
trichonympha, none of which appear to be identical
with P. pristina. The descriptions given of the species,
however, are so scanty that it is impossible adequately to
compare them with the animal described here.

250 D. WARD CUTLER

Livinc CoNnDITIoN.
Movement and General Appearance.

P. pristinaisat once striking because of its great swimming
power, exceeding that of any other protozoon of this termite.

In living preparations it is a very pleasing sight to observe
these animals gliding across the field of view, thrusting away
with their anterior flagella the numerous wood particles and
other protozoa impeding their progress. This gliding move-
ment, too, is characteristic, resembling that of many of the
large ciliates, and doubtless is due to the whole body being
supplied with flagella, the anterior of which are probably the
mnain propelling organs, as in Trichonympha cam-
panula described by Kofoid and Swezy (18). During pro-
gression the whole of the animal’s body revolves on its longitu-
dinal axis, but the direction of revolution is not constant,
sometimes occurring clockwise, at others counter-clockwise.

The whole of the body with the exception of the extreme
anterior and posterior extremities is covered with flagella,
very little differentiated, except that those arising from
the peculiar tube-like organ at the anterior end—to be
described later—are a little longer than the rest, bemg 14-16 p
in length, while the remainder are about 12. Also these
anterior flagella are much more active during progression.
When the animal is stationary, however, the flagella still
show movement, the majority independently, but the anterior
ones in such harmony that they appear as paired thick bands
in whip-like undulation. I was unable to find any indication
of a prehensible function in the posterior flagella as described
by Kent (15) and Porter (24). The continuous movement of
the flagella, even though the animal is at ‘rest’, has been
described in T. campanula by Kofoid; doubtless the
function is to keep the body bathed in the intestinal fluid of
the termites. In shape the animal is almost oval, but there
is a gradual tapering from the anterior to the rounded posterior
extremity. There is no sharp demarcation into ectoplasm
and endoplasm except at the anterior end, where the proto-

THE PROTOZOA PARASITIC IN ARCHOTERMOPSIS WROUGHTONI 251

plasm is clearer than that of the rest of the body. The large
food particles are aggregated at the posterior two-thirds of
the body, and are always found behind the nucleus, as in
Trichonympha. This is, however, in sharp contrast
.to Grassi’s experience, for, in his last paper (7), he states that
in the Pseudotrichonympha the food particles are not
limited to the posterior extremity, but on occasion may be
seen in the region of the anterior organ “ mamella ’. Buscalione
and Comes (1), in their paper, state that when treated with
iodine dissolved in iodide of potassium, the region, near to
the nucleus, in Trichonympha, gives the characteristic
reaction of glycogen, and that this reacting region is sharply
defined from the rest of the body. In P. pristina, however,
the glycogenic reaction is diffused through the whole body,
being greatest behind the nucleus. This reaction and the
results of other microchemical tests will be fully discussed
in a forthcoming paper. As regards the method of food
ingestion I can supply no evidence beyond the fact that I have
been unable to find any trace of the peculiar process described
by Porter (24) in T. agilis. Kofoid and Swezy (18)—
apparently with reluctance—conclude that in Tricho-
nympha campanula the anterior organ (centroblepharo-
plast) may function also as a cytopharynx ; a view also held
by Buscalione and Comes. A grave objection to this conclu-
sion 1s that food particles are never found in the anterior region
of the body ; Kofoid and Swezy themselves say, ‘the anterior
region of endoplasm has, in all individuals observed, been
entirely free from food bodies or vacuoles, with the exception
of small darkly-staining rodlets which may be bacteria or
possibly chromidia’. This has been the experience of all
workers on Trichonympha, and Pseudotricho-
nympha pristina offers no exception to this rule. As
Porter says, “1t seems highly improbable—to say nothing of
the absence of any trace of a permanent oral structure—that
solid food should pass through this anterior region so quickly
that not a single case of its passage, or of its presence in this
part, should have been discovered by any of those who have

952 D. WARD CUTLER

studied these parasites’. One is thus driven to the belief
that the food is incorporated into the body at the posterior
region, though the method is still unknown.

MorpHOoLoGy.

P. pristina is a relatively large animal, its length varying
from 133-9-259-2 4 with a breadth of 60-5-111:2y». The
average size may then be taken as 226-3-99-9 wu. In stained
preparations it is evident that the whole of the body flagella
are arranged in longitudinal series (PI. 10, fig. 1). The extreme
posterior end is, however, naked, and in many preparations
there can be seen a collection of darkly stained bodies, triangu-
larly arranged with the apex directed anteriorly (PI. 10, fig. 2).
These granules are not to be found in every specimen and
are irregular as regards size, never attaining, however, to that
of the numerous food particles formed in other regions of the
body. From their general appearance and from the fact that
they are always confined to the naked posterior region of the
body, it seems possible that they are of an excretory nature
and that this naked region may be regarded as the physio-
logical anus of the animal. This is, however, a pure conjecture, as
IT have found no evidence of granules being ejected from this
region of the body.

Cell Inclusizons.

In preparations fixed by Fleming, as modified by Gatenby (5)
and then stained by Heidenhain’s iron haematoxylin, there
are seen, scattered through the entire plasma, numerous short
deeply-stained rods of a fairly uniform size and thickness
(Pl. 10, fig. 6). In appearance these bodies are very similar to
those found in Ditrichomonas termites and described
in a previous paper (2). On the other hand they in no way
resemble the cytoplasmic inclusions found in the various
animals investigated by Gatenby (5). As I have been unable
to carry out any of the tests requisite for an accurate deter-
mination of the various cell inclusions, I shall content myself
with simply recording their presence in Pseudotricho-
nympha pristina.

THE PROTOZOA PARASITIC IN ARCHOTERMOPSIS WROUGHTONI 253

Anterior Organ (Centroblepharoplast).

Pseudotrichonympha pristina terminates at the
anterior end in the curious organ found in the Tricho-
nymphidae and deseribed under various names by different
observers: thus the Italian workers designate it as ‘la botti-
glia’, ‘il cappuccio’, or ‘il mammiullare’; to it Hartmann has
applied the term ‘ Kopforgan’, and Porter ‘the nipple-like
part ’. Recently, however, Kofoid and Swezy have identified
it as a centroblepharoplast, the name which I prefer to adopt.
In P. pristina it is composed of two portions, an inner
tube-like one surrounded by a sheath which appears to cover
it completely (Pl. 10, figs. 3, 5). This ectoplasmic sheath at
its distal extremity becomes continuous with the rest of the
body, and this is the only region where differentiated ectoplasm
is found. I have been unable to detect any trace of a break
in the tip of the sheath such as one would expect were the
inner region in reality a tube capable of expelling or taking in
liquids as some observers would have us believe. Each anterior
flagellum takes origin from a granule situated on the mner surface
of the ectoplasmic layer of the centroblepharoplast. These
sranules are difficult to detect, but in a few suitable prepara-
tions they are unmistakably demonstrated (Pl. 10, fig. 5, B.G.).
Finally, from the extreme end of the organ there arise two
fine threads, which, taking a parallel course down the centre
of the endoplasm, diverge at their distal ends to reach the
nuclear membrane where they are attached (Pl. 10, figs. 3,
5, s.t.). It seems indubitable that there is such attachment,
for in specimens whose nuclei have been thrust out of position
the threads are still seen running to the membrane. ‘Thus
the nucleus is more or less fixed in position by these threads,
in contrast to the ‘ free’ nucleus described by Grassi.

Striations and Granules.

The striations that are seen crossing the body m a longitu-
dinal series arise from the centroblepharoplast. They consist
of ridges in the body surface, and thus broadly agree with those

954 D. WARD CUTLER

found in Joenopsis polytricha. Just beneath the
surface of these ridges numerous granules are located, from
each one of which a body flagellum has its origin (PI. 10, figs. 1,
4, 6, s.n., B.G.). The flagella origins are in the main similar
to those described by Kofoid and Swezy in T. campanula
and Leidyopsis sphaerica, except that I can find no
trace of oblique fibres running to the granules.

Nucleus.
This body is a large structure situated at the anterior end
of the body and possessing a well-developed membrane, always

TErxt-FIc. 1. TErxt-Fic. 2.

CB

‘Resting’ nucleus of P. pristina Similar to Text-fig. 1, but show-
showing chromatin blocks em- ing the tripartite nucleolus-
bedded in the plastin matrix. Note like body. x 1,800; s.A.,D.H.

the clear peripheral space with the
nucleolus-like body. x 1,880; s.a.,
aL

clearly visible (Pl. 10, figs. 1, 38, and 'ext-fig. 1). Inside the
membrane there is constantly present a clear space, while
the centre of the nucleus is filled with chromatin, in the form
of large irregularly-shaped masses lymg in a matrix of what is
probably plastin. The number of chromatin blocks appear
to be quite indefinite (‘Text-figs. 1 and 2, c.s.). Lying amongst
them there is commonly seen a large body, staming very
deeply with iron haematoxylin, which is sometimes distinetly
tripartite in nature (‘Text-fig. 2, n.). Unfortunately [have been

lor explanation of lettering of text-figures see pp. 263-4,

THE PROTOZOA PARASITIC IN ARCHOTERMOPSIS WROUGHTONI 255

unable to trace its origin and fate, but that it plays no part
in division is shown by its absence in dividing nuclei. Probably
it is cast out of the nucleus before division takes place. It
appears to have no relation with the curious * heterochromo-
some’ described by Kofoid in T. campanula.

DIVISION.

As in Joenopsis polytricha the reproductive phases
of P. pristina are difficult to find, and I have had to

TEXT-FIG. 3.

Early stage in the division of P. pristina; the centroblepharo-
plast has separated into two, leaving a split in the protoplasm.
From one of the centroblepharoplasts the threads still persist,
but with their distal ends free from the nuclear membrane.
x 1,000; S.A., H.I.H.

examine a large number of preparations to obtain those here
described. Division is initiated by the splitting into two of
the centroblepharoplast. This condition is rarely seen, partly
because it is rare to find an animal so orientated as to render
visible the split blepharoplast. Commonly it becomes inflected
on to the body plasma, thus rendering it very difficult to obtain
a clear picture. In the first stage of the process the two
suspensory filaments become detached from the nuclear

956 D. WARD CUTLER

membrane, thus rendering their distal ends free in the plasma ;
subsequently they are absorbed into the body (Text-fig. 3). The
actual divisions of the centroblepharoplast takes place exceed-
ingly rapidly, and I have not seen the intermediate phases.
It seems probable, however, that the splitting originates at
the posterior end and travels forwards, for in a good many
animals the basal region is double, but the anterior one still
single, though obviously much thicker than normal. At
the completion of division the plasma lying between the two
centroblepharoplasts splits, leaving a clear space which is

TEXT-FIG. 4.

Dividing nucleus with the chromatin in the form of a loose spirene.
XALEZDO Is Ss ACH ene

probably the initiation of division of the animal into two
(Text-fig. 3). The whole process recalls that described by Kofoid
and Swezy in T. campanula, and the incomplete descrip-
tion given by Hartmann for his male form of T. hert wigi;
a paradesmose, however, is not formed between the daughter
centroblepharoplasts in P. pristina. As already men-
tioned, the resting—non-dividing—nucleus is composed of large
irregular clamps of chromatin. At the onset of division these
chromatin blocks break up into a number of small rounded
svanules embedded in a matrix (Text-fig. 8). Soon the granules
become arranged to form a long spireme, and at this stage the
clear space between the membrane and the chromatin dis-
appears (‘Text-fig. 4). The nuclear membrane, however, remains

THE PROTOZOA PARASITIC IN ARCHOTERMOPSIS WROUGHTONI 257

intact, and can be seen throughout the whole process of
division. This is contrary to the statement made by Imms.
Directly after its formation the spireme is loosely packed
together, but subsequently its component parts become more
closely aggregated. Finally, it breaks up into a number
of long threads, which separate one from the other to form
the so-called chromosomes (Text-figs. 5 and 7), and the clear
space once more arises. These threads, however, do not
appear to split longitudinally, nor can they be seen to be
lying together in pairs previous to their separation. During
the process just described the nucleus elongates, becoming

TEXT-FIG. 5. TEXT-FIG. 6.

Nucleus in which the spireme Dividing nucleus with the ‘ chromo-
is breaking into individual somes * passing to each pole. Spindle
threads. x950; S.A., H.I.H. fibres or paradesmose not present.

15250 5) SA. (Dd); HH.

oval in shape, with the poles somewhat pointed. The long
chromosome-like threads now separate into approximately two
equal groups, one of which passes to either pole of the elongate
nucleus (Text-fig. 6). Further elongation occurs, and at the
same time the threads begin to aggregate to form a compact
mass, which finally breaks up into irregular chromatin masses
to form the daughter nuclei (Pl. 10, fig. 7). Finally, the
membrane constricts, dividing in the middle.

This process must take place rapidly, for it is common to find
bi-nucleate animals and animals in which the division phase
is being initiated, but it is exceedingly rare to encounter the
intermediate stages.

258 D. WARD CUTLER

As will have been noted, throughout the whole of the divi-
sion there is no development of spindle fibres, centrioles, or
paradesmose.

During the formation of the daughter nuclei the centro-
blepharoplasts migrate from each other, carrying with them
some of the flagella (Pl. 10, fig. 7).

I have been unable to discover the origin of the remaining
flagella or that of the suspensory threads to the nucleus.

TEXT-FIG. 7. TEXT-FIG. 8.

A slightly more advanced stage Early stage of nuclear division
than the one shown in Text-fig. 5. with chromatin blocks resolved
9050 : "5.40, Be. into numerous small granules.

x 9005. s.A. (D.J.), D.H.

The actual division of the animal into two probably does not
occur immediately after the formation of the daughter nuclei,
for binucleate animals are commonly encountered in which
the plasma shows no obvious sign of splitting.

The division is, however, longitudinal, for the daughter
centroblepharoplasts and nuclei always lie in a plane transverse
to the axis of the body. This longitudinal division is a further
indication of the flagellate relationship of P. pristina.

GENERAL CONSIDERATIONS.

Comparing P. pristina with the other species of Pseudo-
trichonympha it is evident that, im many respects, it
differs markedly from them. ‘The species deseribed by Hart-

THE PROTOZOA PARASITIC IN ARCHOTERMOPSIS WROUGHTONI 259

mann is larger than P. pristina, measuring 760-830» by
60-40 », and in shape it is more elongated, with well-defined
ectoplasm and endoplasm, the latter divided into internal and
external zones. As in P. pristina the body is traversed
with longitudinal ridges from which the flagella takes origin,
but basal granules are not definitely described, though Hart-
mann thinks that they may occur. The chief point of differ-
ence, however, is the centroblepharoplast. In Hartmann’s
organism it 1s composed of three distinct regions : (a) a cylin-
drical tube starting in the ectoplasm and extending to the
endoplasm ; (b) cap, covering the tube; (c) a second semi-
circular cap covering the whole of the anterior ectoplasm.
Hartmann suggests that the cap represents the true blepharo-
plast, and that the tube is formed of fused basal granules.
Obviously this ‘ Kopforgan ’ is of a more complicated structure
than its homologue in P. pristina, and the location of the
basal granules in the ectoplasm and not in ‘the tube’ in this
latter organism indicates that Hartmann’s suggestion as to the
origin of the ‘tube’ is not correct. Grassi’s latest description
of the Pseudotrichonympha is as follows: ‘ Body
much elongated and sharpened, with the flagella extending
over the whole of the body, leaving the posterior region naked.
The striations from which the flagella arise are seen running
longitudinally. The nucleus is found in various positions of
the body, and in its * resting ’ stage is composed of a membrane,
peripheral clear zone, and a central mass. The food, consisting
of wood, is not limited to the posterior region of the body, but
is sometimes found in the region of the ‘ mamella ”’.

“The four rods, characteristic of the suspension of the
nucleus in Trichonympha, are not found, and con-
sequently the position of the nucleus is not fixed.’

Grassi distinguishes four species, P. hertwigi var. minor
in Coptotermes Sjosteddi, P. hertwigi var.
major in Coptotermes lacteus, P. magnipapil-
losa nSchedorhinotermes putorius, and P. par-
ripapillosa in 8. intermedius.

The above is sufficient to show that the organism described

NO. 258 uy

260 D. WARD CUTLER

in this paper is undoubtedly a member of the Pseudo-
trichonympha.

The two threads in P. pristina, arising from the centro-
blepharoplast and distally connected with the nucleus, have
not been described in any of the other species, though Hart-
mann believes that he saw them on one occasion. In P. pris -
tina, however, they are conspicuous elements in practically
every animal observed, and undoubtedly function as suspensory
or supporting structures of the nucleus. Rods and threads,
often complicated in their arrangement, have been described
as supporting the nucleus in the Trichonympha, and
it is reasonable to believe that the two threads found in
P. pristina are the homologues of this nuclear * basket ’
deseribed by the Itahan workers.

Foa (4) has suggested that the threads of the Tricho-
nymphidae can be regarded as homologous with the
collar of Joenia, which Janicki regards as the parabasal
body of this animal. There seems to be little justification
for so homologizing the threads of Trichonymphidae,
but until our knowledge of these bodies is greatly extended
it is unprofitable to discuss their possible homologies.
Tt may well be that future research will show that many
of the so-called parabasal bodies are totally unrelated one
to another. As far as the evidence goes the Tricho-
nympha and Pseudotrichonympha do not possess
such bodies.

The nucleus of P. pristina is substantially lke that
described by Hartmann. As Imms states in his paper,
there is not the slightest evidence of it being of a poly-energid
nature; nor have I found any trace of secondary nuclei
scattered through the cytoplasms. It is surprising that such
a wonderful cycle of events as that described by Hartmann
could have been found in such a relatively simple nucleus as
that of the Pseudotrichonympha!

P. pristinais, I think, the first species in which the repro-
ductive phases have been followed : Hartmann describes a few
phases, which agree withsome described here. Thus he states that

THE PROTOZOA PARASITIC IN ARCHOTERMOPSIS WROUGHTONI 261

the blepharoplast (centroblepharoplast) first divides, followed
by a split in the protoplasm. The chromatin blocks of the
nucleus become resolved into granules, which aggregate to
formaspireme. These phases have been foundin P. pristina.
Hartmann’s further account, however, of the degeneration
of the primary nucleus and the formation of secondary nuclei,
with the final appearance of gametes, finds no counterpart in
the animal I have investigated. In one important respect
the nuclear division described by Hartmann differs from that
of P.pristina. In this species there is no trace of parades-
mose or spindle fibres, whereas Hartmann figures both these
structures. This is a point of interest, for in all the protozoa
of Archotermopsis, which I have investigated, the
division centres of the nucleus are either absent or poorly
developed.

Thus in Ditrichomonas termites (2) a paradesmose
is formed, but no spindle fibres, centrioles, &¢., whereas in other
Trichomonads they are described by Kuezynski (19) and
Kofoid and Swezy (16). In Joenopsis polytricha (2)
nuclear division occurs without any obvious centre, which is
not the case in any of the related animals ; for in Joenia (18)
and Parajoenia (18) a spindle is formed. Finally, as
already noted, the Pseudotrichonymphid deseribed
by Hartmann has a paradesmose and spindle fibres; as is
also the case in Trichonympha major and minor
deseribed by Foa (4). In P. pristina such structures are
entirely lacking.

Thus in all the protozoa examined from the gut of
Archotermopsis wroughtoni the nuclear division is
very different from that found in related species.

Further, in D. termites the nuclear division and the
locomotor complex is of a more primitive nature than
that described for other Trichonomads ; a statement probably
true for Joenopsis polytricha and Pseudotricho-
nympha pristina. It appears that the protozoa to
which A. wroughtoni is host are in general more primi-
tive than those inhabiting other species of termites. Imms

ar

262, D. WARD CUTLER

describes A. wroughtoni as ‘one of the most primitive
of living Termites’. The association, therefore, of primitive
parasites or ‘ guests ’, whichever the case may be, with a primi-
tive host is extremely interesting, and is suggestive that the
two groups of organisms have remained associated together
for a long period, neither having developed into more complex
species, as has occurred with other termites and their associated
protozoa.

In conclusion, I wish to express my thanks to Mr. J. B.
Robinson for re-drawing for publication, with the exception
of figs. 4, 5, 6, and 7, the figures illustrating this paper.

REFERENCES.

1. Buscalione, L., and Comes, 8. (1910).—** La digestione delle vegetali
per opera dei Flagellati contenuti nell’ intestino dei Termitidi e il
problema della simbiosi ”’, ‘ Atti Accad. Gioenia Catania ’, vol. 13.

2. Cutler, D. W. (1919-20).—*‘ Observations on the Protozoa Parasitic
in the Hind Gut of Archotermopsis wroughtoni, Desn.”,
parts i, ii, “ Quart. Journ. Micro. Sci.’, vols. 63, 64.

3. Doflein, F. (1916).—* Lehrbuch der Protozoenkunde ’, 4th ed.

4, Foa, A. (1904).—* Ricerche sulla riproduzione dei Flagellati: II.
Processo di divisione delle Triconimfe”’, ‘ Atti Accad. Lincei’,
vol. xiii. ‘

5. Gatenby, J. B. (1917-20).—* Cytoplasmic Inclusions of the Germ
Cells *, * Quart. Journ. Micro. Sci.’, vols. 62, 63, 64.

6. Grassi, B. (1911).—‘‘ Intorno ai Protozoi dei Termitidi’’, ‘ Rend.
R. Accad. Lincei ’, vol. xx (1).

7. —— (1917).—* Flagellati viventi nei Termiti”, ‘Mem. R. Accad.
Lincei ’, vol. xii.

8. Grassi, B., and Foa, A. (1904).—‘‘ Ricerche sulla riproduzione dei
Flagellati: I. Processo di divisione della Joenia e forme affini”’,
“Rend. Atti Accad. Lincei’, vol. xiii.

9. Hartmann, M. (1910).—** Untersuchungen iiber Bau und Entwicklung
der Trichonymphiden (Trichonympha hertwigi, n. sp.)”’, * Fest-
schrift zum sechzigsten Geburtstag Richard Hertwigs ’, vol. i.

10. Hickson, 8. J. (1903).—** The Infusoria or corticate Heterokaryota ”’,
in Lankester, ‘ Treatise on Zoology ’, 1, sec. 2.

11. Imms, A. D. (1919).—** On the Structure and Biology of Archoter-
mopsis, together with Descriptions of New Species of Intestinal
Protozoa and General Observations upon the Isoptera”’, * Phil,
Trans. Roy. Soc, B.’, vol. 209,

ee ps

ee

= «a

THE PROTOZOA PARASITIC IN ARCHOTERMOPSIS WROUGHTONI 263

12. Janicki, C. (1911).—** Zur Kenntnis des Parabasalapparatus _ bei
parasitischen Flagellaten ”, ‘ Biol. Cent.’, Bd. xxxi.

(1915).—** Untersuchungen an parasitischen Flagellaten ”’, * Zeit.
f. wiss. Zool.’, Bd. exii.

14. Kent, 8. (1880-2).—* A Manual of Infusoria ’.

15. (1884).—** Notes on the infusorial parasites of the Tasmanian
white ant”, ‘ Proc. Roy. Soc. Tasmania ’.

16. Kofoid, C. A., and Swezy, O. (1915).—** Mitosis and Multiple Fission
in Trichomonad Flagellates ’’, ‘ Proc. Am. Acad. Arts. Sci.’, vol. 51,

13.

a7. (1919).—** Flagellate affinities of Trichonympha”’, * Proc. Nat.
Acad. Sci.’, vol. 5.
18. (1919).—** Studies on the Parasites of the Termites: Tricho-

nympha campanula sp. nov. and Leidyopsis sphaerica gen. nov.,
sp. nov. ”, “ Univ. Calif. Publ.’, vol. 20.

19. Kuezynski, M. H. (1914).—** Untersuchungen an Trichomonaden ”,
* Arch. f. Protistenk.’, Bd. xxxiii.

(1918).—‘ Uber die Teilungsvorgiinge verschiedener Tricho-
monaden und ihre Organisation im allgemeinen ”’, ibid., Bd. xxxix.

21. Leidy, J. (1877)—* On Intestinal Parasites of Termes flavipes”,
‘Proc. Acad. Nat. Sci. Philadelphia ’.

(1881).—** The Parasites of the Termites’, ‘ Journ. Acad. Nat.
Sci. Philadelphia *, vol. 8.

23. Poche, F. (1913).—** Das System der Protozoen”’, * Arch. f. Proti-
stenk.’, vol. 30.

24, Porter, J. F. (1897).—* Trichonympha and other Parasites of Termes
flavipes ’’, * Bull. Mus. Comp. Zool. Harvard Coll.’, vol. 31.

25. Senn, G. (1900).—** Flagellata ”’, in Engler and Prantl, * Die natir-
lichen Pflanzenfamilien ’, vol. 1, part 1.

Stein, F. (1878)—*Der Organismus der Flagellaten nach eigenen
Forschungen in systematischer Reihenfolge bearbeitet. Der
Organismus der Infusionsthiere. 8. Abth. Naturgeschichte der
Flagellaten oder Geisselinfusorien.’

20.

22.

26

EXPLANATION OF PLATE 10 AND TEXT-FIGURES.

All the figures are drawn from fixed and stained preparations. The
optical apparatus employed was as follows: Zeiss apochromatic oil-
immersion objective 2mm. (N.A. 1-3), and compensating oculars 4, 6,
12, 18. Critical illumination was always employed. The method of
fixing and staining, and the approximate magnification is given below
in the case of each figure. The following abbreviations are employed :
s.A.=Schaudinn’s sublimate-alcohol mixture. s.A. (D.J.) =Schaudinn’s
sublimate alcohol as modified by Dobell and Jepps. Fl. (Gat.) = Fleming's

2964 D. WARD CUTLER

strong fluid as modified by Gatenby. 4.1.H. = Heidenhain’s iron-alum
haematoxylin. D.H.= Dobell’s iron-alum haematein. The lettering of the
figures is as follows: B.G.,= basal granules. o.B. = chromatin blocks.
c.BL, = centroblepharoplast. ©.1.= cell inclusions. ¥.B.=food bodies.
N. = nucleolus-like body. s.R.=striations. s.T. = suspensory threads.

Fig. 1.—Stained preparation of Pseudotrichonympha pristina
showing ‘ resting’ nucleus, striations with basal granules, food particles
behind the nucleus. x 300; S.A. (D.J.), H.I.H.

Fig. 2.—Posterior region of animal with triangular-shaped collection of
granules. Note the region without flagella. 950; s.4., HLH.

Fig. 3.—Anterior region of P. pristina with centroblepharoplast, from
which arise the two threads running to the nuclear membrane. x 950;
SIAL (D25.); H.ieH.

Fig. 4.—Portion of a section through body of the animal, showing the
ridges (striations) under which are situated the basal granules, the origin
of the flagella. x300; FI. (Gat.), 4.1.4.

Fig. 5.—Centroblepharoplast of P. pristina with the threads and
basal granules from which the anterior flagella spring. 1,000; s.a.
(D.J.), H.I.H.

Fig. 6.—Posterior region of the body, showing the basal granules and
flagella. The endoplasm contains unidentified cell inclusions. x 1,000 ;
Fl. (Gat.), H.1.H.

Fig. 7.—Top view of a late phase in the division of P. pristina,
the centroblepharoplasts are situated at either side of the body with the
dividing nucleus between them. Note the absence of any division centre.
x 950; S.A. (D.J.), D.H.

Quart. Journ, Micr. Sci.Vol.65,NS.PL10,

o

The Cytoplasmic Inclusions of the Germ-Cells.'
Part IX. On the Origin of the Golgi Apparatus
on the Middle-piece of the Ripe Sperm of Cavia,
and the Development of the Acrosome.

By
J. Bronté Gatenby, B.A., B.Se., D.Phil. (Oxon.),

Lecturer in Cytology, University College, London, and Senior Demy,
By oJ? 3 y ed
Magdalen College, Oxford ;

AND

J. H. Woodger, B.Se. (Lond.),

Assistant in Zoology and Comparative Anatomy, University College,
London.

With Plates 11 and 12 and 2 Text-figures.

CONTENTS.
PAGE
1. INTRODUCTION . 266
2. Partl. The Dev sieptnente 5 the Mena Middle- piece ; Gale:
Apparatus : é : : : : 200
3. PartIl. Literature . : : : : F : 20
4. Technique brea es
5. GENERAL DESCRIPTION OF THE Ba A\VIOUR OF THE nage SIONS

OF THE CyTOPLASM IN CAVIA SPERMATOGENESIS . 274
6. Pertop I. The Growing Spermatocyte . 5 : : . 2714
7. Pertop Il. Maturation Divisions 275
8. Pertop III. The newly-formed Spermatid 276
9. ON THE SUBSEQUENT BEHAVIOUR OF THE GOLGI eee ARATUS

AND ARCHOPLASM . , ; i c ~~ 2s
10. THE CASE OF THE RAT SPERMATOZOON . : : : oa

11. Discussion.
(a) On the Origin of the Acrosome in Animal Spermatogenesis 279
(b) The Middle-piece of the Spermatozoon after Entry into

the Egg . , : ; . : ‘ : ee eaol
12. SUMMARY . i ‘ ‘ : ; : : F . 284
13. BIBLIOGRAPHY . A ‘ é ; e Zon
14. Description oF PLATES Il AND 2 : - : F ~ 289

1 Part of the materials used for this research was provided by a Govein-
ment Grant of the Royal Society, for which thanks are expressed.

266 J. B. GATENBY AND J. H. WOODGER

1. [yrRopuctTIon.

Iv is well known from the work of Retzius that the middle-
piece of the ripe spermatozoa of many mammals bears around
itself a small clear bead of protoplasmic material which can
be easily recognized in the fresh sperm.

In 1912 Weigl (82) published some comparative studies
on the Golgi apparatus of the somatic- and germ-cells of
different animals, in which he showed that the protoplasmic
bead on the middle-piece of the spermatozoon of the guinea-
pig contained structures possessing all the miucrochemical
characteristics of true Golgi elements.

The work out of which the present paper arose was primarily
undertaken with a view to discovering the mode of origin of
these argentophile structures from the Golgi apparatus of the
spermatid and spermatocyte.

The first part of this paper consists, therefore, of a description
of our results in this field.

The study of the Golgi apparatus of the spermatocytes and
spermatid naturally led, however, to the investigation of the
relations of this structure to other cell constituents, especially
to the acrosome.

The development of the acrosome in Cavia has been the
object of repeated study by Niessing, Moore, Meves, and others,
and quite recently by Papanicolaou and Stockard, but the
exact relation of this body to the Golgi apparatus has not
hitherto been described.

Our observations upon this point form the second part of
the present paper, and we have also attempted to give a general
account of the spermatogenesis of Cavia based upon the
confirmed results of modern workers, together with certain
suggestions for a revised and simplified English nomenclature
of the subject.

2. Parr I. The Development of the Definitive
Middle-piece Golgi Apparatus.

Retzius, as is well known, has published a large number of

drawings of various mammalian and other spermatozoa. If

—

—_—. —_—

CYTOPLASMIC INCLUSIONS OF THE GERM-CELLS 267

we examine his figures (29), we find, as has already been
mentioned, that Retzius has represented In many mammalian
spermatozoa a small bead of protoplasm on some part of the
middle-piece. In our Text-fig. 1 are reproduced six of this
observer’s figures, showing at x the bead of the middle-piece.

TeExt-FiGc. l.

an

Ripe spermatozoa after Retzius (29). a= pig. B = sheep.
c=rabbit. p=cat. E=lemur. F = hedgehog; showing at x the
protoplasmic bead associated with the middle-piece.

In fig. 1, a is the spermatozoon of the pig; fig. B that of the
sheep ; fig. c, the rabbit; fig. p, the cat; fig. B, the lemur
(Lemur catta); and fig. r, the hedgehog. A glance through
the work of Retzius shows that this peculiar bead has been
figured by him in several other mammals, namely: Sciurus

268 J. B. GATENBY AND J. H. WOODGER

J

vulgaris, Cynomys, Myoxus glis, Cavia, Equus, Capra,
Alces, Bos, Canis, and doubtfully in Dicotyles. In the sperma-
tozoa of the following the bead does not appear in Retzius’s
figures : Homo, Didelphys, Talpa, Bradypus, Dipus, Hystrix,
Lemmus, Mus, Myopotamus, Cervus, Rangifer, Globicephalus,
Vulpes, Meles, Halichaerus, Hapale, and Innus. Some of

TEXT-FIG. 2.

A Da Fano (8) preparation of the epididymis of Cavia. At N is
the nucleus, and at GA the Golgi apparatus of the cells of the
epididymis. At x are the middle-piece Golgi apparatus of
the ripe spermatozoa impregnated like the Golgi apparatus of
the epididymis cells. (Original.)

these are, however, doubtful, and may possess the bead in
a very reduced and atypical condition.

If now, as Weigl (82) has shown, the epididymis of Cavia
be prepared by one of the Golgi apparatus techniques (Golgi,
Cajal, or Da Fano), the protoplasmic beads of the free sperma-
tozoa lying within the tubules are all found to contain a number
of little rodlets or elongate platelets as shown in Text-fig. 2
at x. In this figure, drawn from a preparation by Da Fano’s

CYTOPLASMIC INCLUSIONS OF THE GERM-CELLS 269

cobalt-nitrate-silver method (8), the magnification is too low
to show the minute structure of the bead; at Nn is the nucleus
of the cells of the epididymis wall; and at Ga the Golgi
apparatus of these cells. In all preparations we possess, the
Golgi apparatus of the epididymis wall and of the bead-
contents of the middle-piece are the only objects which go
black with the reduced silver. In Pl. 11, fig. 3, a nearly ripe
Cavia spermatozoon is drawn to illustrate the more minute
structure of the bead (GAx) after treatment with Cajal’s method.

The question now arises: What relation does the impreg-
nating middle-piece bead (GAx in PI. 11, fig. 3) bear to the
Golgi apparatus of the spermatid cell (Ga in Pl. 11, fig. 3, and
GE in Pl. 12, figs. 7, &e.) ?

Extensive trials were made with Golgi apparatus techniques,
and our best preparations were examined independently by
both of us. We believe that the conclusion which each of us
has arrived at independently is the correct one, but at the same
time it is recognized that to come to a definite conclusion is
difficult.

In Pl. 11, fig. 2, is drawn a ripening spermatid in which the
Golgi apparatus (Ga) les in the hinder part of the cell. It is
from a preparation made by Cajal’s unmodified Golgi apparatus
method, and the mitochondria appeared as leht golden
spheres (m). The most striking point to be noted is the
undoubted double structure of the Golgi apparatus, which has
a distinct bead projecting from its surface on one side (GAx).
At this stage in the development it is possible to find pockets
of cells within the testis in which every Golgi apparatus has
this double appearance. If the spermatid be examined at
earlier stages such as in Pl. 11, fig. 18, the bead (¢ax) can still
be seen as a swelling on the surface of the Golgi apparatus.

With Cajal, Da Fano, or Kopsch methods, it is found that
this outgrowing bead is not homogeneous—its centre being
formed of a more lightly impregnating material, closely
resembling archoplasm in its appearance. If, moreover, ripe
spermatozoa are fixed in some such mixture as Flemming or
Hermann, and stained in acid fuchsin, it will be noted that the

270 J. B. GATENBY AND J. H. WOODGER

middle-piece bead stains like the archoplasm of the spermatid,
i.e. a deep pink or reddish. We consider, therefore, that the
outgrowing bead figured by us in PI. 11, fig. 2, and Pl. 12, figs. 18
and 14, probably consists of detached portions of both archo-
plasm as well as Golgi apparatus elements.

Tracing now the history of the bead after the stage at which
it still adheres to the main Golgi apparatus, we next find that
it has become separated from the latter in the manner shown
in Pl. 12, fig. 14. Ina large number of cases the bead has been
observed lying in a position intermediate between the main
Golei apparatus and the nucleus, that is, near the letter m
mc Pls ee!

In the majority of cases the Golgi apparatus bead of the ripe
sperm of Cavia hes in the position shown in PI. 11, fig. 3, and less
commonly in the position indicated in PI. 12, fig. 16. Reference
to Text-fig. 1 shows that the middle-piece beads in other animal
sperms vary a good deal in position.

It seems probable that the small Gogh apparatus bead moves
up from its position in Pl. 11, fig. 2, or Pl. 12, fig. 14, to its defini-
tive position near the head centrosome-complex (PI. 12, fig. 15),
the bead becoming applied to the ‘ skeleton’ of the middle-
piece (mp in PI. 12, fig. 14) at a time when the mitochondrial
sranules (mM) are themselves becoming grouped around the
skeleton.

3. Part If. Literattre.

Meves (20), in his classical paper on the spermatogenesis,
has given a detailed review of previous work on Cavia. ‘To
this the reader may be referred. More recently Papanicolaou
and Stockard (26) have gone over the same ground, and also
given a comprehensive review of the results of previous
observers. The work of Papanicolaou and Stockard is chiefly
concerned with the fate of the archoplasm (their ‘ idiosome ’)
and its contents based on a study of material stamed with
a new methylene-blue-acid fuchsin combination after Zenker’s
fixation. The following is a brief résumé of thew account,
using their new and elaborate terminology.

CYTOPLASMIC INCLUSIONS OF THE GERM-CELLS 271

(1) In the Primary Spermatocyte the idiosome is differen-
tiated imto an outer blue-staining ‘idioectosome’ and an
inner purple-staining ‘ idioendosome *. (2) During the prepara-
tion for the First Maturation Division the idioectosome
disappears and, during the division, the substance of the
idioendosome becomes scattered through the cytoplasm in
the form of minute granules called ‘ idiogranulomes ’. (8) In
the Secondary Spermatocyte a new idioectosome is re-formed,
containing the idiogranulomes. (4) During the Second Matura-
tion Division the idiogranulomes are again scattered through
the cytoplasm. (5) In the re-formed idioectosome of the
spermatid each idiogranulome is seen to be surrounded by
a clear vacuole--the ‘ idiogranulotheca ’. (6) The idiogranu-
lomes rapidly fuse to form a single large red-staining ‘ idio-
spherosome ’ enclosed in a large vacuole, the ‘ idiosphaero-
theca ’ formed by the fusion of the idiogranulothecae. (7) The
idioectosome now begins to move away to one side and is
re-named the ‘idiophthartosome’. Meanwhile the idio-
Sphaerosome secretes a crescentic blue-staining ‘ idiocalypto-
some ’, and is itself known henceforth as the ‘ idioeryptosome ’.
(8) In the ripe spermatozoon the idiophthartosome disappears
with the cytoplasm which is lost durmg metamorphosis. The
idiocrypto- and idiocalypto-somes together form a double cap
to the sperm-head called the ‘spermiocalyptra’, and the
idiosphaerotheca * persists through all later stages and develops
into a membranous cover for the cap and head of the sperm ’,
and is then known as the ‘ spermiocalyptrotheca ’.

As we shall mention below, we have not been able to confirm
the statement of these observers as to the scattering of the
‘idiogranulomes ’ during the maturation divisions, but we have
adopted their account for several reasons.

We cannot, however, feel that Papanicolaou and Stockard
have really improved the nomenclature of the subject by the
introduction of these cumbersome new terms.

In the following table we have placed side by side the new
terms of these authors and the corresponding synonyms used
by previous workers. In the third column we have put forward

972, J. B. GATENBY AND J. H. WOODGER

suggested English equivalents based upon those used by
previous English authors, wherever these do not involve any
ambiguity.

We object to the term ‘idiosome’ because it has already
been used by Whitman (88) to mean ‘ an ultimate hereditary
unit’. The term ‘ archoplasm’ has been used by Moore (22),
and we have adhered to it. We have avoided the ‘ archo-
plasmic vesicle’ of Moore because it has sometimes been
applied to the whole of the archoplasm, but we have sub-
stituted ‘ archoplasmic vacuole’ instead. The only new term
we have introduced is ‘ Proacrosomic granules * for the minute
granules (idiogranulomes) of Papanicolaou and Stockard, which
ultimately fuse to form one large * Proacrosome ’, from which
the acrosome is later differentiated. No one can object to this
word for it is self-explanatory. It will be noted that we have
explained all the complicated processes leading to the formation
of the acrosome, without having recourse to the invention
or adoption of a terminology of the type introduced by Papani-
colaou and Stockard.

Papanicolaou Suggested English
and Stockard. Older Authors. Equivalent.
Idiosome. Idiozome (Meves). Sphire Archoplasm (AR).

(Niessing and Meves). Ac-
cessory corpuscle (Brown).
Nebenkern (Hermann). Archi-
plasm (Benda). Archoplasm

(Moore).
Idioendosome. Markschicht der Sphiire (Nies- Inner region of archo-
sing). plasm.
Idioectosome. Rinderschicht (Niessing). Outer region of

archoplasm.
Idiogranulomes. Archosomes (Moore). Kérn- Proacrosomic
chen (Meves). Microsomen- granules (APG).
strata (Niessing).
Idiogranulothecae. Archoplasmic vesicles (Moore). Archoplasmic

Blaschen (Meves). vacuoles (VV).
Idiosphaerosome The archosome (Moore). Das Proacrosome (PRA),
(becomes idio- Korn (Meves), Die stark fiirb-
cryptosome). baren Korper (Benda). Mito-

som (Niessing).
Idiosphaerotheca. Archoplasmic vesicle (Moore). Archoplasmic

Blaschen (Meves). | Vacuole vacuole (v)
(Benda). Helle Membran
(Niessing).

CYTOPLASMIC INCLUSIONS OF THE GERM-CELLS 273

Papanicolaou Suggested English
and Stockard. Older Authors. Equivalent.

Idiocalyptosome. Periphere Zone des Spitzen- Outer zone of acro-
knopfs (Meves). Ausserer Teil some (02).
des Mitosomes ( Niessing).
Idiocryptosome. Innenkorn (Meves). Dunkler Inner zone of acro-
Teil des Mitosomes (Niessing). some (IZA).
Spermiocalyptra. Spitzenknopf or Spitzenkérper Acrosome (A).
of German authors. Acrosom
of v. Lenhossék.
Spermiocalyptro- Kopfkappe of German authors. Covering membrane

theca. of acrosome (CA).
Idiophtharto- Archiplasmarest (Benda). Idio- Golgi elements with
some (Idio- zomrest (Meves). archoplasmic re-
ectosome) mains (GA),

4. TECHNIQUE.

The guinea-pigs used for this work were nearly all supplied
to us by Mr. H. M. Carleton and Mr. J. S. Haldane of New
College, Oxford, to whom our thanks are given.

We used especially the Golgi apparatus techniques of Cajal
and Mann-Kopsch, as well as many other methods. One of
us (J. H. W.) carried out a large number of tests with the
Cajal method in order to ascertain the best time to leave the
testes in the formalin fixative. It was found that twenty-four
hours in the fixative and twenty-four hours in the silver bath
gave the best results, though it was always very difficult to get
really satisfactory preparations with any of the formalin-
silver nitrate methods.

We used the methods of Stockard and Papanicolaou with
fairly satisfactory results, but never got preparations quite
so clear as drawn in their figures. At a later stage in this work
we tried Da Fano’s new cobalt formalin method, which gave
useful results. We also made some excellent Mann-Kopsch
preparations (three hours Mann’s fluid, two weeks 2 per cent.
OsO,), but Flemming without acetic acid and Champy gave
poor results.

274 J. B. GATENBY AND J. H. WOODGER

5. GENERAL DESCRIPTION OF THE BEHAVIOUR OF THE INCLU-
SIONS OF THE CYTOPLASM IN CAVIA SPERMATOGENESIS.

We have compiled the following descriptions, and also
Pl. 12, after a personal study of many preparations of guinea-
pig testes, and also after a careful examination of the literature
of the subject. The works of Niessing (24), Meves (20), Brown
(2), Benda (1), v. Lenhossék (18), Moore (22), and Stockard
and Papanicolaou (26), have been considered especially with
reference to the formation of the acrosome. Regaud (27)
and Duesberg (6) have also been consulted and their various
statements examined. A good many of our results are quite
new, especially with reference to the Golgi apparatus.

6. Periop I. Growing Spermatocyte.

The mitochondria and Golgi apparatus are to be found in
the so-called germinal epithelial cells ; during the growth of
the spermatogonium, the mitochondria, which hitherto tended
to surround the region of the archoplasm, become spread
throughout the cytoplasm, while the Golgi apparatus and
archoplasm increase in size. Some time before the spermato-
cyte has become full-grown the archoplasm becomes distin-
ouishable into two regions—an outer clearer part, and an inner
chromophile part formed by the proacrosomic material.

In Pl. 12, fig. 5, is drawn the spermatocyte just about to
begin the first maturation division. The chromosomes are
appearing within the nucleus and are connected to one another
here and there by chromatic or linin filaments. Throughout
the cytoplasm the mitochondria (m) are scattered haphazardly.
At cup is the enigmatic chromatoid body, which later may be
found in each spermatid, and which apparently therefore may
divide during cell-division. The Golgi apparatus and the
archoplasm are at an. By this stage the inner region of the
archoplasm containing the proacrosomic material has resolved
itself into a large number of discrete granules which have been
figured by Moore, Meves, Niessing, and Stockard and Papani-
colaou, and which we propose to call the proacrosomic granules

CYTOPLASMIC INCLUSIONS OF THE GERM-CELLS 275

(apa), as it is they which ultimately form the acrosome, or
head-cap of the sperm.

In PI. 12, fig. 5, the Golgi apparatus is seen to consist of a large
number of semilunar platelets, rodlets, or dictyosomes (ar),
which le upon the outer surface of the archoplasm. By
Mann-Kopsch technique the Golei apparatus is not a reticulum,
but 1s as drawn in Pl. 11, fig. 1 (ar), and Pl. 12, fig. 5. Examined
after Cajal’s method, or by Da Fano’s modification of Cajal’s
formalin-silver nitrate method, the Golgi apparatus is seen to
be in the form of a reticulum, or of flat plates joined here and
there, as’ shown in figs. 2 and 38 of Pl. 11.

7. Pertop Il. Maturation Divisions.

The periods of division of the spermatocyte are difficult
properly to study. In very little of our material were mitoses
to be found, and this part of our work is the section about
which we feel the most diffident to write. Meves, Niessing,
and Moore all failed to follow the proacrosomic granules
through the phases of the maturation divisions, and we have
been unable to establish Papanicolaou and Stockard’s claim
that these granules retain their individuality and become
sorted out to the daughter cells during cell-division. Meves,
Niessing, and Moore all agree that the proacrosomic granules
soon become visible after the archoplasm is re-formed subse-
quent to division—that is in the late telophase. We have
adopted Papanicolaou and Stockard’s description for two
reasons: firstly, it is extremely unlikely that the proacrosomic
sranules would gradually accumulate and grow, especially
before the first maturation division—only to become disin-
tegrated at the mitotic prophase; and secondly, we are aware
that the Golgi elements or dictyosomes hitherto had not been
followed through division, but we now know that in mammals
as well as invertebrates the Golgi elements may become
sorted out during division and do not lose their individuality.

In Pl. 12, fig. 6, we give a diagram illustrating the inter-
pretations we at present consider to be the most likely to be
correct : the mitochondria are spread haphazardly throughout

NO. 258 U

the cytoplasm, and they offer no remarkable behaviour for
study. Around each mitotic aster are grouped approximately
one-half of the Golgi elements or dictyosomes ; for con-
firmation of this phenomenon in cells other than those of
the guinea-pig testis see Deinecka (8), Golgi (14), Murray (28),
Perroncito (25), Fauré-Fremiet (9), and Gatenby (11, 18).
This behaviour of the Golgi element or dictyosome does not
entail any sort of division of the element itself, but only
a haphazard, though subequal, sorting out of the whole
elements between the daughter cells.t

At apa in fig. 6 of Pl. 12 are the proacrosomic granules,
which become scattered in the cytoplasm during division,
As with the Golgi elements, the individual granules in the
spermatocyte archoplasm are not themselves divided, but
sorted out whole between the daughter cells.

At cup is the chromatoid body whose fate in the maturation
divisions has not been followed out ; one fact, however, may
be mentioned, it is that by far the majority of spermatids
contain a chromatoid body (Pl. 12, fig. 7, cus). In many
animals the spermatocyte and spermatid contain a chromatoid
body of some kind, and in the case of Smerinthus strong
evidence has been accumulated which indicates that this
body has the power of binary fission (10).

8. Periop III. The Newly-formed Spermatid.

In Pl. 12, fig. 7, is a drawing of the newly-formed spermatid ;
it contains the same categories of cytoplasmic elements as the
spermatocyte, only they are approximately one-quarter in
amount. With reference to the fact that the spermatid cell
is generally much more than one-quarter the size of the
spermatocyte, it may be pointed out that between the stages
drawn in PI. 12, figs. 5 and 7, there must be a period during
which the cells are rapidly growing. While it is certain that
the spermatid Golgi apparatus and archoplasm is usually

! Dictyokinesis in the maturation of the germ-cells of Mus, Cavia,
Stenobothrus, Limnaea and Helix is the subject of a forthcoming paper

by Ludford and Gatenby. The process is even more haphazard than
depicted in fig. 6.

CYTOPLASMIC INCLUSIONS OF THE GERM-CELLS OT

more than one-half the size of the same structures in the
spermatocyte, it is difficult to obtain satisfactory evidence of
any increase in size of the individual mitochondria.

With reference to the sorting out of the Golgi elements or
dictyosomes during the maturation divisions, attention is
drawn to recent work on Limax agrestis, where it has
been demonstrated that the number of dictyosomes in the
spermatocyte is eight, and in the spermatid two (18). In
all probability, though no count is possible in Cavia, the
number of platelets or dictyosomes in the spermatid is approxi-
mately one-quarter the number in the spermatocyte.

Within the archoplasm of the spermatid the proacrosomic
sranules have collected (or according to Meves, Niessing, or
Moore, now become visible again) (PI. 12, fig. 7, apa); but very
soon around each proacrosomic granule a clear ring appears,
so that the granule reposes in a vacuole—the archoplasmic
vacuole: the proacrosomic granules together with their
vacuoles in which they lie, now tend to run together, so that
one obtains the appearance of a number of granules, some
larger than others (PI. 12, fig. 7, ape).

At this stage the centrosome is dividing in the cytoplasm,
near, but outside, the archoplasm (PI. 12, fig. 7, c).

In the next stage the proacrosomic granules have run
together so as to form two or three large grains, each surrounded
by the clear vacuolar rmg—the archoplasmic vacuole (Pl. 12,
fig. 8, ape). The whole Golgi apparatus and archoplasm
gradually passes to the anterior pole of the cell, i.e. that part
of the cell which gives rise to the head end of the sperm, and
which most commonly is directed towards the germinal
epithelium. In PI. 12, fig. 8, the Golgi apparatus and archo-
plasm are shifting in an upward direction (according to the way
this cell has been drawn on the Plate). From the posterior
end of the cell, the axial filament grows out from the centro-
somes (c1 and ©).

The next stage in the formation of the acrosome is depicted
in Pl. 12, fig. 9. A part of the nucleus is shown at Nn, and the
Golgi apparatus plus the archoplasm lie nearly in front but to
one side of the nucleus. The whole apparatus lies in contact

U 2

278 J. B. GATENBY AND J. H. WOODGER

with the nucleus at one spot. A change has come over the
proacrosomic structures: these have finally fused to form
a single large bead, the proacrosome, within its vacuole (vy),
and around the entire periphery of the inner granule an outer
rind has been secreted (:za). These two regions are known
as the outer and inner region of the proacrosome (hitherto
proacrosomic granules). The proacrosomic apparatus moves
through the archoplasm and finally becomes stuck upon the
surface of the nuclear membrane, towards the front end of the
nucleus, and hereafter may be called the acrosome (PI. 12,
fig. 10). On the side of the acrosome which touches the nuclear
membrane the outer region of the acrosome is completely
pushed away, so that the inner region of the acrosome alone
touches the nuclear membrane in the mid-region of the
acrosome ; at the edges, however, as shown in PI. 12, fig. 11,
the outer -region of the acrosome les in contact with the
nuclear membrane.

The Golgi apparatus (i.e. all the dictyosomes), and the
archoplasm upon which it lies, keeps its position, partly embrac-
ing both the acrosome and one side of the nucleus (as shown
in Pl. 12, figs. 10 and 11) some considerable time, during which
the two parts of the acrosome grow rapidly. Eventually,
however, the apparatus and the archoplasm break away as
shown in PI. 12, fig. 12, and begin to drift back towards the tail
end of the spermatid (Pl. 12, fig. 13).

The inner region of the acrosome gradually becomes flattened
out on the front of the spermatid nucleus, and the whole
structure undergoes the changes shown in PI. 12, figs. 12-15.

On the Subsequent Behaviour of the Golgi
Apparatus and Archoplasm.

sv the stage drawn in PI. 12, fig. 12, the Golgi elements and
archoplasm have begun to drift down the elongating sperm
cell, and in PI. 12, fig. 18, this apparatus has completely flowed
away from the nucleus. Between the stages depicted in Pl. 12,
figs. 18 and 14, the definitive middle-piece Golgi ooo
appears as deseribed by us on p. 269.

3etween the stages in Pl. 12, figs. 15 and 16, the apparatus

CYTOPLASMIC INCLUSIONS OF THE GERM-CELLS 279

and the archoplasm flow into the bead, which sloughs off,
and take no part in the subsequent development of the sperma-
tozoon. In PI. 12, fig. 16, the apparatus and archoplasm have
undergone degeneratory changes.

10. The Case of the Rat Spermatozoon.

Retzius (29), as we have mentioned above, does not figure
a protoplasmic (Golgi) bead on the ripe spermatozoon of the
rat or mouse, and apparently it would have seemed to be one
of the exceptions to the rule that the mpe mammalian sperm
carries a Golgi apparatus. Our friend Dr. Da Fano of King’s
College, London, who has made preparations of the rat testis
by his new cobalt methods, examined at our request his
preparations of rat epididymis, with the result that he found
that each ripe sperm does carry a small bead which impregnates
with silver nitrate. Retzius, therefore, overlooked this bead
in the rat sperm, and may have done likewise in the other
forms in which he does not draw the characteristic bead.

11. Ditscusston.

(a) On the Origin of the Acrosome in Animal
Spermatogenesis.

The evidence that the Golgi apparatus is in some way
intimately associated with the formation of the acrosome or
perforatorium has accumulated considerably within the last
few years.

In Paludina (12) and in Columbella (80), two molluses, it has
been shown that the Golgi apparatus adheres to the head
end of the nucleus of the spermatid, and before breaking
away deposits or secretes a small granule from which the
acrosome finally develops. In Smerinthus populi,
a moth (10), it has been shown that the acrosome is developed
by changes which take place in crescentic * acroblasts ’, which
we now know as the dictyosomes or individual units of the
Golgi apparatus. In the testis of Stenobothrus viri-
dulus we have endeavoured to follow out the formation of the
acrosome: in this cricket it seems likely that the Golgi apparatus

280 J. B. GATENBY AND J. H. WOODGER

is intimately associated with the formation of the acrosome,
but the form chosen did not provide the very clear evidence
wanted. In the spermatogenesis of the louse, Doncaster
and Cannon (5) observed that the acrosome was formed from
a body which they took to represent the Golgi apparatus.

According to the account given for Smerithus (10) by
Gatenby, and for Pediculus by Doncaster and Cannon, all the
Golgi apparatus is taken up in the formation of the acrosome.
Our recent observations on Stenobothrus, and on several
other moths (e.g. Biston), have shown that in these insects
much of the apparatus finally passes as isolated crescents,
spheres, or dictyosomes into the elongating tails of the sperma-
tozoa: this matter is far from being cleared up, but of one
thing we may feel certain—that the Golgi apparatus of insects
is related to the formation of the acrosome.

Turning now to our observations on the acrosome of the
cavy, we note that the account we give agrees in general with
that previously described for Paludina (12). In both animals
we find a Golgi apparatus (plus archoplasm) which moves
up to the front end of the nucleus of the spermatid, deposits
a granule there, remains for a time, and finally passes away
from the head end of the sperm into the lengthening tail.

Papanicolaou and Stockard describe the proacrosomic
material as appearing inside the archoplasm as a differentiated
area of the latter, which stains specifically im acid fuchsin.
Here we have the crux of the whole matter: is the pro-
acrosomic material, which later forms the acrosome, to be
regarded as a product of the archoplasm, or of the dictyosomes
or Golgi elements ? We believe that this matter may be settled
after the events in the formation of the acrosome of insect
spermatids have been more fully examined: this remark
refers especially to the Smerinthidae.

Another point to which we would like to draw attention
is the fact that in the guinea-pig the Golgi apparatus (the
‘Nebenkern’ of some older authors) embraces the forming
acrosome from the stage when the proacrosomic granule first
touches the nuclear membrane, up to the stage when the

am

—

agen

Pet as ee a te
, —

ve

CYTOPLASMIC INCLUSIONS OF THE GERM-CELLS 281

acrosome has reached almost its greatest size; the natural
inference being that the Golgi apparatus and not the nucleus
is concerned with the growth and perfection of the rudimentary
acrosome. In this connexion it will be remembered that one
of us has shown in Smerinthus (10) that the acrosome may
form completely, while the nucleus lags behind in development,
as occurs in degenerating spermatids.

We conclude at present that the animal acrosome is formed
directly in association with the Golgi apparatus, and that the
nucleus has little if any influence in the process.

(b) The Middle-piece of the Spermatozoon
after Entry into the Egg.

That the middle-piece of the mammalian spermatozoon is
carried into the egg is well known, and it is now established
by the work of van der Stricht (31), Lams (16), and Levi (19),
that excepting the centrosome the entire middle-piece of
Vespertiio and Cavia, after having become carried bodily
into the egg, remains inert and complete, and is passively
borne into one or other of the two blastomeres (or one of three
in Levi’s case), and is ultimately lost sight of, probably de-
generating at a later stage in the cleavage of the egg.

Lams’ (16) work is particularly worthy of mention. Alone,
and also in conjunction with Doorme, he showed that in the
white mouse and the cavy the middle-piece (excepting the
centrosome) remains unchanged after entry into the ovum.
Many of the figures of Lams show the mitochondria lying
upon the middle-piece, but in no case did he find any activa-
tion of these bodies. In both the cavy and the rat we are
aware that the middle-piece bears a Golgi bead, but since
Lams used no methods for the Golgi apparatus, it is hardly
justifiable to use his work as evidence with regard to the
behaviour of the Golgi bead after introduction into the
ovum.

Henneguy, at the discussion following Lams’ communication
to the Brussels congress of 1910, suggested that the blastomere

982, J. B. GATENBY AND J. H. WOODGER

containing the tail of the sperm became transformed into the
embryonic part of the germ, the other blastomere into the
trophoblast.

Meves (21) likewise suggests something similar for the
case of Kchinus. That part of the pluteus contaiming the sperm
middle-piece is supposed to bud off the Echinus rudiment,
a very unlikely suggestion mdeed.

Levi (19) in remarking on these facts and suggestions says :

‘Le ipotesi di. Henneguy e di Meves non furono finora
suffragate da aleun fatto, ed il solo argomento nuovo che 10
adduco, la possibilita della persistenza del pezzo intermedio
dello spermatozoo in uno dei blastomeri provenienti dalla 2*
segmentazione, non contribuisce ancora ad illustrare il si-
enificato del condrioma maschile nello sviluppo ulteriore.’

As one of us pointed out before, the explanation of Henneguy
for the case of the mammal does not accord with the generally
accepted interpretation as to the origin of identical twins,
tor if the presence of a middle-piece was a factor of any sort
of differentiation, the two separating blastomeres would not
produce the identical twins.

We see no reason to suppose that the middle-piece Golgi
apparatus is stripped off the sperm and left outside; there
seems every justification for the supposition that the apparatus
is carried into the egg with the mitochondria. What fate hes
in store for this middle-piece Golgi apparatus is unknown to
us, nor do the works of van der Stricht, Lams, or Levi bring
forward any sort of evidence with regard to this point.

In all probability, the apparatus, like the mitochondria,
first remains complete and inert and ultimately degenerates,
after having fulfilled its function, whatever that may be.

The meaning of the stages in spermatogenesis during which
most of the mitochondria and part of the Golgi apparatus
become applied to the middle-piece of the spermatozoon, is
difficult to understand. If the mitochondria and Golgi
apparatus of the spermatozoon remain inert, unlike those of
Ascaris which persist in the egg and live (15), we are forced to
conclude that the function fulfilled by these bodies is carried

ES ee se

©

CYTOPLASMIC INCLUSIONS OF THE GERM-CELLS 283

out between the time the sperm leaves the spermatic tubule,
and enters the egg.

Two suggestions are obvious and may be set forth: (a) Both
mitochondria and Golgi apparatus are concerned with the
production of the energy used up by the movements of the
sperm tail. (b) Hither the mutochondria or the Golgi
apparatus (or both) carry some active substance which is set
free just as the sperm enters the egg, or after it has penetrated
the egg, and whose function is related in some obscure way to
the phenomenon of heredity.

It seems to be established that every mammalian sperm is
partly formed of mitochondria, and we may find that every
such sperm has a Golgi apparatus. The experimental evidence
which is necessary for the elucidation of the function of these
two categories of cell inclusions within the structure of the
spermatozoon would be very difficult to procure, and it appears
to be very doubtful whether mere observation of the behaviour
of these inclusions during fertilization will provide any con-
clusive facts.

Tt has been said that the animal spermatozoon is merely
a much modified cell, and it has been shown in this paper
that the remark is true to the smallest detail, for a sperm
such as that of Cavia is a complete cell with nucleus, mito-
chondria, Golgi apparatus, and centrosome. In one fact,
however, the two gametes differ widely: While the nuclear
matter (chromosomes) of both gametes is similar in quantity,
the mitochondria and Golgi apparatus of the spermatozoon
are infinitely less in quantity than those of the ripe ovum.
Are we to look upon the presence of the mitochondria and Golgi
apparatus in the animal spermatozoon as being merely of
phylogenetic importance, and indicative of a period when the
two gametes were equal in size and metabolic potentialities,
or should we entertain the view that the mitochondria and
Golgi apparatus are specially concerned with a ‘ cytoplasmic
heredity ’, as apposed to a ‘ nuclear’ one ?

It has never been shown satisfactorily that either the mito-
chondria or the Golgi apparatus can originate from the nucleus,

984 J. B. GATENBY AND J. H. WOODGER

though some indications of this have been noted (see 11,
p. 581), and until such is established we are not justified in
dismissing the hypothesis of a special ‘ cytoplasmic heredity ’.

More than this we cannot at present write ; the very function
of the mitochondria and the Golgi apparatus is not understood,

and those paths which will lead to this understanding are only
now being entered.

UNIVERSITY COLLEGE, LONDON,
April 12, 1920.

12. SUMMARY.

(a) The Middle-piece Golgi Apparatus.

1. The middle-piece of the mammalian spermatozoon is
formed from part of the mitochondria of the spermatid which
become grouped around a central rod or skeleton. Not all
the mitochondria of the spermatid pass into the middle-piece,
a certain proportion always sloughs off.

2. On the middle-piece of many mammalian spermatozoa
there is a protoplasmic bead which can be seen in the fresh,
and which, on fixation, stains in plasma dyes.

3. With formalin and silver nitrate techniques the proto-
plasmic bead is found to contain a number of argentophil
platelets or rods, which impregnate exactly like the Golgi
apparatus of younger sperm cells.

4. The spermatid of Cavia contains a Golgi apparatus con-
sisting of an inner core of archoplasm, and a cortical region
formed of curved plates and rods—the dictyosomes. With
formalin-silver nitrate techniques, the Golgi apparatus either
appears as a reticulum, or the whole cortex of the apparatus
reduces the silver, and then appears homogeneous : with Mann-
Kopsch techniques the individual dictyosomes are often very
clearly marked.

5. At a stage when the spermatid is elongating the Golgi
apparatus buds off a small part of itself. This part becomes

CYTOPLASMIC INCLUSIONS OF THE GERM-CELLS 985

separated from the main Golgi apparatus, and ultimately comes
to lie in the middle-piece bead referred to in paragraph 2.

6. The rest of the Golgi apparatus of the rpening sperma-
tozoon sloughs off.

7. While all the chromatinic substance of the young
spermatid eventually goes to form the nucleus of the sperma-
tozoon, only the majority of the spermatid mitochondria,
and a very small part of the spermatid Golgi apparatus, form
the representatives of these cell organs in the ripe spermatozoon.

8. Attention is drawn to the works of Lams and Doorme,
van der Stricht, and Levi, where it has been shown that the
whole middle-piece of the mammalian sperm (Cavia or Vesper-
tilio) enters the egg at fertilization, but, so far as these authors
could_observe, thereafter remains inert, and is carried whole
and haphazardly into one of the blastomeres.

(6) The Formation of the Acrosome.

9. The acrosome of the spermatozoon of Cavia is formed
from the proacrosomic granules which are differentiated within
the archoplasm during the later growth stages of the spermato-
cyte.

10. The archoplasm in the spermatocyte of Cavia is covered
by the Golgi elements or dictyosomes, which in all probability
are associated with the differentiation within the archoplasm
of the proacrosomic granules.

11. Each of the spermatids derived from the spermatocyte
contain an equal share of Golgi elements, archoplasm, and
proacrosomic granules. According to Papanicolaou and
Stockard the latter granules do not disintegrate during mitosis,
but, keeping their individuality, become scattered in the
cytoplasm, are subequally sorted out among the daughter
cells, and eventually come to lie within the re-formed spermatid
archoplasm.

12. Each proacrosomic granule has a liquid-filled space
formed around it, so that it comes to lie in an archoplasmic
vacuole.

13. The several proacrosomic granules within their archo-

286 J. B. GATENBY AND J. H. WOODGER

plasmic vacuoles approach and fuse into fewer larger granules,
which eventually all come together to form a single large
granule lying in a single archoplasmic vacuole. This structure
is known as the proacrosome.

14. The Golgi apparatus complex now consists of numbers
of dictyosomes lying on the surface of the archoplasm: the
latter contains near its centre the proacrosome. The latter
soon becomes distinguishable into an inner darkly-staining
bead surrounded by a paler cortical zone, the whole lying
in the archoplasmic¢ vacuole.

15. The Golgi apparatus complex has moved up towards
the anterior end of the spermatid nucleus, and it now becomes
applied to the nuclear membrane. Where the complex touches
the membrane the Golgi elements or dictyosomes are pushed
aside, so that the archoplasm comes into direct contact with
the spermatid nuclear membrane.

16. From its more or less central position the proacrosome
passes through the archoplasm and becomes applied to the
nuclear membrane, upon which it becomes flattened so as to
form a hemisphere. The proacrosome is now spoken of as the
acrosome: it has an inner zone, an outer zone, and it is still
covered on its outer side by the archoplasmic vacuole. Where
the latter comes into contact with the archoplasm there is
differentiated the covering membrane of the acrosome, which is
rarely very clear.

17. The acrosome grows rapidly, and at a stage when it
has differentiated to form a conspicuous cap at the anterior
end of the spermatid nucleus, the Golgi elements with archo-
plasmic remains, which hitherto covered and embraced the
developing acrosome, gradually drift away and pass towards
the posterior end of the spermatid.

18. The acrosome now develops by itself. ‘The lower part
of the archoplasmic vacuole spreads down past the equator
of the spermatid nucleus, and the lower edges of the outer zone
of the acrosome cover the equatorial region of the nucleus.
The archoplasmic vacuole becomes less evident.

19. ‘The outer zone of the acrosome grows very rapidly,

CYTOPLASMIC INCLUSIONS OF THE GERM-CELLS 287

becomes cone-shaped, and later flattened and crescentic in
shape when the broad side of the sperm is examined. In the
fully formed acrosome the outer zone of the acrosome is much
greater in extent than the inner zone of the acrosome.

13. BrenioGRAPHY.

1. Benda, C.—‘‘ Neuere Mittheilungen iiber die Histogenese der Siiuge-
thierspermatozoen ”’, ‘ Verh. d. Physiol. Ges. zu Berlin’, 1896-7.

2. Brown, H. H.—‘‘ On spermatogenesis in the Rat”, *‘ Quart. Journ.
Micro. Sci.’, N.S. vol. 25.

8. Deinecka, D.—‘* Der Netzapparat von Golgi in einigen Epithel- und
Bindegewebezellen ”’, ‘ Anat. Anz.’, xli, p. 289, 1912.

4, Doncaster, L.— An Introduction to the Study of Cytology’, Cam-
bridge Press, 1920.

5. Doncaster and Cannon.—‘‘The Spermatogenesis of the Louse ”’,
‘Quart. Journ. Micr. Sci.’, vol. 64, 1920.

6. Duesberg, C.—‘‘ Nouvelles recherches sur l'appareil mitochondrial

des cellules séminales ”’, ‘ Arch. f. Zellf.’, Bd. 6.

“Plastosomen, ‘apparato reticolare interno’ und Chromidial-
apparat ”, ‘Erg. d. Anat. u. Entw.’, Bd. 20, H. 2.

8. Da Fano, C.—* Method for the demonstration of Golgi’s internal
apparatus *’, ‘ Phys. Proceed.’, January 31, 1920.

9. Fauré-Fremiet.—‘‘ Etude sur les mitochondries des Protozoaires
et des cellules sexuelles *’, ‘ Arch. d’Anat. micr.’, t. xi.

10. Gatenby, J. Bronté—‘* The Degenerate (Apyrene) Sperm-Formation
of Moths as an Index to the Inter-Relationship of the Various
Bodies of the Spermatozoon ’’, ‘ Quart. Journ. Micr. Sci.’, vol. 62,
1917.

11. “The Cytoplasmic Inclusions of the Germ-Cells. Part II, Helix
aspersa’’, ibid.

12. “Part III, Notes on the Dimorphic Spermatozoa of Paludina ”’,
ibid., vol. 63, 1919.

13. “Part Il, The Spermatogenesis of some other Pulmonates ”’,

ibid., vol. 62, 1918.

14. Golgi, C—‘ Arch. Ital. de Biol.’, vii (1886), and ‘Arch. Ital. de
Biol.’, xlix (1908).

15. Held, H.—* Untersuchungen iiber den Vorgang der Befruchtung ”,
‘ Arch. f, mikr. Anat.’, Bd. 89,

16. Lams, H.—‘‘ Etude sur l’ceuf du Cobaye aux premiers stades de l’em-
bryogenése ”’, * Arch. de Biol.’, vol. 28.

et Doorme, J.—‘‘ Nouvelles recherches sur la maturation et la

fécondation de lceuf des Mammiféres ’’, ‘ Arch. de Biol.’, t. 23.

aT.

988

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.
30.

31.

32.

33.

J. B. GATENBY AND J. H. WOODGER

v. Lenhossék, M.—‘* Ueber Spermatogenese bei Siiugethieren ”’, ‘ Vorl.
Mitth.’, v. 3, Apr. 97. Tiibingen. .

Levi, G.—‘*Il comportamento dei condriosomi durante i pil precoci
periodi dello sviluppo dei Mammiferi”, ‘Arch. f. Zellf”, Bd. 14,
1915-16.

Meves, F.—‘‘ Ueber Struktur und Histogenese der Samenfiden des
Meerschweinchens ”’, ‘ Arch. f. mikr. Anat.’, 54, 1899.

“Verfolgung des sogenannten Mittelstiickes des Echiniden-
spermiums im befruchteten Ei bis zum Ende der ersten Furchungs-
teilung ”, ‘ Arch. f. mikr. Anat.’, Bd. 80, 1912. See also 1914.

Moore, J. E. S.—‘ The Maiotic Process in Mammalia’, ‘ Cancer
Report, 1906’, by Moore and Walker.

Murray, J. A—‘‘ Contributions to a Knowledge of the Nebenkern in
the Spermatogenesis of Pulmorata ”’, * Zool. Jahrb.’, Bd. xi, 1898.

Niessing, C.—‘‘ Die Betheiligung von Centralkérper und Sphire am
Aufbau des Samenfadens bei Saugethieren ”’, ‘ Arch.f. mikr, Anat.’,
Bd. 48, 1896.

Perroncito, A.—‘‘ Contribution 4 l’étude de la biologie cellulaire’, &c.,
‘ Arch. Ital. Biol.’, liv, 1910.

Papanicolaou and Stockard.—‘* The Development of the idiosome in
the germ-cells of the male guinea-pig”, ‘Amer. Journ. Anat.’,
24, 1918.

Regaud.—‘ Etudes sur la structure des tubes séminiféres et sur la
spermatogenése chez les mammiféres”’, ‘Arch. d’Anat, micr.’,
11, 1910.

Retzius, G.—‘‘ Die Spermien der Vespertilionen’’, ‘ Biol. Untersuch.’,
Bd. 13, 1906.

‘“Spermien der Siiugetiere *’, ibid., no. xiv, 1909.

Schitz, V.—‘‘Sur la spermatogenése chez Columbella rustica ”
‘ Arch. Zool. Expér., Notes et Revue ’, no. 2, t. 1.

van der Stricht, O.—‘*‘ La structure ae Voeuf des Mammiféres ”
3° partie, “Mém. de l’Acad. Royale de Belg.’ Bruxelles, 1909.

Weigl, R.—‘‘ Vergleichend-zytologische Untersuchungen itiber den
Golgi-Kopschen Apparat ’’, ‘ Bull. de l’Acad. Scient. Cracovie ’,
1912.

Whitman, C. O.—‘* The Inadequacy of the Cell Theory of Develop-
ment’, ‘ Wood’s Hole Biol. Lectures ’, 1893.

CYTOPLASMIC INCLUSIONS OF THE GERM-CELLS 289

14. DESCRIPTION OF PLATES 11 anp 12.

Illustrating Dr. J. Bronté Gatenby and Mr. J. H. Woodger’s
paper on the ‘ Cavy Sperm ’.

Explanation of Lettering.

A =acrosome. APG =proacrosomic granules. AR =archoplasm (centro-
sphere). ©, c!, c? =centrosome (first and second). CA =covering membrane
of acrosome. cH=chromosome. CHB=chromatoid body. Ga =Golgi
apparatus plus archoplasm, GAx =middle-piece Golgi apparatus elements.
GAxc =cytoplasmic bead containing the Golgi elements of the middle-
piece. GE =Golgi element or dictyosome. 1zA =inner region of acrosome.
Lev =lower part of archoplasmic vacuole embracing nucleus. M =mito-
chondrium. Mc =manchette. Mp -=middle-piece. Mx =degenerate mito-
chondria coalescing to form von Ebner’s granules. N=nucleus. 0ZA =
outer zone of acrosome. T=tail of sperm. v-=archoplasmic vacuole.
v.v. =archoplasmic vacuoles.

In each plate the scale is on the left-hand side.

Puate 11.

Fig. 1.—Spermatid of Cavia, at the time when the Golgi apparatus
(GE and aR) is still in contact with the forming acrosome: the latter
is distinguishable into two regions, an inner zone applied to the nucleus
(1zA) and an outer zone (024). The mitochondria are scattered throughout
the ground cytoplasm. This cell is drawn from a Mann-Kopsch preparation ;
note the discrete dictosomes, GE.

Fig. 2.—Later spermatid, showing the double appearance of the Golgi
apparatus (GA). At Gax is the bead which later becomes attached to the
middle-piece, as in fig. 3 at GAX. Preparation by Cajal formalin-silver
nitrate method.

Fig. 3.—Ripe sperm just before the residue cytoplasmic bead strips off.
The middle-piece Golgi apparatus bead is at Gax, the definitive Golgi
apparatus, which is cast off, at GA. Preparation by Cajal’s method.

Fig. 4.—Sperm at same stage showing mitochondria heavily impregnated
by reduced OsO,. Preparation by Mann-Kopsch method. The Golgi
apparatus did not impregnate in the region of the testis from which this
cell was drawn. The bead protoplasm is seen at GAXc,

PLATE 12.

This plate is drawn from three separate sets of preparations by (a) Mann-
Kopsch, (6) Cajal’s Golgi apparatus technique, (c) a mitochondrial method.

290 J. B. GATENBY AND J. H. WOODGER

So far as possible we have utilized the work of previous observers. All
the figures are drawn to scale excepting fig. 16, in which the tail of the
sperm is much shorter than it should be.

Fig. 5.—Ripe spermatocyte, I, in the prophases of the first maturation
division, chromosomes appearing in the nucleus. The Golgi complex
contains in the middle the divided centrosome (c). Around the latter are
the proacrosomic granules (aPG) which constitute the inner zone of the
archoplasm. Between the Golgi elements or dictyosomes (GE) which lie
on the surface of the archoplasm, and the inner region of the archoplasm,
is a space free of proacrosomic granules. The space constitutes the outer
zone of the archoplasm. The whole Golgi complex is drawn in optical
section. In the ground cytoplasm lie the mitochondria (M) and the
chromatoid body (CHB).

Fig. 6.—Second spermatocyte division metaphase viewed from side,
The mitochondria lie haphazardly around the spindle. Following
Papanicolaou and Stockard we have drawn the proacrosomic granules
(their idiogranulomes) as preserving their individuality and becoming
distributed here and there in the cytoplasm around the spindle (aP@).

_ Fig. 7.—Newly-formed spermatid showing the same elements as the
spermatocyte in fig. 5, only the proacrosomic granules are now surrounded
by the archoplasmic vacuoles (v.v.). The centrosome is dividing. The
mitochondria tend to pass to the periphery of the cell.

Fig. 8.—Later stage: the Golgi complex begins to move towards the
anterior pole of the cell. The proacrosomic granules have fused one with
another till only three are left, the large main one in the middle (ape)
being surrounded by its archoplasmic vacuole (v.v.). The mitochondria
tend to lie on the periphery of the cell. The centrosome has divided into
two, and from one part the flagellum is growing out.

Fig. 9.—Golgi complex and part of nucleus of later spermatid. The
proacrosomic granules have all run together to form the proacrosome
(PRA), lying in the archoplasmic vacuole (Vv) ; the proacrosome is differen-
tiated into an outer (0zA) and an inner zone (1zA). The proacrosome has
left its position in the middle of the archoplasm and has approached the
nuclear membrane (N).

Fig. 10.—Later spermatid, after the proacrosome has become partly
flattened against the nuclear membrane. The outer and especially the
inner zones (OZA, IZA) of the acrosome have become much larger. The
Golgi apparatus and archoplasm surround the entire acrosome. The
archoplasmic vacuole has begun to grow down on each side of the nucleus
(Lev). The Golgi complex is placed to one side of the nucleus, but in later
stages the acrosome comes to lie at the head end of the nucleus, possibly
by a partial rotation of the latter.

Fig. 11.—Spermatid at a later stage just before the Golgi apparatus
flows away from the acrosome. The front and back parts of the nucleus

CYTOPLASMIC INCLUSIONS OF THE GERM-CELLS 291

show thickenings. The mitochondria have left the periphery and are
collecting towards the middle of the posterior end of the cell.

Fig. 12.—The Golgi apparatus has left the head end of the cell, and is
beginning to regain its spherical shape. The centrosome apparatus and
flagellum have moved up towards the posterior end of the nucleus. From
this region of the latter manchette fibres (mc) begin to grow back. The
acrosome become plastered over the entire front of the nucleus. Nearly
all the mitochondria have left the anterior pole of the cell.

Fig. 13.—Later stage showing great development of the outer zone of
the acrosome (0zA). The manchette has become tubular (mc). From the
spermatid Golgi apparatus (GA) has begun to grow out a small bead
(GAX) which later forms the middle-piece Golgi apparatus. The mito-
chondria are collecting in the region of the manchette. The centrosome
ring is beginning to pass from the posterior part of the nucleus.

Fig. 14.—The acrosome become more oval in contour. The centrosome
ring is passing near the Golgi apparatus (c?). From the latter the middle-
piece Golgi apparatus bead is just separating (GAx). The manchette is
less evident, and around the axial filament or flagellum a distinct thicken-
ing is visible. It was not settled whether the parts in figs. 13 and 14,
Mc and MD, were inter-related.

Fig. 15.—The acrosome is now fully formed. The nucleus has gained
its characteristic shape. The middle-piece Golgi bead (GAx) has become
fixed to the middle-piece (mp) just behind the nucleus. The mitochondria
begin to become attached to the middle-piece skeleton (MD) from before
backwards. The Golgi apparatus is drifting down and undergoes staining
changes.

Fig. 16.—Spermatozoon viewed edgewise, just before skinning off of
residue bead. The middle-piece bead is at GAx, but not all the mito-
chondria (mM) have become applied to the middle-piece skeleton ; in the
residue protoplasm many of the mitochondria run together and undergo
changes, forming von Ebner’s granules (Mx). The Golgi apparatus is
degenerating.

NO. 258

A

| a

Hanae Sree "gnibedin sy ial

vey hal se si sab

i, itm ¥ hd
1 Vine Histesiciy wae yawned

{tah ne riven ars
oot ty feast

ibe hte ekey ‘ory sin ee ba
oy aa waited aes
nen Pony Siren),

‘Mal (es | acres a |
ahs Shins = se ah lat

ea
oduitinl in A .

igre he At lene

7

L* may eee? BAe

wrote af wir baal y i

ripeeta lular? mit y ob

us (fie? ene
ty Reta
tn ea: er aea
; a) ee Oe
ney 3 eee Alia

errr “Har my ei okie

Si
1 \O Pe aN a oh sti ie? : 7
o a a _ _
en or
vi
an - “tp ofa
Suite
Teka Garhi a aft"
om
a iat
iim | strat?
ity “> Qs tbe t Wi alte
ri vr mes og
vik’ ot’ Ae a
(Az rate 4
st Pat 16 to
vy fe a Mazak 3
q ~ ta iris
fF ' Z q
7 Vee 3 Dl
~
*. sf 4
i 7 fil
fi A mG 7
; 7r
* te’ Peet jamie
ie 2;
7 io 7 ,,j y 9 —
ef ems 5 ivy
} hits PONT ‘
1p ®
8 -
fe ».. -
.

Hipivic> “wath '

a? wobry + (anf omer ney ig

¢ 6? oro: hea
tr A se

CV MV eRe: ai a nee
. [+4 2
(ve Wabiy tl .
she ine pre
Alper wet S,
i aelr he Pn e

#:" °# I
0 > Sees
i? e
Pe >» -
= , ou
~~ Fa

Quart. Journ. Micr. Sci Vol.65N.S.PLII.

e7ees

>
L/

Seoniaer

Psa

Quart. Journ. Micr, Sci. Vol. 65, N.S., Pl. 12.

=z

Gnby and Woodger del,

,

Further Studies on Restitution-bodies and free

Tissue-culture in Sycon.

By

Julian S.. Huxley.

With Plates 13 and 14,

CONTENTS.

. INTRODUCTION ;
. MATERIAL AND Metuops

SUBDIVISION OF RESTITUTION-BODIES

. DERMAL BLow-outTs

. DARK-CENTRED MASSES :

. ADHERENCE OF LARVAE TO MASSES . ‘
. ADHESION AND UNIFICATION OF RESTITUTION-BODIES
. RESTITUTION-BODIES AND TISSUE-CULTURE

9, MECHANICAL SHOCK: Toxic AGENTS

. DIscussION :

(a) Dedifferentiation. Position and Fate

(b) Formation of Blow-outs

(c) Size-relations. Viability : :
(d) ‘ Normal’ and * Abnormal’ phenomena .

. SUMMARY
. LITERATURE LIsT

1. INTRODUCTION.

PAGE
293
294.
295
300
305
305
306
307
309

311
315
316
317
318
320

In a previous paper (Huxley, 7) I showed that the remarkable

phenomena of regeneration from dissociated cells, first observed
by H. V. Wilson (15) in monaxonid sponges, later extended
by him (16) and by de Morgan and Drew (4) to coelenterates,
could be studied in a simpler and more satisfactory form in

X 2

294 JULIAN S. HUXLEY

heterocoelous calcareous sponges than in any other types
investigated. I further showed that, by certain methods,
restitution-bodies composed entirely or almost entirely of
collar-cells could be produced, and that these assumed a form
quite unlike anything found in normal sponges, but with
a resemblance to a Choanoflagellate colony. Simple excess
of collar-cells, or, apparently, larger masses composed
almost entirely of collar-cells, led to the formation of what
T called choanocyte blow-outs—a part of the solid mass becom-
ing blown out to form the segment of a collar-cell sphere.

Since then I have continued making observations on the
subject as opportunity offered. Although these cannot pretend
to completeness, they have brought certain new facts to light,
which I publish in the hope that other workers may extend
them by observation on the same favourable material. Some
of the work was done at Wood’s Hole, Massachusetts, and some
at the M.B.A. Laboratory, Plymouth. I have to thank the
authorities at both institutions for their help in getting material
and in other ways. I have also to acknowledge much efficient
help at Wood’s Hole from Mr. I. J. Davies, laboratory assistant
in the Rice Institute, Houston, Texas. :

2. MATERIAL AND METHODS.

A species of Sycon was used at both places. That used at
Plymouth was 8. coronatum, obtained from piles in the
Millbay docks. Orton (12) has recently drawn attention to
the fact that this sponge grows actively during the winter
without reproducing ; but during the summer it reproduces
so long as the temperature is above a certain level, and searcely
crows at all. The same is to be presumed true of other species
of the same genus. If so, it follows that the best time for
conducting similar experiments will be during the cooler half
of the year.

Experiments were tried on homocoelous sponges such as
Clathrina and Leucosolenia, but without much suecess. Restitu-
tion masses are formed, but are small and do not live well.
The collars and flagella are withdrawn on very slight provoca-

RESTITUTION-BODIES AND FREE TISSUE-CULTURE IN SYCON 295
tion, and the organisms and their parts appear to be more
delicate.

The method originally adopted was that discovered by
H. V. Wilson—the squeezing of the chopped-up sponge through
fine-meshed silk bolting-cloth.

In order to procure * pure cultures ’ of collar-cells, the sponge
or a transverse segment of it is held with one needle and briskly
teased with another. By this means large sheets of collar-
cells are obtained. If the pieces are shaken together in a solid
watch-glass, they will cohere and larger masses result.

A method which will give an excess of collar-cells but not
an almost pure culture is simply to perform the teasing process
as above, and then remove the portions of original sponge.
The collar-cells, bemg more easily detached than the others,
will form the bulk of the tissue present.

Finally, simple squeezing of the whole sponge with the
fingers into water will give a thick suspension of single cells
and very small cell-aggregates, which is very similar to the
culture produced by squeezing through gauze. By different
dilutions of this suspension, different results can be achieved.

These methods will be called squeezing through gauze,
choanocyte isolation, teasing, and squeezing without gauze
respectively.

The experiments at Wood’s Hole were done in late July
and August ; those at. Plymouth in July and early August.

3. SUBDIVISION OF RESTITUTION-BODIES.
(Work done at Plymouth.)

A teased culture was made on August 8, 1920. Many of
the restitution masses were of rather large size. They began
to blow out in normal fashion, and after six days a number of
very fine choanocyte blow-outs were present. On the seventh
day they were even better. On the eighth day a certain
quantity of bodies consisting of a number of small spherules,
rather closely packed together, were observed in the dish
(fig. 1). They were attached to the glass, and some force was

996 JULIAN S. HUXLEY

necessary to squirt them free. On examining them I at first
thought that they might be derived from the Sycon restitution-
bodies, but dismissed the idea as improbable. Later in the
same day I took some of them to Miss Lebour, Naturalist at
the Laboratory, to see if she could identify them. On examining
them under pressure with a high power, it was found that they
contained fragments of spicules. Thus the suspicion that they
were of sponge origin was strengthened.

Two days later a restitution-body which had for four days
been isolated for other purposes on a slide in a moist chamber
was examined and was found to have subdivided into six
spherules (fig. 2, a). Thus their sponge origin was conclusively
proved. Meanwhile the original dish was picked over, some
of its contents preserved, and the remainder separated into
divided and normal undivided masses. The normal masses
were examined two days later (the tenth day of the whole
experiment) and found to be still undivided, many with
active flagella and protruded collars still visible externally.
On the thirteenth day, eight out of fourteen masses were still
single, but the remaiming six had subdivided. They were
similar in every way to those observed on the eighth day,
except that they were not so closely packed, and that I could
see no traces of a gelatinous membrane round the spherules.
It would, however, of course be expected that those which
subdivided earlier would be of slightly different composition
from these later-divided ones.

A detailed observation of one of the earlier divided masses
on the ninth day (fig. 1, a) showed that the spherules were
tightly packed and mutually compressed. The whole body
was surrounded by a faint gelatinous membrane, which
apparently caused the whole to adhere to the glass. Under
a higher power (fig. 1, b) the single spherules were seen to
consist of a one-layered epithelium surrounding a central mass.
‘The epithelium was composed of extremely clear cells, with a few
minute granules ; the central mass did not touch the epithelium
at all pomts, and was dense and of a yellowish colour ; cell
outlines were not visible in it. The single spherules did not

RESTITUTION-BODIES AND FREE TISSUE-CULTURE IN SYCON 297

appear to possess separate membranes. My attention, however,
had not yet been drawn to this point, and I cannot be sure of
it. Broken spicules were present in some. Another mass
examined on the same day contained many more fragments
of spicules. It was very similar to the first, but the epithelia
were not so sharply marked off from the central masses,
which in their turn were not quite so dense. When gelatinous
membranes were present, numerous bacteria were usually seen
along their outer edges.

The spherules were of various sizes, as is shown in fig. 4,
which illustrates an isolated specimen on the tenth day. This
same specimen was examined again on the thirteenth day.
The same individual spherules were identified, but their appear-
ance had changed, their outlines being less regular and the
general effect more transparent. On examination with a high
power this was seen to be due to the fact that im the majority
most of the individual cells had separated from each other
(fig. 5). Hach spherule was surrounded by a definite layer of
jelly. Within this, isolated clear cells, all sub-spherical, were
scattered. At one point, either central or at the side, a denser
yellowish mass was seen. This appeared to consist of larger
cells, still adherent, containing many granules (of two types,
large and small). A few of the small clear cells could still
be seen embedded in some of the yellow masses. A minority
of the spherules showed a different appearance (fig. 6). In
them the spherule had simply subdivided into a small number
of pieces, of somewhat irregular shape, each apparently con-
sisting of clear cells round the periphery, yellow cells within.
Finally, one or two spherules intermediate in type were seen,
i.e. with a few large masses and also some isolated clear cells.

The independent gelatinous coverings of the separate
spherules were also seen in other specimens, e.g. in that
shown in fig. 2.

A variant of the types already discussed is shown in fig. 38,
which illustrates a small mass found in the culture-dish,
consisting of an epithelium of dermal cells surrounding a central
mass, presumably mainly of collar-cells, which had subdivided

298 JULIAN S. HUXLEY

into spherules. No gelatinous layer was seen round this mass.
This was paralleled in the development of some other masses ;
e.g. that shown in fig. 2, a, had, three days later, assumed the
appearance shown in fig. 2, b. The smallest spherule was
unchanged. ‘The remaining five, however, were all surrounded
by a well-marked epithelium of dermal cells very different
from the epithelium shown in fig. 1, which I take to be choano-
cytic. The masses had swollen up by the secretion of fluid.
The central portion had in three of the spherules begun to
fragment. The gelatinous layers were of the same thickness
as before. Although no cell-outlmes had been visible in the
spherules when examined three days previously, it had been
noticeable that their outer boundary was very sharp. Other
observations give colour to the idea that this sharp boundary
heralds the formation of a dermal epithelium.

In this case the dermal epithelium is formed after the
spherules have been produced. In the mass shown in fig. 3,
the whole mass forms a dermal epithelium, and the spherules
are then produced internally.

The further history of the spherules was as follows. The
majority showed disintegration of the types shown in figs. 5
and 6. A few degenerated directly. No recovery was observed
in spite of change of water. This, however, may only mean
that laboratory conditions were unfavourable. It is to be
remarked that the general appearance of the tissues in stages
like that of fig. 5 was perfectly healthy.

Another culture was later found, where the same processes
were observed. It was unfortunately not possible to carry
out experiments to determine if subdivision could be initiated
at will.

The subdivision appears to be primarily a reaction to un-
favourable conditions (witness the accumulation of bacteria
round the edges of the subdivided masses). In all the dozens
of dissociation cultures I have made at Naples, Wood’s Hole,
and Plymouth, these two were the only ones where subdivision
was Observed. Both these cultures were from teased, nol
squeezed, material.

RESTITUTION-BODIES AND FREE TISSUE-CULTURE IN SYCON 299

The secretion of gelatinous membranes is of interest. Other
noteworthy points are as follows :—(1) The size of the spherules
produced varies within considerable limits. Those produced by
a single mass might be approximately equal, or of very different
sizes. (2) I at first thought that the phenomenon was deter-
mined by the proportion of dermal cells present, subdivision
continuing until enough surface was formed to accommodate all
the dermal cells in the state of simple epithelium. Appearances
like that of fig. 1, b, however, seem to negative this, for there
the epithelium surrounding the spherules is cuboidal, and quite
unhke any dermal cells seen by me. This epithelium seems to
consist of the healthiest choanocytes present. The difference
of colour between them and the cells of the inner masses, how-
ever, 1s to be noted, and it is possible that they represent
a dedifferentiated condition of the dermal cells. If so, they
would resemble the cuboidal form of the ectoderm cells seen
in dedifferentiated stolons, &c., of the Ascidians, Perophora
and Clavellina-cells which are normally as flattened as the
normal dermal cells of Sycon. At all events the phenomenon
must be determined by some surface-volume relation, the cells
not being able to cohere in large masses when in certain con-
ditions.

In any case, the spontaneous segmentation of the masses
into regularly-arranged portions of smaller size is of interest.
This phenomenon never occurs, as far as is known, in the normal
life-history of Sycon; yet the process is regular, and at first
sight would be taken for a normal occurrence. It is an example
of the determination of physiological processes by the direct
action of external circumstances, without any modification
by way of heredity. A somewhat similar phenomenon was
found by Miller (10) in restitution-bodies of Spongilla, but it
was not so regular, nor, since it only occurred in large masses,
does it seem to have been due to identical causes.

(3) The separation of the clearer cells from each other,
apparently when the circumstances have become slightly more un-
favourable, is also of interest. In Perophora and in Hydroids, a
shght concentration of toxic substances starts dedifferentiation

300 JULIAN S. HUXLEY

in the zooids (results in course of publication). The further
progress of events in these organisms, however, is determined
by the emergence of the cells from the tissue into the blood,
leading to the resorption of the zooid. Here, in these spherules,
the cells emerge from the tissues, but must remain in position,
since there is no means by which they can migrate elsewhere.
Slight mercury poisoning also causes emergence of some of the
endoderm cells from the gut of late Echinoid plutei. It
is probable that total or partial resolution of the tissues into
separate cells is a general occurrence in dedifferentiation,
but that it is masked in many cases, e. g. in Clavellina. A study
of these phenomena, together with that of dissociation of cells
as observed in particular chemical solutions, as, e. g., observed
by Gray (5), will throw light on the problem of cell-coherence
in general.

(4) The production of a definite dermal epithelium late in
the history of many subdivided spheres is to be considered in
relation to the observed fact that restitution-bodies with
dermal epithelium are more viable than those composed of
choanocytes alone (Huxley, 8).

(5) The transition from a state in which no cell-outlines are
visible (figs. 2, a ; 6) to one where the cells are distinct (fig. 2, b)
or separated (fig. 5) is to be compared with the formation of
syncytia in Coelenterate restitution masses, as noted by
Wilson and by de Morgan and Drew, and their subsequent
resolution imto cells. Here again a very important general
phenomenon is made accessible to study.

4. Dermat Buiow-outs.

In my previous paper three types of restitution were
described, leading to: (1) normal regenerates, consisting of
dermal and gastral cells in normal proportion. ‘These formed
spicules, and those that lived long enough produced normal
miniature sponges. (2) Collar-cell spheres: small hollow
spheres, consisting of a single layer of choanocytes, with no, or
very few, other cells. (3) Collar-cell blow-outs : masses con-
sisting mainly of collar-cells, blown out in one or more regions

ee a

Seema

RESTITUTION-BODIES AND FREE TISSUE-CULTURE IN SYCON 301

to form segments of spheres. The external epithelium of the
solid remainder might be formed: (a) by collar-cells only,
(b) by dermal cells only, (c) by patches of both.

Other types have now been observed. The most interesting
are the dermal blow-outs. These appear to be formed
whenever the mass contains a preponderance of dermal cells.
A mass of collar-cells generally fills most of the interior ; it 1s
covered closely with a single layer of dermal epithelium, which
at one point is swollen out to form a segment of a sphere
which thus consists entirely or almost entirely of dermal
cells. Often cells are to be seen adhering to its inner surface ;
these were sometimes rounded and of a fair size, presumably
typical amoebocytes, oftener of the minute elongated type
which I have called finger-cells (see p. 304). A few detached
cells might sometimes occur in the cavity. These were occa-
sionally seen to be forming spicules. A typical mass of this
kind is shown in fig. 12.

Shaking caused contraction and disappearance of the blown-
out regions, as with the collar-cell blow-outs.

One very peculiar mass was seen (at Wood’s Hole). This
was isolated together with a number of others shortly after
concrescence, when they were solid and irregular in shape.
Four days later this was found to have a large hemispherical
collar-cell blow-out, which in its turn showed a small dermal
blow-out on one side. Under the surface finger-cells were
visible. It would thus appear that local as well as general
excess of dermal cells can occur, leading to the formation of
mixed blow-outs.

Previous experience (Huxley, 8) has led me to conclude that
when a small proportion of dermal cells is present in a culture,
they exercise an attraction for each other. This leads to the
production of a few normal regenerates in a culture consisting
mainly of collar-cell blow-outs. In a similar way this con-
gregation of dermal cells can lead to the formation of dermal
blow-outs. This was so in the mass shown in fig. 12.

This and other facts would indicate that the formation
of dermal blow-outs is mainly a matter of the number of dermal

302 JULIAN S. HUXLEY

cells present in a particular mass. That this is not all, however,
is shown by the following experiment (Wood’s Hole).

July 12. Several sponges squeezed through bolting-silk
into a finger-bowl. Three dilutions of the resulting cell-
suspension were made: (1) dense, (2) medium, (3) dilute.

Results :—After one day: (1) large, sometimes irregular,
masses ; (2) medium-sized spheroids, many with good collars
and flagella, many with larvae embedded in them; (8) as (2),
but smaller, and fewer with larvae.

After two days: (1) no blow-outs ; most seem normal restitu-
tion masses; (2) most with collar-cell blow-outs ; (3) some
with collar-cell blow-outs.

After three days: (1) as before; (2) fewer collar-cells than
the previous day ; (3) none seen blown out.

After five days: (1) the smaller masses forming small dermal
blow-outs ; (2) many with large dermal blow-outs ; (8) solid.

After nme days: (1) mostly dead; (2) many attached to
glass ; (3) as before, none attached.

A repetition of the experiment gave similar results, except
that dermal as well as choanocyte blow-outs were formed
early in the middle dilution.

It will thus be seen that dermal blow-outs did not begin
to appear until the fifth (or fourth) day, and that they appeared
most notably in the same culture which had previously pro-
duced the best choanocyte blow-outs. Their failure to appear
in the large masses of (1) may be due to the fact that these
in this experiment were not very healthy. It would appear,
since the only difference between (2) and (3) lay in the size of
the masses formed, that the eventual production of dermal
blow-outs is determined partly by the total, and not only by
the relative number of dermal cells present. It appears that
first of all the collar-cells on the surface protrude collars and
flagella towards the water ; these are, however, very susceptible
to noxious influences, and as culture conditions became less
good they retracted into a spheroidal form. The dermal cells
then migrated to the surface and covered the collar-cells with
an epithelium, which they were apparently unable to do when

————= ehh

i ee ee ee

RESTITUTION-BODIES AND FREE TISSUE-CULTURE IN SYCON 303

the external collar-cells were functional. Since, however, the
total number of dermal cells in a mass is proportional to its
volume, while the number required to form a single external
layer of epithelium is proportional to its surface, there will be
in large masses an excess of dermal cells above those needed
to form the epithelium. This excess apparently forms the
dermal blow-outs. The replacement of choanocytes by dermal
covering is of interest in view of the greater viability or protec-
tive capacity of the dermal cells shown by other considerations
(Huxley, 8).

Fig. 13 shows another type. A number of very large masses
had formed in a culture produced by squeezing without gauze.
The larger masses had first been very irregular in shape, and
had demonstrably been formed by the coherence of original
smaller spheroids. (The culture was made in a finger-bowl.
The flat bottom was covered with small spheroids, while
a ring of the large irregular masses was found at the foot of
the sides. This was due to the opportunity given here for
many masses to come in contact by rolling down the steep
sides.)

These irregular large masses later rounded up, and shortly
after this produced blow-outs. Some were similar to that
seen in fig. 12. Others, however, consisted of a much-
distended sphere surrounded by an epithelium of dermal cells,
the contained gastral cells not forming a well-marked mass,
but spread in layers of varying thickness over part of the inner
surface of the sphere (fig. 13). The majority of the larger
masses in the culture were of this type, while the majority of
the smallest were not blown out at all, but were normal
regenerates. This bears out the conclusion drawn above as
_ to the réle of size of mass.

Wilson, in his experiments with Coelenterates, also found
that size of mass was very important, the larger masses failing
to metamorphose. A study of restitution-bodies from this
point of view would probably throw light upon the reasons
for the sizes of the larvae in many low types.

An interesting feature of most dermal blow-outs examined

304 JULIAN S. HUXLEY

with the high power was the association of the small types of
amoebocytes I propose to call finger-cells with the
dermal cells in the blown-out region. This was observed both
at Wood’s Hole and at Plymouth. In surface view the dermal
epithelium is seen to consist of a number of granular areas
(fig. 12)—cell-bodies—separated by transparent areas, where
the protoplasm is extremely thin. Cell-junctions cannot be
seen in yivo. Below each granular area is seen an irregular
stellate figure. On careful examination this is seen to consist
of a number of finger-cells radiating from below the centre
of the cell-body. Optical section of the periphery gives a profile
view, when the body of the dermal cell is seen to be sharply
marked off from the underlying finger-cells. Similar finger-
cells are seen to protrude, but singly, from the inner mass of
choanocytes. The meaning of this arrangement of finger-cells
remains obscure.

The cultures containing the large dermal blow-outs above
referred to were examined again later. Almost all had produced
some spicules, and a considerable proportion had metamor-
phosed into functional young sponges of the ‘ Olynthus * stage.

In my previous work (Huxley, 7, p. 169) I never obtained
fixed Olynthus stages from restitution masses. Here, however,
some 25 per cent. of them were firmly attached. <A few of
these were of great regularity of form (fig. 7, a), again surpassing
any obtained previously. Others, however, showed marked
irregularities, more pronounced than any seen in 1910, often
appearing as if preparing to form a second osculum. In no
case, however, was a second osculum seen, or even a rudimentary
second oscular crown of spicules. The most remarkable
thing about these forms was the large size shown by many
of them, far exceeding that of a newly-metamorphosed larva,
Thus, although large size is associated with less viability of
restitution masses, yet even very large masses, provided they
remain healthy, can metamorphose into normal-type Olynthi
if the various sorts of cells are present in correct proportions.

Some idea of the normal size of Olynthi from larvae can
be got by comparing the figures of larvae (figs. 9,a; 10, a,

RESTITUTION-BODIES AND FREE TISSUE-CULTURE IN SYCON 305

drawn to a larger seale than figs. 7, a and b). It is possible
that the abnormal-shaped Olynthi were produced by dermal
blow-outs of the types of fig. 12, or by coalesced masses of
irregular shape.

Numerous other masses, however, were seen of the type
shown in fig. 8. Here spicule formation had progressed well,
but no osculum was present. The most noticeable point was
the restriction of the gastral layer to part of the sphere (as
already seen, e.g. in fig. 13). The gastral layer was usually
one cell-layer in thickness, but in some masses was several layers
thick at certain spots only (fig. 8).

The disproportionate number of dermal cells had certainly
delayed development. As I had to leave Plymouth the day
after discovering this type of regenerate, their fate could not
he ascertained.

5. DARK-CENTRED Masses.

These were both seen at Wood’s Hole and at Plymouth.
Typically (fig. 11), they consisted of a dermal epithelium,
surrounding several layers of pale cells, apparently choanocytes,
which in their turn surrounded a central mass of dark yellowish-
brown material, whose cellular nature could not be seen in
vivo. The central mass is separated from the pale cells by
a space. In one or two cases the central body was seen to be
revolving. If this was so, then the collar-cells must have
developed flagella on their inner side.

The meaning of these masses is obscure. There is a strong
resemblance between the inner mass and the yellow-brown
masses seen in the subdivided spherules (fig. 5), and some
resemblance also between the intermediate pale layer and the
isolated cells of fig. 5.

A few specimens were observed where a dark central mass
was present, together with active choanocyte epithelium on the
outside.

6. ADHERENCE OF LARVAE TO MASSES.

Both at Wood’s Hole and at Plymouth it was noticed that
when cultures were made from sponges containing nearly

306 JULIAN S. HUXLEY

mature larvae, these might adhere to and be actually embedded
in the restitution-bodies (figs. 9, a; 10, a). In this situation
their flagella would continue to beat. Transferring masses
with embedded larvae by means of a pipette often resulted
in detaching the larvae (fig. 10,5). Larvae that remained
attached appeared to become resorbed into the masses, finally
disappearing (fig. 9, a-c). Histological investigation of this
has not yet been undertaken.

7. ADHESION AND UNIFICATION OF RESTITUTION-BODIES.

In section 4 an account has already been given of the fusion
of a number of spheroidal bodies to form irregular masses,
which later became spheroidal in their turn. These were all
masses with excess of dermal cells. Some observations on
masses with excess of choanocytes may also be given. Some
four-day restitution-bodies were isolated in a hanging drop.
The chief are shown in fig. 9, a. Most are covered with dermal
cells, but two have dermal epithelium in part. Some have
attached larvae. After two days these were seen to have
fused (fig. 9, b). Three larvae and two other small restitution-
bodies, not shown in fig. 9, a, had not shared in the fusion,

On the next day the larvae were still visible, but the general
form was not so irregular. The day after (fig. 9, c) the larvae
were no longer visible, the blown-out region had increased,
and the traces of the separate original masses have been
almost lost, the mass looking quite unitary, though with
slight irregularities. Two days later (fig. 9, d) slight regressive
changes had set in. The form was more unified, but the blow-
out was smaller, and the collars had been entirely, the flagella
partially, retracted. Gaps in the blow-out appeared, bridged
by dermal cells. Three days later the blow-out, together
with all traces of flagella, had disappeared, and four days later
the mass had still further shrunk, and was apparently covered
entirely with dermal cells, though I could not be quite sure
of this.

The most interesting feature of this is the gradual assumption
of unitary form by artificial aggregation of cell-masses, which

RESTITUTION-BODIES AND FREE TISSUE-CULTURE IN SYCON 307

in their turn are produced in a totally abnormal, artificial way.
The form produced, however, as also in the case of the choano-
eyte spheres, though typical, is not in the least like anything
occurring in the normal life of the species. Here, typical
form and form-changes of an organic type are seen in artificial
ageregates. Once more we see a series of organic forms very
clearly as an equilibrium between external environment and
inner constitution. Here, however, the inner constitution is
simple, the changes are not running in grooves of heredity.
It is probable that many adult forms of simple organisms,
as well as simple developmental forms, are in this way deter-
mined almost entirely by a direct relation of not highly-
differentiated tissues with environment. ‘The blastula, for
instance, may or may not represent an ancestral adult form.
It certainly is a primitive ancestral developmental form ; but
it is this not for any adaptive or eventually ‘ organismal ’
reason, but because it is the simplest way in which
a number of undifferentiated cells can arrange themselves in
a fluid medium. See also Child (2) for an account of the way
in which adult form may be largely determined thus in flat-
worms.

8. RESTITUTION-BODIES AND TISSUE-CULTURE.

It is obvious that the spheres produced by isolation of sheets
of collar-cells are ‘ tissue-cultures ’ in that they consist of one
sort of cells only. Their history in ordinary sea-water is
a history of gradual starvation, followed by involution, since
the fluid does not contain sufficient nutriment. Numerous
experiments were tried with a view to finding a suitable
nutrient medium, but so far without success.

(1) Pure culture of the Diatom Nitzchia, so successfully
employed by Allen and others for feeding Echinoid plutei,
were obtained and mixed with water containing preponderat-
ingly choanocyte restitution-bodies. In a few eases, diatoms
were seen in the collars of collar-cells, or partly embedded in
the cell-body ; but they were apparently too long for con-
venient ingestion.

NO. 258 Wi

308 JULIAN S. HUXLEY

(2) Suspensions of common sea-water Bacteria killed by
heating were added daily.

(3) Masses were put in vessels, together with fresh Ulva, to
see whether they would ingest the swarm-spores.

(4) Sterile solutions of Peptone in sea-water, of various
strength, were prepared. The restitution-bodies were trans-
ferred to this through four changes of sterilized sea-water,
the pipette being sterilized between each operation. Although
somewhat over 50 per cent. of the cultures thus prepared
became contaminated, yet a number remained free of bacteria.
In all these, however, the choanocytes underwent regressive
changes, actually sooner than in normal sea-water, and the
masses died within a few days.

(5) ‘Sponge Broth’. 3c.c. of chopped sponge was
extracted in 20 ¢.e. of sea-water and sterilized, and restitution-
bodies transferred to it as under (4), but again with no success.

(6) Ammonium lactate of 0-1 per cent. concentration
was prepared and a trace of phosphate added (cf. Peters, 14),
and then sterilized. Again some restitution-bodies were
transferred to the medium without infection, but all contracted
and died speedily.

(7) Under gauze, in the circulation (at Wood’s Hole). Unfor-
tunately these experiments had to be discontinued. The
restitution-bodies remained healthy for some time, but growth
could not be detected. The fact that normal regenerates throve
better and actually grew in these conditions, warrants the
belief that some modification of this method might be sue-
cessful.

Although all methods so far tested have proved unsuccessful,
yet I feel sure that choanocyte masses could be supplied with
food. It is possible that experiments in circulation would
succeed best at Naples, where Sycon establishes itself and grows
in the tanks. Further, experiments of this nature would most
profitably be undertaken in the cooler months (see § 1 of this
paper). At Wood’s Hole I found that covered cultures kept
cool in the circulation throve better than those exposed to
air-temperature.

RESTITUTION-BODIES AND FREE TISSUE-CULTURE IN SYCON 309

The cultivation of the collar-cell spheres, if successful,
would open out many points of interest. What, e.2¢., would
happen if considerable cell-multiplication took place? The
resemblance of the collar-cell spheres to colonial protozoa,
and the fact that the collar-cells are the nutritive organs of
the sponge, make the research still more interesting. Finally,
the ease with which sheets of pure collar-cells can be obtained,
and the fact that they will remain healthy, with expanded
collars and active flagella, for one to two weeks without being
fed, renders them very suitable as material. Detached tissues
which assume characteristic form in this way and live for
a considerable period in the normal medium may be termed
free tissue-cultures.

One or two interesting points concerning ingestion by the
eollar-cells may be mentioned here.

(1) Addition of powdered carmine to a culture of choanocyte -
masses was followed within an hour or so by ingestion of some
of the particles. Very many particles adhere to the flagella,
so that the masses appear reddish. Such particles as find their
way within the collar are ingested by a pseudopodial extension
of the intrachoanal protoplasm. No extrachoanal ingestion
was observed.

(2) When Nitzchia was added, very few were ingested, and
these only partially. They were usually caught, like the
carmine particles, by the ends of the flagella, and lashed to
and fro. This adhesive condition of the flagella is of interest.
(In fresh dissociation cultures, finger-cells may often be seen
adherent to the flagella and being waved from side to side with
their beat.) Some were also seen adherent to the inner side
of the collars.

9, MecHAnicaAL SHock. ‘Toxic AGENTS.

Mechanical shock, such as repeated pipetting, or even transfer
to a hanging drop, will cause marked changes in the cells, both
dermal and choanocytic. A collar-cell blow-out treated in this
way shows marked reduction of the size of the blow-out region,

Via

310 JULIAN S. HUXLEY

together with a thickening of its walls. The collars are usually
retracted. (See figs. 10, a, and 10, b.)

This sensitiveness to mechanical stimuli is shown by many
other tissues in culture (cf. Holmes, 6).

Exposure to very dilute solution of mercuric chloride in
500,000 2,000,000
the collar and then, gradually, of the flagellum, together with
slowing of the flagellar beat. The effect is proportional to the
strength of the solution. This retraction of the flagellum is a _
remarkable phenomenon, and the short stumps of the flagella
still beating provide a curious spectacle.

A record of an experiment is appended.

sea-water ( causes retraction, first of

RECORD OF EXPERIMENT.

Five or six collar-cell blow-outs in each solution (100 ¢.c. each).
A. Control. Collarsand flagella remained normal for twenty-four hours.
n
B. 1500,000 H8Ch-

l hr. 25m. Two masses with short collars. Three masses with
very short or no collars, and sharp smooth outline of epithelium,
all with fair to good flagellar movement.

2hrs. 10m. None with more than very short collars. Some
with short flagella. These beating faster than unretracted
flagella of other masses. Smooth outline still visible.

19hrs. No collars. Only two with flagella (moderate length),
two with slight cell-disintegration.

n
©. 750,000 H8Ch-

l hr. 10m. One mass with short collars in most cells, one mass
with vestigial collars, remainder without collars. Flagellar
action moderate, one with shortened flagella. Blow-outs
shrunken in all but one. Some cell-disintegration.

2hrs. No collars. Flagellar action slow and spasmodic.

IShrs. No flagella. All masses with much disintegration into
separate cells.

n

D. 590,000 HeCh-

Lhr. No collars. Flagellar action very slow or nil. Flagella

absent or normal length. Two blow-outs still present.

RESTITUTION-BODIES AND FREE TISSUE-CULTURE IN SYCON 3811

lhr.55m. No flagella. Disintegration starting in all.
18 hrs. 45 m. Masses present, but more disintegrated than C.
. n ‘
E- 100,000 Heh:
50m. Collar-cell blow-outs no longer visible. One dermal blow-
out unaffected. Two masses with a few flagella moving (slow
or spasmodic). Flagella somewhat shortened. Masses with
irregular outlines.
Thr. 50m. All cells rounded off, total disintegration of masses
starting.
18 hrs. 30m. Completely disintegrated into groups of one to twenty
dead cells.
n
50,000 Hele.
40m. No collars or flagella. Blow-outs burst, contracted, or

i

disappeared. Masses with irregular outlines.
1 hr. 50 m, and 18 hrs. 30m. As E.

10. Discussron.
(a) Dedifferentiation. Position and Fate.

Wilson, in his work on dissociation and subsequent regenera-
tion in Monaxonid sponges, left the question entirely open as
to whether regeneration was due wholly to the ‘ totipotent ’
amoebocytes, or whether the differentiated tissue elements
underwent a process of * despecialization ’ (dedifferentiation)
into an ‘indifferent or totipotent state’, after which they
took thei shares in restitution. He does not seem to have
envisaged the possibility of the differentiated cells sharing in
the restitution-process without undergoing total dedifferentia-
tion. In his later paper, on dissociation and restitution in
Hydroids (16), he returns to the subject, and decides that in
these forms, where undifferentiated cells form but a fraction of
the normal body, the differentiated tissue-elements definitively
become despecialized * to form masses of totipotent regenerative
tissue *. The cells in these masses later differentiate in acc or-
dance with their position, the outer cells forming
ectoderm, the central core endoderm. ‘This is, of course, in
accordance with Driesch’s well-known dictum that the fate

312 JULIAN 8S. HUXLEY

of a cell is a function of its position. He finally concludes that,
since the restitution-masses of sponges are so like those of
Hydroids, the processes occurring im them are of the same
nature.

He further stated that the cells mm the early stages of the
restitution-mass formed a syneytium, few or no cell-boundaries
being distinguishable.

De Morgan and Drew, in their later work on restitution-masses
in other Hydroids, while confirming Wilson in some points,
differ from him in others. In the first place, although restitu- .
tion-bodies with perisare, ectoderm, and typical endodermic
coenosarcal tubules were produced and lived for as long as
sixty days, no hydranths were formed. As Orton (12) suggests,
this may be due to the fact that de Morgan and Drew’s experi-
ments were performed from December to March, when the
srowth of Hydroids appears to be at a standstill, while Wilson
worked in the summer.

In the second place, although they describe a syneytial
phase, their figures do not show any such complete cell-fusion
as Wilson’s, and they mention that a small portion of endoderm
cells are always to be recognized as such. They do not pro-
nounce definitively one way or the other as to whether dedifferen-
tiation of all cells to a ‘ totipotent ’ condition occurs.

Miller (10) also believes that collar-cells do not take part
in the redifferentiation of restitution-masses in Spongilla, but
that the amoebocytes and thesocytes form the new gastral
cells. In view of the great réle played by the amoebocytes in
monaxonid sponges, and the specialization, small size, and
relatively small number of the choanocytes, this is not sur-
prising. In gemmule development, for instance, the flagellated
chambers arise from archaeocytes. The same author (11),
in describing dedifferentiation in Spongillidae, notes that the
choanocytes early dedifferentiate and disappear, apparently
ingested by amoebocytes. It would appear that they cannot
maintain themselves as such in unfavourable conditions. In
this connexion, mention may be made of the work of Maas (9),
who found that slow deprivation of calcium led to similar

RESTITUTION-BODIES AND FREE TISSUE-CULTURE IN SYCON 313

dedifferentiation in various calcareous sponges, including
a heterocoelous form (Sycandra). He also describes degenera-
tion and phagocytosis of the collar-cells in late stages of the
process.

In view of my work (see discussion below), it would appear
that in both Calcarea and Monaxonida the choanocytes are
more susceptible than the amoebocytes, and will degenerate
in certain conditions. In Calearea, however, this difference in
susceptibility is less marked, and the choanocytes will remain
capable of maintaining existence in dissociation-masses, while
this is not possible for those of Spongilla.

My own work (7) on Sycon indicated that the conclusions
of Wilson as to the fate of the cells in restitution do not apply
in the case of Sycon. On dissociation the tissue elements all
become dedifferentiated morphologically, e.g. the choanocytes
lose both collar and flagellum and become rounded, the dermal
cells lose their extended flat shape for a spheroidal one; but
this dedifferentiation is not complete in the sense that the
various kinds of cells become physiologically similar, or acquire
the same potentialities of development. After this deditferen-
tiation caused by shock the cells redifferentiate in appropriate
directions, the dermal cells producing an external epithelium
round a central choanocyte mass, which in its turn becomes
hollow with epithelial walls. The normal form of the post-
larval sponge is thus produced by a process exactly the reverse
of that envisaged by Driesch and Wilson. The fate of the cells
is not a function of their position, but their eventual position is
a function of their constitutional differences. The development
of a restitution-body is primarily a process of sorting-out of
different lands of cells, followed by a redifferentiation of the
individual types of cells. We have thus to distinguish sharply
between two types of cellular dedifferentiation: (1) that
which leads to complete loss of the character of the tissue
to which the cell belongs, and a return to a totipotent, or at
least, if I may coin a new word, to a pluripotent condition.
This may be called ultra-typical (or pluripotent)
dedifferentiation; (2) that which leads to a temporary

314 JULIAN 8. HUXLEY

suppression of the characters of the cell, also with the assump-
tion of a simple spheroidal form. Redifferentiation, however,
is only possible in the direction of the original form, and the
cell has not acquired pluripotency by dedifferentiating. This
may be called intra-typical (or unipotent) dedifferentia-
tion.}

The existence of pluripotent dedifferentiation is rendered
probable by various observations which cannot be entered into
here. It has frequently been assumed, however, on insufficient
evidence ; and in view of its theoretical importance, and the
difficulty of proof, very thorough investigation is required to
establish its existence in any particular case.

Further evidence against its occurring In Sycon was afforded
by the artificial production of masses composed entirely, or
almost entirely, of collar-cells. These, though they lived
healthily for a number of weeks, never produced a dermal
epithelium or spicules. This is paralleled by the failure of
endoderm or ectoderm alone to regenerate in Hydra, as has
been shown by various observers.

In a later paper (Huxley, 8) attention was drawn to the
fact that masses composed only of collar-cells were less viable
than those contaming dermal cells also, although both were
kept under identical conditions, and although the collar-cells
are the organs of nutrition.

In the present paper, the converse of the collar-cell blow-outs
is shown to occur in the form of masses with an excess of
dermal cells. These form blown-out vesicles exactly as do the
choanocytes when they are in excess.

It is thus clear that, in Sycon at least, the form and composi-
tion of the restitution-mass depends (apart from questions of
size) upon the proportions of the different types of cells which
entered into its Composition.

It is clear from the observations of Wilson that some process
of dedifferentiation does occur in restitution-masses of Hydroids.

' Since writing the above, I find that a very similar classification of

types of dedifferentiation from the point of view of tumour-growth has
been adopted by Adami and McCrae (1, p. 324. See also pp. 318-22).

CE emt ream § Woh MEN RI Ma Mp SE nN. ian adedcenieendiih

—_—

a

RESTITUTION-BODIES AND FREE TISSUE-CULTURE IN SYCON 315

Ki. g. in Pennaria the endoderm cells enter in large numbers
into the composition of the restitution-masses, and are distin-
cuishable immediately after dissociation by large granules.
Within twenty minutes, however, a syncytial mass has been
formed, in which very few of these granular elements can be
distinguished. Presumably the granules have been resorbed in
the new conditions. On the other hand, neither his observations,
nor those of de Morgan and Drew, in the least exclude the idea
of migration of ectoderm or endoderm cells to thei proper
stations after mtra-typical dedifferentiation.

In this connexion, the facts concerning the possible attrac-
tion of the various types of cells for each other may be men-
tioned (Huxley, 8). In cultures consisting almost entirely of
collar-cells, a small proportion of normal regenerates usually
occurred. In other cultures made at Plymouth, where the
great majority of the masses were choanocyte blow-outs, with
partial dermal covermg, a small proportion were dermal
blow-outs. These facts may be due either to accidental distribu-
tion of dermal cells, or else to an attraction of dermal cells for
each other. This poimt could only be settled by appropriate
experiments. The probable attraction of spermatozoa by
choanocytes was mentioned in the same paper.

(b) Formation of Blow-outs.

The secretion of fluid by epithelia, whether dermal or choano-
cyte, and consequent formation of spheres or segments of
spheres (° blow-outs ’), is an teresting phenomenon.

In this connexion, Mr. J. Gray, of King’s College, Cambridge,
has kindly allowed me to refer to some unpublished observa-
tions of his own, which he is at present investigating, on the
formation of similar blown-out spheroidal masses by fragments
of the gills of Mytilus. The phenomenon would thus seem to
have more than isolated significance. It perhaps imvolves
changes of the same nature as those taking place in the forma-
tion of a blastocoele.

316 JULIAN 8. HUXLEY

(c) Size Relations; Viability.

Wilson (loc. cit.) found that the size of the restitution-
masses produced by Hydroids was of great importance. Large
masses almost invariably died early, while too small masses,
though living for a long time, failed to produce Hydranths
or even coenosarceal outgrowths.

In Sycon also, very small masses, though reaching a two-
layered stage and occasionally forming spicules, fail to meta-
morphose. Similar failure to produce normal structure from
pieces below a certain definite size is well known in studies
on regeneration, both in unicellular and multicellular organisms.
It may be partly due to mere lack of material, but undoubtedly
also, In some way not as yet properly understood, to the
relatively greater surface and the consequences thereon atten-
dant—ditferences of gaseous exchange and difference of stimula-
tion by the environment being prominent.

Sunilarly, in too large a mass, it does not appear that proper
oxygenation for the central cells can be provided, and so
disintegration sets in. Wilson found the interesting fact that
successfully-metamorphosing masses were of the same order
of size as normal planulae. The same is roughly true for Sycon,
although here the upper limit of size for successful masses 1s
much further above the larval size than im Hydroids.

De Morgan and Drew comment on the fact that their restitu-
tion-masses, although not metamorphosing, were much more
resistant to laboratory conditions than the normal colonies,
and regard it as surprising. ‘here should be no ground for sur-
prise in this—the cells of the restitution-masses are definitively,
as we have seen, in a dedifferentiated condition. Experiments
on Perophora and Obelia show that the undifferentiated stolons
and hydrocaulus remain perfectly healthy in conditions causing
dedifferentiation and resorption of the zooids. Clavellina
and other Ascidians hibernate in the form of ‘ winter-buds ’,
which are of somewhat similar nature to restitution-masses ;
and the normal gemmules of sponges have also something
in common with them. In the laboratory the hydriform

~~

es

ce i ll i ia

a

——

—

ee nt oe

RESTITUTION-BODIES AND FREE TISSUE-CULTURE IN SYCON 8317

larva of the medusa Gonionema becomes transformed into
a syncytial, undifferentiated mass, as was shown by Perkins (18).

The obverse of this condition is shown by the failure of
highly differentiated parts of the organism to maintain them-
selves in the restitution-masses. Wilson and de Morgan and
Drew found that portions of tentacles fail to become incorporated
in the masses. This is paralleled by the failure of Hydra
tentacles to regenerate. Apparently, on the one hand they
are too highly specialized to deditferentiate ; and on the other
cannot exist as such in the conditions afforded by the restitu-
tion-bodies. The nematocysts also are gradually resorbed in
the restitution-bodies. ‘

If we seek to embrace the phenomena in one general view,
we may say that Hydroid tissues in unfavourable or abnormal
conditions lose much of their differentiation, come to have
a low metabolic rate (in the general sense in which the term
is used by Child (8)), and are more resistant. In these con-
ditions specialized organs cannot exist. The same tissues
in optimum conditions possess a higher metabolic rate, and
are capable of maintaiming specialized organs such as the
tentacles in existence.

(d) Normal’ and ‘Abnormal’ Phenomena.

Attention has already been drawn to the fact that many of
the processes occurring in restitution-bodies and free tissue-
cultures run parallel with various phenomena of development.
The normal phenomena constitute an interlocking series, each
stage of which is determimed by the preceding and helps to
determine that which comes after. By studying processes
which occur in ‘abnormal’ conditions, e.g. by dissociation
methods, we remove the tissues of the organism from this
developmental chain, where it is often impossible to say what
occurrences are palingenetic, what adaptive, what the direct
consequence of changes in the environment, and what con-
ditioned by previous processes in the series; by varying the
conditions, we may then throw light upon the normal processes.

So far my work has been mainly devoted to elucidating

3158 JULIAN S. HUXLEY

roughly the course of events in restitution-bodies in Sponges.
It is clear, however, that in Sycon we have an admirable
material for qualitative experiment, as to the réle of size of
masses, the proportion of the tissues in the mass, the coherence
of cells, their mutual attraction, &e.

The elucidation of these problems will need many workers,
and it is hoped that others may be mduced by the facts here
set forth to take up the work.

Meanwhile two tendencies should be noticed. The first is
a tendency to discuss the results from a morphological stand-
point. This is shown, e.g., in Wilson’s discussion of results.
He compares the development of the restitution-masses in
detail with that of normal development, and goes so far as to
apply the term ‘ yolk’ to the central syncytial portion which
remains in the middle of the masses while the two layers are
differentiating. This, and indeed his whole discussion, though
of great terest, seems to me to be putting the cart before the
horse. We should rather expect to find some of the causes deter-
mining the presence and form of the normal yolk by examiing
the mode in which the abnormal conditions of a restitution-
mass Influence the internal cells, rather than vice versa.

A word is also in order as to the use of the terms * normal ’
and ‘abnormal’. Abnormal is often used as if it were synony-
mous with pathological. This is not the case in any of the forms
of restitution-mass here described (until we reach degenerative
change at the close of their history, this being due to lack of
nutriment and to laboratory conditions). Dedifferentiation,
ageregation, sorting-out, &¢., are all perfectly healthy pheno-
mena.

11. Summary or REsuLTs.
(Including those recorded in previous papers.)

1. Various methods can be used to dissociate the tissues of
Calcarea Heterocoela.

2. Mixture of the various types of cells in normal proportions
may lead to the formation of normal regenerates, resembling
post-larval Sycon, with spicules, osculum, and pores.

RESTITUTION-BODIES AND FREE TISSUE-CULTURE IN SYCON 319

3. The development of these masses consists primarily in
the sorting-out of the dermal and gastral cells. The former
produce a single-layered epithelium, below which spicules are
subsequently formed, the latter a central mass which later
becomes a hollow one-layered sac, into whose cavity the cells
put forth collar and flagella. Thus their fate is not a function
of their position in the whole, but their position a function of
their nature.

4. The two types of spicules are formed in the same order
as in normal development.

5. Free tissue-cultures consisting only of collar-cells can be
obtained by appropriate methods. These form spheres re-
sembling choanoflagellate colomes with the collars directed
outwards. These lve for a considerable time, but do not
regenerate other forms of tissue or produce spicules.

6. All grades from these to masses contaiming an excess of
dermal cells may be formed. They may be classified as follows :

(a) Collar-cell spheres.

(b) Collar-cell blow-outs. These consist of a solid mass with
one or more portions blown out to form a segment
of a collar-cell sphere.

(b 1) Withactive collar-cell epithelium over the whole surface.

(b2) With mixed collar-cell and dermal epithelium over
the solid portion.

(b3) With dermal epithelium over the solid portion.

(c) Normal regenerates.

(d). Dermal blow-outs, resembling (b1), but with dermal
epithelium over the whole surface.

| 7. In almost pure collar-cell cultures, a few normal regene-
| rates may be found. In cultures consisting almost entirely of
collar-cell blow-outs, a few dermal blow-outs may be found.
| This is probably due to mutual attraction of dermal cells.

8. Normal regenerates are more viable than collar-cell
spheres or collar-cell blow-outs of type (b 1).

9. Dermal blow-outs may be formed from collar-cell blow-outs.
They are in such cases produced more readily from large masses.

10. Numerous methods have been tried for feeding the
collar-cell spheres and blow-outs, but so far without success.

320 JULIAN S. HUXLEY

11. The flagella of collar-cells are adhesive.

12. Larvae may become embedded in the restitution-masses ;
they are gradually resorbed.

13. Restitution-masses, if brought into contact, will cohere.
The irregular masses thus produced gradually round up and
become unified.

14. Mechanical shock causes a contraction of both dermal
and choanocyte blow-outs, and a retraction of the collars and
partial retraction of the flagella in the latter.

15. A peculiar small finger-shaped amoebocyte (‘ finger-
cell’) is numerous in normal sponges and restitution-masses.
These cells are arranged in a remarkable manner below the
dermal epithelium of dermal blow-outs.

16. Spontaneous segmentation of restitution-masses into
small spherules may take place, apparently in unfavourable
circumstances. The spherules usually secrete a gelatinous
covering. They may differentiate a normal dermal epithelium.
The bulk of the component tissue (presumably choanocyte)
usually separates into its constituent cells after a time.

17. A type of restitution-body with dark central mass is
deseribed.

18. Dedifferentiation of all cells takes place after dissocia-
tion, but does not lead to a totipotent condition.

New CoLLeGe, OXFORD.
October 1920.

LireRATURE List.

1. Adami and McCrae (1914).—‘ Text-book of Pathology ’, Macmillan,
London, 1914.

2. Child (1902).—‘*‘ Studies on regulation. I”, ‘ Arch, f. Entw.-Mech.’, 15,
1902, pp. 187, 355.

3. —— (1915).—‘ Individuality in Organisms’, Univ. Chicago Press,
1915 (with references to previous literature).

4, De Morgan, W., and Drew, H. (1914).—‘‘ A study of the restitution
masses formed by the dissociated cells of the Hydroids Antennularia
ramosa and A, antennina”’, “Journ. Mar. Biol. Assoc., Plymouth ’,
10, 1913-15, p. 440.

5. Gray, J. (1920).—‘* The effects of Ions upon Ciliary Movement ”’,
‘Quart. Journ. Micro, Sci.’, 64, 1920, p. 345.

6. Holmes, J. (1914).—‘* The behaviour of the Epidermis of Amphibians
when cultivated outside the body”, ‘Journ. Exp. Zool.’, 17,
1914, p. 281.

RESTITUTION-BODIES AND FREE TISSUE-CULTURE IN SYCGON 321

7. Huxley, J. S. (1911).—** Some phenomena of regeneration in Sycon,
with a note on the structure of its collar-cells”’, * Phil. Trans,
Roy. Soc.’, B. 202, 1911, p. 165.

8. —— (1920).—‘** Notes on certain phenomena observed in regenerates
from dissociated sponges’, * Biol. Bull.’ (in the press).
9. Maas, O. (1910).—** Ueber Involutionserscheinungen bei Schwimmen

etc.’, ‘ Festschr. fiir R. Hertwig ’, 1910, Bd. 3.

10. Miiller, K. (1910).—Versuche iiber die Regenerationsfihigkeit der
Siisswasserschwimme ”’, ‘ Zool. Anz.’, 37, 1911, p. 83.

(1910).—** Beobachtungen iiber Reduktionsvorgiinge bei Spongil-
liden ”’, ibid., p. 114.

12. Orton, J. H. (1920).—** Sea-temperature, Breeding, and Distribution
in Marine Animals”, * Journ. Mar. Biol. Assoc., Plymouth’, 12,
1920, p. 339.

Perkins, H. F. (1902).—** Degeneration Phenomena in the Larva of
Gonionema ”’, * Biol. Bull.’, 3, 1902.

14. Peters (1920)—* Proc. Physiol. Soc.’, February 21, 1920; * Journ.

Physiol.’, 53, p. evili.

13

15. Wilson, H. V. (1907).—*‘ On some phenomena of coalescence and
regeneration in Sponges’, * Journ. Exp. Zool.’, 5, 1907.
16. (1911).—* On the behaviour of the dissociated cells in Hydroids,

Aleyonaria, and Asterias ’’, ibid., 11, 1911, p. 281.

EXPLANATION OF PLATES 13 AND 14.

The figures are all drawn from life with the Abbé camera lucida.
The magnifications are given as follows: 38+40¢., denotes
drawn at table level with a no. 3 (3’’) objective and no. 4
Huyghenian ocular. The objectives and oculars were Reichert
unless otherwise stated.

PLATE 13.

Fig. 1.—A subdivided restitution-mass. (Eight days.) a. The whole
mass. The spherules are mutually compressed and show a definite cubical
epithelium. (3+40c.) 6. A single spherule under higher power. The
central mass is distinct from the epithelium. (6+ 2 oc.)

Fig. 2.—Different stages of another subdivided mass. (3+ 4 0c.)
a. A nine-day mass. The spherules have separate gelatinous layers, and
no sharp epithelia. Dark areas are seen within them. b. Three days later.
All but one possess well-marked dermal epithelia and have somewhat
expanded. The central masses are irregular, and several have fragmented.

Fig. 3.—A small eleven-day mass with dermal epithelium ; the contents
are subdivided into small spherules. No jelly-layer. (3+4 oc.)

Fig. 4.—Ten-day subdivided masses. The individual jelly-layers of
the spherules are not shown, (3+4 0c.)

Fig. 5.—A single spherule of the mass of fig. 4, three days later, under
higher magnification. The layer of jelly, the separation of the clear cells, and

322 JULIAN 8. HUXLEY
the dense mass of yellow-brown cells are seen. (Zeiss jo water-immersion +
4 compens. oc.)

Fig. 6.—Another spherule from the same specimen, same date. The
layer of jelly is thinner. The spherule has subdivided into irregular masses
with clear outer layer and yellow inner centre. From one, cells are beginning
to separate. (Same magnification as 5.)

Fig. 7.—Olynthus stages from restitution-masses. Osculum and oscular
crown are well developed. (3+20c.) a. Large, fixed, of normal shape
(spicules figured at the edge only). 6. Smaller, of abnormal shape (spicules
omitted). A small patch (undotted in the figure) lacks the gastral layer.

PLATE 14.

Fig. 8.—A further stage in the development of the type shown in fig. 13.
The gastral layer is markedly incomplete (spicules only figured at the
edge). (3+ 4 0c.)

Fig. 9.—Successive stage in one hanging-drop culture. (3+4 oc.)
a. The chief masses present in the drop, two hours after isolation (four
days from beginning of experiment). 6. After two days. The masses
shown in (@) have fused together (in addition, in (a) there were three embryos
and two small masses which had not fused). Note three larvae and one
sphere partially attached. c¢. After four days. Larvae no longer visible,
blow-out larger; more unification of the separate masses. d. After six
days. No collars. Flagella shorter and fewer. Still more unification.
Gaps in the blown-out region bridged by dermal membranes with amcebo-
cytes on the inner surface, e. After nine days. Disappearance of blow-
out. No collars or flagella, / After thirteen days. Still further contrac-
tion. A few cells had separated from the mass (not shown).

Fig. 10.—To show the effects of mechanical shock. (8+4 0c.) a. A mass
with good choanocyte blow-out and attached larva. 6. The same mass
after repeated pipetting. The larva is detached, the epithelium of the
blow-out has contracted and thickened, the collars have been retracted.

Fig. 11.—Restitution-mass with dermal epithelium and central dark
yellow-brown sphere, separated from intermediate layers of collar-cells.
(6+ 2 oc.)

Fig. 12.—Small dermal blow-out under high power. (6+20c.) The
dermal cells are granular. Adhering to the inner side of each are a number
of finger-cells, A few dermal cells are figured in surface view. From others,
the subjacent finger-cells have been omitted. Over the rest of the
surface, dermal cells are not figured. The bulk of the interior mass
is composed of choanocytes. From the edge of this, finger-cells protrude
into the blow-out cavity.

Fig. 13.—Very large dermal blow-out, spherical type. (3+ 4 0c.)
Here there is no sharp internal mass, but the collar-cells form irregular
areas-of varying thickness adherent to the dermal epith > Those
on the upper side are represented darker than those below. ,.he cells of
the dermal epithelium have been represented too large.)

Quart. Journ. Mier. Sci. Vol. 65, N.S., Pl. 13.

_ y
=
4 d
_

phi ig

~ a ae _—
ee © ee
> a

=
SS)
~~
>
~
S
3
=
~
~

Quart. Journ, Mier. Sci.

Vol. 65, N.S., Pl. 14,

Ao.) dA
7 ——

oe

The Proboscis of the Syllidea.
Part I. Structure.

By
W. A. Haswell, M.A., D.Se., F.R.S.,

Emeritus Professor of Biology, University of Sydney.

With Plate 15.

I. Divistons or tHE ProBoscis RxEGIon.

THE proboscis of the Syllidea (here taken as comprising
all that part of the digestive tube lying in front of the intes-
tine) is made up of five parts (fig. 6), which are all (with the
exceptions presently to be noted) sharply marked off from
one another. These will be referred to here as (1) the buccal
chamber, (2) the pharynx, (3) the proventriculus, (4) the
ventriculus, and (5) the post-ventriculus with a pair of caeca
appended to it.

Ehlers (1864) recognized in the region (1) ‘ Riisselréhre ’
(buccal cavity), (2) ‘Schlundréhre’ (pharynx), (3) © Driisen-
magen’ (proventriculus), and (4) ‘ Uebergangstheil’ (ventri-
culus plus post-ventriculus).

De Quatrefages (1865) (10, tome 1, p. 3) recognized buccal
cavity, and pharyngeal, dentary, and oesophageal regions of
the proboscis.

Claparéde (1868) distinguished: ‘gaine de la trompe’
(buccal cavity), ‘ trompe ’ (pharynx), ‘ proventricule ’, ‘ ventri-
cule’ with its glands (caeca).

Hisig (1881) describes the ‘ Riisselésophagus ’ as made up
of three sharply-separated regions—the first (pharynx), the
second (° Driisenmagen’), and the third, which he does not
name, but which is the ventriculus: this is followed by

NO. 259 Z

824 W. A. HASWELL

the ‘ Vormagen’ (post-ventriculus) from which the caeca are
siven off.

Malaquin (1893) designates the divisions ‘ gaine pharyn-
aienne ’ (buccal chamber), * trompe pharyngienne * (pharynx),
‘proventricule >, and ‘ventrieule’ (ventriculus plus post-
ventriculus).

MeIntosh (1908) describes the region as consisting of
(1) pharyngeal cavity, (2) protrusible proboscis, (3) proven-
triculus, followed by (4) a short portion which ends in a dilated
region often with two lateral caeea (see Pl. 15, fig. 6).

‘

Il. Tar Buccat CHAMBER.

This is the only part which becomes actually evoluted when
the proboscis is protruded. It is a short chamber with a cuticle
thinner than that of the outer surface : its wall in the ordinary
retracted condition is thrown into a number of folds.

Ill. THe Prarynx.

The pharynx is a cylindrical tube, usually of considerable
length, straight in the majority, sinuous or coiled in the
Autolytidae and in Amblyosyllis. It has a greatly
thickened cuticle, the thickened lining terminating abruptly
in front in an entire, lobed, or denticulate edge. In front of
this is a circlet of papillae on the surface of which open the
numerous fine ducts of the pharyngeal glands. In most cases
the pharynx contains a single triangular tooth (or rather
stylet) with the base embedded in its dorsal wall. This is nearly
always situated at the anterior end, and is so placed that,
when the proboscis is protruded, its apex projects freely in
front. In some cases the single tooth is replaced by a paired
crescentic group of several teeth (Odontosyllis), or by
a circlet (Trypanosyllis, Autolytus).

The cellular layer of the pharynx in the anterior part of
its extent is a simple epithelium complicated only by being
perforated by the system of splanchnic nerves. Posteriorly
it becomes greatly modified by the development of numerous
gland-cells, so that it virtually assumes the character of

PROBOSCIS OF SYLLIDEA 325

agland. Inthe Kxogoneae this gland, which I have termed
the anterior proventricular, is more conspicuous than
in the other groups of the Syllidea owing to its being more
distinctly marked off; but in the latter it is quite as important
so far as relative development is concerned (fig. 7). The
cuticle in this region is as thick as it is throughout, and appears
quite imperforate, so that the secretion of the gland must
find its way out elsewhere. As in Exogoneae, in fact,
the ducts of the gland-cells run back through the epithelium
to the anterior region of the proventriculus, where the cuticle
is very thin and, apparently from its staining reactions, not
strongly chitinized. Here most of the ducts terminate, though
some appear traceable for some distance in the region behind
the chitinous plates.

Hisig describes the structure of the pharynx correctly as
regards the greater part of its length. The change which
takes place at the posterior end he describes rightly as regards
the epithelium, but he falls into an error in stating that in this
region the structure corresponds closely with that of the ventri-
culus, not only in the modification of the epithelium, but
in the development of radial muscular fibres.

Malaquin gives a more exact account of the structure as
far as the Syllidae and Husyllidae are concerned. He
recognizes the glandular modification of the epithelium at the
posterior end, but assumes that this has to do with the growth
of the pharynx and the formation of additional chitin. In
Amblyosyllis and Autolytus. with elongated coiled
pharynx, he places the glandular region towards the middle
instead of at the posterior end, i.e. instead of at the opening
into the proventriculus, where it occurs exactly as in the
Syllidae and Exogonae.

In connexion with the pharynx and its papillae mention
has been made of the pharyngeal glands, the secretion of
which is discharged on the surface of the latter. As I have
pointed out (7, p. 229), these glands were referred to by
Claparede (2) and De Saint-Joseph (11) and were fully de-
seribed by Malaquin (9). They consist, in most cases, of about

Z2

326 Ww. A. HASWELL

ten narrow cylindrical bodies of varying length surrounding
the pharynx, with which they run parallel, ending blindly
behind, and in front terminating in the pharyngeal papillae.

They are solid bodies each of the nature essentially of a group
of greatly elongated cells, the anterior end of each of which is
produced into a narrow duct termimating in a very minute
aperture on the surface of the corresponding papilla.

In Odontosyllis the arrangement of these glands is, as
pointed out by Malaquin, somewhat modified by their restric-
tion to the ventral side. In Amblyosyllis and in certain
species of Autolytus, as also observed by Malaquin, they
are fused together into a pair of irregular masses of considerable
size. These divide up in front into narrow lobes running for-
wards to the papillae.

IV. Tur PRoventricutus: GENERAL STRUCTURE.

The proventriculus is an exceedingly conspicuous and very
characteristic structure to which reference is made by all
writers who have dealt with this group of the Polychaeta.
But it was not till, in 1881, Eisig published his paper entitled
‘Ueber das Vorkommen eines schwimmblasenihnlichen Organs
bei Anneliden’ that an approximately correct interpretation
was given of the structure of this complex organ.

In Eisig’s account, though it marks a distinct advance in
our knowledge, there are certain omissions and certain mis-
statements. Of the former one of the most important is the
failure to recognize that the muscular tissue of the radial
columns, which make up the bulk of the substance of the wall
of the organ, is of the striated type. The true nature of this
tissue was pointed out by the present writer in a short paper
published in this journal in 1886; and the subject, as regards
the histology of the muscular tissue, was further developed
in 1889 (6).

In 1893 was published Malaquin’s * Recherches sur les
Syllidiens ’. In this comprehensive work the author gives
a very full account of the proventriculus, summarizing pre-
viously published results and adding numerous observations

PROBOSCIS OF SYLLIDEA 327

of his own. He gives many details, more especially regarding
the radial columns of striated muscle and the variations which
they undergo in different families and genera. Since the
publication of Malaquin’s valuable work there has not, so far
as 1 am aware, been any further contribution to the subject
with the exception of the brief reference to it contained im
a paper on the Exogoneae contributed by me to the Linnean
Society (7).

On approaching this subject anew, with a wider command of
material, I have found that Malaquin’s account, excellent
though it is, with many new observations, is yet not altogether
correct in some respects, and leaves untouched several structural
features that seem to be of some importance in connexion with
the study of the proboscis as a mechanical system.

The proventriculus is of cylindrical or sub-cylindrical form,
usually with a small degree of lateral compression, and varies
greatly in length in different members of the group. The
surface is marked by a series of rings, an appearance which
examination with a low power of the microscope shows to
be due to the presence of annular fine lines and rows of dots.
The fine lines correspond to annular bands of non-striated
muscular fibres: the dots, which are frequently coloured in
the living animal, are the outer ends of the cores of the radial
columns of striated muscle. Along the dorsal side of the organ
runs a longitudinal light or coloured line, the dorsal raphe,
and a similar ventral raphe runs along the ventral
side.

A comparison of the pattern on the surface of the proven-
triculus in representatives of different groups of the Syllidea
reveals the occurrence of three main types. In one of these
the annular lines alternate with the rows of dots. In a second
the lines run through the dots. In the third type, which is the
prevalent one in the Syllidae and in the Exogoneae,
while the lines perforate the dots in all the lateral regions,
they leave that position in the neighbourhood of the raphes,
and pass to the latter in the intervals between the rows of dots.

These three types of pattern arrangement mean respectively :

328 W. A. HASWELI:

(1) that the annular bands run throughout in the imtervals
between the radial muscle-columns; (2) that the annular
bands perforate the muscle-columns throughout ; and (3) that
the same arrangement as in (2) holds good except in the neigh-
bourhood of the raphes, where the annular rings pass to the
position they occupy throughout in (1)."

The lumen of the proventriculus may be described as
a vertical slit the upper and lower ends of which lie near the
dorsal and ventral raphes respectively. This is the form
assumed in the contracted state ; in complete contraction the
sides of the slit are in close contact: when dilated the slit
expands till in transverse section its outline becomes ellipsoidal.

The thick wall of the proventriculus (P]. 15, fig. 1) consists
of the following layers: (1) splanchnic layer of coelomic
epithelium ; (2) outer fibrous membrane; (8) layer of radial
muscle-columns and annular muscle-bands ; (4) inner fibrous
membrane ; (5) enteric epithelium ; (6) cuticle.

The coelomic layer is a very thin one, recognizable by its
infrequent flattened nuclei. The outer fibrous membrane is
the layer described by Malaquin, and earlier by myself, as
a layer of non-striated muscle. Of its contractile character
IT am by no means certain. It is a thin layer, only about
()-003 mm. in thickness in the largest forms, and is made up of
two strata in the outer of which the fibres run transversely
and in the inner longitudinally: the fibres are exceedingly
fine and there are no nuclei. The chief function of this layer
seems to be to serve for the insertion of the radial fibres and the
fibres of the annular bands. he inner fibrous membrane. is
a similar layer, also composed of outer transverse and inner
longitudinal fibres: it has the inner ends of the radial fibres
inserted into it. At the raphes paired trabeculae pass at regular
intervals from the mner fibrous membrane to the outer and bind
the two layers firmly together.

The enterie epithelium and the cuticle need not be specially

' Towards the anterior end of the proventriculus the regularity of the

rings on the surface is broken owing to a modification in the arrangement
of the radial muscles associated with the presence of the chitinous plates.

PROBOSCIS OF SYLLIDEA 329

deseribed here. They both become specially modified towards
the anterior end of the organ in connexion with the valvular
apparatus to be described later.

V. THe Proventricutus: Muscutrar ELEeMeEnts.

The greater part of the substance of the thick wall of the
proventriculus (figs. 1-5) is made up of the radial musele-
columns and the annular bands. The former are hollow fibres,
squarish or polygonal in cross-section, arranged in annular
rows, and extending radially from the outer fibrous membrane
to the inner.

The hollow of each column is occupied by a protoplasmic
core. In the columns which are perforated by the annular
bundles the protoplasm is divided into anterior and posterior
halves, and this division may extend to the imner end, but not
to the short portion of the core outside the annular bands,
the two halves being here continuous. In the Exogoneac
and in certain members of the other groups each core contains
only a single nucleus. But in the rest the structure is more
complicated and the number of nuclei increased. The maximum
of complexity is reached in the case of Syllis coruscans.
In this species (fig. 4), in which the arrangement of the muscles
is of type 2, the core is permeated by a system of exceedingly
fine fibrils—forming an irregular meshwork with a prevailing
longitudinal arrangement : this is more condensed towards the
outer end. Communications occur between adjoining cores of
the same row along the lines of the annular bands, and there
are also communications, irregularly arranged, between the
columns of neighbouring rows by means of processes which
perforate the cortex. Fibrils from the meshwork of each core
radiate outwards and penetrate through fissures into the sub-
stance of the cortex. Such communications are most numerous
opposite the Z membranes (Krause’s membranes) of the cortex,
if they are not entirely restricted to such an arrangement.

Nuclei are present in large numbers in each core. These
are of two main varieties—larger, clearer nuclei of about
0-0075 mm. in diameter, and smaller, denser, of a diameter of

330 Ww. A. HASWELL

about 0-005 mm. ‘The former are less numerous, mainly
situated towards the outer end, but occurring throughout the
core to its inner extremity. The smaller nuclei are extremely
numerous, distributed fairly uniformly throughout the length
of the core. In addition there are a comparatively small
number of nuclei belonging to what appear to be distinct
cell-elements with fine-grained cytoplasm embedded in the
core. Surrounding the cortex is a layer contmuous with the
core at the longitudinal fissure, composed apparently of similar
material, and containing an occasional nucleus: the imvest-
ments of contiguous columns coalesce completely.

Slightly less complex than Syllis coruscans are the
cores in Trypanosyllis zebra. In this species the
arrangement of the muscles conforms to type (3). The cores
here consist of two kinds of material—an axial part, split into
two in the perforated fibres, and a peripheral part. The former
is loaded with rounded granules which are strongly coloured
by haematoxylin ; the latter appears as a meshwork of delicate
threads, prolongations of which pass into the substance of the
cortex. Strands of granules similar to those in the central part
of the core run longitudinally between the fibrils of the cortex,
and the latter is enclosed in an investing layer which encloses
similar granules. The central part of the core contains numerous
nuclei.

In Syllis variegata (figs. 2 and 8), in which also the
arrangement of the muscles conform to the third type, the
core is greatly simplified. In the perforated columns it is
split longitudinally into anterior and posterior halves which
unite together only at the extreme outer ends outside the annular
bands. The substance of the core and the layer investing the
cortex is a finely granular homogeneous material which does
not become very readily stamed. In this are embedded some
five or six nuclei, one (or two) of which are larger than the others
(about 0-008 to 0-01 mm. in long diameter), and are situated
usually about the middle of the length of the fibre, while the
rest are mostly towards the outer end. The core has a thin
investment of what looks like fibrillated material.

PROBOSCIS OF SYLLIDEA 331

As regards the cortex of the column. ‘This consists of a bundle
of fibrils among which penetrate branching processes from the
protoplasmic core. Each column or fibre is characterized,
except in the Exogoneae, by the presence of one (Typo-
syllis variegata, T. closterobranchia, T. trun-
cata), or more ‘striations’. In all essentials these fibres
resemble the striated fibres of Arthropods and Vertebrates.
‘The fibrils of each are bound together by one or more transverse
membranes (Krause’s membranes, telophragms) which pass
through the fibrils, and, through the interfibrillar substance,
bind all the fibrils intimately together. The fibre itself is
composed of alternating zones of singly and doubly refracting
material, the telophragms passing through the latter. More-
over, gold-chloride methods reveal systems of J-granules
(sarcosomes) and transverse networks in the neighbourhood
of the telophragms, exactly as is the case in the striated muscles
of Arthropods and Vertebrates.t

At their outer and inner ends the fibrils of the striated
muscular fibres are firmly fixed into the outer and inner
fibrous membranes.

Occupying much less bulk than the radial fibres are the
annular bundles of non-striated fibres. The extent of this
system, its relations and the part which it plays in the move-
ments of the proventriculus, have not hitherto received adequate
attention. Malaquin, a little misled by his idea of a system
of transverse septa separating the annular rows of muscle-
columns from one another, pays little heed to these bundles.
He says in his account of the proventriculus of the Autolytea
(p. 217), ‘Comme nous aurons l’occasion de le voir plus loin
pour d’autres types, il est des points du diaphragme ou les
fibrilles, s'arrangeant en faisceaux, ont tout a fait l’apparence
de fibres musculaires, et on peut croire alors que ce tissu con-
jonetif fibrillaire passe au tissu musculaire proprement dit.
Nous reviendrons sur ce point & propos d’un autre type’.

' Mesophragms and Q-granules I have not hitherto succeeded in detecting,

except somewhat doubtfully in the case of Syllis (Typosyllis) varie-
gata.

332 W. A. HASWELL

The only further mention is under Syllis) hyalina
(p. 227): ° Les diaphragmes transversaux ont la méme disposi-
tion et la méme structure, a part ce fait que le tissu fibrillaire
qui les compose présente vers la périphérie un arrangement
en faisceau tres marque.’

But these annular muscles, as they may be termed, are of
much greater importance than such casual mention as that
given above would imply.

Each annular muscle is a bundle of non-striated fibres, com-
pressed in the antero-posterior direction, running (in the
prevailing third type) transversely between two adjoiming
rows of radial striated fibres in the immediate neighbourhood
of the raphes, and, farther on. passing through the outer ends
of the radial fibres. At the raphe the annular muscle is con-
tinued straight across the middle line to the opposite side.
Irom the raphe the muscle runs in an annular way in the
position indicated above, and is inserted at intervals imto
the outer fibrous membrane. These insertions occur between
the radial fibres of the row, around the corresponding
accessory fibres (non-striated radial fibres) described below.

It will thus be seen that the annular muscles are so arranged
as to form a system of constrictors by means of which the
lumen of the proventriculus, dilated by the action of the radial
fibres, 1s contracted.

The striated fibres, though the most important, are not the
only radial fibres in the wall of the organ. Another set of
radial elements, hitherto entirely overlooked, play a part
which must be of some consequence, since their occurrence
seems to be universal, and their arrangement varies little.
These elements, which for the sake of distinction may be
called the accessory or non-striated radial fibres,
like the striated, run from the outer fibrous membrane to the
inner. They are single fibres (usually bifurcated close to the
outer endin $8. variegata, usually branched in $. corus-
cans), placed at regular intervals between the striated fibres,
as shown in figs. 1 to 4. As mentioned above, the main rela-
tions of these fibres are with the annular strands of non-striated

>A

PROBOSCIS OF SYLLIDEA 333

muscle, and their chief function would seem to be to provide
a series of ‘ points d’appui ’ for the latter. It may be mentioned
here that it is largely to the presence of these fibres in transverse
sections in certain planes that the illusion of regular partitions
between the rows of striated fibres is due.

VI. Tue ProventricuLus: CHiTiInous PuLaTEs.

In the interior of the proventriculus towards its anterior
end is an elaborate structure which has hitherto failed to
receive the notice which its importance in the mechanism of
the proboscis seems to demand. It occurs in essentially the
same form in all the members of the group which I have
examined for it—not only inthe Syllidae and Eusyllidae,
but in the Exogoneae and Autolytidae.

De Saint-Joseph seems to have been the first to direct atten-
tion to the appearance presented by this structure, though he
misunderstood its significance. In his description of Ty po-
syllis alternosetosa, he says, ‘le proventricule, avec
30 rangées de poimts gris, qui a a sa partie supérieure un
anneau chitineux ’, with a foot-note, ‘ Cet anneau, quiseremarque
souvent chez les Syllidiens, me parait étre la continuation de
la trompe qui pénétre dans le proventricule’. On the other
hand, he refers to the same structure in Pterosyllis
spectabilis as‘ deux valves cornées’ (pp. 65 or 189). Mala-
quin (p. 213) gives a much more consistent and complete
description: ‘Dans la région antérieure de l’organe, 1|’épithé-
lium prend un autre aspect, il devient en quelque sorte fibril-
laire ; les cellules en sont trés allongées avee noyau médian
(Pl. v, fig. 7, Hp. pr.). Cette structure correspond a une disposi-
tion particuliere, a un épaississement de la cuticule formant
en avant du proventricule un anneau chitineux. Cet anneau
chitineux, visible sur le vivant (Pl. iv, A. ch., figs. 1, 2,8, 4, 5),
peut surtout s’étudier dans une coupe horizontale du proven-
tricule (Pl. v, fig. 6). Dans la région antérieure de l’organe
lépithélium est beaucoup plus épais et les parois se touchent
a l’état de repos de maniére a fermer totalement la lumieére.

384 W. A. HASWELL

lin arriére de cet épaississement existe lanneau chitineux
auquel correspond une disposition particuliere des colonnes
musculaires; celles-ci. au heu d’étre régulierement radiaires,
sont obliqnement disposées, au moins dans le plan horizental
médian du proventricule, de fagon a agir dans deux sens per-
pendiculaires. Cette disposition est destinée probablement 4
faire glisser et au beso a comprimer fortement les aliments
avalés par l’animal.’

A short distance behind the abrupt posterior edge of the
thickened cuticle of the pharynx (fig. 5) is a deep transverse
(circular) groove in the thick epithelium, and a little farther
back a second similar groove, the two separated from one
another by a prominent band of thickened epithelium. Just
behind the posterior groove the cuticle is developed on either
side into a dense chitinous plate. These plates are of no great
length im the direction of the long axis of the body, but con-
siderably elongated vertically, extending downwards so as
to bound almost the whole of the slit-like lumen (PI. 15,
fig. 1). At the dorsal and ventral edges of each run grooves
in the epithelium. Dorsally and ventrally these plates pass
into the unmodified cuticle which bounds the lumen of all the
rest of the organ: anteriorly the same holds good, but pos-
teriorly each plate projects a little beyond the general level
of the surface, as the free edge of a finger-nail, elsewhere lying
close on its bed, projects beyond the general surface of the
digit. The radial muscle-columns of the wall of the proven-
triculus in the belt through which these chitinous plates
extend, depart from their arrangement in regular annular
zones (fig. 5), and, as observed by Malaquin, run obliquely
inwards and forwards or inwards and backwards. ‘The object
of this oblique direction would seem to be to enable the two
plates to be tilted up so that their edges may be brought mto
contact.

VII. THe VENtTRICULUS.

The ventriculus is a small chamber, reduced or absent in
some. It has fairly thick walls with a correspondingly reduced
lumen. Numerous thin bundles of muscular fibres run radially

PROBOSCIS OF SYLLIDEA $35

through the substance of the wall ; but the chief space is taken
up by the epithelium modified into a mass of gland-cells similar
to those composing the anterior proventricular glands. They
are apparently syncytial, and in most specimens present the
appearance, in the aggregate, of a mass of sinuous and anasto-
mosing tubules and vacuoles with thin walls and without distinct
contents: more rarely the spaces are filled with a secretion
capable of taking a strong stain with haematoxylin. In the
EKxogoneae the‘ ducts’ from this mass of unicellular glands
do not seem to open—in great number at least—into the cavity
of the ventriculus itself, but run forwards to open into the
recess at the extreme posterior end of the proventriculus. It
is in very few preparations that this destination is traceable :
the specimen must happen to have been fixed when the secre-
tion was actually being discharged, and the strands of secretion
by which alone the course of the ‘ ducts’ is traceable, must
have become differentially stamed. In the other sections of
the Syllidea I have not been able to trace this connexion,
and I am led to conelude it is not universal.

VIII. Tue Post-vENTRICULUS.

Sharply marked off both from the ventriculus in front and
the intestine behind is the chamber from which are given off
laterally the two caeca present in most of the Syllidea
with the exception of the Autolytea.

This, as already noticed, is recognized as a separate chamber
by Hisig, and he gives prominence to it as the second main
division of the alimentary canal—the first being the whole
proboscis-oesophagus (Riissel6sophagus) and the third the
intestine. De Saint-Joseph, on the other hand, and Malaquin
do not recognize the distinctness of this chamber from the
proventriculus. Its walls have only a thin layer of muscle!
without radial fibres. Its epithelium is ciliated and is loaded

1 It may be pointed out here that Malaquin was in error in stating that
the intestine is devoid of a muscular layer. There is a thin layer of flattened
fibres, not placed in close contact with one another, composed of outer
longitudinal and inner circular elements.

386 W. A. HASWELL

with unicellular glands. The caeca are of essentially the same
structure, with numerous unicellular glands which discharge
their secretion into the lumina. A name is needed to designate
the small but distmet part of the digestive canal from which
the caeca are given off. The term oesophagus is in general
use for a corresponding part in the Nereids ; but, whatever its
claims, it seems inappropriate to a compact glandular chamber
with a ciliated epithelium. I propose instead the term post -
ventriculus as-not involving any doubtful homologies
and indicating simply position.

Though the post-ventriculus, with the caeca, resembles the
intestine in its ciliated epithelium, it differs from the latter in
the presence of the very numerous and characteristic uni-
cellular glands. It is also sharply constricted off from it, and
the narrow aperture of communication between the two is
guarded by a valve composed of folds of the intestinal epithe-
lium which must prevent the passage of liquid forwards from
the intestine.

LITERATURE.

1. Benham, W. B.—‘‘ Notes on some New Zealand Polychaetes ”,
‘Trans. N.Z. Inst.’, vol. 47, 1914.

2. Claparéde, R.—*‘ Annélides chétopodes du Golfe de Naples’, “Mém.
Soc. Phys. Hist. nat. Genéve ’, t. 19, 1868.

3. Ehlers, E.—‘ Die Borstenwiirmer’. Leipzig, 1864.

4, Kisig, H.—‘* Ueber das Vorkommen eines schwimmblasenahnlichen

Organs bei Anneliden ’’, ‘ Mittheil. Zool. Stat. zu Neapel’, II. Bd.,

1881.
5. Haswell, W. A.—** On the structure of the so-called glandular ventricle
of Syllis ’, ‘ Quart. Journ. Micro. Sci.’, N.S., vol. 26, 1886.
6. “A comparative study of striated muscle ”’, ibid., vol. 30, 1889.
at * On the Exogoneae ”’, ‘ Journ. Linn. Soc. Zoology’, vol. 34, 1920.
8. McIntosh, W. C.— A Monograph of the British Annelids’, vol. 2,
part 1. Ray Society, 1908.
9. Malaquin, A.—** Recherches sur les Syllidiens”’, ‘Mém. Soc. Sci.
Lille ’, 1893.
10. Quatrefages, A. De.—‘ Histoire naturelle des Annelés ’, 1865.
11. Saint-Joseph, Baron de.—‘‘ Les Annélides polychétes des ecdtes de
Dinard ’’, ‘ Ann. sci. nat.’, 7° série, t. i, 1887.

PROBOSCIS OF SYLLIDEA 8387

EXPLANATION OF PLATE 15.

Letlering common to all Figures.

an., annular bands of non-striated muscle ; a.pr.g., anterior proventri
cular glands ; @.7., accessory radial fibres ; ¢.p., chitinous plates of pro-
ventriculus ; ¢u., cuticle; e.m., external membrane; ep., epithelium ;
i., intestine; 7.m., internal membrane; p., pharynx; pa., pharyngeal
papillae; p.g., pharyngeal gland; pr.v., proventriculus; p/., post-
ventriculus ; 7., radial muscle-columns.

Fig. 1.—Diagram of a transverse section of the proventriculus of Syllis :
about one quadrant shown. The section is represented as passing through
the chitinous plates; but the typical arrangement of the muscles is
illustrated—not the modified arrangement in the chitinous plate region
(see fig. 5). The coelomic epithelial layer is not represented in this or the
other figures.

Fig. 2.—Part of a transverse section of the proventriculus of Syllis
variegata (Grube). x330. The dark transverse lines passing across
the radial muscle-columns indicate the telophragms. The accessory
radial (a.r.) fibres are drawn in black, as in the other figures, for the sake
of contrast.

Fig. 3.—Portion of a tangential section of the proventriculus of
S. variegata internal to the annular bands. x 330. The pattern of
the transverse sections of the cortex (“Cohnheim’s areas’) is not repre-
sented in this or in the following figure.

Fig. 4.—Portion of a tangential section of the proventriculus of
S. coruscans (Haswell), internal to the annular bands. x 330.

Fig. 5.—Part of a horizontal section of the proventriculus of Syllis
closterobranchia (Schmarda), passing through the chitinous plates.
x 330.

Fig. 6.—Diagrammatic general view of the proboscis of a Syllis
from above: part of the dorsal wall of the pharynx and proventriculus
removed to show the region of the anterior proventricular glands and the
chitinous plates. Only one of the pharyngeal glands is represented.
a.pr.g., anterior proventricular glands ; c., caeca; ¢.p., chitinous plates ;
cu., thickened cuticle of the pharynx; 7., intestine; p., pharynx ;
pa., pharyngeal papillae; p.g., pharyngeal gland ; pr.v., proventriculus ;
pt., post-ventriculus.

Fig. 7.—Semi-diagrammatic view of a horizontal section through the
junction of the pharynx and proventriculus of Grubea to show the
position and relations of the anterior proventricular glands. x 780.

rv ie

oe aoa hy Ole ah ae pues:
‘ foe) evel» ital de) Sailseaie ts Te teaaa ams brviagl’
pt tin, Aes Mt Palas #2 65
- ura ero deere alas +f? pe
. i Ave ce S Unningtnttert leaeAT he
wy a ee ee. Me fT nate"!
anit wey inde

J Lagi, od? le molt aerial at 0am,
m™ ‘nee eye) Aas ae mt JTW ONte ts

pe =pineallerm Wy “yeh itt had =
ioe tl AD a fore aga seryt? a a
é : : a = wend (OPE hay aT Srp ni aii fs “odie
evs hie ; be

5
leu | rrr oe eon ae a
bad: pom efi ialy ad} > AN = dale
epee yt set ob iTV aru y :

~

ne wit @ ye bel at a ts we,
ant ‘ } wil ¥ cl ibegste a lo pu
baa : ial sala 1® ailf oa} Lana?

5 : “boera & miodnne r- mri nlp Wi ilo
uml eal! yea

a rey ett WM aAdS aS Letts giant awn
: dnul alieee tio beeen ieee
oe ‘oo rerny iti ee tonite i ee oe “i
(1) wl © deeply Sankey Join eesti eee
a ven 5%
- jpinli tT) We err. Heemiog ot
: ' ed xiveterliy wet WO Taw Jpn ola ta
ime at alge Malet rt eevee ear nlante FAS be suche Yi

usite~uwme: < Qpiely engag tig wlth. Oi cle
ee Pe ee
ipiitey wrietine A etme le hee

np feng 0 athe Lenya ag |

: ‘

jeune Loteoxteel & lo. este, ghia

ae A nat to eilenaiirering, tyre. Fag

vrenhg, eeleagalaanicad mente mld

>< ye [29 i pee (oF

, a A ‘> a mn i ay
7 z .- >
:

5 ae
a4

ag

ped 8

y

s
i

ae

Haswell. del.

Quart. Journ. Mrer: Sea. Vol. 65, NS., PL.15.

The Life-history of Melicertidium octo-
costatum (Sars), a Leptomedusan with
a theca-less Hydroid Stage.

By

Prof. James F. Gemmill, University College, Dundee.

With Plate 16.

CoNTENTS.
PAGE
DEVELOPMENT OF EGGs OF MELICERTIDIUM . ; : . 3840
DESCRIPTION OF THE TANK HyDROID . : ‘ : : . 41
GENERAL . ‘ : : : , : E ; : . 344
REFERENCES . ’ : E : ‘ : . 346

Tuts well-known medusa (fig. 19) is classified among the
Thaumantiadae, and is characterized by the presence of
eight ‘radial’ canals on which the gonads are developed.
The marginal tentacles are numerous (up to 140) and of
unequal size, larger and smaller ones alternating more or less
regularly. There are no lithocysts, cordyli or ectodermal
ocelli. The manubrium is short, the mouth four-angled and
without oral tentacles. The medusa has a fairly wide distribn-
tion in the North-east Atlantic ranging from Bergen to
Falmouth. (See EK. T. Browne, 4, for details and a discussion
of the nomenclature.) The hydroid, as I have ascertained by
rearing the eggs, proves to be a hitherto undescribed species
identical with one which has been noted for several years
(with a year’s interval of absence) growing abundantly and
spontaneously in the tanks at the Millport Biological Station.

An allied form Melicertum campanula (Agassiz)
occurs in West Atlantic Canadian and U.S. waters. In 1868

NO. 259 A a

840 JAMES F. GEMMILI,

A. Agassiz (1) described the young hydroid reared from the
eggs of Melicertum, but this hydroid has not up to the
present been recorded in nature from the American coasts.

Development of Eggs of Melicertidium.

Ripe examples appeared in the tow-nettings at Millport
towards the end of June 1918. By keeping specimens in
aquaria in the Research Fellowship Laboratory at Glasgow
University [ obtained numbers of fertilized eggs. These are
small (0-08 mm. in diameter), homogeneous-looking, faintly
yellowish in tinge, and with delicate closely-adberent mem-
brane. They are ripe before extrusion and pass outwards
through the mouth, as also do the spermatozoa in the males.
No membrane of fertilization is formed. Segmentation is total
and equal (figs. 1-5), the two-celled stage beginning with
a notch or groove on one side of the egg. A blastocoele cavity
is recognizable even at the eight- or sixteen-celled stage.
Karly blastulae are irregular in outline, the blastula wall being
a single layer, but exhibiting folds and inpocketings which soon
straighten out and do not seem to have any subsequent forma-
tive importance (figs. 6 and 7). The larva now becomes pear-
shaped, and, having acquired cilia, progresses with the blunt
end in front and rotates in the solar direction as viewed from
the blunt end (figs. 8 and 9). At this stage the endoderm
arises by inward budding from the blastula wall (figs. 8, 9, 10).
The budding occurs first near the pointed end, and then all
round, gradually fillimg up the blastocoele cavity, the last part
of this cavity to be filled being at the blunt end (fig. 11). The
endoderm cells are rounded, slightly granular, and less trans-
parent than the ectoderm. The planula now elongates, becom-
ing almost worm-like, and swims vigorously through the water
at any depth. Later it seeks the bottom and becomes attached.
The mode of attachment presents certain peculiarities which
I hope to elucidate later. The free end becomes swollen and
rudiments of the first tentacles appear (fig. 12). Figs. 12-14
illustrate four-tentacled and eight-tentacled stages. Both
show a delicate perisare covering hydrorhiza and hydrocaulus,

LIFE-HISTORY OF MELICERTIDIUM 341

and ceasing at the base of the hydranth without forming even
a rudimentary hydrotheca. At no stage are the bases of the
tentacles united by a web or membrane. The sixteen-tentacled
stage is entirely similar to young polyps (fig. 15) of the tank
hydroid deseribed later in this paper, though the latter are
relatively rather larger, no doubt because they could draw
during growth on a nutritional reserve greater than was at the
disposal of the parent of the colony. This year (1919) T have re-
peated the rearing experiments and obtained the same results.

Description of the Tank Hydroid.

In the early spring of 1916, 1917, and 1919 colonies of an
apparently new theca-less hydroid appeared on stones and
on glass in several of the tanks at the Millport Biological
Station. Dr. James Ritchie, Royal Scottish Museum, Edin-
burgh, to whom I sent a specimen in February 1917, made the
conjecture, which has proved right, that it might turn out
to be the hydroid of some Leptomedusan. A little later in
the same year young medusae budded off from a colony were
obtained. They had four radial canals, eight tentacles, no litho-
cysts, and no ectodermal ocelli or oral tentacles. I tried to rear
them, but without suecess. The matter remained there till
July 1918, when the results (given above) of rearmg Melicerti-
dium eggs unexpectedly connected the tank hydroid with
this medusa, and made me undertake more careful experi-
ments (see below) on rearmg the young medusae, when these
were budded off from the tank colonies in the spring of the pre-
sent year (1919). The characters of the hydroid are as follows :

Hydranth: entirely theca-less. Tentacles: long,
slender, tapering, with solid core of endoderm cells in a single
row, studded with nematocysts, not united at their bases by
a membrane, arranged in a single circle but tending when
fully extended to curve upwards and downwards alternately,
commonly sixteen in number, but often more numerous
especially in sterile colonies, in which individuals with as
many as thirty-two may be noted. Hypostome: conical
when closed, shaped like a shallow wide-mouthed urn when

Aa2

342 JAMES F,. GEMMILI

fully opened, lined for a very short distance downwards from
fhe margin by close-set columnar cells having the characters
of ectoderm. Body of Hydranth: sometimes slender,
elongated (1-7 mm. in length), sometimes short (0-9 mm.) or
vase-shaped according to contraction, usually showing con-
striction below hypostome, furnished with stinging cells
near middle, merging insensibly into hydrocaulus, except in
contracted condition, when junction becomes evident.
Hydrocaulus: short but varying in length (1 mm. to
1:7 mm.), often irregularly bent, evidently weak, unbranched
except in giving off the stalk of a medusa bud. Hydro-
rhiza: creeping, branching but not anastomosing, 0-1 mm.
across (including perisare). The distinction between hydro-
caulus and hydrorhiza is not always sharply apparent. In the
thicker parts of a colony hydrorhizae may intertwine, and
leaving the surface of attachment become equivalent to low
irregular branching hydrocauli. When, however, the hydro-
rhizae are not too crowded they remain adherent and give off
unbranched hydrocauli. Perisare: thin, wrinkled irregu-
larly but not ringed, enclosing hydrorhiza and hydrocaulus
and separate from these except at occasional points of
‘anchorage ’, thinning away at distal end of hydrocaulus and
fusing with ectoderm at base of hydranth which is entirely
theca-less. Medusae: Gonophore production takes place
from the beginning of Iebruary till the end of March. Parts
of the colony were isolated, kept in filtered sea-water, and in
course of time a number of young medusae were collected.
The buds appear at the end of short stems arising from the
hydrocaulus well below the base of the hydranth, each hydro-
caulus only producing a single medusa. The medusa buds,
especially at full size, are more elongated than the free medusae,
but the characteristic shape is acquired during the period
immediately prior to detachment when vigorous pulsations
may be noted. The young medusae have four rather wide
radial canals, four tentacles opposite these, four small tentacles
or tentacle buds in the interradii, and no lithocysts or ecto-
dermal ocelli (figs. 16, 17, 18). The bell is dome-like and
moderately deep: the stomach is quadrangular and_ the

LIFE-HISTORY OF MELICERTIDIUM 343

manubrium short, showing four blunt, radially-placed, grooved
angles. At first the bell shows a small pit m the middle of the
aboral surface, to the bottom of which a cone-like projection
of the stomach is anchored. Later this remnant of the con-
nexion between bud and stalk becomes severed, and the summit
of the dome shows an upward convexity (fig. 17). Over the rest
of the bell, the mesogloea superficial to the plane of the stomach
and radial canals forms a relatively thin layer. At their bases
the tentacles are hollow and slightly swollen, the endoderm
here containing yellowish intracellular pigment. The measure-
ments of the young medusa at rest are: height 1-2 mm.,
breadth 1:3 mm., interradial diameter of stomach 0:45 mm. ;
breadth of radial canal 0-06 mm., depth of superficial meso-
gloea 0-075 mm. The surface of the bell shows numerous
minute glancing-points which do not disappear on treatment
with acid. ‘The medusae were kept alive for a time, and increased
in size; the four imterradial tentacles grew almost as big as
the radial ones, and new tentacle buds appeared in irregular
sequence, one for each interspace between a radial and an
interradial tentacle. Stages with ten, twelve, fourteen, and
sixteen tentacles were thus obtamed. Medusae four weeks old
and with c. ten tentacles showed a single blunt outgrowth
from the stomach in each interradius (fig. 18, b). A week later
(c. twelve tentacles) these outgrowths had extended over the
summit of the bell, becoming pointed at their ends. In another
week or fortnight (c. fourteen to sixteen tentacles) the out-
growths had extended downwards along the sides of the bell
and become continuous with slender corresponding upgrowths
from the ring canal (fig. 19). I failed to rear the medusae
further, but they had already reached the eight-rayed condition
characteristic of Melicertidium.

I have not obtained the early four-rayed medusae in tow-
nettings off the Millport Station, but they were moderately
abundant during April 1919 in plankton from the Gareloch,'
an inlet farther up the Firth of Clyde.

! Since this paper was written, I have found the intermediate stages
described above in May plankton from this locality, and the adults
at the end of June.

344 JAMES F. GEMMILL

General.

As far back as 1865 A. Agassiz (1, p. 130) inferred from the
results of tow-nettings that the eight-rayed condition in
Melicertum campanula was reached by the formation
of four new interradial outgrowths from the stomach in an
originally four-rayed young medusa.

Mayer (7, p. 208) thinks that Melicertum campanula
(Agassiz) and Melicertidium octocostatum (Sars) are
probably identical species, and that Melicertum should have
priority as the generic name. However, there are suflicient
reasons (especially under (1) and (2) in the following com-
parison) for keeping Melicertum and Melicertidium as distinct
genera, at least in the meantime.

Melicertum Melicertidium
(hy droid) (hy droid)
(1) Tentacles united at their (1) Tentacles not united at
bases by a membrane. their bases by a mem-
brane.
(2) A small theca at base of (2) No theca.
hydranth.
(3) Tentacles up to ten in (8) Tentacles sixteen or more
number. (up to thirty-two) in
number.
Melicertum Melicertidium
(medusa). (medusa).
(4) Karliest free stage with (4) Earliest free stage with
only two marginal ten- four marginal tentacles
tacles. and four intervening ten-

tacle buds.
(5) No ‘radiating lines’ on (5) Numerous ‘radiating lines’

sub-umbrellar surface. onsub-umbrellar surface.
(6) Marginal tentacles, in adult (6) Marginal tentacles in the
equal or sub-equal in largest specimen exam-
g1ze. ined consist of about sixty

small tentacles and about
eighty much larger ones.

LIFE-HISTORY OF MELICERTIDIUM 345

I agree with Romanes’ opinion (9, p. 527) that the * radiating
lines’ referred to under Melicertidium (medusa) above are
bands of muscle fibres, and not of nematocysts as 1s thought
by Browne (4, p. 764) and others.

Additional instances in which theca-less hydroids have been
reared from Leptomedusae are recorded by Claus (5), Metchni-
kotf (8), and Brooks (2). The medusae concerned belong to
the genus Eutima (McCrady), the species bemg Eutima
campanulata (Claus), Octorchis gegenbauri(Haeckel),
in the first two cases, and Hutima mira (MeCrady) in the
third. Hutima differs from Melicertum and Melicertidium,
among other things, in having marginal lithocysts, and in
having the stomach mounted on a long peduncle. In the
hydroid of E. campanulata, described by Claus and
named by him Campanopsis, the tentacles are up to twenty-
four in number and are united at their bases by a membrane.
A theca is entirely absent, and the young medusae are formed
near the middle of the hydranth body. Brooks (2) describes
the hydroid of EH. mira as small, Perigonimus-like, with
eight tentacles united at their bases by a membrane.

KE. Stechow (10) has described a theca-less hydroid, with
short hydrocaulus having definitely ringed perisare, with
hydrorhizae forming a network, and with fourteen to eighteen
tentacles which were not, so far as could be made out in the
preserved material, united at their bases by a membrane.
‘he specimens were in a tube left by a former assistant at
Munich and were labelled ‘ Polyp of Octorchis’. Stechow
names it Campanopsis dubia and considers the medusa
to have been an Octorchis Eutima.

On the whole, the life-history of Melicertidium supports
the generally-accepted view that Leptomedusan hydroids are
derived from Anthomedusans. ‘The hydroid is theca-less, the
medusa is deep and has no lithocysts or ectodermal ocell, and
though the gonads are on the eight radial canals in the adult,
the mode of development of the second four radial canals by
outgrowths from the stomach makes it clearly possible that
ontogenetically or phylogenetically the gonad tissue of the

346 JAMES F. GEMMILL

other four originates in the region of the stomach or manubrium.
Indeed, in the earhest stage of the Melicertidium medusa
identified by Browne (4, p. 763) the gonads extended outwards
from the stomach only along the proximal halves of the radial
canals.

The Leptomedusan Family ‘haumantiadae, to which
Melicertum and Melicertidium belong, contains other twelve
typical genera. The hydroid stages of only three of these,
viz. Thaumantias (Wright, 11), Laodicea (Metchnikoff, 8),
and Dipleurosoma (Browne, 8), are known, and, curiously
enough, they all possess complete thecae. In having a rudi-
mentary theca Melicertum recalls the Anthomedusan Peri-
gonimus, while Melicertidium having no theca is in line with
Kutima (Campanopsis) and Tima, which are members of the
Leptomedusan Family Eucopidae. Dr. James Ritchie com-
pares the general facies of the Melicertidium hydranth to that
of Halecium. The just liberated medusa of Melicertidium
resembles that of Podocoryne carnea except in having
a slightly shorter manubrium and no oral tentacles. It is
evident that on the borderland between the Antho- and the
Leptomedusae there are numerous forms which, whether in
their hydroid or their medusoid stages, exhibit features charac-
teristic of better-defined members of either group.

I have to thank the Trustees of the Carnegie bequest for
a grant in aid of the expenses of this investigation, and the
staff of the Millport Station for help in obtaimg material and
rearing the young medusae.

REFERENCES.

1. Agassiz, A~—‘‘ North American Acalephae ”’, * Cat. Mus. Comp. Zool.
Harvard ’, ii, 1865, p. 130, figs. 202-4.

2. Brooks, W. K.—** On the Life-history of HEutima and on Radial
and Bilateral Symmetry in Hydroids ”’, * Zool. Anz.’, Bd. 7, 1884.

3. Browne, E. T.—‘* Fauna and Flora of Valencia Harbour”, * Proc.
Roy. Irish Acad.’, ser. iii, vol. 5, 1898-1900, p. 696.

4, —— “ A Report on the Medusae found in the Firth of Clyde ”, * Proc.
Roy. Soc. Edinburgh ’, 1905, p. 72.

LIFE-HISTORY OF MELICERTIDIUM 347

Claus, C.—‘* Arbeit. Zool. Inst. Wien’, Bd. 4, 1881, Heft 1. p. 589.

. Hartlaub.— Wiss. Meeres-Untersuch.’, Bd. 1, 1894, p. 192.

. Mayer, A. G.—‘*‘ Medusae of the World’, * Public. Carnegie Inst.

Washington *, L09, 1910, vol. i, pp. 207-8.

8. Metchnikoff, E.—‘* Embryologische Studien an Medusen*. Wien, 1886.

9. Romanes, G. I.—‘‘ An Account of some new Species, Varieties, and
Monstrous Forms of Medusae ”’, ‘Journ. Linn. Soe. Zoology ’, vol. 12,
1876, pp. 524-31.

10. Stechow, E.—** Kin thecenloser Hydroid der mit einer Leptomedusa
in Generationswechsel steht’, ‘Zool. Anz.’, Bd. 41, 1913, pp. 582-6.

11. Wright, T. S.—‘* On the Reproduction of Thamantias incon-

spicua’’, ‘ Quart. Journ. Micr. Sci.’, vol. 2, 1862, pp. 221, 208.

ADH

EXPLANATION OF PLATE 16.

Figs. 1-15.—Development of Melicertidium.

Figs. 1-7.—Stages in segmentation and blastula formation of the egg
of Melicertidium.

Figs. 8-]1.—Change to the planula, formation of endoderm (end), &c.
The arrow and circle between 9 and 10 indicate respectively the direction
of progression of the larva, and its rotation as viewed from the narrow
end,

Figs. 12-18.—Fixation of the larva: formation of first tentacles.

Fig. 14.—Stage with eight tentacles.

Fig. 15.—Portion of a colony, (a) hydranth; (b) medusa ready for
liberation ; (c) young hydranth and young medusa bud; (/¢) medusa
bud almost fully grown; (e) hydranth fully stretched out; (/) young
polypite arising from a hydrorhiza.

Fig. 16.—Just liberated medusa.

Fig. 17.—Aboral part of medusa, two weeks old, showing mesogloeal
projection on summit of bell.

Fig. 18, (a), (b), (c)—Stomach and radial canals viewed from above
in two days, three weeks, and six weeks’ old medusae respectively, showing
the formation, by interradial outgrowths from the stomach, of four new
“radial ’ canals. .

Fig. 19.—Medusa, seven weeks old, showing interradial outgrowths from
the stomach which have met corresponding upgrowths from the ring
canal. R, one of the four original radial canals ; LR., one of the four new
interradial canals formed in the manner described above.

(Ae obMe 2G ROTEL:

Fy HET, G3 a ‘ f ans ; ’ As fe
y one . 9 ° Teh) 1 |
. bet t il’) ‘ a! Mai (8 ie bat.
ri ut " ary ' Me dy Pst ;
7 7 iy fay ‘ 7 i sshd war?
4 A AA fi Wi
of ie - - : > _ u lia
‘ 5 ay if po a ? ’ J
i Sina :
« r
—
;
:
s *

!

A

Al

J.F. Gemmill del.

Huth,London.

Luart. Iourn. Mor Sct VA.6. SMS, Pt. 76.

—_

- os rer ea OS he feta ees FS

On the Blood-Vascular System of the Earthworm
Pheretima, and the Course of the Circu-
lation in Earthworms.

By

ry

Karm Narayan Bahl, D.Se..

Of the Muir Central College, Allahabad, India.
(From the Department of Comparative Anatomy. Oxford.)

With 11 Text-figures.

TABLE OF CONTENTS.
PAGE
1. INTRODUCTORY : 5 . ; F : : ; eit}
2. THE TypricAL ARRANGEMENT OF THE BLOOD-SYSTEM IN THE
INTESTINAL REGION OF THE BoDY BEHIND THE FOoOuUR-

TEENTH SEGMENT . 352
(a) The Longitudinal Trunks 352
(b) The Intestinal Blood-plexus . 857
(c) The Commissural, Integumentary, and Nope idial Vous se ealne
(d) The Dorso- and Ventro-intestinals . : : : . 0368

3. THE BLoop-sySTEM IN THE First FouRTEEN CEPHALIZED
SEGMENTS . : : - j : . 369
(a) The reel Pranks 5 : ; : 5 2! a7
(6) The *‘ Hearts’ and the Anterior Loops. : eSilics
(c) The Blood-vessels of the Oesophagus and Pane) ynx . . 378

4, COMPARISON OF THE BLOOD-SYSTEM OF PHERETIMA WITH
THAT OF LUMBRICUS AND ALLOLOBOPHORA . . 318

5. THE VALVES IN THE BLOOD-VESSELS AND THE COURSE OF THE
CIRCULATION OF THE BLOOD ; : £ : . 380
6. SUMMARY , 3 : ; 3 i ; : : ool
7. List oF REFERENCES . y ; ; : F ; » 3892

350 KARM NARAYAN BAHL

1. INTRODUCTORY.

Tue blood-vascular system of earthworms has engaged the
attention of many distinguished observers. _Lankester (12)
described the blood-vessels of Lumbricus in one of his
memoirs on the ‘Anatomy of the Karthworm’, which forms
about the earliest contribution to this subject. Jaquet (9)
gives a comparative account of the vascular system in Annelids,
describing the system in typical genera of the various classes
of the group. Of the Olgochaeta, he selects Lumbricus
as a type. Perrier (18) and Benham (5), also working on
Lumbricus, describe the course of flow in all the blood-
vessels from a study of the disposition of the valves; to
Benham we also owe our knowledge of the blood-supply of
the nephridium in Lumbricus (6). Harrington (8) gives
a detailed account of the anatomy of the blood-system in
Lumbricus with elaborate diagrams, and was the first
to describe the arrangement of blood-vessels in the integument.
Recently, Johnstone and his student, Miss Johnson (10 and 11),
have published two papers on the course of blood-flow in
Lumbricus demonstrating the course in various vessels by
a series of interesting experiments and observations. ‘The
blood-system has thus been thoroughly studied in Lumbricus
since that is the form studied as a type in Europe and America.
Amongst the Oriental forms of Oligochaeta, Bourne (1) has
described the blood-system in some detail in the Perichaete
worm Megascolex and also in Moniligaster grandis
(2, 1894), a huge worm about two feet long placed by Beddard
in the group Microdrili. Besides Bourne’s work on Mega-
scolex, very little attention has been paid to the blood-
system of the Perichaetidae, the largest family of earthworms.

The earthworm Pheretima (the genus Perichaeta sensu
stricto) is now studied as a type of the Oligochaeta in
Northern India and also at the Universities of Bombay and
Calcutta, and it has become necessary, therefore, to have as
complete a knowledge as possible of the anatomy of this form.
An attempt has been made in this paper to present an account

VASCULAR SYSTEM OF PHERETIMA 351

of the blood-system of Pheretima and the course of blood-
flow about which, even in Lumbriecus, there has been
a great divergence of opinion amongst the various observers.
Some of the observations were made in India, but in this
country, besides having an opportunity of examining the
two English genera Lumbricus and Allolobophora,
I was able to complete my work on Pheretima, having
been lucky to obtain specimens of this Oriental form in the
Lily-house of Kew Gardens.

The work was carried out in the Department of Comparative
Anatomy at Oxford. Iam indebted to Professor E. 8. Goodrich
for his keen interest in my work; he has made valuable
suggestions, and has also found time to read through and correct
the manuscript of the paper.

Although essentially the blood-systems of both Lum-
brieus and Pheretima can be reduced to a common
type, there are important differences in the system in the two
genera, which I have indicated in the text. Pheretima
resembles Allolobophora rather than Lumbricus so
far as the blood-system in the general body-region 1s concerned,
while the system differs in important respects from that of
Megascolex. As regards the course of the blood-flow studied
by holding the vessels with fine forceps, by cutting the vessels
and observing the direction of blood-flow, and by a study of
the valves, I am led to confirm the observations and con-
clusions of Johnstone (10 and 11) and to reject part of Bourne’s
theory of the course of the circulation (1).

The typical arrangement of the blood-system in P here-
tima is found behind the fourteenth segment, bemg meta-
merically repeated behind that segment. In the first fourteen
segments, on the other hand, this typical arrangement is con-
siderably modified, this modification, together with that shown
in the digestive, reproductive, and nervous systems, being spoken
of as cephalization. It will be convenient, therefore, to
describe, as Harrington (8) does in the case of Lumbricus,
first, the typical arrangement as it occurs in the region of the
body of the worm behind the fourteenth segment, and then the

$52 KARM NARAYAN BAHL

blood-vessels in the first fourteen cephalized segments, and

finally to diseuss the course of the cireulation in the system.

29. Tur Typtcan ARRANGEMENT OF THRE BLOOD-SYSTEM IN
THR INTESTINAL REGION or THE Bopy BEHIND THE
FOURTEENTH SEGMENT.

The blood-system in this system in this region of the body
consists of (a) three longitudinal trunks running parallel to
one another, namely, the dorsal, the ventral, and the sub-
neural vessels; (b) the intestinal blood-plexus, situated in the
wall of the gut, is directly connected with the dorsal and
ventral vessels, and indirectly with the subneural; and
(c) the commissural, integumentary, and nephridial vessels.

(a) The Longitudinal Trunks.

1. The dorsal vessel.—tThe dorsal vessel is the most
prominent of all the blood-vessels in the worm and is rhythmi-
cally contractile. It runs along the mid-dorsal line immediately
beneath the body-wall, between the latter and the intestine,
and is at once seen lying on the gut, when the worm is opened
by a mid-dorsal incision. In Lumbricus the dorsal vessel
is heavily covered over with ‘ yellow cells’, which must be
removed before the vessel is seen; but in Pheretima
the * yellow cells’ do not cover the dorsal vessel, so that the
latter is at once prominent on dissection. Although lying close
upon the gut, the dorsal vessel is not actually attached to the
wall of the former in any portion of its course. It is single
throughout its length and has thick muscular walls which are
responsible for its contractility. The average diameter of this
vessel is about 220 ~ ; it is narrowest at places where it pierces
the intersegmental septa. On opening a narcotized worm,
we can easily see the wave of contraction in this vessel travelling
from behind forwards and consequently driving the blood
in that direction. During its course through the body, the
dorsal vessel, on piercing each septum, has a pair of forwardly-
directed valves (figs. 7 and 10) in its lumen. These valves,

A

VASCULAR SYSTEM OF PHERETIMA 353

as I shall show later, prevent the flow of blood backwards
when the vessel contracts. There are also valves (vide
infra) at the orifices of the dorso-intestinal and commissural
vessels.

[t will be seen from fig. 1 that the dorsal vessel is connected
with the intestine by two pairs of dorso-intestinal vessels
(di.v.) in each segment ; these vessels serve to establish a com-
munication between the internal intestinal plexus and the
dorsal blood-vessel (fig.2). The anterior pair of dorso-intestinals
come off from the dorsal in the anterior third of the segment,
while the posterior pair he in the posterior third, nearing the
hinder septum of the segment, in close association with the
so-called ‘ lymph-glands ’ which lie on each side of the dorsal
vessel in every segment here. These dorso-intestinals are
very short vessels, being only about 450, in length, on an
average. They soon enter the imtestinal wall, in which they
are continued as ‘ transverse vessels’ (vide infra).

Again, just before piercing each septum from behind, the
dorsal vessel receives a commissural vessel (the dorso-lateral
or the parietal vessel), which is connected ventrally with the
subneural (comm.v., figs. 1 and 2). This commissural vessel
runs along the posterior face of each septum very near and
parallel to its outer edge, i.e. the edge joining the body-wall ;
and is connected with capillaries of the nephridia and the
body-wall.

As I shall show later on, both the dorso-intestinal and the
commissural vessels bring blood into the dorsal vessel and
replenish its supply. No blood leaves the dorsal vessel in this
region of the body.

2. The ventral vessel.—The ventral vessel, like the
dorsal, is single throughout its length and extends from the
anterior to the posterior end of the body. In the region of
the intestine it has an average diameter of 115 uw and gives
off a pair of ventro-tegumentary branches in each
segment. Each of these branches leaves the ventral vessel
just anterior to the septal wall in each segment and, after
running alongside the anterior face of each septum for a little

“‘yassoa Are WOUINF94-OI}UIA ="a"7a £ [OSsOA [VUTYSIZUT-O.QUGA =—"A"72a £ [OSSOA [®4jU0A ="a'a * JesSeA OSIOASURIy
=‘a'suns ‘ ajosopydsy="h) f ows [ejos=s's fyouviq jerpraydou-oydos="qrus {ossoa [wnouqns=<a°us $ [essen [RUTySseqUT
-oydos="ars Ssjyonp Arojoloxo [euTsoyur-vadns="p'ays : pxoo-earou="9"'u { sjassoa ArvyuoUNSezuI="a-yua £ sprvaroy
peu umgdes [ejzueusesi9ojuI=,'s"? tungdes [ezyUSUSes1oyUI="s"2 { OUTZSEQUT= "yur £ JaSSOA [BUI}8eqUI-OS1Op =*a"ap
S[OSs9A [BSIOp="a'p fF [assed [BINSSTUITUOD="awMuor {4no peyooHor pue yno [[eM-Apoq ="m-q £[eM-Apog ="a'q
‘dorpIsod UT sjasseA-poolq oY} esodxo 04 Jopso UI “yno poyooper pUk yNO UaEq SLY s}UEWBes JO110}Ue MOF ay} Jo opts
4J9[ OY} JO UTYS oY} Jo 4avd 10}veI3 oY} pUv ‘UMOYsS ore syUEUZeS OAT “vUUNYYZsod vuIyoI9qY gq Jo UOIZoI rejosopyd Ay
oy} UI QueUIdes YZOO}ZANOJ oY} puryeq Apoq 944 Jo UOIFeI oY} UT Wazshs-poojq ay} Jo UOTRYUESeIdor O1VRUURABeIp WF

S's AQUI

NARAYAN BAHL

KARM

CE bot eee G

Ee S=

Ai y \
avsueay nip st A wwOD qus

354

'T ‘Ola-Lxay,

SS

VASCULAR SYSTEM OF PHERETIMA 355

distance, it pierces the septum and gets into the succeeding
segment (vt.v., fig. 1). Here it lies on the inner surface of the
body-wall near the middle line of the segment just in front of
the row of setal sacs, going right up near the mid-dorsal line
(figs. 1 and 2). As it ascends along the body-wall transversely,
the ventro-tegumentary vessel (vt.v.) gives off backwards and
forwards capillaries that supply blood to the body-wall
(epidermis and the muscles) and the integumentary nephridia.
Besides, the septal nephridia and the prostates also receive
their blood-supply from the ventro-tegumentaries. The septal
nephridia are supplied by a septo-nephridial branch (sn.b.,
fig. 1) of the ventro-tegumentary given off in each segment
at the place where it pierces the septum; while the prostate
glands in the segments sixteen to twenty-one receive small
branches from the ventro-tegumentary in each of these
segments.

Besides the paired ventro-tegumentary branches the ventral
vessel gives off dorsally a single unpaired ventro-intestinal
vessel in each segment (vi.v., fig. 1). This vessel originates
from the ventral a little behind the middle of each segment,
and runs forward to enter the ventral wall of the imtestine,
by three or four branches, close to the anterior intersegmental
septum. The ventro-intestinal, though generally overlooked
in this worm, is, however, an important vessel, and measures
as much as 1-5 mm. in length in some worms from its place of
origin on the ventral vessel to its place of entrance into the
intestinal wall. It puts the ventral vessel into communication
with the intestinal plexus. There are no valves anywhere along
the course of the ventral vessel.

The ventral vessel is the main and, in fact, the only distribut-
ing channel in the intestinal region of the body. All parts
in this region get their supply of blood from the ventral
vessel.

3. The subneural vessel.—The subneural vessel runs
along the mid-ventral line of the body-wall, being intimately
attached to it, and lies, as its name indicates, beneath the
nerve-cord. It is a very slender vessel and extends from the

NO, 259 Bb

356 KARM NARAYAN BAHL

posterior end of the worm to the fourteenth segment anteriorly,
being absent from the first fourteen segments. The com-
missural vessel, connecting the subneural with the dorsal in
the septal regions, has already been referred to above. At
about the middle of each segment just in front of the line

TEXT-FIG. 2.

Dr

Wm

Mimo

WX

SILV.

A diagrammatic transverse section through the region of the intes-
tine, the right half showing a section through the intersegmental
region and the left half through a segment proper passing through
one of the dorso-intestinals. b.w.=body-wall; c¢.e.p.=capil-
laries of the external plexus; c.i.p.=capillaries of the internal
plexus; comm.v.=commissural vessel; d.v.=dorsal vessel ;
di.v.=dorso-intestinal vessel; si.v.=septo-intestinal vessel ;
s.v.=subneural vessel; trans.v.=transverse vessel; ty.v.=
typhlosolar vessel; v.v.=ventral vessel; vt.v.=ventro-tegu-
mentary, vessel,

of setal sacs, the subneural receives a pair of very small branches
from the ventral part of the body-wall. One also finds in
sections the subneural receiving a branch on its ventral side
from the body-wall in the mid-ventral line (fig. 2).

The subneural is connected with the intestinal plexus

r

VASCULAR SYSTEM OF PHERETIMA 357

through the septo-intestinal (si.v., figs. 1 and 2), a
vessel which I describe below along with the commissural
vessel.

This vessel collects blood from the small ventral part of the
body-wall and the nerve-cord; and as the area over which
its branches ramify is very small and the quantity of blood
received is also small, the vessel itself is very slender as com-
pared with the other longitudinal trunks.

There are no supra-intestinal vessels in this region
in this worm: a pair of longitudinal ducts attached to the mid-
dorsal line of the gut and described as supra-intestinal blood-
vessels by Stephenson (14) have already been shown by me to
be excretory ducts (7).

There are also no lateral neural vessels as found in Lum-
bricus.

(b) The Intestinal Blood-plexus.

The intestinal blood-plexus (fig. 3) consists of a close network
of capillaries and blood-vessels in the walls of the intestine.
In Pheretima as in Megascolex (1) there are two
capillary networks in the alimentary canal, ie. (1) an internal
deep-lying network, and (2) an external more superficial one.
The internal network lies deep in the wall of the gut inside the
layer of circular muscle-fibres, between it and the internal
epithelial lining ; while the capillaries belonging to the external
network lie on the surface of the gut-wall amongst or even
outside the yellow cells (chloragogen cells) which form the
splanchnic layer of the peritoneal lining of the coelom. When
a freshly-killed worm is opened in saline solution it is at once
seen that the blood-plexus on the gut is marked out into three
distinct regions—the first region is from the fourteenth to the
twenty-sixth segment, where the intestinal capillaries are very
thickly set and lie at right angles to the longitudinal axis of
the body (transverse capillaries) ; the second is the longest
portion and extends from the twenty-sixth segment to twenty-
three to twenty-eight segments in front of the anus, the main

Bb2

858 KARM NARAYAN BAHL

portion of the plexus in this region consisting of longitudinal
capillaries lying parallel with one another along the intestine
all round the circumference ; and the third region comprises
the last twenty-three to twenty-eight segments of the animal,
where the blood-plexus differs markedly from what we have
in the first two regions. The difference in appearance of the
blood-plexus in the three regions is illustrated in fig. 3, where
at the pomt marked z there is a sudden change in the arrange-
ment of capillaries from the second to the third region. While
there is a regular, almost rectangular arrangement of the
capillaries in the anterior two regions of the gut, the capillaries
in the posterior region (last twenty-three to twenty-eight
segments) branch off in a tree-iike fashion from the dorso-
intestinal vessels. That the three regions mentioned above are
distinct from one another will be evident from the fact, ascer-
tained by a study of sections passing through the three regions,
that in the first region (fourteenth to twenty-sixth segment)
the intestinal capillaries form only the internal plexus, the
external plexus being absent, that in the second region
(twenty-sixth segment onwards) there are both the internal
and external plexuses well developed, while in the third region
(last twenty-three to twenty-eight segments) we have no
internal plexus at all, all the capillaries belonging to an external
plexus.

Besides the difference in the arrangement and position of
capillaries in the three regions there is another feature which
also distinguishes these three regions from one another, and
that is the presence and absence ot a typhlosole and the
typhlosolar vessel. Taking the last region first, we have to
note the entire absence of a typhlosole in this region. Bed-
dard (8) describes the absence of typhlosole in the last few
segments of Acanthodrilus, and ealls this last part of
the gut without a typhlosole the ‘rectum’. Similarly, the
typhlosole is absent in the gut in the last thirty-six segments
of Lumbricus, and we can apply the term ‘ rectum’ to
these last thirty-six segments of Lumbricus and the last
twenty-three to twenty-eight segments of Pheretima.

VASCULAR SYSTEM OF PHERETIMA 359

It seems reasonable to suppose that by the time the earth
reaches the last rectal portion of the gut there is hardly any
nutriment left in it for absorption, and hence we have the
absence of the typhlosole as well as of internal blood-plexus
in this region, both of these structures being the likely media
for absorption of nutriment from the earth. A well-developed
external network of capillaries is, however, present in the

TEXT-FIG. 3.

ya TTY
\ir \\\

; iM iy A mile i ah A

an a wo jy

Ist region 2nd region xX dv. 3rd region

Semi-diagrammatic representation of the intestinal blood-plexus
in the three regions of the intestine. The Ist region extends from
the fourteenth to the twenty-sixth segment; the 2nd region
from the twenty-sixth to tw enty-three to twenty-six segments in
front of the anus and the region includes the last tw enty -three
to twenty-six segments (ree ‘tal region). d.v.=dorsal vessel ;

=the place where there is a change from the regular geometrical
Geiss to the branching tree-like plexus of the rectum.

rectal region and serves to supply blood to the wall of the gut,
and also, being distributed amongst the chloragogen cells,
allows the latter to take up the excretory products from the
blood capillaries.

In the second region, which is the most extensive (twenty-
sixth segment to twenty-three to twenty-eight segments in
front of the anus) of the three regions, we have a typhlosole
as well as both the internal and the external plexus equally

360 KARM NARAYAN BAHL

well developed. The internal plexus is a dense network of
capillaries appearing as a sort of blood-sinus interrupted at
places by the foldings of the gut epithelium (fig. 2). The
typhlosolar vessel, which should be regarded as part of the
internal plexus, communicates with it at two places in each
segment. ‘The external blood-plexus, which is not continuous
from segment to segment, has capillaries of varying diameters.
The blood apparently passes from the external to the internal
plexus, as, like the case in Megascolex (1), we can see the
capillaries of the external network communicating with the
capillaries of the internal network at numerous places in
sections.

In the first region we have only a well-developed internal
plexus but no external one. Neither is there a typhlosole,
although, of the specially large mid-dorsal and mid-ventral
capillaries, the mid-dorsal one simulates the typhlosolar
vessel.

(1) Alimentary plexus in the first region (fourteenth to
twenty-sixth segment).

In this region of the gut the internal blood-plexus is best
developed. The network is very dense, almost a blood-sinus
interrupted at certain places; the interspaces in the dorsal
half of the plexus are very small indeed, even less than one-
fourth the size of the vessels which surround them. The
capillaries run parallel to one another transversely to the
length of the gut, and towards the ventral half break up into
capillaries of smaller calibre, so that in the ventral half of the
gut a continuous blood-sinus gives place to a coarse network
of capillaries. In a freshly-opened worm this region of the
gut presents a very bloody appearance.

Besides the richness in capillaries of this region we have
a pair of well-marked vessels lying on the dorso-lateral aspect.
These begin ventrally in the intestinal plexus about the four-
teenth segment, and incline gradually dorsalwards up to the
twenty-sixth segment, where they join the posterior pair of
dorso-intestinal vessels of that segment at the inner angles of
the roots of the intestinal coeca, and also communicate at that

VASCULAR SYSTEM OF PHERETIMA 361

place with the other blood-vessels on the walls of the coeea
themselves.

An equally well-developed vessel runs along the mid-dorsal
line of the gut, being only a specialized capillary of the internal
plexus and being also continuous with the typhlosolar vessel
behind.

The external blood-plexus is almost completely absent in
this region. ‘There are, however, a few capillaries present,
which ean be seen attached to the outside of the gut; for
example, at places where the ventro-intestinal and septo-
intestinal vessels join the wall of the intestine. But they soon
enter the intestinal wall and pour their blood into the internal
blood-plexus ; so that a regular external plexus such as we find
in the second and third regions (vide infra) is absent in this
part of the gut, the internal plexus being very strongly developed.

(2) The alimentary plexus in the second region (twenty-sixth
segment to twenty-three to twenty-eight segments in front
of the anus).

In this region we have both the external and internal plexuses
well developed. The external plexus consists of capillaries of
various sizes which are continuous on the ventral wall of the
gut but not on the dorsal. They are connected with the septo-
intestinals and the ventro-intestinals which apparently form
their source of blood-supply. They open into the capillaries
of the internal plexus as shown in fig. 2.

The internal plexus in this region of the gut presents a very
regular geometrical arrangement, as shown in fig. 3. This
network consists of (a2) Longitudinal capillaries, which
are very Closely set around the wall of the gut, extending all
along its length. They are continuous from segment to seg-
ment and number about forty all round. ‘These capillaries
form the main portion of the plexus and in transverse sections
are seen to lie in the folds of internal gut-epithehum.

(b) Transverse Channels.—We have already men-
tioned that in each segment the dorsal vessel is connected with
the gut by means of two pairs of dorso-intestinal vessels.
These dorso-intestinals on leaving the dorsal vessel enter the

362 KARM NARAYAN BAHL

intestinal wall about }mm. from their origin and go round
the wall of the gut to its ventral side. I propose to apply the
term dorso-intestinal to the vessel from its point of
origin from the dorsal to the poimt of its entrance into the
intestinal wall. The continuation of the dorso-intestinal
on the wall of the gut I propose to call a transverse
channel. Corresponding to the two pairs of dorso-intes-
tinals there are two pairs of transverse channels in each
segment ; each of these transverse channels is joined at its
point of junction with the dorso-intestinal by a branch from
the typhlosolar vessel (vide infra) (fig. 2, left half): so that
these transverse channels serve to connect not only the longitu-
dinal capillaries with each other but also the whole plexus
with the typhlosolar vessel.

(c) Oblique Channels.—These begin at the mid-ventral
line of the intestine at the intersegmental plane and run
forwards and dorsalwards, passing through three segments
before reaching the mid-dorsal line, where they join the
typhlosolar just in front of the septa (fig. 3).

(d) Typhlosolar Vessel.—The typhlosolar vessel runs
along the free edge of the typhlosole all down the second region
of the gut (fig. 2). The typhlosole itself cannot be compared
to the structure of the same name in Lumbricus, for in
Pheretima it is really a bigger fold of the gut-epithelium
containing not yellow cells, like those which fill up the typhlo-
sole of Lumbricus, but only connective tissue which has
the same staining qualities as the connective-tissue matrix
in the layer of circular muscle-fibres of the body-wall. The
typhlosolar vessel does not seem to possess a definite wall like
the capillaries of the external plexus in Pheretima or
the typhlosolar vessel of Lumbricus, but is only a part
of the blood-sinus like the longitudinal capillaries, being, like
them, in communication with the two pairs of transverse
channels in each segment. We can therefore think of these
transverse channels as circular ring-vessels which collect blood

* T have called these channels as they are thicker than the longitudinal
capillaries,

VASCULAR SYSTEM OF PHERETIMA 3638

from the longitudinal capillaries and the typhlosolar vessel
(which we may regard as a specialized longitudinal capillary
lying in the mid-dorsal line), and convey it to the dorsal vessel
by means of the two pairs of dorso-intestinals in the same
way as the ring-vessels of the oesophagus convey its blood to
the supra-oesophageal vessel there (vide infra). It would
be interesting to note here that, although the typhlosole is
absent in the segments fourteen to twenty-six, there is
a prominent blood-vessel in the mid-dorsal line of the gut-
epithelium, the vessel corresponding to the typhlosolar behind,
with which it is directly continuous.

(3) The blood-plexus in the third region (last twenty-three
to twenty-eight segments).

In the last twenty-three to twenty-eight segments of the
worm where the typhlosole in the gut is absent, and which
region Beddard (8, p. 18) has referred to as the ‘ rectum’,
the intestinal plexus is different from what we have seen in the
first two regions. The whole of the plexus is external, i.e. lies
outside the muscular coats, there being no internal plexus.
The regular and rectangular arrangement of capillaries in the
typhlosolar (second) region at once changes into a branching
tree-like plexus as shown in fig. 3. There is only one pair of
dorso-intestinals in this rectal region in place of two pairs
in the first two regions. Since there is no internal plexus
the dorso-intestinals change their connexions and communicate
in this region with the external blood-plexus.

The blood coming to the rectum from the ventro-intestinals
and septo-intestinals goes to the external plexus, from where
it passes to the dorsal through the dorso-intestinals, the part
of the course involving the internal plexus having been cut
out (vide infra).

(c) The Commissural, Integumentary, and
Nephridial Vessels.
1. The Commissural Vessel.—As already mentioned,
there is a pair of commissural vessels (parietal vessels) in each
segment connecting the dorsal with the subneural vessel

364 KARM NARAYAN BAHL

(figs. 1 and 2). The commissural lies in the most anterior
position in each segment, since the posterior face of a septum,
on which this vessel hes, forms the anterior boundary of a seg-
ment. In its ventro-lateral part each commissural vessel is
joined by a ‘ septo-intestinal’ branch (figs. 1 and 2) which
puts the commissural vessel in communication with the
intestinal plexus, so that the commissural joins the dorsal
and subneural vessels at its two ends, while in its ventral third
it gives the septo-intestinal branch to the imtestinal blood-
plexus. It is interesting to note the Y-shaped places of junc-
tion (fig. 2) one comes across in sections, where the three limbs
of the Y represent the branches of the commissural going to
the dorsal and subneural vessels and the intestinal plexus
respectively. All along its length the commissural vessel is
joined by branches coming from the septal nephridia and the
body-wall. In segments sixteen to twenty-one the com-
missural vessel also receives the efferent capillaries from the
prostates which get their blood-supply from the branches
of the ventro-tegumentaries. As shown in fig. 1, I could count
in one preparation as many as eight branches entering the
commissural, each of these branches being formed by the
union of several branchlets.

The commissural vessel of Pheretima is a very interesting
structure when we compare it with similar structures in other
earthworms. Bourne (1) describes in Megascolex two
vessels, which he calls ‘ intestino-tegumentary ’ and ‘ dorso-
tegumentary ’, as follows: ‘ The main portion of the intestino-
tegumentary vessel lies closely adherent to the body-wall just
behind a septum, i.e. in the anterior portion of a segment ’,
and ‘the dorso-tegumentary arises in all segments regularly
from the dorsal vessel immediately posterior to the septum
which forms the anterior boundary of the segment in which
it lies’. It is clear from this description and also from his
diagram (Pl. [X, fig. 7, in his paper) that these two vessels
of Megascolex run in the same transverse plane, and would
thus correspond exactly to the commissural vessel of Phere-
tima minus its small ventral portion, since the commissural

VASCULAR SYSTEM OF PHERETIMA 365

also lies in exactly the same position. Its dorsal part with its
connexions with both the dorsal vessel and the body-wall
would correspond to the ‘ dorso-tegumentary ’, and its lateral
part together with the septo-intestinal having connexions
with the body-wall on the one hand and the intestinal plexus
on the other would correspond to the * intestino-tegumentary ’
of Megascolex. ‘There being no subneural vessel in the
latter genus, there is nothing in its blood-system corresponding
to the ventral part of the commissural of Pheretima.

Again, the ‘ dorso-tegumentary’ of Moniligaster (2)
and Lumbricus (8) corresponds to the commissural vessel
of Pheretima minus the septo-intestinal. Unhke Mega-
scolex, these two genera (Moniligaster and Lumbricus)
possess a subneural vessel like Pheretima, and we have
a loop or commissural vessel connecting the dorsal with the
subneural, which has been described by Jaquet (9) in Lum-
bricus as the ‘ branche dorso-sous-nervienne ’, a term adopted
by Bourne for the same structure in Moniligaster. Jaquet
also describes a ‘branche tégumentaire’ from the dorso-
tegumentary ; but I have examined the tegumentary (com-
missural or parietal) of Lumbricus and do not find
a special ‘ branche tégumentaire’ as Jaquet makes out. Of
course, there are several branches from the body-wall (tegu-
mentary branches) joining the commissural all along its
course as in Pheretima, to which the term ‘ branche
tégumentaire’ can be applied; but the real point in which
the commissural of Lumbricus and Moniligaster
differs from that of Pheretima is that in the former two
genera it has no connexion with the intestinal plexus, there
being nothing corresponding to the ‘septo-intestinal’ of
Pheretima.

From the comparisons made above it seems reasonable to
deduce that the commissural vessel of Pheretima is a com-
pound vessel which combines in itself the ‘ dorso-tegumentary ’
(commissural or parietal) of Lumbricus and Moniligaster
(the dorso-tegumentary of Megascolex corresponding only
to one of the tegumentary branches joining the commissural

366 KARM NARAYAN BAHL

in the other earthworms) and the * intestino-tegumentary ° of
Megascolex. The probable homologies are set out in the
following table :

1. Lumbricus Branche tégumen- | Branche dorso-sous-| Absent
taire nervienne

2. Moniligaster 5 :. rf =~ is

3. Megascolex Dorso - tegumen -| Only partially re- Intestinal
tary presented by the| part of

tegumentary part) intestino-

| of the ‘ intestino- tegumen-
tegumentary ° tary”

¢, Pheretima One of the capil- Commissural ves- Septo-intes-
laries from the! sel tinal,
body-wall joining
the dorsal por-
tion of the com-
missural

In deseribing the ‘ventro-intestinals’, of which there is
a pair in each segment in Moniligaster (2, 1894, p. 330),
Bourne remarks: ‘ They are the sole afferent vessels of the
intestinal walls. There are no such vessels in Megascolex
coeruleus, their function being performed by the “ intes-
tino-tegumentary ”’ vessels... In Pheretima we have both
the ‘intestino-tegumentary (represented by the septo-
intestinal) as well as the ventro-intestinal vessel in each
segment ; and if both are afferent vessels of the gut-wall, as
I believe they are, there is a double source of supply of blood
to the gut in Pheretima.

As I shall diseuss later on, I believe that the course of blood
in the commissural is towards the dorsal vessel. The blood
from the subneural goes to the intestinal plexus through the
septo-intestinal, and the branches joining the commissural all
along its course bring blood into it from the body-wall and the
septal and integumentary nephridia.

2. The Integumentary Vessels.—The body-wall,
consisting of its muscular layers, and the epidermis receives its
supply of blood from the ventro-tegumentary branches, a pair
of which comes off from the ventral vessel in each segment.
I have already stated that these ventro-tegumentary branches

VASCULAR SYSTEM OF PHERETIMA 367

supply the body-wall of the segment succeeding the one in
which they arise from the ventral vessel (e.g. the ventro-
tegumentary arising from the ventral in the fortieth segment
runs along and supplies the body-wall of the forty-first segment
and so on). The ventro-tegumentaries give off numerous
branches backwards and forwards (fig. 1), which are distributed
over the body-wall and also supply blood to the integumentary
nephridia (vide infra). The ventro-tegumentaries grow
thinner and thinner along their course towards the mid-dorsal
line near which they end in the body-wall.

TREXT-FIG, 4.

comm.v.b.

cirm. ep. long. m.

A diagrammatic reconstruction of three serial sections showing
the close parallelism of * arterial’ and * venous’ capillaries in the
body-wall. ep.=epidermis ; cir,m.=layer of circular muscle-
fibres ; long.m.=layer of longitudinal muscle-fibres ; vt.b,=
a branch of the ventro-tegumentary vessel ; comm.v.b,=a branch
of the commissural vessel,

The efferent vessels of the body-wall are the paired branches
of the subneural in each segment and the numerous branches
joining the commissural vessel in each segment.

The afferent and efferent capillaries run side by side in the
substance of the body-wall, and can always be followed from
the coelomic epithelium through the muscular layers to the
epidermis. I can confirm for Pheretiia Bourne’s state-
meni (2) with regard to the peripheral capillaries in Monili-
gaster, that ‘the most striking feature of these networks
(he is speaking of capillaries in the body-wall)

368 KARM NARAYAN BAHL

is the strict parallelism which obtains throughout between
“artery? and‘ vein’’’. In serial sections it is very interesting
to follow pairs of parallel capillaries in the body-wall, and one
can invariably trace them to their afferent and efferent vessels.
Fig. 4, reconstructed from three sections of 6 thickness,
serves to illustrate the parallelism obtained in sections, while
fig. 4a gives an accurate camera lucida drawing of part
of the body-wall mounted flat after the removal of longitudinal
muscles. The strict parallelism between an ‘artery’ and
a vein together with the capillary loops connecting them are
very clearly displayed.

3. The Nephridial Blood-system.—tThe blood-supply
of the three kinds of nephridia in Pheretima has already
been described by me elsewhere (8), and I have nothing further
to add here.

(d) The Dorso-intestinals and the Ventro-
intestinals.

The Dorso-intestinals.—l have referred to these
vessels already in describing the dorsal vessel. The dorso-
intestinals form, so to speak, the efferert vessels (veins) of
the intestinal blood-plexus, as all the blood in the intestine
is returned to the dorsal vessel through these dorso-intestinals.
There is a single pair of them in the fourteenth segment and in
all the segments of the rectal (post-typhlosolar) region, while
in the remaining large part of the intestine we have two pairs
to each segment. We have already noted that the dorso-
intestinals communicate with the external plexus in the rectal
region but with the internal plexus in the first and second
regions. At the place where the dorso-intestinal leaves the
cut, it also receives a branch from the typhlosolar vessel
(fig. 2).

The Ventro-intestinals.—These single unpaired
vessels in each segment have also been referred to above.
They form the afferent vessels (arteries) of the gut, and are
present in all the three regions.

VASCULAR SYSTEM OF PHERETIMA 369

8, THe BLoop-SystEM IN THE First FourtTEEN SEGMENTS.

In the first fourteen segments the blood-system is highly
modified on account of the cephalization of this region, and
differs a good deal from the system in the general body-region.

Amongst the longitudinal trunks the subneural as such is

TEXT-FIG. 4 A.

eff.v.c. aff.v.c.

int.Lc.

aff.v.c.
eff.v.c.

Disposition of blood-capillaries in the body-wall from a whole
mount of a portion of the body-wall treated with caustic potash,
showing how a ‘ venous’ capillary passes into an ‘ arterial’ one.
aff.v.c.=capillary of the afferent vessel ; eff.v.c.=capillary of the
efferent vessel; int.l.c.=capillary loop connecting the afferent
and efferent vessels,

370 KARM NARAYAN BAHL

absent ; it bifureates in the fourteenth segment, and the two
branches curve round (fig. 5) the nerve-cord to be continued
into the two lateral oesophageal vessels. A new large vessel
in this region limited in extent is the supra-intestinal vessel,
which is closely attached to the oesophagus in the mid-dorsal
lme and communicates freely with the blood-plexus of the
oesophagus. Besides these there are the big pulsating ‘ hearts ’
in many of the segments of this region, by means of which the
dorsal vessel pumps out all the blood it receives either into the
ventral vessel to be distributed by it or directly to the various
organs in this part of the body.

(a) The Longitudinal Trunks.

1. The Dorsal Vessel.—The dorsal vessel continues
in front up to the third segment, where it divides into three
branches near the cerebral ganglion, these branches being
distributed over the pharyngeal mass and the wall of the
buceal cavity. While in the region of the intestine the dorsal
vessel lies close upon the gut, being connected with it by two
pairs of dorso-intestinals ; in this anterior region it is removed
considerably away from the oesophagus. Except in the four-
teenth segment, where the dorsal vessel is connected by a single
(not two) pair of dorso-intestinals, there are no such venous
branches at all in the anterior cephalized region. Since there
is no subneural vessel in this region the commissural vessels
connecting the dorsal with the subneural in the intestinal
region are absent in this anterior region. However, the dorsal
vessel here gives off, in many segments, pulsatile vessels called
the ‘hearts’. ‘These structures I shall describe separately below.

The intersegmental valves present in the posterior part of
the dorsal vessel are present here also, and have the same
structure and disposition, making the blood flow in the anterior
direction. But the valves at the orifices of the dorso-intestinals
and commissurals into the dorsal (vide infra) in the posterior
region have no counterpart here ; in their place there are other
valves away from these orifices, leading the blood outwards
from the dorsal vessel.

371

VASCULAR SYSTEM OF PHERETIMA

*, Sdooy dorzUe | 94} UreUOD syUoWZes YQUaAa[a pur YWUAI AY} ayry Mm
‘S}USWUB98 YUU PUL YZUOAVS oY} UT sjAvoY , [VIO , oY) ‘syUOUTZOS YQU9e}IYY PUR YAO} oY} UT ore syavaY
, [(RU{SOZUT-O107R] , OYJ, ‘Tossed ArvjUOUINS}-01},U0A ="a"70 £ [SSA [RIJUOA="2"'A STOSSAA [RINoUqns ="a°us
:xudivyd ="yd ‘ snseydoseo ="sa0 ‘[assea [voseydosoo [v.19yR] ="0"990 "7D] 2 OUTYSOYUL = "yw $ paezzIB="z0h {asso
[PSIOp="A'p “RUTZALIY JO SJUOUTs W99}XIS 4SAY VY UT STassaa-poorq Jo UOIZLIUdSaIdor O1VRULUTRADLIp-Tutes W

A’Sa0 Je

WD) yy »))) f)

cl =) . “i
nates
, ' el

bl

Ww

[BimI\. ets \
| \

Ce

NO, 259

872 KARM NARAYAN BAHL

2. The Supra-intestinal Vessel.—The supra-intes-
tinal vessel, which is confined to the oesophageal region
behind the gizzard, occupies the same relative position with
regard to the gut as the dorsal vessel does in the region of the
intestine. It lies beneath the dorsal vessel rather closely
attached to the dorsal wall of the oesophagus, while the dorsal
vessel itself is removed considerably away from the gut. It is
usually double along its whole extent, but the two halves come
together and communicate with each other at several places.
The supra-intestinal vessel extends from the tenth to the
thirteenth segment. In the tenth and eleventh segments it
communicates with the lateral oesophageal vessels by large
commissural vessels or ‘loops’ that go round free from the
wall of the oesophagus ; while in the twelfth and thirteenth
segments it communicates with the ventral vessel through the
‘hearts. The vessel ends anteriorly by breaking up into
capillaries in front of the tenth segment, and these capillaries
are distributed over the walls of the oesophagus and the
gizzard. Posteriorly the vessel ends by joining the posterior
pair of * hearts’ in the thirteenth segment, although a slender
branch very often continues backwards on the mid-dorsal line
of the gut for a segment or two.

The supra-intestinal is the efferent vessel for the gizzard
and the oesophagus, and all the blood brought in it from
these structures is no doubt carried into the ‘ hearts’ of the
twelfth and thirteenth segments.

3. The Ventral Vessel.—The ventral vessel extends
anteriorly up to the second segment, and in each segment
gives off a pair of ventro-tegumentary branches as in the
posterior region, with the difference that the branches from
a particular segment are spread over and distribute blood to
the body-wall, the septa, and the nephridia in the same
segment and not the succeeding one, as they do behind.
All the special organs in this part of the body, e.g. the sperma-
thecae, the seminal vesicles, the ovaries, and the oviducts
are supplied with blood by little branches from the ventro-
fegumentaries. The vessel ends anteriorly in a pair of branches

VASCULAR SYSTEM OF PHERETIMA 373

in the second segment. There are no ventro-intestinals in this
region of the body.

4. Lateral-oesophageal Vessels.—These are a pair
of fairly large vessels in the first fourteen segments of the animal
situated on the ventro-lateral aspect of the oesophagus. They
are always found full of blood and can be easily seen. Behind
the gizzard, i.e. in segments ten to thirteen, they are very
intimately attached to the wall of the oesophagus and, as can
be seen in sections, communicate with the oesophageal ring-
vessels throughout these four segments by as many branches
as the number of ring-vessels. In the region of the gizzard
and in front, however, they are free from the wall of the gut,
but receive a branch in each segment from the wall of the gut.

The lateral oesophageals receive in each segment a pair of
branches that bring back blood not only from the body-wall
and septa of this region but also from the seminal vesicles and
the spermathecae. They thus function here like the branches
of the subneural and commissural vessels behind, which collect
blood from the body-wall, the nephridia and other organs in
coelom like the prostates.

It only remains to be added that the lateral oesophageals are
a continuation forward of the subneural vessel. In the four-
teenth segment the subneural vessel forks into two, and each
of the two branches loops round the nerve-cord and comes to
lie dorsal to it and is continued forward along the ventro-
lateral aspect of the oesophagus as the lateral-oesophageal
vessels.

(b) The ‘Hearts’ and the Anterior Loops.

Tt will be seen from what we have described above that there
is no direct communication between the dorsal and ventral
vessels in the region of the body behind the thirteenth segment,
but in the anterior thirteen segments the dorsal vessel com-
municates directly with the ventral through the * hearts’ in
the seventh and ninth, and twelfth and thirteenth segments.
It is only these four pairs of ‘ hearts’ that are connected with
the ventral vessel ; but, besides these, there are other ‘ hearts ’

G/G12

374 KARM NARAYAN BAHL

which are also pulsatile but supply blood to some of the organs
directly, e.g. the gizzard and the pharyngeal nephridia. I have
adopted Bourne’s suggestion (1, p. 64 n.) of naming all rhythmi-
cally contractile, circularly disposed vessels as ‘hearts’, which
term thus includes even the anterior branches of the dorsal
vessel which do not join the ventral vessel.

TEXT-FIG. 6,

dv.
7

SU.LV.

n.c

is

A diagrammatic transverse section of the earthworm through the
region of the ‘ latero-intestinal ’ hearts. In the right half is shown
the intersegmental septum just behind the ‘heart’. d.o.=
dorsal vessel; /¢.=latero-intestinal heart ; ¢.s.=intersegmental
septum ; dnt.v.=integumentary vessels taking blood (venous) to
the lateral oesophageals and the supra-intestinals ; /at.oes.v.=
latero-oesophageal vessels ; 7.v.=a ring-vessel in the oesophagus ;
su.i,v.=supra-intestinal vessel ; v.v.= ventral vessel.

Again, Bourne (1, p. 64n.), following Perrier, distinguishes
‘lateral hearts ’ from the ‘ intestinal hearts ’ according as they
are connected dorsally with the dorsal or supra-intestinal
vessels. The ‘ hearts’ in the twelfth and thirteenth segments
in Pheretima communicate dorsally with both the dorsal
and supra-intestinal vessels and are therefore * latero-intestinal ’
hearts, while the ‘ hearts’ in the seventh and ninth segments
belong to the category of ‘lateral hearts’. Coming to the
‘loops’ of the tenth and eleventh segments, we find that they

VASCULAR SYSTEM OF PHERETIMA 375

communicate dorsally with the supra-intestinal vessel, while
ventrally they are connected with the lateral-oesophageal
vessels. They might have been called ‘ intestinal hearts’ but
for the fact that these ‘loops’ do not pulsate, have non-
muscular walls unlike those of the ‘ hearts’, and I believe that
the flow of blood in them is from the lateral oesophageals to
the supra-intestinal, a fact which I refer to agam below. On
these considerations | exclude these vessels from the category
of ‘hearts’ and call them ‘anterior loops’, since they have
nothing in common with the so-called ‘ hearts’ and ‘ anterior
loops ’ in greater detail below ; they are shown in fig. 5.

Thirteenth and Twelfth Segments.—In each of
these two segments there is a pair of * latero-intestinal ’ hearts.
In systematic accounts of the genus Pheretima if is only
these two pairs that are described, and no mention is made
of the anterior pairs of ‘ hearts’. Even if the term ‘ hearts ’
be restricted to those commissures which communicate with
the ventral vessel below it should include the ‘ hearts’ of the
seventh and ninth segments. ‘This diagnostic character for
the genus Pheretima is thus generally erroneously described,
and the genus should be recognized to possess at least four
pairs of ‘ hearts ’, two ‘ lateral’ and two ‘ latero-intestinals ’.

The ‘ hearts’ of the twelfth and thirteenth segments (fig. 5)
are situated in the posterior parts of these two segments, and
their walls are intimately attached to the septa behind them.
They have thick muscular walls and a spacious cavity, and at
their dorsal ends communicate anteriorly with the supra-
intestinal and posteriorly with the dorsal vessel. At the
places where the branches from the dorsal and supra-intestinal
meet to enter the ‘heart’, each has a pair of valves leading
to the ‘ heart’, and similarly there is a pair of valves at the
ventral end of each ‘heart’ just above the place where it
joins the ventral vessel (fig. 11). ‘The dorsal valves prevent
the blood from going back to the dorsal or supra-intestinal
vessels during systole, while the ventral valves prevent the
blood from entering the ‘heart’ from the ventral vessel
during diastole.

376 KARM NARAYAN BAHL

Eleventh and Tenth Segments.—These two seg-
ments contain no ‘hearts’, but each of them has a pair of
commissural vessels connecting the supra-intestinal with the
lateral oesophageal of each side. These vessels lie in the
posterior parts of these segments near their posterior septa,
and are partially covered by the latter. Unlike the ‘ hearts’
these ‘loops’ of the tenth and eleventh segments are thin-
walled, their walls being non-muscular, and they have no valves
anywhere along their length.

The blood, by means of these ‘loops’, flows from the lateral-
oesophageals into the supra-intestinals. The latter collect
blood from the gizzard and oesophagus and also receive blood
in these two segments directly from the lateral oesophageals.
All this blood they carry into the ventral vessel through the
‘hearts ’ in the twelfth and thirteenth segments.

We may note here that the lateral oesophageals in Lum-
bricus pour their blood into the dorsal vessel in the tenth
segment and into the large parietal in the twelfth.

Ninth Segment.—In the ninth segment there is a pair
of ‘lateral hearts’ connecting the dorsal with the ventral
vessel. This pair of ‘ hearts’ is generally asymmetrical, the
left ‘ heart ’ bemg large and well developed as compared with
the small thin-walled and ill-developed one of the right side,
which, however, sends a branch to the oesophagus in this
segment. The ‘heart’ on the left side has valves pointing
downwards along the greater part of its length, and there is
also a pair near the point of opening of the ‘ heart © into the
ventral vessel. There are altogether four pairs of valves and
their position and arrangement is illustrated in fig. 11 a.

Highth Segment.—In the eighth segment the dorsal
vessel gives off a pair of large thick-walled branches which
do not join the ventral vessel but on account of their contrac-
tility are still called ‘ hearts’; each of them presents a bulb-
like dilatation at some distance from its origin and immediately
forks into two (fig. 5), the posterior branch gomg to the septum
and body-wall, and the anterior dividing and distributing
blood over the wall of the gizzard in a large number of capil-

VASCULAR SYSTEM OF PHERETIMA ott

laries which run longitudinally parallel to one another. These
branches of the dorsal vessel have a series of paired valves
along their length between the point of their origin and the
place where there is the bulb-like dilatation. The bulb-like
dilatation which occurs at the distal end of all the ‘ hearts ’
contains a pair of thick valves pointing away from the dorsal
and towards the ventral vessel, as shown in fig. 11.

The blood to the gizzard, therefore, is supplied from the dorsal
vessel by the pair of branches in this segment ; while the eapil-
laries of the supra-intestinal vessel, which has its beginnings
here, collect blood from the gizzard and take it into that
vessel.

Seventh Segment.—In the seventh segment there is
a pair of ‘lateral’ hearts, each of which is joined below both
with the ventral and the lateral oesophageal vessels, which
latter are themselves jomed together by a cross channel.
In its upper part each of this pair of * hearts’ is thick-walled
and has valves leading blood outwards, but in its ventral part
each ‘heart’ is thim-walled and has also no valves in it.
There is no doubt that the blood flows from the dorsal to the
ventral vessel ; but it seems probable that the supply of blood
in the ventral vessel, which is very thin in this region and
contains little blood, is also replenished from the lateral
oesophageals, which are always large and full.

Sixth, Fifth, and Fourth Segments.—In the
sixth segment, and also in the fifth and fourth, there is a pair
of branches given off from the dorsal vessel each of which
has a pair of valves leading outwards near its origin, and
supplies blood to the masses of pharyngeal nephridia in each of
these three segments. These branches are also pulsatile and
can therefore be named * hearts ’.

Third Segment.—lIn the third segment before the dorsal
vessel breaks up anteriorly, it gives off a pair of branches
to the pharyngeal mass behind the cerebral ganglion. These
branches also possess valves near their origin which direct
the flow of blood outwards.

378 KARM NARAYAN BAHL

(c) The Blood-vessels of the Gut in the
first Fourteen Segments.

In segments ten to fourteen there are in the oesophageal
wall a series of very definite and striking transverse vessels,
about twelve pairs per segment, joining the supra-intestinal
above and the lateral oesophageals below; the breadth of
these vessels is at least equal to the intervals between them.
They are not united by longitudinal connexions and are con-
tinuous across the mid-ventral line. These ring-vessels (fig. 6)
are very characteristic of the oesophagus behind the gizzard,
and are situated inside the muscular coats of the oesophagus.
In this region both the lateral oesophageals and the supra-
intestinals are intimately attached to the oesophagus, and the
blood flows from the former into the latter through these
transverse ring-vessels, the latter receiving no supply at all
from the ventral vessel.

In the eighth and ninth segments the gizzard receives its
supply of blood from the ‘hearts’ of the eighth segment,
the branches of which divide and run along the outer wall of
the gizzard in about fourteen parallel longitudinal capillaries.
There is a second set of parallel capillaries which collect blood
from the gizzard and join the supra-intestinal vessel.

In front of the gizzard, i.e. in the first seven segments,
the pharynx and the oesophagus get their supply of blood from
the ‘ hearts’ of the dorsal vessel, and branches of the lateral
oesophageals collect blood and take it to the latter from this
part of the gut.

4, COMPARISON WITH THE BLOOD-SYSTEM OF THE
LUMBRICIDAE.

In main outline the arrangement of blood-vessels in Phere-
tima resembles that of Lumbricus and Allolobophora,
the latter more than the former. The main longitudinal
trunks—the dorsal, the ventral, and the subneural—are the
same in the three genera, but in Lumbricus there are also
in addition the two lateral neurals which are absent in the

VASCULAR SYSTEM OF PHERETIMA 379

other two genera. Moreover, while in Lumbricus and
Allolobophora the subneural goes right up to the anterior
end of the body, in Pheretima it passes into the lateral
oesophageals in the fourteenth, as it also does in Monili-
gaster (2), being absent in the first thirteen segments. The
venous branches of the dorsal vessel bringing blood into it
behind the ‘hearts’ are the ‘ dorso-intestinals’ and_ the
“comiissurals ’, The latter, while they le completely in one
segment in Pheretima, occupy two segments in Lum-
bricus and Allolobophora. In these the ventral
portions of the commissurals lie on the posterior face of
a septum in one segment, while the dorsal portions lie on the
anterior face of the same septum in the segment in front.
In this way, while the commissural vessel enters ‘the dorsal
vessel in front of a septum, if enters the subneural imme-
diately behind that septum; but in Pheretima, both the
ends of the commissural and, in fact, the whole of the com-
missural, lies on the posterior face of a septum.

The ventro-tegumentaries in all the three genera arise in the
segment anterior to the one they supply ; but while in Phere-
tima and Allolobophora the ventro-tegumentary runs
along the middle line of a segment (fig. 3), it runs very near
the anterior septum alongside the commissural in Lum-
bricus. The parallelism between an artery and vein shown
in fig. 4 in Pheretima in the body-wall is not found in
Lumbricus, in which the arterial branch lying inside the
muscular layers of the body-wall takes a dip towards the
epidermis, runs beneath this layer for a short distance, and
runs back to the muscular layers to be continued as a venous
branch to the commissural into which it enters (6).

As regards the blood-vessels in connexion with the gut we
may notice the absence of septo-intestinal vessels in the
Lumbricidae, whereas in Pheretima the gut has
a double source of blood-supply (the ventro-intestinals and the
septo-intestinals) ; in the other two genera it gets all its blood
from the ventral vessel only. ‘The typhlosopar vessel of
Pheretima, unlike that of the Lumbricidae, is only

380 KARM NARAYAN BAHL

a specially developed mid-dorsal portion of the gut-plexus,
and has no definite walls of its own, nor does it communicate
directly with the dorsal vessel as 1t does in Lumbriecus.

In the anterior cephalized region of the body besides the
differences in the number and position of the ‘ hearts’, there
is the presence in Pheretima of an additional * supra-
intestinal vessel * which receives all the blood from the lateral
oesophageals and pours it into the ‘hearts’; while in the
other two genera, the blood from the lateral oesophageals
goes directly to ‘hearts’, and there is no ‘ supra-intestinal ’
vessel.

5. Tur Course oF THE CIRCULATION OF THE BLoopD.

All observers are agreed upon the fact that the blood-current
in the dorsal vessel has a forward direction. I have already
stated that just in front of each septal plane, where the dorsal
vessel is very much constricted and has the narrowest lumen,
there are forwardly-directed valves which, when the vessel
contracts, prevent the flow of blood backwards. These inter-
segmental valves, as we may call them, form an incom-
plete cireular ridge on the internal wall of the vessel at their
point of origin; but it can easily be seen that the valves
consist of two large dorso-lateral valves, while there are
small dorsal and ventral ones (figs. 7 and 10). These valves
are more or less continuous with one another, so that we can
regard them as constituting one valve with small dorsal and
ventral lobes and large lateral lobes. The large dorso-lateral
lobes project forwards into the lumen of the vessel for some
distance, and are seen as two masses lying free in the dorsal
vessel in transverse sections. lig. 10 (a, b, and c) shows the
disposition of this intersegmental valve in serial sections.
In Lumbricus, on the other hand, there are two large
lateral valves, as shown by Johnstone (9), in the same position
and having the same function.

The dorsal vessel receives two pairs of dorso-intestinals
and one pair of commissurals (‘ parietals’ or ‘ dorso-sous-
neryiens’) in each segment behind the fourteenth. The

VASCULAR SYSTEM OF PHERETIMA 381

question is, what is the course of blood in these two kinds of
vessels? Does the blood come into the dorsal from both or
from only one ? According to Bourne (1, p. 74) and Vejdovsky
(11, p. 115), the blood flows from the intestimal capillaries into
the dorsal vessel through the dorso-intestinals, and in this
I agree with them. In recently-killed worms I have cut these
dorso-intestinals to see from which of the cut ends the blood
flows, and I have invariably found blood oozing out from the
side of the intestinal capillaries. Moreover, the arrangement
of valves which I refer to later confirms this view. With
regard to the course of blood in the commissural vessel (‘ dorso-
tegumentary ’ of Bourne in Moniligaster), | believe with
Perrier (as quoted by Bourne in 1) and Benham (1, p. 255)
that blood enters the dorsal vessel from these commussurals.
Bourne (1, p. 75), however, believes that blood leaves the dorsal
vessel by the dorso-tegumentaries. But later on in his paper on
Moniligaster, after discussing the pomt m an elaborate
manner (2, p. 335) and concluding that Benham’s view is
incorrect and that blood flows outwards from the dorsal by
the dorso-tegumentaries, he adds (2, p. 336), ‘ the peripheral
capillaries in the region of the body behind the hearts are
also supplied, to an extent which probably varies from time
to time and is, I expect, never very great, from the dorsal
vessel by means of the dorso-tegumentary vessels.’ Further
on in the same paper (p. 850), while generalizing on the vascular
system of earthworms, Bourne refers again to the course of
blood in the dorso-tegumentaries (commissurals) and says,
‘T have again and again returned to the course taken by the
blood in these vessels (dorso-tegumentaries). I cannot help
thinking that primitively they are efferent vessels, and that
both they and the dorso-intestinal vessels
bring blood to the dorsal vessel. In this case
they can only have, in worms otherwise well provided with
a venous system, the function suggested above for Monili-
gaster grandis of regulating the pressure in the peripheral
capillaries, and have practically no flow in them in one direc-
tion or the other.’ Bourne here seems to give away his case

382 KARM NARAYAN BAHL

for the course of blood in the dorso-tegumentaries, and I am
convinced that his statement with regard to the primitive
condition that I have quoted above holds for adult Phere-
tima, and in fact all earthworms. Both by a study of
the disposition of valves, and by cutting the commissurals
and observing from which of the cut ends the blood flows,
I am convinced that blood flows into the dorsal vessel from the
commissural vessels as it does in the case of the dorso-intestinals.
In fact I believe that the dorsal vessel all along the body of the
worm behind the first thirteen cephalized segments is a channel
only for collecting blood and propelling it forwards. It gives
out no blood at all behind the thirteenth segment as it receives
none in the first thirteen segments; so that we have two
clearly marked divisions of the dorsal vessel—the large pos-
terior division of it behind the * hearts’ being the collecting
channel, and the anterior short division of the first thirteen
segments being the channel for distribution of all the blood
collected behind.

As regards the disposition of the valves situated at the
entrance of the dorso-intestinals and the commissurals into
the dorsal, they are easily seen in transverse sections projecting
into the lumen of the dorsal vessel. In two lucky preparations
of the dorsal vessel, in which the latter was torn open and fixed
with the valves projecting out into the open lumen, I have
been able to see the valves displayed in an admirable manner.
They are shown in fig. 7. The valves are seen in two condi-
tions, i.e. either protruding inwards into the lumen of the dorsal
vessel or flush with the wall of the vessel. In the former
condition they are more or less conical in shape, the blunt
apex of the cone forming the projecting end into the dorsal
vessel, and the base being continuous with the wall of the
vessel; in the latter condition there is nothing projecting
into the lumen of the dorsal vessel, and the valves look like
closed sphincter muscles in the wall of the vessel, the actual
valves being contained in the upper ends of the dorso-intestinals
or commissurals. There can be no doubt that these two
conditions of the valves represent them as they are during the

VASCULAR SYSTEM OF PHERETIMA 388

diastole and systole of the dorsal vessel projecting inwards
when the dorsal vessel is fillmg and the blood is coming in
through both the dorso-intestinals and the commissurals, and
lying flush with the wall with the apertures closed when the
dorsal vessel contracts.

Bourne (2, p. 384) says, ‘In Moniligaster asin Mega-
secolex, while there are valves which would mechanically

Ne We Iie. 8. Fia. 10.

Fra. 9.

Text-Fig. 7.—Portion of the dorsal vessel cut open along its median
dorsal line showing the valves in its lumen. v.=the valves at the
intersegmental septa ; v’.=valves at the entrance of the dorso-
intestinals into the dorsal; »”.=valves at the entrance of the
commissural vessels into the dorsal.

Text-Fig. 8.—Section of the dorsal vessel passing through the region
where the dorso-intestinals enter the dorsal vessel showing the
valves at the entrance. d.v.=dorsal vessel; di.v.=dorso-intestinal
vessel.

Text-Fig. 9.—Section of the dorsal vessel showing the valves at the
entrance of the commissural vessels into the dorsal. d.v.= dorsal
vessel ; comm.v.=commissural vessel.

_ Text-Fig. 10.—Three sections of the dorsal vessel showing the inter-
segmental valves. a.=about the place of origin of the valve ;
b.=a little in front ; ¢.=still further forward.

prevent blood flowing into the dorso-intestinal vessel from the
dorsal vessel, there are no such valves where the dorso-
tegumentary vessels join the dorsal vessel. I have, however,
observed in Moniligaster and some other worms a sphincter
muscle in the wall of the dorso-tegumentary vessel close to its

384 KARM NARAYAN BAHL

origin.” As a matter of fact the valves at the point of entrance
of both the dorso-intestinals and the commissural vessels
(dorso-tegumentaries) look lke sphincter museles when they
are not in the protruding position and are flush with the wall
of the dorsal vessel. It is not unlikely that the sphineter
muscles seen in Moniligaster by Bourne are really the
valves in the closed condition, which, like those of the dorsal
vessel, have the form of circular ridges. In transverse sections
of Pheretima they are seen as small club-shaped structures,
attached to the inner wall of the commissural vessel just
where the latter narrows to join the dorsal vessel, and having
their broad ends projecting freely into the cavity of the dorsal
vessel (fig. 9). Johnstone (8 and 9) describes a similar disposi-
tion of valves in Lumbricus both in the dorso-intestinals
and the commissurals, and I have verified it from my sections
of Lumbricus. The disposition of valves and the course
of blood-flow in these two vessels are therefore similar in both
the worms (Lumbricus and Pheretima) and probably
in all earthworms.

Another fact, which confirms my view with regard to the
flow of blood into the dorsal vessel from the commissural
(dorso-tegumentary) and not vice versa, is that in dissec-
tions of the fresh worm when the flaps of body-wall are pinned
down after a mid-dorsal incision, the commissural vessels are
almost always torn off from the dorsal vessel near their point
of entrance into the latter, and the blood oozes out not from
the dorsal vessel or the portion of the commissural left attached
to it, but always from the cut end of the commissural near
the outer edge of the flaps. This shows that the direction of
hlood is towards the dorsal and not away from it. If the flow
of blood were from the dorsal to the commissurals, we should
see the dorsal emptying itself through the upper cut pieces
of the commissurals, especially since the dorsal vessel keeps
pulsating for some time after the worm is opened in the salt
solution. As a matter of fact no blood oozes out of the dorsal,
which remains full.

Moreover, leaving aside the question of valves and the

VASCULAR SYSTEM OF PHERETIMA 385

flow of blood from cut ends, I think Bourne’s view that blood
in the commissural vessel comes out of the dorsal and
flows towards the subneural is untenable even on theoretical
grounds. He is agreed on the fact that branches joining the
commissural vessel are veins bringing blood to it from the
hody-wall and the nephridia, and shows them as such in his
diagrams (Pl. 26, fig. 34, 2); but he believes that all the
blood is collected in the subneural and passes forwards along
the lateral longitudinals (lateral oesophageals) to enter the
posterior pair of ‘hearts’. Assuming for a moment that
Bourne’s view is correct (although I do not agree with it)
and that the blood from the subneural goes all the way to the
hearts, why should any part of this blood come from the
dorsal in each segment via the commissurals ? If the commis-
sural is a collecting channel for all the blood from the body-
wall and the nephridia, why should it get any blood at all
from the dorsal vessel? There is no meaning in the blood
coming from the dorsal into the subneural in each segment
and then entering the ‘ hearts’, while it could do so by going
into the ‘hearts’ straight along the dorsal vessel. It is to
obviate this difficulty that Bourne takes the view that the
commissurals have practically no flow in them in one direction
or the other and that they regulate the pressure in the peripheral
capillaries—a supposition which is easily disproved by cutting
the commissurals and seeing that blood does flow in them
towards the dorsal vessel.

As a matter of fact, so much blood leaves the dorsal vessel
anteriorly through the ‘hearts’, of which there are four in
Pheretima connected with the ventral vessel and others
supplying the organs directly, that it is difficult to conceive
on a priori grounds that any blood leaves the dorsal vessel
at all behind the thirteenth segment.

Having decided that the dorsal vessel all along the body
behind the thirteenth segment is only a channel for collection
and propulsion forwards of the blood which enters it from the
intestinal network and the commissural vessels, the rest of the
circulation in the worm becomes easy to follow.

386 KARM NARAYAN BAHL

The ventral vessel is the chief distributing channel and, so
to speak, the arterial trunk of the body. All observers are
agreed that blood flows backwards in this vessel in the region
of the body behind the ‘ hearts ’, and that the blood is distri-
huted to the body-wall and the other organs lying in the body-
cavity (nephridia (septal and integumentary), nerve-cord,
prostates, &c.) by means of the pair of ventro-tegumentaries
in each segment, and to the gut by means of a single unpaired
ventro-intestinal. Every structure in the body region im fact
gets its supply from the ventral vessel.

The subneural vessel collects blood from the ventral part
of the body-wall and the nerve-cord by means of a pair of small
branches it receives in each segment. All this blood goes into
the commissural vessels, from which part of it goes to the
intestine through the septo-intestinal and the rest to the dorsal
all along the commissural, the latter receiving the greater
part of its blood-supply from the capillaries that enter into it
from the body-wall and the nephridia all along its length.
The flow in the subneural is therefore from in front backwards.
This can be easily seen by pinching or cutting the vessel in
a nareotized worm and watching the direction of blood-flow.

It should be noted that the intestine has a double supply—
one from the ventral through the single ventro-intestinal,
and the other from the subneural through a pair of septo-
intestinals in each segment; this is what we should expect
considering the large amount of blood in the extensive network
of capillaries on the gut-wall. In Lumbricus the only
source of blood for the gut is the ventral vessel; but there
the gut receives two or more ventro-intestinal branches in
each segment, while in Pheretima, there being only
one unpaired ventro-intestinal vessel in each segment, the
amount of blood supplied to the gut from the ventral vessel
is comparatively small, and I suppose it is to supplement this
that we have blood brought to the gut by the septo-intestinals.
30th the ventro-intestinals and septo-intestinals bring blood
to the external intestinal plexus from which the blood passes
into the internal intestinal plexus. From the internal plexus

VASCULAR SYSTEM OF PHERETIMA 387

he blood finally passes into the dorsal vessel through the two
pairs of dorso-intestinals in each segment. In the posterior
region of the gut—the post-typhlosolar or the rectal region,
however, the blood brought to the external plexus passes
directly into the dorsal vessel through a single pair of dorso-
intestinals in’ each segment, which, as already mentioned,
communicate with the external plexus, the imternal plexus
being absent in this region. The course of blood in the intestinal
region can be shown diagrammatically as follows :—

Dorsal vesse
Hearts

Ventral vesse/

Ventro -tegumentaries Ventro-intestinals
Supplying supplying
Body-wall ¢ Nephridya, Intestine,

Branches trom the Branches from the
ventral body -wal/ Lody-wall & nephridia to
to the subneural. the commissural vessel.

Subneural ————» Commissural Este intestinal
lexus.

Commissural Sh eat
vesse/

Internal intestinal
plex us

Dorso-intestinals

Dorsal vesse/

It will be seen that the ventral vessel and its branches, the
ventro-tegumentaries and ventro-intestinals, form the arterial
vessels, while the subneural, the commissurals, the dorso-
intestinals, and the dorsal vessel itself are the chief veins
(using the word in an anatomical sense) in the worm. ‘The
blood in the dorsal vessel in a certain segment must go to
the ‘ hearts ’, and return by the ventral vessel into that segment

NO. 259 pd

388 KARM NARAYAN BAHL

again—so that the blood-flow is not self-sufficient in one
segment ; the blood must circulate in the whole body.

In the first thirteen segments (fig. 5) the blood-system is
different, and so is the course of blood. The dorsal vessel is
no longer a receiving channel; it has no dorso-intestinals
and commissurals opening into it and feeding it with blood—in
fact it receives no blood at all, but behaves instead as a great
arterial trunk, pumping out all the blood it has received in its
posterior region. Of course the greater part of its blood,
together with the whole of the blood in the supra-intestinal
vessel, is pumped into the ventral vessel through the two pairs
of ‘latero-intestinal hearts’ in the twelfth and thirteenth
segments. But a quantity of blood flows forwards anteriorly
and this is pumped into the ventral vessel by means of the
‘lateral hearts’ of the ninth and seventh segments, and is
supplied to the gizzard and the pharyngeal nephridia by the
‘hearts’ in the eighth and fourth, fifth and sixth segments,
until the dorsal vessel ends by branching on the pharyngeal
mass. In accordance with the change of function of the dorsal
vessel we have the change in the disposition of the valves.
In this region there are no valves projecting into the lumen
of the dorsal vessel ; on the other hand, the valves are present
in all the ‘ hearts’ at a little distance away from their origin
from the dorsal vessel. These valves point in the direction
away from the dorsal vessel, and lead the blood from the dorsal
vessel outwards, preventing any blood taking the reverse course.
There are also valves at the distal ends of the * hearts ’ (fig. 11)
which allow blood to flow out of ‘hearts’ durmg systole,
but do not let the blood come back during diastole. The dorsal
vessel is therefore a distributing channel here; most of its
blood it pumps out into the ventral vessel for distribution,
but a small quantity it distributes itself to the gizzard,. the
pharyngeal nephridia and the pharynx.

With regard to the flow of the blood in the ventral vessel,
L agree with Bourne (1, p. 77) in thinking that the blood coming
from the ‘ hearts’ flows both forwards and backwards. There
are no valves in the ventral vessel preventing blood from flowing
anteriorly, and in addition to the ‘ hearts’ of the twelfth and

VASCULAR SYSTEM OF PHERETIMA 389

thirteenth segments there are ‘hearts’ in the ninth and
seventh segments also to take blood into the ventral vessel.
T also agree with Bourne (1) when he says, ‘ All the blood which
enters the ventral vessel comes from the ‘“ hearts ”, and that
all the ventro-integumentary branches—those anterior to the
hearts’, as well as those posterior to them--are efferent
vessels. So far as the ventral vessel is concerned, they carry
blood away from it.’ The ventral vessel, therefore, here as in

Texr-rie. Il.

Semi-diagrammatic representation of ‘hearts’ in longitudinal
sections. A is one of the ‘lateral’ hearts of the ninth segment
with the valves in its lumen and a bulb-like dilatation at its
ventral end before it joins the ventral vessel. d.v.=dorsal
vessel; ht.=heart; v.—valves ; si.v.=supra-intestinal vessel.

the region of the body behind the thirteenth segment, is the
distributing vessel and supplies blood through the ventro-
tegumentaries to the body-wall, the integumentary nephridia
as well as the spermathecae and seminal vesicles, the ovaries,
and the oviducts. But it does not supply blood to the gut
as it does in the hinder region ; there are no ventro-intestinals
here, and the function of supplying blood to the gut here is
taken over partly by the dorsal vessel which supplies blood to the
gizzard in the eighth segment and the pharynx and oesophagus
in front, and partly by the lateral oesophageals. These vessels
Dd2

590 KARM NARAYAN BAHL

’

in this region of the body are the counterpart of the subneural
the commissural, and the septo-intestinals of the hinder
region, and bring blood from the periphery to the main stream
and to the gut. They receive a pair of branches bringing blood
from the body-wall, the septa, and other organs of the body,
e.g. the nephridia (pharyngeal and integumentary) and the
reproduction organs. The part of the oesophagus behind the
gizzard is supplied with blood by the lateral oesophageals
which he intimately attached along the ventro-lateral aspect
of the oesophagus. The blood from the oesophagus (ten to
thirteen) (*‘ ring-vessels ’) and the gizzard is collected by the
supra-intestinal vessel, which also receives blood directly
from the lateral oesophageals through the ‘ anterior loops’ of
the tenth and eleventh segments, and is conveyed to the hearts
in the twelfth and thirteenth segments. The course of blood
can be represented as follows :—

ee part
¥ Dorsal, 2 Dorsal vessel.

:

Pharynx Hearts — Pharyngeal Nephridia Gizzard
ant. ae (12th € 13th, (4th, ath g¢ 6th). (8th).

7th | 9th),
it vesse/
Ventro-te ee
throughout this la
Ea -wall £ septa
fee lh ee)
Latera/- a
co eae loops”
(10 - 13 segmts) of 10th ¢ Ith.
“Ting - vesse/s”

Subneural vesse/. Supra- ihe vesse/s.

ne
(12 ¢ 713th)

ee a ee es ee el eee ECC Cr ee eee ee

VASCULAR SYSTEM OF PHERETIMA 39]

6. SUMMARY.

1. The typical arrangement of the blood-system in Phere-
tima occurs in the region of the body behind the fourteenth
segment, the first fourteen segments forming the cephalized
region. The main longitudinal trunks are the same as in
Lumbricus, except that the lateral neurals are absent as
in Allolobophora. ‘The dorsal vessel receives two pairs
of dorso-intestinals and one pair of commussurals in each
segment behind the cephalized region.

2. The intestinal blood-plexus is both an external and an
internal one, and three regions can easily be distinguished.
The first is internal, and extends from the fourteenth to the
twenty-sixth segment; the second is both external and
internal, is co existent with the typhlosole, and extends over
the larger part of the gut ; and the third is only external, and
is confined to the rectal or post-typhlosolar part of the gut
(last twenty-three to twenty-six segments).

3. The commissural vessel of Pheretima is a compound
vessel, and represents both the ‘ dorso-sous-nervien’ of
Lumbricus and the intestino-tegumentary of Megascolex.
The capillaries of the integument are not like those of Lum-
bricus but like those of Moniligaster, and there is
a close ‘ parallelism ’ between an ‘ artery ’ and a‘ vein’ in the
body-wall, in which the two pass into each other through
a number of capillary loops.

4. There are four pairs of ‘ hearts ’ which connect the dorsal
with the ventral vessel, and five pairs which supply blood
directly to the various organs in the cephalized region. ‘There
are two pairs of non-contractile ‘ anterior loops’ connecting
the lateral oesophageals with the supra-intestinals, these loops
being the counterpart of the connexions of the lateral oesopha-
geals with the dorsal and the parietal in the tenth and twelfth
segments respectively of Lumbricus. The subneural
vessel is absent in the first fourteen segments, and is con-
tinuous with the lateral oesophageals of the anterior region.

5. As regards the course of circulation of the blood, the chief

392 KARM NARAYAN BAHL

fact is that the dorsal vessel is wholly ‘ venous’ behind the
‘hearts ’ and wholly ‘arterial’ in the region of the ‘ hearts’
and in front (the whole of the cephalized region). The examina-
tion of valves and experiments by cutting and pinching the
blood-vessels in Pheretima confirm the results of Johnstone
for Lumbricus as regards the course of blood in dorso-
intestinals and commissurals and make Bourne’s theory unten-
able. The ventral vessel is the arterial trunk throughout,
while the venous function of the dorsal and subneural behind
is taken up by the lateral oesophageals in the cephalized region.
The thin-walled and non-contractile ‘loops’ of the tenth and
eleventh segments must be distinguished from the thick
walled and contractile ‘ hearts’ of the other cephalized seg-
ments, the ‘loops’ being the channels for conveying blood
from the lateral oesophageals to the supra-intestinals.

List oF REFERENCES.

1. Bourne, A. G.—‘* On Megascolex coeruleus and a Theory of
the Course of the Blood in Earthworms”’, ‘ Quart. Journ. Micro.
Sci.’, vol. 32, 1891.
- —— “On Moniligaster grandis”, ibid., vol. 36, 1894.
3. Beddard, F. E.—‘‘ On the Structure of a New Genus of Oligochaeta,
and on the Presence of Anal Nephridia in Acanthodrilus ”’, ibid.,
vol. 31, 1890.

4, —— ‘A Monograph of the Oligochaeta’. Oxford, 1895.

5. Benham, W.—‘‘ Studies on Earthworms”’, ‘Quart. Journ. Micro. Sci.’,
vol. 26, 1886.

6. —-— ‘“*The Nephridium of Lumbricus and its Blood-supply”’,
ibid., vol. 36, 1891.

7. Bahl, K. N.—‘“* On a New Type of Nepbridia in Indian EKarthworms
of the genus Pheretima’”’, * Quart. Journ. Micro. Sci.’, vol. 64,
1919.

8. Harrington, N. R.—** The Calciferous Glands of Karthworms, with an
Appendix on Circulation’, “Journ. of Morphology ’, vol. 5, Supple-
ment, 1899.

9. Jaquet.—‘‘ Recherches sur le systéme vasculaire des Annélides ”’,
* Mitth. Zool. Stat. Neap.’, Bd. VI, 1885-6.

10. Johnstone, J. B., and Johnson, Sarah W.—‘* The Course of Blood-flow
in Lumbricus’”’, *‘ American Naturalist ’, vol. 36, 1902.

11. Johnstone, J. B.—** On the Blood-vessels, their Valves, and the Course
of the Blood in Lumbricus”’, © Biological Bulletin ’, vol. 5, 1903.

Xe)

a eT a ree

VASCULAR SYSTEM OF PHERETIMA 393

12. Lankester, E. Ray.—‘‘ The Anatomy of the Earthworm ’”’, ‘ Quart.
Journ. Micro. Sci.’, 1865.

13. Perrier.—‘“ Etudes sur lorganisation des lombriciens terrestres ”,
‘Arch. Zool. Exp.’, t. ix, 1881.

14. Stephenson, J.—‘* On Intestinal Respiration in Annelids: with Con-
siderations on the Origin and Evolution of the Vascular System in
that Group”. ‘ Trans. Roy. Soc. Edin.’, vol. 49, 1914.

15. Vejdovsky, F.—‘* System und Morphologie der Oligochaeten’, Prag,
1884.

eee “i “a ;
penta hy
A J: = Talent .

is ae pert creat
Teal iY ate = Okan

LA Hey A ab:
es ie
orn “y Me cs

The Development of the Ovary and Ovarian
Ege of a Mosquito, Anopheles maculi-
pennis, Meig.

By

A. J. Nicholson. M.Se. ‘B’ham.).
(From the Zoological Laboratory of the University of Birmingham.)

With Plates 17-20.

CONTENTS.
PAGE
INTRODUCTION . : : ; : , : : » 396
MATERIAL AND Maiaans : : . ey
HIBERNATING MosQurrogEs AND First Bagron Or Hed Desmedanacenn 399
FEMALE GENITAL ORGANS . : . ‘ . 401
GENERAL Lines 0F DEVELOPMENT OF Cree AND eee ; . 405
ANATOMY OF THE MatuRE Eae . ; 407
DIFFERENTIATION OF GERM-CELLS AND FIRST Pia JOD OF anos TH OF
THE EGG-FOLLICLES : : : ‘ er Al 2
SECOND PERIOD OF GROWTH OF THE eee FOLLICLES ‘ ‘ . 417
I. Branching of the Oocyte Nucleus and Segregation of Vegeta-
tive and Germinal Parts : , . 418
Il. Yolk Formation and the Nutrition of the fascete : 422
III. Discussion concerning the Oocyte Nucleus and N Hae
of the Oocyte in A. maculipennis . ; : . 425
IV. Development of the Outer Wall ; : : E . 485
V. Development of the Micropyle Apparatus . : : . 436
VI. Development of the Inner Wall . . : : : . 439
DEGENERATING EGG-FOLLICLES . ; . 440
PRESENCE OF SPOROZOA AND SAT ERTA IN rhe FOLLICLES . . 441
SUMMARY . ; ‘ P P : F : : ‘ . 442
List oF LITERATURE . : : : : ‘ ; : . 444

EXPLANATION OF PLATES . , . ‘ : * : . 446

396 A. J. NICHOLSON

INTRODUCTION.

THE examination oi mosquito ovaries was first commenced
with the idea of finding out at what period the ovaries of the
hibernating females commence to develop, so that an accurate
knowledge of the time at which the mosquito lays the first
eggs of the season might be determined. From an examination
of sections of the ovaries, it soon became evident that the
oocyte nucleus behaved in a somewhat unusual manner during
the period of yolk formation. I therefore decided to examine
this in detail and at the same time observe what might be
termed the grosser anatomy of the developing ovary and
oocyte. An immeuse amount of work has been done on the
oogenesis of insects, but most of this has been confined to the
detailed examination of the complicated nuclear changes which
take place during this period. The mosquito, however, is
peculiarly unsuitable for the study of the differentiation of the
oocyte and of the prophases of maturation which takes place
in the end chamber. As this is very small, in order to examine
some of the stages, it would be necessary to cut very thin
sections of ovaries containing oocytes with large yolk-masses
and in some cases chorion as well. This is an operation which
I found quite impossible to perform. The finer structure of
the oocyte nucleus has therefore only been studied where it is
rendered necessary in order to give a connected account of the
development of the oocyte.

To the best of my knowledge the only references to the
development of the ovarian egg of the mosquito are contained
in two short papers by Christophers. In one of these (2) he
gives a very general description of the ovary and egg-follicles,
while in the other (8) he describes the development of the
egg-follicle from the examination of fresh material. As the
information in both these papers is of a very general nature it
has been found necessary to repeat portions of it, as otherwise
a connected account of the development of the ovary could
not be given.

I will take the opportunity here of expressing my deep

Ee ee

OVARY AND OVARIAN EGG OF ANOPHELES 397

indebtedness to Mr. A. J. Grove for the suggestion that I should
take up this line of research, for much useful advice during
the earlier stages of my work, and for my first supplies of
material. For my later supplies I was entirely dependent
on the kindness of Mr. R. F. Burton, to whom I wish to make
grateful acknowledgement.

The work was done under the supervision of Professor
F. W. Gamble, I'.R.S., whom I have to thank for assistance in
obtaining the very considerable, and not always easily acces-
sible, literature of the subject.

MATERIAL AND METHODS.

The majority of the mosquitoes were taken during the latter
part of their hibernating period and the first few weeks after
they had regained their activity and had commenced to feed
normally.

In order to eliminate the possibility of beimg misled by
artefacts due to fixation, the following method was employed.
Each batch of material was divided into three different parts ;
two of these were fixed in different re-agents, and in the case
of the third the ovaries were dissected out in salt solution and
one ovary of each insect was rapidly transferred to one fixative
and the other to another. In this way the effect of different
fixatives on ovaries in the same stage of development could
easily be compared.

In the cases where the ovaries were not dissected out, the
abdomen alone was fixed, and this was slit along each side with
a fine needle in order to allow the easy entrance of the fixative.

It was found that in the case of the less-developed ovaries
much the best results were obtained with those which were
dissected out, but far less distortion was produced in more
mature ovaries fixed while still in the abdomen. This was
probably due to the fact that the surrounding tissues only
allowed the fixative to reach the ovaries gradually and so
prevented rapid osmosis.

A number of different fixatives were used, the principal of
which were Flemming, both with and without acetic, Petrunke-

398 A. J. NICHOLSON

witsch, Zenker, and alcoholic Boum. Of these, Petrunkewitsch
was by far the most useful for general purposes as its penetra-
tion is very good, a most important consideration when dealing
with oocytes containing a large yolk-mass, particularly when
the egg-walls are present. For the finer cytological details
Flemming with acetic gave the best results, though Flemming
without acetic appeared to give a more perfect fixation, but
the latter had the disadvantage that the chromatin did not
stain as distinctly as it did with unmodified Flemming.

Another fixative of which I made considerable use was the
moditied Bouin described by Sheppard (27). This I used in
conjunction with the method of staming described by the same
author, i.e. bulk staining with carmalum and counter-staiming
with Griibler’s hight green. Using this method the fixation was
excellent, and the double staining gave very beautiful prepara-
tions—yolk and chorion staining bright green and the proto-
plasmic structures red. This property was very useful in
following the branching nucleus through the yolk-mass and
in following the production of chorion by the epithelial cells.
The fixative had the disadvantage, however. of making the
material brittle.

The staim principally used was Grenacher’s haematoxylin
counter-stained with dilute Lichtgriin picric. For the latter
the ordinary Lichtgriin picric solution (0-2 grm. Lichtgrin
dissolved in 100 ¢.c. of a saturated solution of picric acid in
absolute alcohol) was diluted with about ten times the bulk
of 90 per cent. alcohol. The counter-staining was done under
observation, as Lichtgriin appears to displace the haematoxylin
and the reaction requires to be stopped when all the yolk has
become green and the protoplasmic structures are still blue.
Using this method the branching nucleus can easily be followed
amongst the yolk granules.

Heidenhain’s iron haematoxylin counter-stained with eosin
or orange G was also extensively used and was particularly
useful for the finer nuclear details, but as it stained the yolk-
mass dense black it was not satisfactory for the more developed
oocytes.

—r

> aaa ey S=aam

OVARY AND OVARIAN EGG OF ANOPHELES 399

Most of the material was embedded in paraffin in the ordinary
way, but this was not very satisfactory, as oocytes contaiming
large yolk-masses broke up easily and it was difficult to cut
uninterrupted series. ‘Towards the end of my work I obtained
much more satisfactory results using the double-embedding
collodion and paraffin method described by Newth (21). Using
this method uninterrupted series of thin sections were easily
obtaimed.

HIBERNATING MosquItToEs AND First Periop or Eae
DEVELOPMENT.

During the winter, female A. maculipennis may be
found in cowsheds, church towers, and in fact in almost any
dry and comparatively warm place. They pass the winter
in a semidormant state, but they are found to feed a little
during this period, as occasionally an insect with a little blood
in the abdomen may be observed. A microscopic examination
of an insect at this stage shows that the fat bodies are
relatively very large, the ovaries are always very small, in
the ‘resting stage’, and the spermatheca is full of sperms.
In cases where the first batch of eggs had already been laid
and the second was developing, the spermatheca was seen to
contain sperms, though they were not in such a compact mass
as in the hibernating insects. As the first males of the season
had not emerged at this period, it would appear that one
fertilization of an insect is sufficient for more than one period
of oviposition.

The period at which the ovaries of mosquitoes first commence
to develop depends on the warmth of the season and also on
the locality. Thus in 1919 the majority of the insects taken at
the end of March in Kent showed considerable development
of the ovaries, while a similar degree of development of the
ovaries was not found till about three weeks later in the
Shrewsbury district. On March 238, 1920, however, insects
with the ovaries in the resting stage were only found with
sreat difficulty in the Shrewsbury district. This is no doubt
due to the early spell of fine weather in that year.

400 A. J. NICHOLSON

Warm weather acts merely as a stimulus to the activity of
the insects and causes them to go out and seek food. The
stimulus which gives rise to egg development appears to be
a good meal of blood. Numerous experiments have been
carried out to determine whether blood is necessary for the
production of eggs in mosquitoes. ‘To the best of my knowledge
in only one case have mosquitoes been induced to lay when
fed on any substance other than blood. §S. K. Sen (26) sue-
ceeded in inducing Stegomyia scutellaris to oviposit
by feeding with milk or peptone sweetened with cane-sugar,
and in two instances was successful when the insect had fed
on nothing but cane-sugar. I carried out a number of feeding
experiments on A. maculipennis, feeding them on sugar
and water, with and without the addition of peptone, and on
dates, bananas, and other fruit, all of which the mosquitoes
consumed very greedily, but in no case did any development of
eggs take place. In all my sections of abdomens in which the
eggs are developing, the gut is found to contain blood. with
the exception of the final stage, in which the eggs are fully
developed, when the gut is alwaysempty. As all these speci-
mens were collected in cowsheds, this does not prove that blood
is always necessary for the production of eggs, but it appears
to me certain that this normally is the case.

In hibernating mosquitoes the abdomen is very narrow and
flattened dorso-ventrally, but when they take their first meal
of blood in the spring the abdomen becomes almost globular
and distended to its limits with blood. In an insect which has
recently fed, the abdomen shows a large semi-transparent
uniform mass of blood, while a small whitish mass is seen
through the cuticle at the anal extremity. his consists prin-
cipally of Malpighian tubules but also contains the ovary.
On the second day the posterior portion of the blood is very
dark red and opaque, while the remainder is as before. In
sections the dark-red portion is seen to consist of partially-
digested blood containing very distorted corpuscles, while the
remainder appears to be quite fresh and might easily be
mistaken for a fresh meal of blood. The white mass at the

A BA NR RN NE Ms Me SOO A CR MONS Nt Nee

OVARY AND OVARIAN EGG OF ANOPHELES 401

anal extremity has enlarged somewhat owing to the growth
of the ovary. By the third or fourth day the blood-mass
is seen to be much reduced, and the whitish mass, the ovaries,
about half fills the abdomen. The blood-mass is reduced to
a mere spot or is entirely absent by the sixth or seventh day,
and, if the weather is warm enough, the eggs are then laid
during the night. In cold weather, however, the insects may
wait several days before oviposition.

These observations were made on a number of insects
collected in a calf-pen. At the time of collection their abdo-
mens contained semi-transparent blood-masses and they had
only recently fed. They were kept in jars in the laboratory
and the eggs were laid though they received no further food.

The period elapsing between the time of feeding and ovi-
position appears to be about a week, which agrees very closely
with Christophers’s observations on A. rossi in India (8), in
which case the period is given as six days.

I only sueceeded in observing insects durig the process of
laying in two cases, and in both the eggs were laid within an
hour of darkness setting in. The insects floated on the water
by spreading their long legs over it and frequently dipped their
proboscides into the water. When disturbed they flew off the
water with ease and seemed in no danger of drowning. The
actual oviposition I was unable to observe as the mosquitoes
refused to lay in the light.

FEMALE GENITAL ORGANS.

In the ‘ resting stage © the genital organs of the adult female
mosquito consist of two small ovaries lying ventro-laterally
in the posterior portion of the abdomen. Each of these com-
municates posteriorly with an ovarian tube, and the two ovarian
tubes unite to form a common duct, the gynaecophoric canal,
which opens to the exterior at the posterior end of the eighth
segment. A spermatheca, consisting of a thick perforated
chitinous shell and surrounded by a layer of large clear cells,
gives off a very thick-walled sperm-duct to the gynaecophoric

402, A. J. NICHOLSON

canal, which it enters a short distance anterior to the genital
aperture. A mucous gland, which consists of very large goblet
cells, also communicates with the gynaecophorie canal, close
to the entrance of the sperm-duct (fig. 7).

The ovary is surrounded by two sheaths, an outer bag-like
structure, the investing membrane, and an inner membrane,
which is closely applied to the egg-folhcles and fits them lke
a glove; the fingers of the glove are the follicular tubes and the
portion joining up the fingers encloses the lumen of the ovary.
The investing membrane passes anteriorly into a tubular
suspensory filament, which is fixed to the hypodermis at the
junction of the fourth and fifth segments, m a dorso-lateral
position. This filament is very long in the young ovary,
but it becomes quite short when the ovary is fully developed.

The two sheaths are identical in structure, and consist of
a structureless membrane, over one surface of which large
nuclei are found. From these radiating muscle-bands pass
over the membrane. These nuclei and muscle-bands are on
the inside of the investing membrane and on the outside of
the follicular tubes, and muscle-bands pass from the nuelei
of the one to the other, thus traversing the cavity between
the two sheaths and linking the investing membrane and the
folheular tubes together, so forming a very complicated
muscular system (fig. 24).

The muscle-bands of the sheaths are striped in the normal
manner, thus differmg from those of most insect ovaries (see
J. Gross, 9). They form broad bands close to the point of
origin from the nuclei and taper away from here and branch,
some of the finer branches appearing to consist of only a few,
or even a single muscle-fibre, as the ‘ striations’ consist of
bead-like, deeply-staining nodes on a fine thread (fig. 24).

It would probably be more correct, in many eases, to consider
that the nuclei are placed at intervals on the muscle-fibres,
rather than that they are the origin of the fibres. From an
examination of fig. 28 it will be seen that many muscle-bands
pass through the cytoplasm of the cells, and merely become
slightly indefinite there. The ‘ striations’, though somewhat

OVARY AND OVARIAN EGG OF ANOPHELES 408

distorted, are still placed at regular intervals. In other cases,
however, the nuclei certainly appear to be the origin of the
muscle-bands.

Over the greater portion of the surface of the investing
membrane the muscle-bands radiate in the normal manner
(fig. 28), but towards the junction with the oviduct they
gradually become reduced to two laterally-placed bands which
pass transversely to the long axis of the ovary (fig. 29). Finally,
the imvyesting membrane passes over the oviduct and the
muscle-bands now form the circular muscles of the oviduct.
In a similar manner the muscle-bands of the follicular tube
membrane pass insensibly into the longitudinal muscles of the
oviduct, inside which is found a layer of columnar cells surround-
ing the lumen of the oviduct.

If an ovary of a living insect is dissected out in salt solution,
& vigorous rhythmatic peristaltic movement is noticed. This
may be produced by the stimulus of the salt solution, but there
is little doubt that this movement takes place in the living
insect, at least when the eggs are being laid. The movement is
undoubtedly due to the muscular system described, and the
basket-work arrangement of the muscle-fibres is ideal for
compressing the ovary and so pressing the eggs into the
oviduct. The muscle-bands which pass from the investing
membrane to the follicular tubes are probably of use in drawing
the latter off the eggs, a process which takes place some time
before the eggs are laid.

A number of very characteristic cells are found in the space
between the two sheaths, and also between the follicular tubes
and the egg-follicles ; one or more is almost always to be found
in the region of each terminal chamber, between it and the
follicular tube membrane (fig. 25). These cells consist of large
nuclei embedded in a mass of very much vacuolated protoplasm,
from which fibres are frequently seen to pass. The exact
nature of these cells I have not been able to determine, but
T am of the opinion that they have some relation to the tracheal
system. The fibres seen passing from them are probably
tracheal endings, but they are so fine that it is difficult to

NO. 259 Ee

404 A. J. NICHOLSON

determine their nature and they might equally well be proto-
plasmic strands. In several cases, however, I have succeeded
in tracing some of these fibres to the bundles of tracheal endings,
so that some at least are tracheal in nature.

It is possible that these vacuolated cells may be leucocytes
us they agree in structure and size with Vaney’s (81) deserip-
tion and illustrations of the leucocytes in the larva of Gastro-
philus equi, but the fact that fibres enter them throws
considerable doubt on this theory.

The tracheal system in the ovaries is very highly developed.
‘l'racheae from the fourth and fifth segments go to the ovaries
and branches of these penetrate the investing membrane.
These tracheal trunks branch repeatedly in the space between
the two ovarian sheaths. The final branches consist of exceed-
ingly fine tubes, in which no spiral filament can be distinguished.
‘These pass to the various parts of the ovary im bundles, the
tubes being joined together by the tracheal cells which occur
at intervals along the bundles. When such tracheal cells are
cut transversely they appear to be very much vacuolated,
owing to the numerous tubes passing through the cytoplasm.
The individual tubes eventually become free from the bundles
and end in the tissues of the ovary.

A moderately large tracheal branch passes into the base of
each follicular tube and gives rise to numerous bundles of
tracheal endings. In the young ovary these have a very
characteristic appearance, and are seen as a prominent coiled
mass at the base of each follicular tube. This allows for
expansion when the ovarian follicles increase in size. The
ultimate endings of these tubes are difficult to discover, but
I have noted some entering cells of the ovarian follicles and
isolated tracheal endings may be seen in almost any part of
the follicular tubes.

Inside the follicular tube is an egg-string consisting of an end
chamber followed by two or three egg-follicles in various stages
of development. These follicles are joined together by cellular
stalks consisting of a single row of cells (fig. 26). The last
“stalk ’ or funicle runs from the posterior and most-developed

OVARY AND OVARIAN EGG OF ANOPHELES 405

follicle to the portion of the follicular tube membrane which
invests the lumen of the ovary, with which it fuses.

The whole of the egg-string is invested by a thin structureless
membrane, the tunica propria, which may also be regarded
as the basement membrane of the follicular epithelium. I am
here using the term ‘tunica propria’ in the sense defined by
J. Gross (9). The term has been used by many authors as
synonymous with ‘ peritoneal membrane’, a practice which
has led to much confusion. The peritoneal membrane is repre-
sented in the mosquito by the two ovarian sheaths.

Normally the tunica propria can scarcely be observed, as
it is very closely applied to the follicular epithelium ; but it
frequently happens that the follicles degenerate, and then the
tunica propria can easily be seen as a somewhat wrinkled,
structureless bag surrounding the remnants of the follicle
(fig. 24),

Each follicle consists of an oocyte and seven nurse-cells
completely surrounded by a single layer of cubical cells, the
follicular epithelium.

GENERAL LINES OF DEVELOPMENT OF OvaRY AND Eaas.

Before giving a detailed description of the various changes
which take place during the oogenesis of A. maculipennis,
I will first give a general outline of the development of the egg-
follicles and of egg formation, as a comprehensive view of the
whole subject will render it more easy to follow the detailed
descriptions of the different processes which together produce
the mature egg, but which, for sake of clearness, have to be
dealt with separately. Also a description of the anatomy of
the mature egg will be given, as with a knowledge of this it
will be possible to understand the object of the various processes.

The earliest stages of oogenesis are to be found in the end
chamber. This consists of a central mass contaiming compara-
tively large nuclei, which often vary considerably in appear-
ance but are not definitely divided into nurse-cells and oocytes,
and of a peripheral layer containing smaller nuclei which give
rise to the follicular epithelium.

E02

406 A. J. NICHOLSON

At intervals a mass of cells is cut off from the end chamber
and consists of seven nurse cells and an oocyte surrounded by
a follicular epithelium. ‘This follicle increases in size till it
reaches the resting stage (fig. 26), which is characteristic of
the ovaries of hibernating females. When the most-developed
follicles of the ovary are at this stage the ovary is very smal]
and quite transparent.

If the ovary of an insect which has just had a meal of blood
be examined in a fresh condition, it will be found that a white
opaque cloud is visible surrounding the nucleus of the oocyte.
This consists of fine yolk. In living ovaries at a slightly later
period it will be found that the whole oocyte is opaque white
and occupies about half the follicle. In sections this opaque
inass is found to consist of both coarse and fine yolk, and the
oocyte nucleus is no longer spherical but sends out blunt
processes into the yolk (fig. 18).

At a still later stage the follicles are elongated instead of
almost spherical, and are quite opaque except for a small
transparent cap, consisting of nurse-cells, and a thin investing
layer of follicular epithelium. The nucleus has now become
very much branched, branches passing throughout the yolk-
mass and appearing to be in some connexion with the nurse-
cells, which are evidently in a state of activity. A new structure
has now appeared between the follicular epithelium and the
yolk-mass. This consists either of globules or of a layer of
gelatinous material, and is the commencement of the mner wall.

Shortly after this stage the nurse-cell nuclei are extruded
from the yolk-mass and come to le in the follicular epithelium,
forming a cap over the anterior end of the egg. The oocyte
nucleus has now reached its maximum condition of branching
and shortly afterwards breaks down. The inner wall is
thick but still gelatinous. The follicular epithelium becomes
modified in two lateral areas and gives rise to the floats. The
rest of the epithelium secretes the chorion over the whole
surface of the egg, that portion which contains the extruded
nurse-cell nuclei giving rise to the micropyle apparatus.

Finally, the follicular epithehum degenerates into a mere

OVARY AND OVARIAN EGG OF ANOPHELES 407
membrane surrounding the fully-formed eggs; these lie in
the lumen of the ovary, as the follicular tubes have contracted
and merely cover the less-developed follicles and a small
portion of the anterior end of the fully-formed eggs. The eggs,
however, still lie in the position in which they developed (fig. 8).

When the eggs are being deposited they appear to break
through the remains of the follicular epithelium and then
pass down the oviduct to the exterior, the sperms entering the
micropyle immediately before the eggs are laid.

ANATOMY OF THE Mature Kaa.

The mature egg is more or less cigar-shaped and is provided at
each side with a float (fig. 9). It is, however, noticeably thicker
at one end thanthe other, and I consider the thick endas anterior,
ag it is anterior in the ovary. The portion of the egg which
is uppermost when floating I shall refer to as the dorsal surface.

The egg can be divided into three main parts—the outer wall
or chorion, the inner wall, and the yolk-mass.

The outer wall of the egg of A. maculipennis appears
to be identical in nature to that of other insect eggs, that is, it
is formed of chorion, which closely resembles chitin, but differs
from it in that it is soluble in warm KOH solution, whereas
chitin may be boiled in concentrated caustic potash for hours
without effect.

The structure of the chorion of the mosquito egg shows a high
degree of specialization. It consists essentially of a thin
envelope surrounding the whole egg, two floats placed dorso-
laterally, and a very beautifully-formed micropyle apparatus
situated immediately below the extreme anterior end of the egg.

The envelope is completely covered with processes of four
kinds. The ventral surface is thickly covered with short
knob-like processes (fig. 41), and some of these are slightly
larger than the remainder and are so arranged that they divide
the whole of the ventral surface of the egg into polygonal areas
(fig. 9). These areas probably have some connexion with the
form of the epithelial cells, but they appear to be too large to
be produced by individual cells.

408 A. J. NICHOLSON

The dorsal surface and the portion of the envelope lying under
the floats are covered with very different processes. They are
longer, and thin sheets of chorion radiate from a central axis,
so that in section the processes are star-shaped. At the top
of each process a cap-like structure joins all the radiating thin
sheets together (fig. 41).

At the line of division between the dorsal and ventral types
of processes there is a single row of much longer processes which
extend as a band from the terminations of the floats to the tips
of the egg (fig. 9).

The fourth type of process only occurs in small numbers and
is found at each end of the egg. This type is a comparatively
large boss-like structure consisting of a solid mass of chorion,
and seven or eight are found at the extreme anterior and
posterior ends of the egg (fig. 40).

The floats consist of a single sheet of chorion attached to
the chorion envelope along its ventro-lateral surface only.
The sheet curves round till it almost touches the dorso-lateral
surface so enclosing a considerable cavity. The whole of the
chorion sheet which forms the float is highly corrugated
(fig. 9).

The whole of these structures, with the possible exception of
the ‘ bosses’, appear to serve the purpose of supporting the
egg on the surface of the water. The ventral processes enclose
a film of air, which cannot be expelled by the water owing to
its surface tension and the closeness of the processes to one
another. The floats enclose a relatively large volume of air,
and again surface tension prevents the entrance of water. The
band of long processes, from the floats to the tips of the egg,
probably helps to support the egg by making use of surface
tension directly, i.e. by lyimg on the surface film of water.
The comparatively long dorsal processes do not help to
support the egg normally, but if an egg is sunk it will be
found that the relatively thick film of air enclosed by these
always causes the egg to regain the surface with its dorsal
surface uppermost.

If a drowning mosquito lays its eggs under the water if is

OVARY AND OVARIAN EGG OF ANOPHELES 409

found that they all sink, so it is obvious that the buoyancy of
the eggs is entirely due to the entrapped air.

The micropyle apparatus consists of a very thin disk-like
membrane surrounded by a thick supporting rmg. The central
portion of the membrane is produced into a funnel, which
passes through the inner wall to the interior of the egg, and the
cavity of the funnel is the micropyle (fig. 41).

The supporting ring is somewhat irregular on the outer side,
but the inner edge is very regularly scalloped, and the top portion
of the ring in each scallop is produced towards the micropyle
so that it overhangs the rest and together with the disk forms
a Shallow pocket (figs. 40 and 41).

Radiating out from the region of the micropyle to the point
of junction of each ‘scallop’ is a very fine ridge. ‘These are
thickenings of the disk corresponding to the divisions of the
cells which give rise to the apparatus. These ridges, together
with the ‘scallops’, mark the apparatus off into well-defined
areas. There are normally eight of these areas, but I have
found examples of the apparatus with from seven to ten.

The funnel continues right through the inner wall and ends
at the inner edge of the latter. It does not, however, com-
municate direct with the cytoplasm of the egg, but is sealed
up by a small globular portion of the inner wall which for
convenience I shall term the stopper (fig. 41).

A consideration of the structure of the micropyle apparatus
and of the genital aperture leads me to the following theory
as to the function of the former.

While the egg is passing through the gynaecophoric canal
it is no doubt considerably compressed by the muscular walls
of this canal. This would cause the thin membranous disk
of the micropyle apparatus to be forced outwards, and it would
probably lie level with the top of the supporting ring. By the
time the micropyle apparatus has reached the region where the
spermathecal duct opens into the gynaecophoric canal, the bulk
of the egg has left the genital aperture, and thus the pressure
on the egg is released. The membranous disk is now able to
resume its original position, and in so doing would probably

410 A. J. NICHOLSON

draw sperms into the saucer-shaped cavity of the apparatus.
Here the sperms would be directed by the radiating ridges to
the micropyle, and would then pass down the funnel and
between the stopper and the inner wall to the protoplasm of
the egg.

When an egg is freshly laid on water it is nearly white,
but after a few hours it becomes grey, and by morning it is
usually, if not always, quite black. The whole of this change
of coloration is due to the inner wall, which is transparent
when laid but later becomes opaque black. In several cases
insects in captivity have laid their eggs on dry media instead of
on water, and in none of these cases did the eggs become black :
they merely turned dirty yellow. It would thus appear that
water has something to do with the productionof the dark colora-
tion, though how the water gets to the inner wall is not clear.

Besides changing colour the inner wall changes in character
after the deposition of the egg. Ifa freshly-laid egg is placed in
strong acid or alkali rapid expansion of the inner wall takes
place, and it is seen first to become rapidly wrinkled and finally
to burst through the chorion with explosive force.

An egg which has become black, treated in the same way does
not appear to be acted upon. Also if a freshly-laid egg is
crushed under a cover-slip the inner wall is seen to be gelatinous,
and oil-like globules may be broken off if a little pressure is
applied. If treated with osmic acid these globules become
brown, so that there may be a chemical, as well as a physical,
resemblance between the inner wall and oil. If an egg which
has become black is crushed it is found that the inner wall is
no longer gelatinous, but is hard and somewhat brittle as it
cracks with pressure.

The yolk-mass occupies the whole of the egg inside the inner
wall, It consists of an alveolar protoplasmic mass in the
vacuoles of which yolk granules of two kinds are found (fig. 11).

The more obvious form of yolk consists of comparatively
large granules 0-008 mm. to 0-01 mm. in diameter, which are
proteid in character, as the following reactions show.

If treated with copper sulphate solution followed by excess of

tag (Gna ) Cup <1 age, Gae- eer

OVARY AND OVARIAN EGG OF ANOPHELES 411

caustic potash, i.e. the ‘ Biuret Reaction’, the granules turn
a beautiful violet colour.

Nitric acid turns the previously white yolk light yellow, and
when excess of ammonium hydroxide is added the yolk turns
a brilliant orange colour. (Xanthoproteic Reaction.)

With osmic acid the granules turn yellow or yellow brown.

These granules appear to be homogeneous and solid, as they
can be broken by the pressure of a cover-slip.

The other type of yolk consists of small granules 0-001 mm.
in diameter, and these are found surrounding the granules of
coarse yolk (fig. 11). They are chemically quite different from
the large granules as they do not respond to any of the above
reactions. They are certainly not fat globules, as might be
expected, as they are not coloured in any way by osmic acid,
either alone or in the presence of chromic acid in the form of
Flemming without acetic: a test described by Gatenby (7).
[ have not succeeded in making the fine yolk granules respond
to any chemical reaction or stain in any way. In sections they
appear as clear vacuoles, but they must be more than mere
drops of watery fluid, as, if an egg is broken in water, they are
scattered through the liquid as minute spheres which exhibit
a very pronounced dancing movement. This movement comes
to rest after a few hours, so that it is probably due to diffusion
currents from the granules.

It may be noted here that the fine yolk is a definite con-
stituent of the mature mosquito egg, and is not an inter-
mediate substance produced during the formation of yolk,
as in the case of the ‘ granules adipeux ’ of Pholeus phalan-
gioides according to Van Bambeke (1).

The protoplasmic portion is very inconspicuous in the
mature mosquito egg. In sections it is seen as a uetwork of
fine threads, the meshes of which are occupied by the yolk
granules (fig. 11). At the periphery the protoplasmic threads
are slightly thicker than in the remainder of the egg, and they
form an ill-defined cortical layer. This layer is thickest at the
two extremities of the egg, in each of which it forms a small area
of granular protoplasm free from yolk. Oceasionally a few

412 A. J. NICHOLSON

disconnected fragments of the branching nucleus can be distin-
guished in the yolk-mass, even in eggs which have been laid.

The only other protoplasmic structure visible is a small mass
of granular protoplasm situated about a quarter of the length
of the egg from the anterior pole and in the centre of the yolk-
mass. This appears to be the remains of the chromatin residue,
but before the egg is laid no nuclear structure can be distin-
guished in it. A short time after oviposition, however, this
mass is found to contain minute chromosomes.

DIFFERENTIATION OF GERM-CELLS AND THE First PERIOD
oF GROWTH OF THE HKGG-FOLLICLES.

The growth of the egg-follicle falls naturally into two periods.
The first period is from the time when the follicle is separated
off from the end chamber up to the formation of the ‘ resting
stage’. Growth is arrested at this point and only recommences
after the insect has had a good meal of blood, when the folhele
enters upon the second period of growth, which culminates in
the formation of the mature egg. In the second and later
generations of eggs, however, the two periods run concurrently,
i.e. while the primary follicles are undergoing the second period
of growth the secondary follicles are undergoing the first,
and when the former have formed the mature egg the latter
have reached the ‘ resting stage ’.

If an end chamber in an early stage of development be
examined it will be found to consist of a central mass containing
comparatively large nuclei surrounded by a layer in which
smaller nuclei are found scattered somewhat irregularly.
Cell divisions cannot be distinguished and the mass is probably
a syncytium (fig. 24). The larger central nuclei are those of
the oogonia and oocytes, and the smaller peripheral ones give
rise to the follicular epithelium.

The nuclei of the oogonia vary considerably in appearance
even in the same mass, but in the earlier stages of the end
chamber I have been unable definitely to separate the true
oocyte from the nurse-cells. Occasionally mitotic figures are
found in the central mass (fig. 31). This is no doubt the stage

OVARY AND OVARIAN EGG OF ANOPHELES 413

at which the oogonia divide to produce the oocytes, all but
one of which become modified to form nurse-cells.

In many end chambers, where the foregoing process has no
doubt already taken place, the proximal nucleus is quite
distinct from the remainder. It is clearer and contains a well-
defined spireme, while the other cells are somewhat darker,
and, though they often also appear to contain a spireme,
this is never so sharply defined and usually can only be made
out with difficulty. ‘This proximal nucleus is the true oocyte
nucleus, while the cells lying above it are the nurse-cells, and
at the distal end of the chamber the remaining oogoniaare found.

The epithelial layer grows between the mass of nurse-cells
and the oogonia, so that the former, together with the oocyte,
are completely enclosed in a follicle. The fact that there are
seven nurse-cells and one oocyte suggests that they are produced
by three successive divisions of a single oogonium, seven of the
daughter cells becoming nurse-cells and the eighth the true
oocyte. Occasionally an aberrant number of cells are included
inside the follicular epithelium: eight large cells, and at the
distal end a number of smaller cells, apparently eight in number.
In this case the epithelial layer has surrounded two masses
of daughter cells instead of one. I have only found such
follicles in young ovaries, so that the mass of smaller cells
evidently does not take part in the development of the follicle
and probably degenerates. This further supports the theory
that the nurse-cells and oocyte are the daughter cells of a single
oogonium.

When the follicles are first formed the follicular epithelium
consists of a comparatively small number of cells. These
multiply rapidly by mitotic division (fig. 32), and at this stage
no clear cell divisions are visible. Also there is only a small
quantity of cytoplasm, the epithelium consisting principally
of a large number of closely-packed nuclei. This mitotic
division takes place throughout the first period of growth till,
when the resting stage is reached, the full number of epithelial
cells is attamed and also the nuclei have reached their full size.
During the second period of growth the epithelium increases

414 A. J. NICHOLSON

very considerably in area, but I am convinced that this is
entirely due to the increase in size of the dividual cells. In
no case have I seen any sign of mitotic division during this
period. Further, counts were made of the epithelial cells in
median longitudinal, and transverse sections of follicles in
various stages of development, and the average number of
cells visible in a section was found to be practically constant
irrespective of the size of the follicle.

The ‘ funicle ’ arises by the local proliferation of the epithelial
cells of the septum between the end chamber and the young
follicle. Shortly after the young follicle is cut off from the end
chamber a number of nuclei are found closely packed one above
the other in the form of a short rod. This rod-like structure
is found in one side of the septum, and this asymmetrical
position is retained throughout the growth of the follicle, as
will be readily understood when it is considered that the
micropyle apparatus is formed immediately under the ‘ funicle ’,
and this is not terminal but ventral in position.

The nuclei in this rod-like structure continue to divide,
forming a long string. The rest of the septum splits and gives
rise on the one hand to the follicular epithelium and on the other
to the epithelial layer of the end chamber. The only portion of
the septum which does not split is that which contains the rod-
like series of nuclei, and this now forms the funicle.

When the egg-follicle is first formed the nurse-cell nuclei
often contain a more or less indefinite spireme and there is very
little cytoplasm. As the follicle grows the nuclear contents
become arranged in much convoluted bands which appear to
be directly derived from the spireme. These bands are some-
what peculiar in structure. They consist of a non-staining
ribbon of linin, across which lie a large number of chromatin
bands giving an appearance somewhat resembling that of
a striped muscle (fig. 80). These convoluted bands lie round
the periphery of the nucleus with the result that individual
sections give a very wrong idea of the appearance of the
nucleus. This is due to the fact that the nurse-cell nuclei are
gigantic (0-02-0:08 mm.) in diameter, so that each nucleus is

OVARY AND OVARIAN EGG OF ANOPHELES 415

cut into a number of sections and the idea of the continuity
of the convoluted bands is lost.

Towards the centre of the nucleus a large irregular nucleolus
is found imbedded in a mass of linin. This is joined to the con-
voluted bands by linin threads. The nuclear membrane is
very thick and is always plainly visible.

When the follicle has reached the resting stage each nurse-
cell nucleus is found to be surrounded by a large mass of
cytoplasm which is limited by a definite cell-membrane. The
cytoplasm is slightly granular and takes up rather more stain
than that of the younger cells.

The earliest stages of the oocyte nucleus which I have been
able to distinguish contain a number of deeply-staiming
chromatin loops all of which arise together from one side of
the nucleus (fig. 12). This is evidently the ‘ bouquet stage ’
of the prophases. Usually a small nucleolus is also seen, and
in the few cases I have observed in which this is not visible it
is possible that it was hidden by the chromatin threads. At
this early stage the nucleolus is a spherical vesicle and only
stains very lightly.

As the nucleus grows the nucleolus becomes relatively larger,
takes up stain rather more readily, and soon several vacuoles
become visible in it. While this is taking place the chromatin
threads wind themselves round the nucleolus and invest it
tightly (fig. 13). The nucleolus continues to grow at a more
rapid rate than the rest of the nucleus, while the chromatin
strands do not appear to grow at all. The result is that the
enlarging nucleolus gradually pushes the investing chromatin
strands off itself, and these are then seen as a small mass of
closely-woven threads at one side of the nucleolus (fig. 16).
These threads concentrate into a closely-packed mass in which
the individual chromatin threads can no longer be distinguished,
and the whole has the appearance of a small dark nucleolus at
the side of the true nucleolus. For want of a better term
I shall refer to this mass as the ‘ chromatin residue’. This is
embedded in a mass of linin which also invests the nucleolus and
thus holds the two closely together (fig. 17). This arrangement

416 A. J. NICHOLSON

persists until after the oocyte has left the resting stage.
l'requently, however, the chromatin residue is very difficult
to find in oocytes in the resting stage. Close examination
reveals the fact that it is not only very closely applied to the
surface of the nucleolus, but is situated in a shallow depression
in the latter and thus does not disturb the spherical contour.
Nuclei in which this arrangement exists appear on a cursory
examination to contain one large nucleolus and nothing else
but nuclear sap. This is the characteristic appearance of the
vocyte nucleus in the resting stage.

We have noted that the nucleolus, which is at first vesicular
und has little affinity for stam, soon becomes vacuolated and
stains more deeply. As it creases in size the vacuoles increase
rapidly im number and the affinity for stain becomes more and
more marked, till, when the resting stage is reached, the
whole surface is covered with vacuoles and the nucleolus stains
as deeply as chromatin. Though the nucleolus has so great
an aftinity for chromatin stains I do not consider that it is
formed of chromatin. I regard it rather as a composite struc-
ture, consisting of a plasmosome in which chromatin, or some
similar basiphil substance, is present. This is indicated by the
fact that when stained with eosin the nucleolus is stained
bright red, whilst the rest of the ovary is hardly perceptibly
tinted by it, but when stamed with haematoxylin the red
colour is completely masked. These staining properties
are confined to a cortical layer, in which the above-mentioned
vacuoles lie. This layer surrounds a large central cavity the
contents of which appears to be nuclear sap. This structure
of the nucleolus is easily seen as it is so large, about 0-015 mm.
in diameter, that it may be cut into three or four sections with
ease. In such sections the cortical layer appears as a deeply-
staining ring surrounding a cavity which contains non-staiming
material, in which irregular strands of another substance,
possibly linin, can be seen (fig. 17).

Though the nucleus normally contains only one large
nucleolus this sometimes appears to undergo fragmentation.
Tu the more usual cases of this, one or more small nucleoli

i = ere ee

C

OVARY AND OVARIAN EGG OF ANOPHELES 417

may be seen in the nuclear sap surrounding the large nucleolus,
or even inside the central cavity of the latter. The smallest
examples of these appear homogeneous, but the larger fragments
show a similar vacuolation to that of the large nucleolus, and
are evidently produced by the fragmentation of the cortical
layer. In one ovary examined all the oocyte nuclei contained
numbers of small nucleoli. In some cases, however, the large
nucleolus was indicated by a sort of phantom, as if it had given
up practically all its substance, but the very small quantity of
material remaining still traced its original form. In the cases
where the original nucleolus had completely disappeared it
is obvious that the fluid which was in its central cavity must
have mixed with the surrounding nuclear sap, but there was
no indication of two different fluids inside the nucleus. It
therefore seems probable that the nucleolus is merely a hollow
sphere, with nuclear sap both inside and surrounding it.

The nucleus is surrounded by a thick nuclear membrane
which is stained black by iron haematoxylin and frequently
shows numerous local thickenings.

When the oocyte has reached the resting stage it has a thick
layer of cytoplasm round the nucleus. ‘This usually stains the
same as the cytoplasm of the nurse-cells, but sometimes it
appears slightly more granular.

Seconp PEriop oF GROWTH OF THE HGG-FOLLICLES.

The second period of growth extends from the time when
the egg-follicle leaves the resting stage up to the formation of
the mature egg. During this period several different processes
take place and it will be convenient to consider these separately.
These processes may be divided primarily into the nutrition
of the egg and the formation of the egg-walls. The complexity
of the changes of the oocyte nucleus renders it advisable
first to treat with these thoroughly and then to deal with yolk
formation and the nutrition of the oocyte in general. The
subject of the formation of the egg-walls will be divided into
the production of the inner wall, of the outer wall, and of the
micropyle apparatus.

418 A. J. NICHOLSON

I. Branching of the Oocyte Nucleus and
Segregation of Vegetative and Germinal] Parts.

During the second period of growth of the egg-follicle of
A. maculipennis the oocyte nucleus undergoes a most
remarkable development. I have not been able to find
a detailed description of a similar development in the case of
any other sect, but, as will be seen later, it is probable that
this particular form of development of the oocyte nucleus
is by no means confined to A. maculipennis.

After the mosquito has fed on blood the first indication of
alteration in form of the oocyte nucleus is observed in the
nuclear membrane. Previously this was spherical in form,
but now it is seen to be somewhat irregular in outlne. This
irregularity becomes more and more marked as development
proceeds, till the nuclear membrane is seen to send out a few
blunt processes into the cytoplasm and the cavity enclosed by
the membrane appears somewhat larger than it was previously.

While this has been taking place the nucleolus has also been
altering somewhat in shape. It loses its spherical form, first
becoming ovoid and later slightly flattened in a plane at right
angles to the axis of the egg-follicle, at which stage it begins
to send out blunt processes (fig. 18). The structure, however,
is still the same as in the resting stage, that is, it is vacuolated
and contains a large non-staiming central mass.

The chromatin residue commences to separate from the
nucleolus at this stage. Its subsequent history will be dealt
with separately. From this point the more obvious nuclear
changes are confined to the nucleolus and the nuclear mem-
brane, the nucleolus and its products forming by far the greater
part of the bulk of the nucleus.

It has been noticed that both the nucleolus and the nuclear
inembrane have begun to send out blunt processes. These
processes rapidly elongate and take on the form of branches,
which in their turn send out secondary branches. The branching
of the nucleolus and the nuclear membrane is intimately
connected, as the nuclear membrane surrounds the nucleolar

OVARY AND OVARIAN EGG OF ANOPHELES 419

branches. The progressive stages of the branching are seen
in figs. 1-6, which are reconstructions made from serial sections.
In individual sections the branching nature of the nucleus
cannot be seen, as only sections of the branches are found
and these appear to be fragments of the nucleus, as described
by S. R. Christophers (2).

Besides altering in form the nucleolus undergoes an altera-
tion in structure. Shortly after the branching has commenced
the vacuoles of the outer crust become indistinct, and the
central mass, which up to this point has not taken up stain,
now becomes darker and the whole becomes very granular
(fig. 19). Later the whole of the products of the nucleolus
form a homogeneous granular mass which readily takes up
nuclear stains. The branches of the nucleus are entirely
formed of this mass and, in the earlier stages at least, are
surrounded by the nuclear membrane.

As the branching proceeds the branches become finer and
finer, and pass throughout the whole of the rapidly enlarging
oocyte. They have, however, a very definite arrangement.
It will be seen in figs. 22 and 23 that the main branches occupy
approximately a median position between the centre of the egg
and the periphery, forming a cup-like structure roughly
following the contours of the egg. It must be pointed out,
however, that in these two sections the nuclear branches
appear much more continuous than is normal, though indiea-
tions of this arrangement can be seen in all sections of this
and later stages. The thickenings of the ring-shaped nuclear
mass in fig. 22 are the main branches cut transversely, and
the thin portions joining them are smaller lateral branches ;
these appear to connect the larger branches together, so that
probably the nucleus forms a reticular structure from which
thin short branches pass towards the centre of the oocyte,
while others go towards the periphery. The reconstructions
do not show this reticular structure of the nucleus, but this
may be accounted for by the fact that only the very finest
branches appear to join the main branches together, and I found
it impossible to reconstruct the course of such fine branches with

NO. 259 Ff

42.0 A. J. NICHOLSON

accuracy. ‘They have, therefore, been omitted from the’
figures of the reconstructions. ‘Thus fig. 6 only shows a large
number of more or less longitudinally placed branches, but
I consider that these were joined together by a number of much
finer branches.

As the branching proceeds the nuclear membrane becomes
less conspicuous, but it is easy to see in the earlier stages of
nuclear branching. Later it becomes closely surrounded by
yolk and evidently les closely applied to this. The appearance
ofa membrane 1n this position can usually be observed, but this
by no means proves that a membrane is present. I find that
if a crack appears in the yolk-mass, the edges of the crack
often appear to be limited by a membrane, and this I believe
to be due to the refraction of the transmitted light by the
spherical yolk granules. In such cases the apparent membrane
always closely follows the contour of the closely-packed yolk
granules. In cases of the branching nucleus, therefore, in which
the appearance of a membrane can be observed in a position
separated from the yolk-mass, I consider that this is actually
the nuclear membrane, while, on the other hand, if there appears
to be a membrane closely following the limits of the yolk-mass,
it cannot be definitely stated that a membrane is, or is not,
present. Bearig these considerations in mind, I find that
portions at least of the nuclear membrane cover the branches
up to a late stage, as when the nurse-cells are breaking down
the nuclear membrane can still be seen in places. Whether it
is continuous or not at this stage it is impossible to say, but
I favour the view that it does not exist over some portions
of the branches.

When the nurse-cells are breaking down large deeply-staining
slobular masses are found in the nuclear branches (fig. 21).
These appear to be formed of substance derived from the
degenerating nurse-cells. The globular masses are probably
absorbed by the nuclear substance as they cannot be observed
in later stages.

After the extrusion of the nurse-cell nuclei the main function
of the branching nucleus, that of the nutrition of the oocyte,

OVARY AND OVARIAN EGG OF ANOPHELES 491

appears to be completed. It does not immediately degenerate,
however, but continues to branch, the branches becoming
finer and finer till finally they merge imperceptibly into the
cytoplasmic reticulum, when all trace of the nucleus is lost.
Occasionally there are still some vestiges of the branches
remaining when the egg is laid.

It will be seen later that all this complicated branching of
the nucleus may be regarded as a mechanism for the transfer-
ence of nutritive material to the egg. As has already been
noted, this nutritive mechanism is mainly the product of the
nucleolus, the nuclear sap and nuclear membrane participating
but being only of secondary importance. ‘The nucleolus may
therefore be regarded as the vegetative portion of the nucleus.
The chromatin residue does not take any part in the nutrition
of the egg, but from it the female pronucleus and the polar
bodies appear to be produced, so that it.is the germinal portion
of the nucleus. $. R. Christophers (2) refers to this chromatin
residue as the ‘female pronucleus’, but as the polar bodies
have not yet been separated from it this is obviously a misuse
of the term.

When the chromatin residue first begins to leave the side of
the nucleolus, it is found to be no longer a deeply-staining mass
of chromatin, as only portions of it take up stain readily. Us
appearance at this stage varies considerably, but it is usually
formed of a non-staining matrix in which a deeply-staiming
round spot is found, and commonly several other parts take
up stain often appearing to be portions of the coiled threads
of which it was originally composed (fig. 19). The whole of
this is embedded in a mass of lining from which radiating
strands pass to various parts of the nucleus.

During the growth of the oocyte the chromatin residue
travels progressively farther away from the nucleolus, and
as it does so its staining properties decrease. The round spot
mentioned above is the last portion to lose its power of taking
up chromatin stains, but finally the chromatin residue can only
be recognized as a small lightly-staining mass situated a little
below the nurse-cells. This is the last stage I have been able

Ff2

422, A. J. NICHOLSON

to discover, and it occurs when the egg-follicle is about a third
of its full size. It now becomes lost im the yelk-mass from
which stains will no longer differentiate it. It does not follow,
however, that because it is ne longer visible it has therefore
ceased to exist as a separate entity.

A little later, after the nurse-cells have been extruded,
a small mass of protoplasm which is free from yolk is found
situated a little behind the anterior extremity of the egg,
that is, approximately in the position in which the chromatin
residue disappeared. The central portion of this mass is rather
denser than the remainder, and this I regard as the derivative
of the chromatin residue. Some eggs were sectionized which
had been fixed about an hour after laying. In these a number
of minute chromosomes were found situated in the denser
central portion of the above-mentioned mass. It would there-
fore appear that the reconstruction of the chromosomes takes
place shortly after the fertilization of the egg, a process which
frequently occurs in insect eggs.

As the chromatin residue was derived from the chromatin
of the spireme, and as after fertilization the reconstruction of
the chromosomes takes place in the anteriorly-placed mass of
protoplasm which is free from yolk, it seems a reasonable
assumption that this mass contains the derivative of the
chromatin residue.

Il, Yolk Formation and the Nutrition of the
Oocyte.

Shortly after the egg-follicle enters the second period of
srowth the oocyte commences to enlarge and soon occupies
about half of the egg-follicle, the nurse-cells occupying the
other half. At this stage yolk begins to make its appearance.
First the cytoplasm is seen to contain a number of small
globules which do not stain. These enlarge and are then recog-
nized as fine yolk. Immediately after this granules of coarse
yolk make their appearance, usually forming a zone midway
between the nucleus and the periphery of the oocyte. These
granules are very small, but they increase in size as the oocyte

OVARY AND OVARIAN EGG OF ANOPHELES 498

grows, and more small granules or ‘ young yolk ’ appear in the
cytoplasm till this is completely filled with yolk. As the
oocyte grows, therefore, it is natural that the young yolk
should appear at the point where the cytoplasm is increasing
most rapidly, that is round the periphery and more particularly
at the proximal end of the oocyte. This actually is the case, us
will be seen from fig. 27, in which the small granules of young
yolk can be plainly seen around the periphery of the oocyte
and a much larger mass is visible at its proximal end. Young
yolk may also be observed amongst the larger and older
granules in the central mass of the oocyte, and no doubt
growth is by no means confined to the peripheral portion of the
cytoplasm.

At the same time that this coarse yolk is appearing fine yolk
is also being laid down in the cytoplasm, the production of the
two substances thus taking place simultaneously.

As the oocyte is growing rapidly and large quantities of yolk
are being laid down, the question arises as to how the nutrition
of the oocyte takes place. The fact that most of the young
yolk is laid down in a peripheral position might lead one to
suppose that nutritive material passed by diffusion through
the follicular epitheium. This probably does take place to
some extent, but only in the earliest stages, as later the folheular
epithelium begins to secrete the inner wall and then no doubt
requires all the nutritive material which passes into it.

The greater part of the nutritive material undoubtedly
reaches the egg through the medium of the nurse-cells, and
these in their turn must receive it from the ‘ rosette cells ° as
these are the only portion of the epithelium which is not
secreting the inner wall. The mpushing of the rosette-cells
and their close application to the nurse-cells (figs. 86 and 37)
may assist in the transference of the nutritive material.

That the nurse-cells are in a state of activity during this
period is indicated by the fact that the cytoplasm stains
irregularly, more deeply on one side than the other (fig. 27),
an appearance which seems to be characteristic of cells which
are secreting.

494 A. J. NICHOLSON:

3etween the inner side of the nurse-cells and the branching
nucleus a mass of cytoplasm is found which stains more deeply
than the remaining cytoplasm of the oocyte. This forms a con-
nexion between the nurse-cells and the oocyte nucleus, and
I regard it as the path of the nutritive material from the former
to the latter (fig. 27).

The branching nature of the nucleus, and the general arrange-
ment of the main branches in a medium position between the
centre of the oocyte and its periphery, form an ideal distribu-
tion system for carrying nutritive material to all parts of the
oocyte.

The path of the nutritive material would therefore appear
to be from the surrounding fluid to the rosette-cells, through
these into the nurse-cells, which in their turn pass it to the
branching nucleus through the medium of the above-mentioned
more deeply-staining mass of cytoplasm. The branches carry
the fluid to all parts of the oocyte, and the cytoplasm of this
uses it in the formation of yolk granules.

When the oocyte is approaching full size the cytoplasm of
the nurse-cells begins to disappear (fig. 37) till finally the
nuclei are only surrounded by the cell membrane. Simultane-
ously large globular masses of deeply-staming material appear
in the branches of the oocyte nucleus (fig. 21) and obviously
have some connexion with the degenerating nurse-cells. These
slobular bodies are by no means confined to the region of the
nucleus near the nurse-cells, but are found in all parts of the
main branches, and it is therefore only reasonable to suppose
that they have travelled along the branches. This gives
considerable support to the view that the branching nucleus
is a mechanism for the transference of nutritive material.

It should be noted that the nurse-cells degenerate when the
period of nutrition is practically completed, and that in so
doing part of their substance is used directly for the nutrition
of the oocyte. This is further proof of the nutritive character
of the nurse-cells.

As there is no longer any nutritive material for the branching
nucleus to carry, it is obvious that if this is its only function

ey ee eee ee eee a

OVARY AND OVARIAN EGG OF ANOPHELES 495

it should now degenerate. This it does, as we have seen, by
continuing to branch till the final branches merge into the
reticulum of the cytoplasm, when nutrition is completed.

III. Discussion concerning the Ooeyte Nucleus
and Nutrition of the Oocyte in A. Maculi-
pennis.

I have unfortunately not found it possible to examine the
whole of the literature dealing with the nutrition of insect
eggs, but, in the literature I have consulted, I have not
discovered a case in which the mechanism of nutrition is to my
mind as clearly demonstrated as it is in the developing egg-
follicle of A. maculipennis. I do not believe, however,
that the insect under consideration is unique in having this
particular mechanism of nutrition. From several series of
sections which I cut of the closely allied insect Theobaldia
annulata, I am convinced that the same mechanism is
present here. Also Soyer (28) makes a short reference to the
nucleus of the developing oocyte of a Staphylinid, and from
his description it would appear closely to resemble that of
A. maculipennis. He remarks, * Le noyau, trés irrégulier
déja a ses phases les plus jeunes, se ramifie et se déchire en
une multitude de franges dans toute l’étendue du vitellus.
Cette ramification finit par étre poussée si loin qu’on n’a plus
sous les yeux qu’une sorte de long filament avec quelques
branches latérales, a peine visibles, dont les extrémités se
ramifient et se perdent entre les vésicules lécithiques qui
emplissent & ce moment la masse ovulaire’. According to
Korschelt (18) Stuhlmann has observed the branching of. the
nucleus throughout almost the entire oocyte of Necrophorus
vespillo and of Silpha sp.

Branched nuclei are by no means uncommon, particularly
in insects. They are commonly found in nurse-cells, gland-
cells, fat body-cells, and the cells of Malpighian tubules, in all
of which cases they appear to have some relation to the secretory
activities of the cells. Thus in many gland-cells the nucleus
is only branched during the period of secretion. In only two

426 A. J. NICHOLSON

cases have I found references to branched nuclei in cells which
are not obviously secretory in function, but both of these con-
cern embryonal structures which are undergoing rapid growth
and are therefore in a state of great activity. Korschelt (18)
cites cases of branched segmentation nuclei, and Seeliger (25)
describes the branching of the nuclei in the muscle-bands of
young Oikopleura. In the latter case the branching
reaches an extraordinary high state of development, becoming
finally a complicated reticular network of very fine threads.
Korschelt (18) regards the formation of nuclear branches as
a method of increasing the surface of the nuclei to aid secretion.
Thus, speaking of egg-cells, he remarks, ‘ Die Bildung der
Fortsiitze stellt eine Oberflichenvergrésserung des Kernes
dar, vermdge welcher dessen Beriihrungsfliche mit der
Zellsubstanz erhebich vergréssert wird. In aihnlicher
Weise wurde die Bildung von lingeren oder
kiirzeren Fortsitzen des Kernes bei secerniren-
den Zellen verschiedener Art beobachtet. Hier
waren die Fortsitze nach demjenigen Theil der
Zelle gerichtet, wo die Secretion stattfand.’

It will thus be seen that the form and position of the nucleus
of the oocyte in A. maculipennis indicates that it is
secretory in function and comparable to the nuclei of secreting
cells. This similarity is further shown by the fact that during
the process of branching the nuclear contents break down and
form a granular mass, a process which normally takes place
in secreting cells during the period of activity. The close rela-
tion of one end of the branching nucleus to the nurse-cells and
the other to the area of maximum activity of the growing
oocyte, i.e. the posterior end, together with the relatively
deeply-staining mass of cytoplasm between the nucleus and the
nurse-cells, renders it difficult to imagine that the branching
nucleus can be other than secretory in function. It is from the
somewhat similar arrangement in other cells that Korschelt (18)
draws the conclusion that the nucleus takes an active part in
the nutrition of a cell. Thus he observes, ‘ Das Aussenden von
Fortsitzen und Anniherung des Kernes an diejenige Seite der

OVARY AND OVARIAN EGG OF ANOPHELES 427

Zelle, von welcher derselben Nihrsubstanz zugefiihrt wird,
die Umlagerung des Kernes mit einer von fern her angezogenen
Niihrmasse,—diese Vorgiinge konnten einzig und allein als
eine Hinflussnahme des Kernes auf die ernihrende Thatigkeit
der Zelle gedeutet werden.’ Also Doncaster (5), in his recent
work, makes the following assertion: ‘ The nucleus—in some
way controls the metabolic activities of the cell, and its peculiar
behaviour in the growing oocyte can only be ascribed to its
activities ir this connexion.’

Chubb (4), on the other hand, denies that the oocyte nucleus
takes an active part in yolk formation. Thus he says, ‘ The
actual formation of the yolk spherules must therefore be
regarded as an automatic process, which commences as soon
as the accumulated materials in the cytoplasm attain the
requisite degree of concentration, and which does not entail
either increased nutrition of the ovum or increased activity of
the nucleus’. The amoeboid movements of the germinal
vesicle described by various authors, e.g. Bambeke (1), which
are considered as an indication of nuclear activity, Chubb
regards as probably being artefacts due to fixation. He
observed oocyte nuclei in Antedon which were apparently
amoeboid, but he shows that these are purely artefacts as
‘In the first place the nuclear irregularity shows no spatial
relation whatever, either to the other cell structures, to com-
mencing yolk formation or to the position of the nucleus in
the cell. In the second place it is only in radial section that the
nuclear irregularity presents the appearance of Pseudopodia;
in tangential sections these nuclear “ processes ”’ are found to
invariably resolve themselves into a coarse wrinkling of the
nuclear membrane. Finally, the artificial nature of the nuclear
irregularity is strongly indicated by the variable behaviour of
the nucleus with varying fixation.’

It is very probable that this explanation does apply to many
cases where amoeboid structure has been deseribed, but it
certainly does not apply to the oocyte nucleus of A. maculi-
pennis. ‘The high degree of branching of the nucleus in this
case could not possibly be regarded as an artefact due to

498 A. J. NICHOLSON

fixation, and in addition the branching has a definite spatial
relation to other cell structures, yolk formation, and the
position of the nucleus in the cell. The branched appearance
of the nucleus is not confined to any type of section and it is
a perfectly constant character in no way dependent on the
fixative. Also it is not possible to regard the branches as
the result of the pressure of the yolk-laden cytoplasm, so that
the only possible explanation is that the oocyte nucleus of
A. maculipennis is in a state of great activity during the
period of yolk formation.

It has been shown that the oocyte nucleus only commences
to branch when the yolk begins to appear, and that when all
the yolk has been produced and the nutrition of the oocyte is
complete, the branching nucleus breaks down and its substance
is absorbed directly by the cytoplasmic portion of the yolk-
mass. Immediately before the final disappearance of the
branching nucleus this structure rapidly loses its power of
taking up stain. This is a further mdication of the close
similarity existing between the oocyte nucleus of A. maculi-
pennis and the nuclei of secretory cells. Thus Bambeke (1),
speaking of glandular cells, points out that after the secretion
has lasted for a certain time the power of the nucleus to take
up nuclear stains diminishes.

At this point it will be convenient to examine some of the
various mechanisms which have for their object the nutrition
of the rapidly-growing oocyte. In each ease it will be found
that the main object of the mechanism is to increase the surface
in contact with the cvtoplasm of the oocyte, in order to facilitate
the passage of nutritive material into the latter.

The activities of the oocyte nucleus in Colymbetes
fuseus as described by Will (84) are in many ways not unlike
those of the insect under consideration. When the oocyte
enters on its period of rapid growth the nuclear membrane
becomes irregular and finally many small branches pass into
the cytoplasm. Later these become separated from the rest
of the nucleus, and are used directly as nutritive material by
the cytoplasm. A fresh nuclear membrane develops behind

be ee Le eee

th
.

aaa ad

OVARY AND OVARIAN EGG OF ANOPHELES 42.9

the separated branch, and the nucleus then produces more
branches which in their turn become separated, so that ‘ der
protoplasmatische Leib der Eizelle auf Kosten des Hikernes
wichst >. This, however, is not a case of the degeneration of
the nucleus, as it continues to increase in size while it is giving
up these portions of its substance. Therefore this mechanism
of nutrition is practically the same asin A. maculipennis,
except that in this case the nucleus continually passes portions
of its substance into the growing oocyte as nutritive material
instead of merely conducting nutritive fluid to the oocyte.

In Calliphora erythrocephala Lowne (16) describes
another manner by which the oocyte receives portions of the
nucleus as nutritive material. He says that “When the egg
is enlarged to about two-thirds of its maximum size the granules
in the largest nucleus appear to stream out, the nucleus itself
shrivels and is ultimately lost, whilst the whole protoplasm
of the cell assumes a granular yolk-like appearance, in which
the nuclear granules can no longer be distinguished’. The
‘largest nucleus ’ is evidently the oocyte nucleus, the remainder
being those of the nurse-cells. A similar passage of granules
from the nucleug has been observed in the oocytes of many
insects.

A modification of this process of nutrition has been observed
by Gatenby (8) in the oocyte of Apanteles. Inthis, minute
solid chromatoid granules first appear, and later a nuclear
membrane appears round each of these. These grow and a lirin
network appears, and the larger nuclei so formed resemble the
true oocyte nucleus to the smallest details. These secondary
nuclei disappear when nutrition is complete.

In Rhizotrogus solstitialis Rabes (24) describes
a very different mode of nutrition. In this the nutrition of the
oocyte is not confined to the nucleus and nurse-cells, the
follicular epithelium playing an important réle. As the oocyte
grows the epithelium forms folds which penetrate into the
volk-mass, often as far as the middle of the oocyte, an excellent
example of the tendency to increase the surface of contact
between the oocyte and secretory structure.

4380 A. J. NICHOLSON

Finally, we may consider the cases in which ‘ yolk nuclei ’
form part of the nutritive mechanism. It is evident that this
collective term includes several distinct types of structures,
and | will only deal with one of these, the Corpuscles of Bal-
biani. The origin of this body is obscure in most types which
have been examined, but Chubb (4) shows very clearly that
in the oocyte of Antedon this body arises in the nucleolus
as a series of deeply-basophile spherules which are passed
into the cytoplasm. These form a mass just outside the
nucleus, and eventually they fuse to form the yolk nucleus.
MeGill (19) describes a similar aggregation of granules close
to the nucleus in the oocyte of the dragon fly, and this gives
rise to the yolk nucleus. Though she has been unable to
demonstrate the origin of the granular mass she shows that it
is very probably nuclear in origin, and in support of this theory
remarks that “ Hennegay (1893) believes that the Corpuseles of
Balbiani in Vertebrates are either parts of the nucleolus or the
entire nucleolus which passes through the nuclear wall into
the cytoplasm ’.

Similarly Bambeke (1) observes that the ‘corps vitellin ’
of Pholeus phalangioides arises close to the germinal
vesicle, and he considers that it is nuclear in origin. He shows
that this grows into a large and somewhat branched structure
which takes an active part in the nutrition of the oocyte.
This structure bears a considerable superficial resemblance
to the branched nucleus of A. maculipennis, and a careful
consideration of Bambeke’s very excellent paper has led me
to the conclusion that the resemblance is not merely superficial
but that the two structures are both morphologically and
physiologically comparable. It should be noted here, however,
that Chubb (4) considers that the yolk nucleus of Antedon
has no connexion with yolk formation though it is almost
identical in every respect with the yolk nucleus of Pholeus.
He gives a perfectly simple physical explanation for the
changes undergone by this structure, which he regards as waste
material forming a purely passive body.

We have seen. that the branched nucleus of A, maculi-

-_

eA | Oe

4
.
°
.
i
al
4
.

OVARY AND OVARIAN EGG OF ANOPHELES 431

pennis is almost entirely the product of the nucleolus. Now
Bambeke considers that the yolk nucleusof Pholeuws is nuclear
in origin, and other authors are of the same opinion with regard
to other animals. Thus Korschelt (18) observes, ‘ Wenn man
sieht, welche complicierte Gestaltung dem aus concentrischen
Schichten gebildeten Dotterkern mancher Spinnen zukommt,
méchte man ihn fiir einen bedeutungsvollen Bestandtheil des
Kernes halten und ihn gewiss nicht mit demsoeben besprochenen
* Dotterkern ’’ der Amphibien zusammenwerfen.’

In further support of the theory of the nuclear origin of the
Corpusceles of Balbiani, Bambeke remarks: ‘ Des que la forme
de batonnet a fait place a celle de bourrelet ou de cupule,
la constitution du corps vitellin se montre trés semblable, voire
méme identique, a celle de la tache germinative. . . . Cette
frappante analogie entre la constitution de ces éléments ne four-
nit-elle pas un argument de plus en faveur de lorigine nucléaire
du corps vitellin ? ’

Having shown that the body with which he is dealing is
probably nuclear in origin and is comparable to the nucleolus,
Bambeke proceeds to give his reasons for believing that the
body is a true ‘corps vitellin de Balbiani’. These may be
summarized as follows :

1. Situation near gerniunal vesicle.

2. Affinity for colours similar to that of the nucleolus.
3. Presence of vacuoles.

. Constancy of the character.

. Appearance at commencement of growth.

6. Final degeneration.

All these characters are also true of the branching nucleus
in A. maculipennis except that in nos. 1 and 2 snmilaritiy
of position and character is replaced by identity. he presence
of vacuoles is only found in the earliest stages of the nucleus,
but this is not actually an important difference from the yolk
nucleus of Pholeus, as in the latter the vacuoles disappear
before it degenerates, so that actually this is a further indication
of the similarity existing between the two structures. Is it

482 A. J. NICHOLSON

not reasonable, therefore, to consider the yolk nucleus of
Pholeus and the branching nucleus of A. maculipennis
as being homologous structures which only differ in that the
one passes to the outside of the nuclear membrane while the
other remains inside ?

I have already shown that the branching nucleus of the
oocyte in A. maculipennis can only be regarded as
a structure the function of which is to carry nutritive material
to the various parts of the developing oocyte. After an
exhaustive consideration of the various theories as to the fune-
tion of the yolk nucleus Bambeke comes to the conclusion that
the only one which can be adopted in the case of Pholeus
‘est celle qui considére ce corps comme centre de formation
des éléments nutritifs du vitellus ’.

Tor these reasons I have come to the conclusion already
stated that the branching nucleus of Anopheles
maculipennis and the yolk nucleus of Pholeus
phalangioides are morphologically and physio-
logically comparable. These structures are homologous
with other types of oocyte nuclei and Corpuscles of Balbiani
respectively. It would therefore appear that the Corpuseles of
Balbiani may be considered as portions of the oocyte nucleolus
which have escaped through the nuclear membrane in order
to carry on the nutritive portion of the nuclear functions.

In Pholcus the division of the nucleus into two portions,
one nutritive or vegetative and the other germinal, is only
partial, as the germinal vesicle itself appears amoeboid and
evidently takes part in the nutrition of the oocyte.

In A. maculipennis it has been shown that from an
early stage the nuclear contents are sharply divided into
aw vegetative and a germinal portion, the nucleolus and
chromatin residue respectively. During the resting stage there —
may be an apparent fusion of the two, but actually they are
only closely applied together, the chromatin residue lying
in an identation of the nucleolus. A close parallel is found in
the ovary of the dragon fly according to McGill (19). In this
case the thick spireme of the young oocyte surrounds the

OVARY AND OVARIAN EGG OF ANOPHELES 433,

nucleolus, giving rise to a ‘ double nucleolus’. Later one side
of the nucleolus is formed of chromatin and the other is the
plasmosome.

Gatenby (8) shows that in Apanteles glomeratus the
division of the oocyte nucleus into germinal and vegetative
parts takes place in a very different manner. Secondary nuclei
are produced, apparently arising from material which has
escaped from the true oocyte nucleus, aud these are found round
the periphery of the oocyte. Then ‘some time before the
ovarian oocyte has become ripe the secondary nuclei disappear
by a process of degeneration or chroniatolysis ’. The secondary
nuclei are considered to influence the production of yolk.
Discussing this subject Gatenby remarks: ‘The egg nucleus
of many insects, of which A panteles is an example, becomes
partly decentralized ; this is te say, the nucleus, instead of
influencing various processes of oogenesis from afar, sends
pieces of itself into the furthermost regions of the egg, which
carry on part of the vegetative functions at least of the
chromatin of the ordinary nucleus.’ This statement applies
equally well to the oocyte nucleus of A. maculipennis,
though the pieces sent ‘into the furthermost regions of the
egg ’ remain attached to the rest of the nucleus.

It has already been shown that, though there is good reason
to believe that the ‘chromatin residue’ gives rise to the
segmentation nucleus, there is a period in which no chromatin
matter can be distinguished, and the oocyte of the mosquito
then appears to be without a nucleus. A similar phenomenon
has been encountered in the oocytes of other insects by many
observers. Will (84) states that the oocyte nucleus of Dytis-
cus becomes a mass of fine granules from a small portion
of which the ‘ definitive Kern’ is later produced. Lowne (16),
speaking of Calliphora erythrocephala, remarks,
‘In the ripe unimpregnated ovum I have entirely failed to
find any nuclei or cellular elements of any kind, and I feel sure
that if any such elements were present they would readily
be distinguished in wy sections’. Lubosch (17) states that
this disappearance of the staining portions of the oocyte

43 A. J. NICHOLSON

nucleus for a certain period is the rule rather than the exception
in animal eggs, and Doncaster (5) makes the following observa-
tion on the subject : *‘ Very commonly the chromosomes .. .
disappear, and the chromatin becomes scattered through the
nucleus in the form of fine particles, or for a time it may
vanish altogether, at least in the sense that it ceases to take
up stain.’

The production of the segmentation nucleus at about the
period when the egg is laid is the normal occurrence in insect
eggs, and it is quickly followed by the polar divisions. Don-
caster (5) observes that ‘in some animals the act of laying
seems to be the stimulus and in others the polar division only
occurs when a spermatozoon enters the egg’; but as in
A. maculipennis oviposition and fertilization are simul-
taneous, it cannot be stated which acts as the stimulus.

In conclusion, the more important points with regard to the
oocyte nucleus of A. maculipennis may be summarized
as follows :

1. Froin the earliest stages separate vegetative and germinal
portions can be distinguished in the oocyte nucleus.

2. During the second period of growth the nucleus branches
throughout the entire oocyte.

3. The branching nucleus, in conjunction with the nurse-
cells, takes an active part in the nutrition of the oocyte.

4, The branching nucleus is almost entirely the product of
the nucleolus.

5. The branching nucleus is morphologically and physio-
logically comparable to the Corpuscles of Balbiani of other
animals.

6. The germinal portion of the nucleus, the ‘ chromatin
residue’, is the product of the condensation of the spireme
threads,

7. The ‘chromatin residue’ becomes invisible for a short
period and reappears after oviposition as the segmentation
nucleus.

se ee ee

OVARY AND OVARIAN EGG OF ANOPHELES 435

IV. Development of the Outer Wall.

The first portion of the outer wall to appear is that which
forms the floats. This is secreted between two layers of
epithelial cells which come to he one above the other by a very
specialized form of folding of the epithelium.

During the earlier stages of the growth of the follicle the
epithelium is of a typically cubical form, but later the cell
divisions in two lateral areas become oblique, the obliquity
being more marked towards the centre of each area. This
process continues with further growth of the follicle (fig. 34)
till one much elongated cell hes over the top of several (fig. 35).
The underlying cells, however, do not lose their connexion with
the tunica propria, but remain attached to it immediately in
front of the end of the overlying cell. Finally, it is found that
in the two lateral areas there are groups of very much elongated
cells which lie almost parallel to the tunica propria. The float
is secreted between the outermost of these and the one lying
immediately under it (fig. 35). Hach corrugation of the float
is produced by the secretion of the chorion over the outer surface
of one of the much elongated underlying cells.

It will be seen that this overlapping arrangement of the
follicle cells is practically a fold of the epithelium. It is not
an ordinary epithelial fold, however, as the basement mem-
brane, i.e. the tunica propria, is not disturbed and does not
take any part in the folding.

The remainder of the wall makes its appearance shortly
after the commencement of the formation of the floats. It is
first seen as a simple and very thin membrane lying immediately
under the follicular epithelium. Soon lecal thickenings are
found on this membrane (fig. 35). These are the commence-
ment of the processes. The thickenings become larger and
grow into the cytoplasm of the epithelial cells. Numbers of
such thickenings are formed under each epithelial cell, and
the shape of the processes cannot therefore be determined by
the form of the secreting cells in the manner which frequently
occurs, e. g. in the corrugations of the floats.

NO. 259 Gg

436 A. J. NICHOLSON

The thin membrane of the outer wall does not appear to
increase appreciably in thickness, but the processes grow far
into the cytoplasm of the epithelial cells till they reach their
final size and form. ‘The bosses, in spite of their large size,
arise in exactly the same manner as the rest of the processes.

The epithehal layer now undergoes degeneration and becomes
separated from the processes till it forms a layer lying over the
top of these. Degeneration proceeds till only irregular masses
of flattened nuclei can be seen attached to the inner side of the
tunica propria (fig. 41), which forms a thin sheath round the
whole egg.

V. Development of the Micropyle Apparatus.

The first indication of a special structure being produced
for the formation of the micropyle apparatus appears when
the egg is about a third of its full size, at the period when the
inner wall is beginning to form as a definite layer. At this
stage the epithelial cells immediately surrounding the point
where the funicle of the secondary follicle joins the primary
ovarian follicle become somewhat larger than their neighbours
and protrude shghtly inwards towards the nurse-cells (fig. 36).

As the egg increases in size this inward protrusion becomes
more marked, particularly in the case of the peripheral cells of
the group. Finally, the latter are pushed completely inside
the epithelial layer and lie between the nurse-cells and the
epithelium (fig. 37).

If examined from a surface view these extruded cells are seen
to radiate from a common centre, in the form of a rosette,
and for that reason | propose to refer to them as rosette-cells
(fig. 38).

At this period the cytoplasm of the nurse-cells is seen to be
rapidly breaking down and disappearing, and also the contents
of the nuclei are degenerating. The chromatin strands lose
their definite structure and gradually become a_ shapeless
mass and the nucleoli undergo fragmentation (fig. 37).

The cytoplasm of the rosette-cells becomes very closely

OVARY AND OVARIAN EGG OF ANOPHELES 437

applied to the nurse-cells and gives the appearance of ingesting
them.

The nurse-cells, which consist merely of degenerating nuclei
invested by the cell membrane, now pass into the epithelium,
in which they lie till they become completely degenerated.
It will be seen from fig. 39 that they have every appearance
of being ingested by an epithelial cell, i.e. a rosette-cell, though
I have been unable to demonstrate that they are completely
surrounded by the cytoplasm of the rosette-cells. his is not
surprising as, owing to the large size of the nurse-cell and the
comparatively small size of the rosette-cell, the layer of
cytoplasm of the latter surrounding the former would of
necessity be exceedingly thin, and would be very difficult
to distinguish from the nurse-cell membrane or from the
surrounding epithelial cells.

Whether the degenerating nurse-cells are completely ingested
by the rosette-cells or not, it is certain that there is a very
intimate relation between the two, and the latter invest a
considerable portion at least of the former. The degenerating
nurse-cell nuclei would appear to form a general food reserve
which is used by the rosette-cells while forming the micropyle
apparatus.

The large size of the nurse-cells causes the radial arrangement
of the rosette-cells to appear distorted, though indications of
this arrangement can always be made out.

The micropyle apparatus arises under the rosette-cells at
the same time that the rest of the chorion appears. The whole,
with the possible exception of the narrow portion of the
funnel, is secreted by the rosette-cells, and there is no obvious
mechanism to account for the secretion of the thick supporting
ring by part of the surface and the thin disk by another.

The bases of the epithelial cells which are surrounded by the
rosette-cells pass as fine threads down the funnel, and it is
probably these that secrete the funnel, though the bases of the
rosette-cells certainly reach the top of the funnel and may pass
down it (fig. 39).

As the stopper appears to be a definite portion of the micro-

Gg2

438 A. J. NICHOLSON

pyle apparatus it will be convenient to describe its origin
here.

When the rosette-cells are arising from the epithelial cells
and are just protruding slightly towards the nurse-cells, globules
of matter are appearing between the epithelium and the oocyte
over the whole follicle with the exception of this one point.
These globules are the commencement of the mner wall.

If the protruding group of cells is examined carefully it will
be found that there are globules opposite the central cells of
the group (fig. 36). These are the beginning of the stopper and
are exactly the same as those which are giving rise to the inner
wall. The only point in the egg, therefore, where this secretion
is not taking place is a ring corresponding with the rosette-
cells (fig. 37).

As the egg grows this secretion continues till a well-formed
inner wall and a definite mass of similar matter, the stopper,
has appeared.

After the extrusion of the nurse-cells the inner wall narrows
the hole through which they have passed, only leaving sufficient
room for the passage of the funnel, and in so doing passes
over the stopper, so that this now takes up a position imme-
diately beneath the micropyle (fig. 39).

A very similar process of development is described by
Gross (9) for the micropyle apparatus of Xanthogramma
citrofasciata. In this a special group of epithelial cells
is detached from the anterior pole of the egg, and this travels
between the nurse-cells and finally comes to rest immediately
under them. ‘The follicle epithelium grows inward and separates
the group of nurse-cells from the oocyte except in the region
of the detached group of cells. By the time this is completed
the nurse-cells have passed most of their cytoplasm into the
egs-chamber, so that a mass consisting practically only of
nurse-cell nuclei lies over the anterior end of the egg. The
croup of cells secretes a *‘ polsterformiges Gebilde ’, and the rest
of the follicular epithelium secretes the exo- and endo-chorion.
This * polsterf6rmiges Gebilde > comes to lie immediately under
the micropyle apparatus, and is perforated by the micropyle.

a

OVARY AND OVARIAN EGG OF ANOPHELES 439

It is interesting to note the different manner in which the
specialized group of epithelial cells are produced, and the
degenerating nurse-cells passed out of the egg chamber in this
insect and in A. maculipennis.

The ‘ polsterf6rmiges Gebilde’ of Xanthogramma andthe
‘stopper’ of A. maculipennis are probably homologous,
as they are produced in a similar manner by a specialized
group of epithelial cells, and they are also similar in appearance
and position. There is one noticeable difference, however: in
Xanthogramma the structure is pierced by the micro-
pyle, while in A. maculipennis it appears to be solid, the
micropyle terminating immediately above it.

VI. Development of the Inner Wall.

When the egg-follicle has reached about a third of its ultimate
size small globules of matter are found between the follicular
epithelium and the oocyte. These are deeply stained by
haematoxylin and can be readily distinguished from the yolk
granules. The globules increase in number and size and finally
fuse, forming a coat investing the entire oocyte, with the
exception of a ring-shaped area under the rosette-cells.

It has already been shown that the inner wall is gelatinous
in nature till some time after the egg has been laid, and when
in this state rapidly swells in the presence of acids. It is
therefore not surprising that this structure becomes very
much distorted during fixation. In fig. 36 the inner wall is
shown as a fibrillar structure, the fibrils stretching across the
space between the oocyte and the follicular epithelium. This
is a very common appearance of the inner wall in follicles
of about this stage of development, and I regard the fibrils as
being produced from globules which adhere to both the oocyte
and follicular epithelium and become stretched into threads
when these become separated. In eggs nearing maturity the
inner wall appears to be a thick homogeneous layer lying under
the follicular epithelium and in it large vacuoles are frequently
seen, but the layer does not show any signs of fibrillar structure.

44() A. J. NICHOLSON

I consider that the substance of the inner wall is no longer
in globules but has formed a continuous gelatinous layer.
Obviously a fibrillar structure could not be produced from
such a layer im the manner described above.

When the egg is freshly laid the inner wall is still a thick
gelatinous structure, but after some hours it hardens and in
sections is seen to form a thin dark-coloured membrane lying
immediately under the outer wall.

As the inner wall appears between the oocyte and the
follicular epithelimm the question arises as to which of these
secretes it. The cytoplasm of the oocyte is already occupied
in the production of yolk and the follicular epithelium secretes
the outer wall at a later period, so that whichever of these
structures form the inner wall is also capable of producing an
entirely different substance.

Over the greater part of the egg it is impossible to determine
whether the inner wall is secreted by the follicular epithelium
or the oocyte; but the stopper, which is merely an isolated
portion of the inner wall, is formed between the follicular
epithelium and the nurse-cells. The inner wall must therefore
be secreted by the follicular epithelium, and after this has
been produced the epithelium changes its form of activity
and secretes the outer wall.

DEGENERATING EGG-FOLLICLES.

The degeneration of a certaim number of egg-follicles seems
to be a normal occurrence in the ovary of A. maculipennis.
Commonly this degeneration takes place when the follicles are
just entering on the second period of growth, but not infre-
quently at a much earlier stage the primary follicles are found
to be represented by a small mass of degenerated cells sur-
rounded by a loose and much-folded tunica propria. ‘The
significance of this degeneration is not clear. I have been
unable to detect the presence of any bacteria or other organisms,
and the fact that degenerating follicles are almost invariably
to be found in small numbers in ovaries, but that all, or even

OVARY AND OVARIAN EGG OF ANOPHELES 441

a large part, of the follicles of an ovary have never been found
affected, suggests that the phenomenon should be considered
as one of atrophy or auto-digestion rather than as a disease.
When an ovary is developing the follicles are very crowded
and are obviously under compression, and it is probable,
therefore, that the removal of several of the follicles from the
more crowded parts would benefit the remainder. This may
account for the degenerating follicles, but there is nothing but
the above consideration to support the theory.

Fig. 33 shows part of the degenerating epithelium of a follicle
which has Just commenced to produce yolk. It will be noticed
that the appearance of degeneration is confined to the epithe-
hum. This is normally the case, and it is only after the
epithelium has almost broken down that the central mass of
cells degenerates. Each epithelial cell produces one or more
large globular masses inside the inwardly-directed portion of
its cytoplasm, so that it closely resembles a goblet cell. The
masses are very variable in appearance as they stain very
irregularly. They are commonly very granular but are other-
wise structureless. The thin protoplasmic investment of the
globules soon breaks down, so that the globules form a mass
which penetrates amongst the nurse-cells.

The mass of cells and globules appears gradually to enter into
solution, as it decreases in size till nothing but a few degenerat-
ing nuclei and a very loose tunica propria remain to indicate
the position of the original follicle (fig. 24).

PRESENCE OF SPOROZOA AND BacTERIA IN EGG-FOLLICLES.

As has already been observed by 8. R. Christophers (2), the
yolk of a mosquito egg is frequently entirely displaced by
a mass of sporozoa. These appear as transparent spherical
eysts 0-005 mm. in diameter, approximating in size to the
coarse yolk granules, in which eight small bodies which take
up stain are found (fig. 10). In sections this number is not
constant, but there are never more, and the reduced number
is probably due to the removal of part of the cyst. This is the

449, A. J. NICHOLSON

only stage of the organism which I have observed and, though
a number of insects were found affected, the cysts were only
observed in mature oocytes.

The nurse-cells of the ovary of one insect were found to be
heavily infected with diplococei. The follicles were nearly
fully developed, and T could observe no harmful effect of the
bacteria. The infection appeared to be entirely confined to the
nurse-cells.

SUMMARY.

1. The period at which the ovaries of A. maculipennis
commence to develop depends on the season and. locality.
Normally this is from about the middle of March to the begin-
ning of April.

2. A meal of blood appears to be necessary for the production
of eggs.

3. One meal of blood is sufficient to cause eggs to be produced.
After the lapse of a day the large blood-mass in the stomach
shows two zones: a posterior partially-digested portion and
an anterior portion of apparently fresh blood. ‘This appearance
has sometimes been taken as evidence that more than one meal
of blood has been consumed.

4, The eggs are fully developed six days after the insect has
fed on blood.

5. In the case of two insects which were observed at the time
of oviposition the eggs were laid immediately after dark.

6. The muscle-bands of the ovarian sheaths are striped ;
not unstriped as is usual in insects.

7. A large number of vacuolated cells are found in the ovary.
The nature of these is not clear, but they appear to have some
relation to the tracheal system.

8. The chorion of the egg is highly specialized to retain
air round the egg, and the buoyancy of the egg is entirely due
to the entrapped air.

9. The floats are produced by a very specialized form of
folding of the follicular epithelium.

+ 0) at agg es:

OVARY AND OVARIAN EGG OF ANOPHELES 448

10. The micropyle apparatus is produced by specialized cells
of the epithelium, the ‘ rosette-cells ’.

11. Immediately below the micropyle is a specialized portion
of the inner wall, the ‘ stopper ’.

12. The inner and outer walls of the egg, though formed of
entirely different substances, are hoth secreted by the follicular
epithelium.

13. The inner wall is first gelatinous in nature and trans-
parent ; but, after the egg is laid, becomes brittle and dark
in colour, causing the egg to appear black. This change in
character only takes place when the eggs are laid on water.

14. The mature egg contains two distinct kinds of yolk, one
of large granules which are proteid in nature, and the other
of small granules the nature of which I have been unable to
determine.

15. There are two distinct periods of growth of the egg-
follicles, the first culminating in the ‘ resting stage’ and the
second only commencing after the mosquito has bad a meal of
blood.

16. Each egg-follicle consists of a follicular epithelium
surrounding seven nurse-cells and an oocyte. These appear
to be the product of a single oogonium.

17. The cells of the follicular epithelium multiply by mitotic
division during the whole of the first period of growth. In the
second period, theugh the follicular epithelium increases
greatly in area, this is due purely to the increase in size of the
individual cells.

18. From the earliest stages separate vegetative and germinal
portions can be distinguished in the oocyte nucleus.

19. During the second period of growth the oocyte nucleus
branches throughout the entire oocyte.

20. The branching nucleus, in conjunction with the nurse-
cells, takes an active part in the nutrition of the oocyte.

21. The branching of the nucleus may be regarded as
a mechanism for the purpose of increasing the surface.

22. I have observed a similar method of branching of
the oocyte nucleus in Theobaldia annulata, and it

444 A. J. NICHOLSON

probably also exists in Necrophorus vespillo and
Silpha sp.

23. The branching nucleus is almost entirely the product
of the nucleolus.

24. The branching nucleus is morphologically and physiologi-
cally comparable to the Corpuscles of Balbiani of other animals.

25. The germinal portion of the nucleus, the ‘ chromatin
residue ’, is the product of the condensation of the spireme
threads.

26. The ‘ chromatin residue’ becomes invisible for a short
period and reappears after oviposition as the segmentation
nucleus.

27. The chromatin of the active nurse-cells consists of
minute bars situated on a much convoluted band of linin.

28. Degeneration of a certain number of egg-follicles is
normal during the development of the ovary.

29. Sporozoa are frequently found in the eggs, often -com-
pletely replacing the whole of the yolk.

List oF LirERATURE.

1. van Bambeke.—“ Recherches sur Vodcyte de Pholcus phaian-
gioides’”’, ‘ Arch. de Biol.’, t. 15, 1897.

2. Christophers. S. R.—‘ The Anatomy and Histology of the Adult
Female Mosquito’’, ‘Roy. Soc.. Reports to the Malaria Com-
mittee ’, Fourth Series, London, 1901.

3. —— “The Development of the Egg-follicle in Anophelines ”’, ‘ Palud-
ism’, Simla, pt. 2, 1911.

4. Chubb, Gilbert C.—‘‘ The Growth of the Oocyte in Antedon”,
‘Phil. Trans. of the Roy. Soc. of London’, Series B. vol. 198, 1906.

5. Doncaster, L.—‘* An Introduction to the Study of Cytology’, Cam-
bridge, 1920.

6. Dublin, L. J.—‘‘ On the Nucleoli in the Somatic and Germ Cells of
Pedicellina americana”, ‘ Biol. Bull.’, vol. 8, 1905.

7. Gatenby, J. Bronté.—* The Identification of Intracellular Structures ”’,
‘Journ. Roy. Micr. Soc.’, 1919.
8. —— ‘The Cytoplasmic Inclusions of Germ Cells’, pt. iv. ‘ Quart.

Journ. Micr. Sci.’, vol. 64, pt. 2, 1920. ;
9. Gross, J.-—‘‘ Untersuchungen iiber die Histologie des Insecten-
ovariums ”’, ‘ Zool, Jahrb. Jena’, vol. 18, Heft 1, 1903,

10.
a;

12.

13.

14,

15.

16.

Li

18.

19.

20

21

22.

23.

24.

25.

26.

OVARY AND OVARIAN EGG OF ANOPHELES 445

Hennegay.—* Les Insectes ’, Paris, 1904.

Kepner, William A.—‘‘ Nutrition of the Ovum of Scolio dubia”,
* Journ. Morph. Philadelphia, Pa.’, vol. 20, 1909.

Korschelt und Heider.—* Lehrbuch der vergleichenden Entwick-
lungsgeschichte der wirbellosen Thiere’, vol. 1.

Korschelt, E.—*‘ Beitrage zur Morphologie und Physiologie des
Zellkernes ’’, “ Zool. Jahrb. Jena’, vol. 4, 1889-91.

—— “Ueber die Entstehung und Bedeutung der verschiedenen
Zellelemente des Insectovariums’’, ‘Z. fiir wiss. Zool.’, Bd. 43.
1886.

Lécaillon, A.—*“‘ Recherches sur la structure et le développement
postembryonnaire de l’ovaire des Insectes, 1. Culex pipiens”,
* Bull. Soe. Ent. France ’, 1900.

Lowne, B. T.—*‘ On the Structure and Development of the Ovaries
and their Appendages in the Blowfly (Calliphora erythro-
cephala)’’, ‘ Journ. Linn. Soc. Zool.’, vol. 20.

Lubosch, Wilhelm.—‘‘ Uber die Hireifung der Metazoen, insbesondere
iiber die Rolle der Nucleolarsubstanz und die Erscheinungen der
Dotterbildung”’; ‘‘ Ergebnisse der Anatomie und Entwicklungs-
geschichte ’’, “ Anat. Hefte ’, Wiesbaden, Abt. 2, 1902.

Marshall, Wm. S.—*‘ The early history of the cellular elements of the
ovary of a Phryganid, Platyphylia designatus Walk.’
‘Z. fiir wiss. Zool. Leipzig’, Bd. 86, 1907.

McGill, Caroline.—** The behaviour of the nucleoli during the oogenesis
of the dragon fly, with special reference to synapsis ”’, “ Zool. Jahrb.
Jena *, Abt. fiir Anat. 23, 1906.

Montgomery, Thomas, H.—** Comparative Cytological Studies, with
especial regard to the Morphology of the Nucleolus”’, ‘ Journ,
Morph. Philadelphia, Pa.’, vol. 15, 1898-9.

Newth, H. G.—*‘ On the Orientation of Minute Objects for the Micro-
tome ”’, ‘ Quart. Journ. Micr. Sci.’, vol. 63, pt. iv, 1919.

Paulcke, W.—‘‘ Uber die Differenzirung der Zellelemente im Ovarium
der Bienenkénigin (Apis mellifica)”, ‘Zool. Jahrb.’, vol. 14,
1900.

Payne, Fernandus.—*‘ The nucleolus in the young oocytes and origin of
the ova in Gelastocoris”, ‘Journ. Morph. Philadelphia, Pa.’,
vol. 23, 1912.

Rabes, O.—‘‘ Zur Kenntniss der Eibildung bei Rhizotrogus
solstitialis L.”, ‘ Z. fiir wiss. Zool.’, Bd. 67, 1900.

Seeliger, Oswald.—-“* Einige Bemerkungen iiber den Bau des Ruder-
schwanzes der Appendicularien ”’, * Z. fiir wiss. Zool.’, Bd. 67, 1900.

Sen, 8. K.—** Beginnings in Insect Physiology and their Economic
Significance”, ‘ The Agricultural Journal! of India’, vol. 18, pt. 4,
1918,

446 A. J. NICHOLSON

27. Sheppard, E. J.—“‘ Two Valuable Methods of Staining in Bulk and
Counter-Staining ”, ‘ Journ. of the Roy. Micr. Soc.’, 1918.

28. Soyer, C.—‘‘ Nouvelle série de faits cytologiques relatifs 4 ’ovogenése
des insectes ”’, ‘C. R. Soc. Biol. Paris’, vol. 63, 1907.

29. —— ‘“‘Considérations sur les cellules folliculeuses et certaines homo-
logies de lovaire des insectes ”’, ibid., vol. 63, 1907.

30. Taylor, Monica.—“ The chromosome complex of Culex pipiens.
Pt. 2. Fertilization’, ‘ Quart. Journ. Mier. Sci.’, vol. 62, pt. iii,
1917.

31. Vaney, C.—‘ Contributions 4 l'étude des larves et des métamorphoses
des diptéres ’’, “ Ann. de Univ. de Lyon ’, nouvelle série, 1. Sciences.
Médecine, fas. 9, 1902.

32. Vaney, C., et Conte, A.‘ Evolution du vitellus dans l’ceuf du ver
a soie’’, “C. R. Soc. Biol. Paris ’, vol. 67, 1909.

33. Von Wielowieyski, R.—“‘ Morphologie und Entwicklungsgeschichte
des Insektenovariums ”’, ‘ Arb. aus dem Zool. Inst. der Univ. Wien ’,
t. 16, 1906.

34, Will, L.—‘‘ Oogenetische Studien. 1. Die Entstehung des Eies von
Colymbetes fuscus, L.”, * Z. fiir wiss. Zool.’, Bd. 48, 1886.

EXPLANATION OF PLATES.

REFERENCE LETTERS.

b.=‘ Bosses’. c.c.=Central cavity of nucleolus. c¢.l.=Cortical layer of
nucleolus. ¢7.=‘Chromatin residue’. ¢.y.=Coarse yolk granules.
D.= Dorsal surface. d.=Disk of micropyle apparatus. d.p.= Dorsal
processes. /.=Float. /e.=Follicular epithelium. /.m.= Follicular tube
membrane. /fu.=Funicle. /y.=Fine yolk granules. g.=Gynaecophorie
canal. 7.m.= Investing membrane. 7.w.=Inner wall. m.=Mucous gland.
m.a.= Micropyle apparatus. m.b.= Muscle bands. ».=Nurse-cells. 2.m.=
Nuclear membrane. 0.c.= Oocyte cytoplasm. 0.n.=Oocyte nucleus.
o.t,= Ovarian tube. 0.w.=Outer wall or chorion. .f.= Primary follicle.
r.c.= Rosette-cells. s.=Spermatheca. s./.=Secondary follicles. s.p.=
Suspensory filament. s.7.=Supporting ring of micropyle apparatus.
sl.=*" Stopper’. ¢.=Tracheae. t.p.=Tunica propria. V.= Ventral surface.
v.c.= Vacuolated cells. v.p.= Ventral processes. y.= Yolk.

Figs. 1-6.—-Reconstructions of progressive stages of branching of oocyte
nucleus. Same scale. Note.—Branches overlie, and do not enter nurse-
cells.

Fig. 1.—Resting stage.

Fig. 2.—Nucleus becoming irregular.

OVARY AND OVARIAN EGG OF ANOPHELES 447

Fig. 3.—Commencement of branching. Nucleus still vacuolated.

Fig. 4.—Later stage of branching, vacuoles have disappeared.

Fig. 5.—Branching nucleus in half-developed follicle.

Fig. 6.—Branching nucleus in full-sized oocyte, after extrusion of nurse-
cells,

Fig. 7.—Adult female genital organs. Ovaries in resting stage.

Fig. 8.—Ovary containing full-sized oocytes.

Fig. 9.—Egg after deposition. Anterior end at top of figure. «A. Dorsa!
view. 8. Lateral view. c. Median transverse section.

Fig. 10.—Sporozoa from yolk-mass.

Fig. 11.—Section of yolk-mass.

Figs. 12-21.—Progressive stages of oocyte nucleus. 12-16 scale of 15.
17-21 scale of 17.

Fig. 12.—‘ Bouquet stage’.

Fig. 13.—Spireme surrounding nucleolus.

Fig. 14.—-Nucleolus becoming free from spireme.

Fig. 15.—Nucleolus becoming vacuolated.

Fig. 16.—Spireme condensing.

_ Fig. 17.—Resting stage. Spireme condensing to form chromatin residue.

Fig. 18.—Commencement of second period of growth. Chromatin residue
losing staining properties.

Fig. 19.—Slightly later stage. Chromatin residue separating from
nucleolus which has practically lost vacuolated structure.

Fig. 20.—Portion of nuclear branch in half-developed follicle.

Fig. 21.—Portion of nuclear branch containing globular masses at
period when nurse-cells are breaking down.

Fig. 22.—Transverse section of half-developed follicle, showing ring-like
formation of branching nucleus.

Fig. 23.—Longitudinal section of follicle at same stage, showing posi-
tion of nuclear branches.

Fig. 24.—Longitudinal section of secondary follicle and end chamber.
Folded tunica propria left by degenerated primary follicle.

Fig. 25.—Longitudinal section of secondary follicle. One nurse-cell
nucleus contains spireme.

Fig. 26.—Longitudinal section of follicle in resting stage.

Fig. 27.—Longitudinal section of follicle at beginning of second period
of growth, showing denser cytoplasm between nurse-cells and oocyte
nucleus, commencement of inner wall and yolk production.

Fig. 28.—Musculature of investing membrane.

Fig. 29.—Muscles of investing membrane and follicular tube membrane,
showing transition to circular and longitudinal muscles of oviduct.

Fig. 30.—Nurse-cell nucleus in resting stage. Tangential section.

Vig. 31.—Transverse section of end chamber containing mitotic figure.

448 A. J. NICHOLSON

Fig. 32.—Mitotic division in follicular epithelium cells during first period
of growth.

Fig. 33.—Degenerating follicular epithelium.

Fig. 34.—Early stage of epithelial folding for float formation. Some-
what distorted section chosen as it clearly shows limits of epithelial cells.

Fig. 35.—Later stage of folding. Float and outer wall with commence-
ment of processes secreted.

Fig. 36.—Commencement of differentiation of rosette-cells and produc-
tion of ‘stopper’. Longitudinal section.

Fig. 37.—Later stage of same, Nurse-cells degenerating. Longitudinal
section.

Fig. 38.—Transverse section of rosette-cells at same stage as 37.

Fig. 39.—Longitudinal section. Degenerating nurse-cell nuclei shown
partially surrounded by rosette-cells.

Fig. 40.-Surface view of micropyle apparatus.

Fig. 41.—-Longitudinal section of anterior end of egg, showing section
of micropyle apparatus and position of ‘ stopper ’.

Quart. Journ. Miere Sc Y. Vot.09. NSCs

fmm.

o-ol|mim.

-O] MIN.

0

—__—_

il

Nicholson del.

Vol OO NS PLIS

cz

Quart Journ. Micr §

Wu sO-0

s "9?
s seatiae

Ose lila:
eH K |

O os mm.

Quart.Jowmn. Micr. Sci.

~~ mb.

Vol 69,NV-S8:, PLIPD

+2 STUER Man.

32

o-os5smm

33

Quart. Jowrn. Mier Sci. Vol 69 NS, Pt.20

0-o5mm,

35

o0-o5mm,

0-OS min,

o-osmm

41

= A a — i
| aE _
ie A, ®
5 ee n
* ;

On the Bionomics and Post-Embryonic Develop-
ment of certain Cynipid Hyperparasites of
Aphides.

By
Maud D. Haviland,

Research Fellow of Newnham College.

With 11 Text-figures.

CoNTENTS.

PAGE
INTRODUCTION . : : : : . ‘ : Z + 452
MATERIAL . P : : : F 5 , : . ~ 18)
Brotogicat Note ON THE Host . : ‘ ; : F . 453
PATRING . : H : : : : : : . 454
_OvrIposITION : ; : ; : : ‘ : . 455
THE Kee . F : : ; : : - : : | 456
THE EMBRYONIC MEMBRANE : ‘ é ; j F . 458
THE First INstTar : : ; ‘ ; : : : - 461
THE SECOND INSTAR . : s 2 : : : : 46
INTERMEDIATE STAGES : : , : F : : . 464
THE FULL-GROWN LaRva . : : : ; A : . 465
PUPATION AND EMERGENCE . F : ; ; : : . 468

CoMPARISON OF THE LARVAL CHARACTERS OF Charips WITH
THOSE OF OTHER ENTOMOPHAGOUS CYNIPIDAE . ; F aoo

CoMPARISON OF THE Larval CHARACTERS OF Charips WITH
THOSE OF Parasitic HYMENOPTERA IN GENERAL . P . 470
REACTION OF THE Host... i : , ; : ‘ . 472
Economic Status iy . . ; : : ? . 474
‘SUMMARY . : 3 : : j 3 i j , s WATS
BIBLIOGRAPHY . F : . : ‘ , : i - 476

NO. 259 Hh

452 MAUD D. HAVILAND .

INTRODUCTION.

Tue biology of the entomophagous Cynipidae, which include
the sub-families of Encoilinae, Figitinae, and Chari-
pinae, has been little studied. The Knceoilinae and
Figitinae are known to be parasitic chiefly upon Diptera.
The Charipinae have hitherto been reared from Aphididae,
and occasionally from Coccidae; but no account of their
development has been published, and systematic workers have
described them indifferently as parasites or hyperparasites.
It is probable that the latter view will prove correct for the
majority of the sub-family.t

The following is an account of the bionomics of certain of
these Cynipidae, of the genus Charips. This was formerly
known as Allotria, but in 1910, Kieffer (19) reverted to
the name originally given by Haliday in 1870, and his termino-
logy has been followed here. The genus is divided into twe
sub-genera, Bothrioxysta, Kieff., and Charips, Hal.
The majority of individuals reared from material collected
in the field in the course of this work were of the species
Bothrioxysta curvata, Kieff.; but a few examples
of Charips victrix, Westw., and of another genus,
Alloxysta erythrothorax, MHartig, were obtained.
No distinction was observed between the larval forms, which
is not surprising where the specific distinctions of the adults
are variable and slight. It is even possible that certain forms,
now ranking as species, may not be physiologically distinct ;

for in one instance, in captivity, a male of Alloxysta

erythrothorax appeared to mate witha female of Charips
vietrix, which afterwards oviposited.

Hence throughout this work it has been thought most con-
venient to use the generic name, Charips, when speaking

1 Silvestri (23), in a foot-note to his work on Encyrtus aphidi-
vorus, remarks that Allotria (Charips) is a hyperparasite of
aphides through Aphidius (Braconidae); and he adds that it lives
upon the host internally, an observation which has been neglected by
writers, both before and since.

© oe ee Ci.

DEVELOPMENT OF CYNIPID HYPERPARASITES 453

generally, and to indicate the particular sub-genus or species
where necessary.

I would here express my sincere thanks to Professor J. Stanley
Gardiner for giving me facilities to carry out the work in the
Zoological Laboratory at Cambridge; and my obligations to
Professor J. J. Kieffer, who kindly determined the examples
of Cynipidae submitted to him.

MATERIAL.

The material used was obtained in Cambridge in the
summer of 1920. Charips (Allotria) has been reared
from various Aphidiidae in different aphides, but through-
out this work, Aphidius ervi, Hal., a parasite of
Macrosiphum urticae, Kalt., was used, as the com-
paratively large size of the cocoons rendered them con-
venient for dissection. The parasite and its host were
common and widely distributed round Cambridge in June
and July. Moreover, the food plant of this aphid, the common
nettle, usually grew in isolated patches along the roadside.
This was an advantage, since the Aphidius, after parasitiza-
tion by the Cynipid, is liable to secondary parasitization by
certain ecto-parasitic Chalcids and Proctotrypids, which kill
both the host and the first hyperparasite. Collections made
from one spot showed that almost every Aphidius, whether
attacked by a Cynipid or not, might bear one or more of these
external parasites ; while collections made fifty yards away
were free from secondary infestation, and contained Cynipid
larvae in all stages of development.

The rearmg methods were the same as those used when
studying Lygocerus (10). Camera lucida drawings and
measurements were made from living specimens, mounted in
salt solution or dilute glycerine. ‘The larva, and the host
when necessary, were also studied in serial sections.

Brotoetcan Nore on THE Host.

The development of the Braconid, Aphidius, within the
aphid has been described by Seurat (21), Timberlake (25), and
others.

Hh 2

454 MAUD D. HAVILAND

The egg is deposited in the haemocoele of the host, and in
the course of development a pseudo-serosa or trophic membrane
of hypertrophied cells is formed round the embryo. The
first larval stage is a transparent caudate form, which varies
somewhat in different genera, the cauda of Aphidius being
single, whereas, according to my observations, it is bifid in
Ephedrus and Praon. This appendage diminishes in
the succeeding instars, and the larva, which les curved head
to tail in the body of the host, gradually assumes the apodous
maggot-shaped form usual among hymenopterous larvae. At
first the presence of the parasite makes little difference to the
aphid, which feeds and reproduces as usual; but, as develop-
ment proceeds, degeneration of the host’s tissues sets in. The
embryos are affected first, and then the fat-body. The ‘ pseudo-
vitellus ’ or symbiotic organ is not attacked until a later stage,
and the nervous system and alimentary canal remain unchanged
until just before the Aphidius transforms, when they, in
common with the rest of the fluids of the body, are ingested by
the parasite. The tissues break down into large globules,
which in stained preparations appear as a vacuolated mesh-
work of connective tissue containing droplets of fat, while
there is often a mass of degenerating nucleoplasm in the centre
of the mass. By what means the parasite thus breaks down
the surrounding tissues is not known, but although the larva
possesses powerful mandibles, chemical rather than mechanical
action seems probable.

As soon as the Aphidius has completely emptied the body
of the aphid, it changes apneustic for peripneustic respiration,
and weaves a cocoon inside the dry skin with silk secreted by
the salivary glands.

The meconium is then voided and metamorphosis takes place.

PAIRING.

In Bothrioxysta curvata, reproduction was either
sexual or parthenogenetic according to whether a male was
introduced into the rearing-tube or not. All observed ovi-
positions of Charips victrix took place after mating, but
the ovipositions of Alloxysta were not determined.

DEVELOPMENT OF CYNIPID HYPERPARASITES 455

OVIPOSITION.

The female Charips oviposits in the Aphidius larva
only while the aphid is alive. In this it differs from other
hyperparasites, such as Lygocerus (Proctotrypidae) and
Asaphes (Chalcidae), which insert their eggs only after the
host has woven its cocoon. My observations in this respect
are opposed to those of Gatenby (8), who says: ‘ The Cynipid
parasitic forms associated with aphids apparently never attack
live Aphidae, but seek out the dried skins of those already
parasitized by an Aphidius.’

Subject to the condition that the Braconid larva shall still
be bathed in the body fluids of its aphid host, the Cynipid
has considerable latitude in its choice of a victim. The
Aphidius usually selected is in the third or early fourth
instar, but a second instar larva may be chosen (Text-fig. 3),
though in such eases there is no evidence to show whether the
hyperparasite can complete its development. The number of
eggs laid by one female appeared to be about thirty. Only
one egg is inserted at each oviposition, and others, when
found, are probably the result of subsequent attacks.

The female Cynipid runs over the plant in an excited manner,
vibrating her wings and tapping the aphides with her antennae.
Healthy specimens are ignored, but the Charips seems
to detect the presence of the primary parasite unerringly.
When she finds an aphid containing a suitable host, she leaps
on to its back, facing the head, and clings there firmly, despite
its struggles, like a rider controlling a restless horse. Sometimes
she is thrown off, but returns repeatedly to the attack until
the aphid is exhausted into passivity. The actual insertion of
the ovipositor takes from two to six minutes. This leisurely
procedure is not surprising when it is remembered that the
cuticle and body-wall of the aphid must be pierced before the
probing for the host can begin, and as the Aphidius larva
lies among the mass of aphid embryos its location can be no
easy matter. Even when found the mesenteron is so distended
with food that the body cavity is correspondingly reduced ;

456 MAUD D. HAVILAND

and if the ovipositor of the hyperparasite were to be thrust
the smallest degree too far, the egg would be inserted in the
host’s gut, and be lost at evacuation of the meconium.

THe Kee.

The egg (Text-fig. 1) is an oval body, 0-010 mm. x 0-006 mm.,
with a short peduncle continuous with its long axis. The
oogenesis was not observed, but immediately after oviposi-
tion a cloud of deeply-staining granules was visible at the
posterior pole. This may represent the germ-cell determinant,
or, as it has reeently been termed by Silvestri, the oosoma.
An oosoma in the eggs of phytophagous Cynipidae was
first described by Weismann in Rhodites rosae as
the *‘ Furchungskern’. Hegner (12) has demonstrated it in
Diastrophes nebulosus, and Hogben (14) in Synergus.
The latter says of the last-named species that the oosoma
appears as ‘a cloud of granules, more and more heavily stain-
ing, until the determinant resembles a spherical ball at the
end of the egg’. On the other hand, an oosoma has so far
not been seen in other forms, such as Rhodites ignota,
Neuroterus, Andricus, and Cynips Kolleri.

The described eggs of Cynipidae are all pedunculated, and _

in certain gall-forming species the peduncle may be five or
six times the length of the egg-body. Adler (1) first pointed
out that the peduncle is situated at the anterior pole of the egg,
which, according to him, differs in this respect from the eggs
of other Hymenoptera Parasitica. He supposed that the
function of the peduncle is respiratory, and he was supported
in this view by Cameron (8), who observed that the species
which have long peduncles are those which place their eggs
where they cannot receive much oxygen from the plant, while
in the spring generations of the same species, which oviposit
in the leaves, it is usually short. Hegner considers the peduncle
analogous to the two anterior processes of the egg of Ranatra
linearis, described by Korshelt, which float out in the water
from the plant-tissues within which the egg is placed.

A MOG tan

DEVELOPMENT OF CYNIPID HYPERPARASITES 457

The observations of Riley, quoted by Sharp (22), suggest,
however, that the peduncle may have another function. He
found that in the ovipositions of Callirhytes clavula
and Biorhiza nigra only the peduncle is inserted into the
plant at first, and that the fluids collect at the posterior end
of the egg. ‘ The fluids are then gradually absorbed from this
exposed position into the inserted portion of the egg, and by
the time the leaves have formed .. . the egg-contents are all
contained within the leaf-tissue.’

Pedunculated eggs also occur in certain Chalcids. The egg
of Leucospis gigas is furnished with a hooked process,
whose purpose is evidently to suspend it from the cocoon of the
Chalicodoma bee upon which the larva is parasitic. Imms (17)
found that the egg of Blastothrix britannica, a parasite
of Lecanium capreae, has a peduncle which protrudes
through the body-wall of the host. The tip of the process
disappears, thus putting the cavity of the chorion into com-
munication with the outside air like a siphon. Timberlake (26)
says that the egg of Microterys, parasitic upon Cocecus
hesperidium, is formed by two bodies connected by
a hollow stalk. The stalk, together with the smaller body,
projects through the body-wall of the host, and apparently
serves for the respiration of the egg and of the larva in the
early stages. The egg of Aphelinus mytilaspidis,
parasite of Lepidosaphes ulmi (16), has also a process
which, however, never projects outside the body of the host ;
and this is also the case with the egg of Comys infelix,
a parasite of Lecanium hemisphaericum, deseribed
by Iimbleton (4) as possessing a bifid process. Howard and
Fiske (15) state that the peduncles of the eggs of Schedius
kuvanae protrude through the chorion of the eggs of the
gipsy moth in which they are deposited. It may be remarked
that four of these cases are parasites of Coccidae, sedentary
animals whose metabolism and oxygen content must be low
in comparison with that of other insects. Hggs approaching
the pedunculated form occur in Eneyrtus aphidivorus,
Ageniaspis fuscicollis, &e., and here perhaps the

458 MAUD D. HAVILAND

increase of the egg’s surface, in proportion to its mass, may
bear some relation to oxygen absorption.

There is no reason why the peduncle should not in some
cases be respiratory, as supposed by Adler, and in others for
attachment, as suggested by Riley. In certain instances it
possibly serves both functions ; but its reduction in Charips
probably indicates that it has lost its use, whatever that may
have been.

TrExt-Fic. 1. TExT-FIG. 2.

Fig. 1.—The egg immediately after oviposition. x450. n=
nucleus ; g.c. = cloud of granules.

Fig. 2.—Cells of the trophic membrane with degenerating nuclei.
Afrom above; Binsection. x 350.

THe EmBryonic MEMBRANE.

In Charips, as in certain other hymenopterous parasites,
a trophic membrane or pseudoserosa is formed round the
developing embryo as a globular sphere of large eosinophil
cells, with definite nuclei and well-marked walls, polygonal in
surface view and crescentic in section (‘l'ext-fig. 2). Membranes
in this stage may be found up to the point of the hatching of the
larva, after which they soon degenerate and disappear, though
sometimes degeneration sets in at an earlier stage. A similar
degeneration can be seen also in the membrane of the
Aphidius host.

A membrane, resembling that deseribed above, has been
observed in certain Chalcids, but it does not appear to arise

we Bt ¢ G8 G4, We 9 >.4 quale tan)

DEVELOPMENT OF CYNIPID HYPERPARASITES 459

in the same manner throughout the group. Silvestri (28,
p- 67) has described its formation in Encyrtus aphidi-
vorus, Mayr., where it originates as a delamination of the
peripheral cells of the blastula. In the same work he gives an
account of its origin in Oophthora semblidis, where at
a certain point, the central protoplasm of the blastocoele
breaks out through the blastoderm, bearing with it some free
nuclei from the interior. This extruded protoplasm extends
round the egg and forms the membrane.

In 1917 Gatenby (6) eriticized the conclusions of Silvestri
with regard to the latter species. Working on the development

TEXT-FIG. 3.

Larva of Aphidius containing two embryos of Charips. x70.

of Trichogramma evanescens, a form which he later
recognized as con-generic with Oophthora, Gatenby showed
that during the formation of the blastula small masses of
nuclear matter are extruded into the blastecoele. Later,
these, with the surrounding cytoplasm, move towards the
periphery and ultimately stream out through the blastoderm.
If the chorion is ruptured, the mass floats out into the host
and soon perishes. If the chorion remains intact the extruded
mass is flattened and extended by its pressure, until it surrounds
the embryo, and the nuelei which it contains give it a fictitious
cellular appearance.

Owing to the limited material at my disposal I originally
intended to make no reference to the embryology of Charips ;

460 MAUD D. HAVILAND

but in the course of this work three stages in the formation of
the blastula were observed (Text-fig. 4), and therefore a partial
description of them is now given.

A shows the egg soon after segmentation has begun.
B represents the blastula already formed, and comparison with
the figures of Silvestri and Gatenby shows no essential differ-
ence, save that in Charips the germ-cells are indistinguish-
able from the rest of the primary layer. In c the egg is seven
hours old, and it will be seen that the nucleoplasmic masses

TEXT-FIG. 4.

Early stages in the segmentation of the egg. x 900. ch.m. = extruded
chromatin; 6l. = blastoderm; t.m. = trophic membrane.

in the blastocoele have disappeared, and that there has been
considerable displacement of the nuclei on the right-hand
side. Certain nuclei are arranged in a manner that suggests
that we have here a stage similar to that which Gatenby has
indicated as the first appearance of the endoderm. Moreover,
an involucre, apparently of cellular structure, surrounds the
egg, and contains nuclear staining elements distinct from the
degenerating chromatin masses shown in the previous figure.
As intermediate stages are lacking it is impossible to say with
certainty how this involuere arose.

Nearly all my available material was in the stage figured
as B, but the membrane did not appear in it and there was no
sign of the delamination described by Silvestriin Encyrtus.

DEVELOPMENT OF CYNIPID HYPERPARASITES 461

Moreover, the arrangement of the cells does not suggest that
they have arisen by division from the peripheral nuclei. The
disappearance of the chromatin masses seems to indicate that
there has been a recent escape of the contents of the blasto-
coele, but this matter does not appear in the involucre. It
may be represented by a small mass found in the host’s tissues
opposite the point marked z in the figure. In any case, though
Gatenby’s explanation accounts for the appearance of the
membrane in his own and in Silvestri’s figures, it does not seem
possible that the extruded matter could, under compression of
the chorion, take an outline such as that shown in Text-
fig. 4 c.

The data are too scanty to permit of our forming a definite
opinion on the origin of the involucre in these Cynipidae, but
I hope to pursue this subject later when more material is
available. Gatenby, however, remarks that in some cases
living nuclei are carried out with the extruded material :
‘Curiously enough these fragments seem to live a good while,
and nuclear changes, such as those undergone in the blastoderm,
take place in some cases.’

Without hazarding an opinion on the different views of
these observers as regards the Trichogrammatinae, a sugges-
tion may be made that if the expulsion of live nuclei were to be
carried further in Charips than it isin Trichogramma,
these might by division give rise to the membrane. But
either this division must be very rapid, to develop the nvolucre
in the space of two or three hours, or else the initial expulsion of
the living nuclei must be larger than it appears to be from
an examination of the material.

Tue First Instar (Text-fig. 5).

Dimensions, 0:38 ~ 0-13 mm. The embryonic membrane is
ruptured two or three days after oviposition. The newly-
hatched larva is heavily armoured with dark segmental plates
of chitin, which render it easily visible through the tissues
of the host. It possesses a distinct head and thirteen body-
segments, the last of which terminates in a caudal appendage.

462 MAUD D. HAVILAND

In the living larva the twelfth segment is somewhat telescoped
into the eleventh, so that only twelve segments appear to be
present. The mouth parts are produced into a proboscis,
within which lie two long slender mandibles. The head bears
three pairs of chitinous nodules on the ventral side, and, in
addition, a fourth pair dorsally. These processes are each
furnished at the extremity with a transparent spot which

TEXT-FIG. 5. TEXT-FIG. 6.

Fig. 5.—Larva of the first instar. x 150.
Fig. 6.—Anus and caudal appendage of the newly-hatched larva.
x 350. a.= anus; S. 9-12 = chitinous plates of segments 9-12.

is possibly sensory in function. ‘The body-segments diminish
in diameter from the thorax posteriorly. Hach appears as
a circular band of chitin, somewhat averlapped by the one
immediately preceding it. This overlap is so pronounced on
the ventral side of the thorax in some examples as to give the
effect of short processes; and as the latter actually appear
after the first ecdysis it is possible that they may already
exist under the chitinous plates, but at this stage it is not
possible to demonstrate their presence definitely. The anus,
which hes dorsal to the cauda, is a large and conspicuous
structure surrounded by a chitinous ring (‘Text-fig. 6). From the

ae 2 i ee

DEVELOPMENT OF CYNIPID HYPERPARASITES 463

periphery transverse bands of chitin extend into the lumen,
and give it a spiracle-like appearance. Owing to the opacity
of the chitinous coat the internal organs cannot be seen, but
the outline of the gut, which already contains food globules, is
faintly visible by transparency.

The larva is curved ventrally with the tail bent round to
form an angle with the abdomen. Its usual position is between
the nerve-cord and gut of the host, either in the anterior or
posterior third of the body. Owing to the manner in which
the Aphidius lies in the aphid these are the parts most
accessible to the ovipositor of the female Charips, and
thus the earliest larval stage is presumably found where the
ege has been deposited. The chitinized stage persists for
a variable time. In one case observed the skin had been cast
and left behmd when the larva emerged from the trophic
membrane. In others it lasted from two to four days. In the
later stages the chitin can be found among the host’s tissues.
In ecdysis the skin usually splits transversely across the thorax,
and the larva slips out. I have occasionally found examples
in the second instar in which the moult had been incom-
plete, and the body of the larva was still encircled by one or
more of the chitinous bands, like a rolled napkin enclosed
by a ring.

Tur Seconp Instar (Text-fig. 7).

The second instar resembles the first in size and general
form, but is white and transparent without thickened chitin.
The mouth is transversely oval, and furnished with two large
simple mandibles. Below it is a pair of ventro-lateral lobes
surmounted by sensory papillae. Hach of the three first body-
segments bears a pair of protuberances on the ventral surface,
and the segmentation of the body is less marked.

The internal structure is visible through the transparent
integument. ‘The salivary glands lie latero-ventrally on either
side of the midgut as two straight tubes. The nerve-cord
appears as a broad unconstricted band. The two Malpighian
tubules are very short, and immediately behind their orifices

464 MAUD D. HAVILAND

the proctodaeum is much enlarged with a bulb-shaped lumen,
communicating with the exterior by the wide anus. In some
examples newly removed from the host a transparent mem-
branous substance was seen extruded from it. When larvae
at this stage were stained with carmine or methylene blue,
it was found that the stain readily entered through the anus,
and was taken up by the lining epithelium of the hind-gut
before any other part of the body was affected.

TEXT-FIG. 7. TEXT-FIG. 8.

Fig. 7.—Larva of the second instar. x 150.
Fig. 8.—Intermediate stages of the larva. x 50.

INTERMEDIATE StaGes (Text-fig. 8).

As the larva increases in size the tail and cephalic papillae
become reduced, and the thoracic processes disappear. It
was not ascertained whether there was a moult between this
and the previous stage, or whether the change of form was due
merely to growth and absorption of the appendages ; but it
is probable that there was at least one ecdysis about this time,
though it was not actually observed. The body becomes much
distended as the gut is filled with food matter, until the tail
and processes finally vanish. After the disappearance of the
cauda the anus gradually shifts back until it is at last terminal,
and at the same time it becomes proportionately smaller.

The egg, as previously mentioned, is usually deposited in

.

DEVELOPMENT OF CYNIPID HYPERPARASITES 465

the ventral side of the Aphidius at either extremity of the
body. The chitinized larva, and subsequently its cast skin,
are found in the same position, and orientated indifferently
in any direction, but the later stages invariably lie along the
dorsal side of the gut of the host with the head towards the head
of the latter. Hence at some intermediate stage the hyper-
parasite must change its position. How this takes place
was not observed, but, in view of the fact that the cauda of
analogous forms is sometimes regarded as locomotory, it may
be remarked that in Charips the first tailed larva does not
move at all, while at some later stage, when the cauda is
reduced, a definite, and frequently elaborate, change of position
occurs.

THe Funu-crown Larva (Text-fig. 9).

When the larva is full grown it makes its way out behind
the head of the host, whose remains it devours within the next
few hours. The gut may then be evacuated and metamor-
phosis ensue speedily, but frequently there is a resting period
of several days. ‘Thus, in one case, eleven days elapsed between
emergence and transformation, and in another case, eight.

The full-grown larva is an apodous form measuring 1-70 x
0-90 mm. The body of thirteen distinct segments tapers some-
what to the anus. The skin is smooth, and there are no
appendages except to the mouth parts. The crescentic labrum
is furnished with eight small papillae. ‘The mandibles are
large, bidentate, and strongly chitinized. Hach maxilla bears
a disk, upon which are three papillae, one of which terminates
in a short seta; and the labium, which is large and oval, bears
laterally two pairs of papillae (‘Text-fig. 10).

The salivary glands, which in this form never secrete silk,
extend forward from the seventh segment on either side of
the gut ventrally. Each gland is a long straight tube composed
of polyhedral cells, and, in the first segment, enters a duct
which immediately behind the head unites with its fellow of the
opposite side to form the short dilated common salivary duct

466 MAUD D. HAVILAND

opening on the floor of the mouth under the U-shaped hypo-
pharynx.

The mid-gut is shut off from the oesophagus by a valve.
The former, which is greatly distended, is lined with flattened
polyhedral cells with large nuclei. As in other hymenopterous
larvae at this stage there is no communication between the

TEXT-FIG. 9,

Text-Fic. 10.

Fig. 9.—The full-grown larva. x 35,
Fig. 10.—Head of full-grown larva. x100. lab. = labium; lbr. =
labrum: md. = mandible: ma. = maxilla; sal. d. = salivary duct.

mesenteron and proctodaeum. The structure of the latter
merits description. In larvae of the first and second instars the
lumen is wide, and lined with a columnar epithelium of hyper-
trophied hypoderm cells with conspicuous nuclei. As develop-
ment proceeds the anus becomes proportionately smaller,
and an outgrowth from the antero-ventral wall of the procto-
daeum projects backwards into the lumen. This outgrowth
is shaped like a shovel, shortest on its dorsal aspect, and has

nl Aitas

DEVELOPMENT OF CYNIPID HYPERPARASITES 467

lateral expansions over-arching the cavity inside. In effect,
it partly divides the proctodaeum into two compartments, one
within the other, and the Malphigian tubules communicate
ventrally with the inner of the two. The outgrowth or process
itself is formed of two layers of elongated basophil cells, with
well-marked nuclei, similar to those of the wall of the hind-gut,
and in the later stages it almost fills the lumen. If it contained
muscular fibres it would be easy to suppose that this outgrowth
functions as a valve, shutting off the orifices of the Malpighian
tubules from the general proctodeal cavity ; but as the presence

TExt-FiG. 11.

Proctodaeum of the young larva, 4 in sagittal, B in transverse section.
x 350. c.ep.=columnar epithelium: c.m.=circular muscles ;
J.b.=fat-body ; h.p.=hypoderm; J.m,=longitudinal muscles ;
I.pr.=lumen of proctodaeum ; mes.ep.= epithelium of mesenteron ;
8.p.= process projecting into the lumen of the proctodaeum.

of muscular tissue cannot be demonstrated, its only purpose
appears to be to increase the surface area of the columnar
epithelium of the hind-gut (Text-fig. 11).

The two Malpighian tubules are exceedingly short. Hach
is composed of eight or nine large cells only, but these
surround a lumen of considerable diameter. The nervous
System appears as a broad slightly-constricted band. The
Supra- and sub-oesophageal ganglia, and the three ganglia of
the thorax, are well marked; but those of the abdominal
region are indistinctly separated, with the exception of the
last two, which are fused and form a distinct bulb-like swelling.

NO, 259 Il

468 MAUD D. HAVILAND

The rest of the internal strueture demands no particular
comment.

The tracheal system becomes functional when the parasite
leaves the host. The two main lateral trunks are united by an
anterior and a posterior commissure. Dorsal and ventral
lateral branches are given off in each segment 1-10. There
are six pairs of open spiracles. The first is placed between
segments 1 and 2, and the remainder on segments 3, 4, 5, 7,
and 9. Of the considerable number of examples examined
only two departed from this rule in possessing, in addition,
a pair of spiracles on segment 8.

PUPATION AND HMERGENCE.

Pupation lasts from twenty-two to twenty-six days, and
at the end of this time the Cynipid gnaws an irregular hole
on the dorsal side of the cocoon and creeps out. In captivity
the adults lived from three to eight days. They fed upon the
sap oozing from cut leaves and upon the honeydew of the
aphides. They sometimes sipped the latter from the anus of
the living animal, and were occasionally observed to scrape
the dried sugar from empty skins with their mandibles.

It is not known how many broods may be reared in the
season, nor how far these Cynipid hyperparasites are specific
for different Aphididae, but as far as it goes the evidence
suggests that they have a considerable range of hosts. ‘Thus
the number of broods is probably determined by the number
of Aphididae available.

Also at present there is no evidence as to how the parasites
and hyperparasites of Aphides pass the winter. J have found
living larvae of Aphidius salicis, Hal., in Aphis
saliceti, Kalt., in cocoons collected in July, and opened in
the laboratory in January. This suggests that a few may pass
through the winter in this stage; but, although I paid parti-
cular attention to this point, I could find no indication that
Aphidius ervi had not all emerged by the end of August,
for, of the considerable number of cocoons from different
localities that were examined, all were empty.

DEVELOPMENT OF CYNIPID HYPERPARASITES 469

ComMPaARISON oF THE LARVAL CHARACTERS OF Charips witH
THOSE OF OTHER HNTOMOPHAGOUS CYNIPIDAE.

Our knowledge of the larval forms of the other entomo-
phagous Cynipidae is limited to three species.

In 1834 Bouché (2) described the full-grown larva of Figites
anthomyiarum, Bouché, found in the puparia of Antho-
myia (Diptera).

In 1886 Handlirsch (9) gave an account of the corresponding
stage of another Figitine, Anacharis typica, Walker,
parasitic upon Hemerobius nervosus, Fabr.

In 1913 Keilin and Pluvinel (18) described the post-
embryonic development of an Encoiline, Encoila keilini,
Kieff., parasitic upon the Dipteron Pegomyia.

In comparing the full-grown larva of Charips with these
three forms, we find certain structural differences between
them. Charips and Anacharis possess thirteen segments,
whereas Figites and Encoila have but twelve. The
tubercles of Anacharis are distinctive, and Enecoila
alone possesses simple mandibles. Figites, Anacharis,
and Eneoila all have nine pairs of spiracles, a character
they share with the phytophagous forms. In Charips
there are but six pairs of spiracles (exceptionally seven),
and these are not arranged upon consecutive segments.

As regard the early stages, the only form available for com-
parison with Charipsis Encoila. Jn the first instar the
larvae are of the same general type, but Charips differs
from Encoila in the absence of pronounced thoracic processes,
and in the possession of a chitinized skin, mandibles, and an
enlarged anus. ‘he embryonic membrane does not seem to
occur in Encoila, and, so far, has not been recorded in the
Cynipidae.

470 MAUD D. HAVILAND

COMPARISON OF THE Larval CHARACTERS OF Charips
WITH THOSE OF PARAsItTiIc HYMENOPTERA IN GENERAL.

The early larvae of the hypermetamorphic Hymenoptera
Parasitica may be referred to three main groups:

The first, or cyclopoid, type so far as has been found only
in Platygaster (Proctotrypoidea), and is known chiefly
through the researches of Ganin (5) and Marchal (20).

In the second type the last segment is furnished with an
appendage, and thus may be called caudate. It includes,
for example, such forms as Limnerium (Ichneumonidae)
Aphidius (Braconidae), Comys, and certain Agenias-
pids (Chaleidae) and Teleas (Scelionidae).

The third type was first observed by Wheeler (27) in the
myrmecophagous Chaleid, Orasema, and has since been
described by Smith (24) in another Chalcid, Perilampus.
This larva, known as a planidium, is elongated and testudinate,
furnished with imbricated plates of chitin.

The caudate type is the most frequent. The function of the
tail has been supposed by different authors to be either loco-
motory or respiratory, but may possibly be both. In the
early stages of such forms the tracheal system is apneustic
and respiration is cutaneous. The cauda, by increasing the
body-surface, may assist in the absorption of oxygen, and the
thoracic processes of Encoila may have a similar function.
At the same time the setae with which the cauda is furnished
in some Aphidiidae suggest that it may sometimes serve for
locomotion.

The first-stage larva of Charips is caudate, but I can
find no other instance of heavy chitinization in this type.
Indeed, the only parallel instance appears to be the planidium
of Perilampus, whose life-history is somewhat different.
Perilampus is hatched as a free living form, and later seeks
out the caterpillar which contains the proper hymenopterous or
dipterous host. It then lives as an endoparasite without growth
or ecdysis fora variable time. After metamorphosis of the host,
it emerges, sheds its chitinized skin, and completes development

> nual =< eee ibm, Cd) hee

DEVELOPMENT OF CYNIPID HYPERPARASITES 471

as an ecto-parasite upon the pupa. Here presumably the chitin
protects the larva during the search for the host. Charips
is an endoparasite throughout larval life, but certain facts
suggest that this may be a later adaptation, and that the
chitinous armour may be a survival of a life-cycle not unlike
that of Perilampus.

For instance, the chitin does not now seem to be of vital
importance to the young larva, since it may either be thrown
off at hatching and left behind in the embryonic membrane
or persist for a variable number of days afterwards. Smith (24)
suggests that the histolysis of the surrounding tissues is the
stimulus that impels the Perilampus to change its mode
of life and moult. Something of the kind may occur in
Charips, though in this form metamorphosis of the host
does not actually take place. The host larvae may be in
different stages of development at oviposition, and yet those
younger than the third instar could scarcely contain enough
food material to enable the Cynipid to reach maturity. It is
doubtful whether in such a case as that shown in fig. 3, where
the gut is already displaced before the hyperparasites have left
the embryonic membrane, the Aphidius can survive. But
even in ovipositions in third-instar Braconids it would be fatal
to the Cynipid if the development of the host were arrested
too soon, for instance before the cocoon was woven. Thus it
is possible that the chitinized stage is in some sort a resting
phase, and I now regret that I did not pay more attention to
this point in the material at my disposal.

Another point is that Perilampus is endoparasitic only
in the first instar, whereas Charips lives internally until
larval development is completed.

But a parallel may be drawn if the internal habit of the
latter is a comparatively recent adaptation, and the demolition
of the host’s remains after emergence is a survival from a time
when it made its way out of the host at an earlier stage and
completed development as an ectoparasite.

The metabolism of Charips presents certain problems.
The thick chitin must prevent cutaneous transfusion of oxygen

472 MAUD D. HAVILAND

from the host’s tissues. It is possible that the structure of the
anus and proctodaeum is correlated with this, and that some-
thing analogous to rectal respiration exists in this form. The
hind-gut has a large lumen enclosed by modified hypoderm
cells. In the later stages the proctodaeum is proportionately
smaller, and, when the chitin is cast off, respiration is pre-
sumably carried on through the cuticle, as in such forms as
Aphidius, though mention should be made of the tongue-
shaped process of large deeply-staining cells, which, like
a typhlosole, projects into the lumen of the proctodaeum as
development proceeds, and, if the view suggested here is correct,
would increase the respiratory area.

A peculiar modification of the hind-gut occurs in the larvae
of certain Braconids, suchas Apanteles and Microgaster.
The body terminates m a hollow bladder or vesicle of hyper-
trophied cells ; and Gatenby (8), who has recently re-deseribed
this structure, makes the interesting suggestion that this is
morphologically the proctodaeum, which has become everted
for respiration. The enlarged, though uneverted, hind-gut
of Charips may be intermediate between the highly-special-
ized structure found in these Microgasterinae and the
unmodified proctodaeum of most hymenopterous larvae.

It is noteworthy that in these Cynipidae great development
of chitin is associated with unusually short Malpighian tubules.
If the chitin persisted throughout larval life we might be
tempted to regard it as a means of disposing of such nitrogenous
waste material as could not be dealt with by the tubules.
But as the chitinized plates are lost early, while the tubules do
not increase in size in the later stages, it is improbable that the
two characters are correlated.

REACTION OF THE Host.

Aphidius reacts very differently to Charips and to
Lygocerus. In parasitization by the latter, as described
elsewhere (10), the host dies, and speedily deliquesces into
a mass. Nothing of this kind happens where the Braconid
contains a Charips larva. The Aphidius demolishes

DEVELOPMENT OF CYNIPID HYPERPARASITES 473

the viscera of the aphid, and then secretes silk and weaves the
cocoon as usual. ‘he tissues retain their tone and colour, and
irritation excites slight movement. On close examination,
however, it can be seen that the body is somewhat contracted.

At this time the Cynipid larva, its head orientated with
that of the host, lies above the mesenteron of the latter, which
it constricts into a dumb-bell form. By some means the further
development of the Aphidius is arrested, and always at
the same point, namely, after the weaving of the cocoon.
The meconium is never evacuated, and metamorphosis, which
normally takes place soon afterwards, never occurs. The con-
dition of the Braconid larva resembles in fact that of the prey
that certain Hymenoptera store in their brood-cells.

Two explanations of this phenomenon suggest themselves.
Hither the female Charips at oviposition may inhibit the
final changes of the host, possibly by injection of some secre-
tion; or the Cynipid larva itself, during development, may
affect the Aphidius by chemical or physical means.

The evidence is not conclusively in favour of either view.
In support of the first one particularly marked instance came
under notice.

A Charips female was observed to oviposit on June 26.
The aphid was isolated, and four days later the Aphidius
within began to spin silk. On July 4 the cocoon was opened
in order better to follow the development of the hyperparasite,
a plan that was adopted successfully in several instances. The
Aphidius remained without change until August 7, a period
of five weeks. The meconium was not voided, but beyond some
contraction the larva looked healthy. In replacing it in the
tube after examination it fell from the brush, and must have
received some injury, for next day a discoloured patch appeared
at the hinder end of the body. The larva was dissected care-
fully, but no hyperparasite could be found, and the organs
showed little signs of histolysis. As oviposition had been
observed, the facts suggest that some accident had prevented
the development of the Cynipid larva, and this leads to the
inference that the agent arresting the metamorphosis of the

474 MAUD D. HAVILAND

host comes into force, if not at oviposition, at least at an
early stage in larval life.

In support of the view that the larva itself may inhibit
the development of the host is the parallel case of Peri-
lampus. As the larva is hatched as a free-living form and
subsequently enters the host, there can be no question of the
inhibition dating from oviposition. Yet, according to Smith
(24), ‘ The development of the host ... invariably ceases at
the time of exit of the planidiuam. Whether or no it is actually
killed is not evident. In any case decomposition does not
take place immediately, the host being left in a condition
somewhat comparable to that of the prey of certain aculeate
Hymenoptera.’

Perilampus differs from Charips in that metamor-
phosis has taken place before the exit of the planidium ; but
when the latter begs to live as an ectoparasite upon the
newly-formed pupa, it is found that the growth of the head
and appendages, with their setae and pigments, is arrested,
and development is not completed.

Nothing resembling phagocytic reaction against the hyper-
parasite was observed, either as regards the living larva or
the cast skin, which could sometimes be found unchanged
among the host’s tissues up to the time of emergence of the
full-grown Cynipid larva.

EcoNoMIc STATUS.

Charips checks the Aphidius in its destruction of plant-
lice, and thus, from the economic standpomt, must be con-
sidered an injurious insect. But throughout its development
it shares the vulnerability of its host to ectoparasitic Chaleids
and Proctotrypids, and when secondary parasitization occurs it
perishes with the Aphidius. From observations made in
the course of this work it would seem that where the incidence
of Chaleid and Proctotrypid hyperparasitization is high, the
chances of Charips larvae attaining maturity are corre-
spondingly reduced. For instance, if, ofa hundred Aphidius,
twenty-five are parasitized by Charips, and thirty-two

DEVELOPMENT OF CYNIPID HYPERPARASITES ATS

parasitized by such a form as Lygocerus (Proctotrypidae)
by chance, 8 per cent. of the former should be destroyed ;
while where the incidence of parasitization by Chaleids, such
as Asaphes, is as high as that of Lygocerus, this rate
of mortality must be doubled. The above figure for Cynipidae
is hypothetical, though, as it is based on examination of much
material, it is probably not too low. That for Lygocerus
was found to be the actual rate in certain instances (10). It
is difficult to estimate the mortality accurately, because the
host, if subsequently reparasitized, rapidly decomposes, and
any endoparasite that it may contain soon becomes unrecogniz-
able. Moreover, the bionomical relations of the different
hyperparasites are so intricate that the chances of survival
of any particular case are difficult to compute. Thus Charips
actually lessens its own chance of survival, for the effect of
its parasitization is to arrest the metamorphosis of the host,
and thus maintain it in the optimum condition for oviposition
by Lygocerus or Asaphes. Hence in the hypothetical
case given above the number of Aphidius larvae parasitized
by Charips and reparasitized by Lygocerus would
probably be larger than that parasitized by Ly gocerus only,
and the mortality of the first parasite would actually be higher
than the figure given. ‘To this mortality from reparasitization
I attribute the fact that from collections of parasitized aphides
made in the field there were proportionately more Cynipid
emergences in June than in July. Most of the hyperparasites
obtained from later collections were Chalcids or Proctotrypids
(Lygocerus) ; and the inference is that the later broods of
Cynipidae suffered from a second parasitization of their hosts
by other hyperparasites.

SUMMARY.

1. Bothryoxysta curvata, Kieff., Charips victrix,
Hartig, and Alloxysta erythrothorax, Westw., are
hyperparasites of aphides through Aphidius (Braconidae).

2. Reproduction may be either sexual or parthenogenetic.

3. The egg is laid in the haemocoele of the host larva before

476 MAUD D. HAVILAND

the death of the aphid, and post-embryonic development is
internal.

4, A trophic membrane of hypertrophied cells is formed
round the embryo.

5. The larva is, at first, hypermetamorphic ; and exhibits
sreater development of the chitimous cuticle than is usual
in endoparasites ; but in the succeeding stages it approximates
more closely to the general hymenopterous type.

6. The development of the Aphidius is arrested at
a certain point, and metamorphosis does not take place.

7. The Cynipid, when ready to pupate, makes its way out
of the Aphidius, whose remains it devours, and undergoes
metamorphosis within the cocoon previously woven by the
latter in the skin of the aphid.

8. These forms differ in certain particulars from the ento-
mophagous Cynipidae previously described, and the chief
differences are discussed.

9. Comparison is also made of the larvae of other Hymeno-
ptera Parasitica, particularly of Perilampus.

10. Certain problems of metabolism are pointed out, and it
is suggested that respiration may be partly rectal.

11. These Cynipidae are economically injurious as they
check the Aphidius in its destruction of plant-lice; but
there is high mortality among the larvae owing to secondary
parasitization of the Braconid by other hyperparasites.

BIBLIOGRAPHY.

1. Adler, H. (1881).—‘‘ Ueber den Generationswechsel der Eichen-Gall-
wespen ”’, ‘ Zeit. wiss. Zool.’, Bd. xxxv, pp. 151-246, Taf. x-xii.

. Bouché (1834).—* Naturgeschichte der Insecten.’

3. Cameron, Peter (1890).—* A Monograph of the British Phytophagous
Hymenoptera ’, Ray Society’s Publications, vol. ili.

4. Embleton, Alice (1904).—‘‘On the Anatomy and Development of
Comys infelix, Embleton”’, ‘Trans. Linn. Soc.’, vol. ix, pt. 5,
p. 231. Pts. 11-12.

5. Ganin, M. (1869).—“ Beitriige zur Kenntniss der Entwickelungs-
geschichte bei den Insecten”’, ‘ Zeit. wiss, Zool.’, Bd. xix, pp. 381-
448, Taf. xxx-xxxiil,

bo

=

so

18.

19.

20

21.

22.
23.

24.

DEVELOPMENT OF CYNIPID HYPERPARASITES ATT

Gatenby, J. Bronté (1917).--“The Embryonic Development of
Trichogramma evanescens, Westw.’’, ‘ Quart. Journ. Micro.
Sci.’, vol. 62, pt. ii.

(1917).—“ Note on the Development of Trichogramma

evanescens ”, ibid.

(1919).—‘** Notes on the Bionomics, Embryology, and Anatomy
of certain Hymenoptera Parasitica ’’, “Journ. Linn, Soc.’, vol. xxxiii,
pp. 387-416.

Handlirsch, A. (1886).—‘* Die Metamorphose zweier Arten der Gattung
Anacharis”’, ‘Verh. zool.-bot. Ges. Wien’, pp. 235-7, pl. vii,
figs. 1-4.

. Haviland, Maud D. (1920).—-“* On the Bionomics and Post-Embryonic

Development of Lygocerus cameroni, Kieff.”, ‘ Quart.
Journ. Micro. Sci.’, vol. 65, pt. i.

—— (1921).—* Preliminary Note on a Cynipid hyperparasite of
Aphides ”’, ‘ Proc. Phil. Soc. Camb.’, vol. xx. no. 2.

- Hegner, R. W. (1915).—‘‘ Studies in Germ-cells—Protoplasmic

Differentiation in the Oocytes of certain Hymenoptera’’, ‘ Journ.
Morph.’, vol. 26.

. Hennegay, L. F. (1914).—‘ Les Insectes ’, Paris.
. Hogben, Lancelot T. (1920).—-‘‘ Studies in Synapsis: Oogenesis in the

Hymenoptera ”’, “ Proc. Roy. Soc.’, Series B, vol. 91.

. Howard, L. O., and Fiske, W. F. (1911).—** The Importation in the

U.S.A. of the Parasites of the Gipsy Moth and Brown-tail Moth”,
“U.S. Dept. Agr. Bur. Entom.’, Bull. xci, 312 pp.. 74 text-figs., and
28 pls.

. Imms, A. D. (1916).—‘‘ Observations on Insect Parasites of some

Coccidae ’’. “ Quart. Journ. Micro. Sci.’, vol. Ixi, no. 243.

. —— (1918).—‘‘ On Chalcid parasites of Lecanium capreae”’,

ibid., vol. lxiii, no. 251.

Keilin, D., and Baume-Pluvinel, C. de la (1913).—*‘ Formes larvaires
et biologie d’un Cynipide entomophage”’, ‘ Bull. Sci. France’,
septiéme série, t. xlvii. fase. i.

Kieffer, J. J., and Dalla Torre, K. W. von (1910).—*‘ Cynipidae ”’ in
* Das Tierreich ’, Berlin.

Marchal, P, (1906)—‘‘Les Platygasters’*, ‘Arch. Zool. Expét.’,
quatorziéme série. t. iv.

Seurat, L. G. (1899).—‘‘ Contributions a l'étude des Hyménoptéres
entomophages ”’, ‘ Ann. Sci. Nat.’, huitiéme série. t. 10.

Sharp, D. (1899).—‘ Camb. Nat. Hist.’, “ Insects ”’, pt. i.

Silvestri, F. (1909).—‘ Contribuzioni alla conoscenza biologica degli
Imenotteri parassiti’’, ‘ Boll. Lab. Scuola Agric. Portici’, vol. iii,
Smith, Harry (1912).—“ The Chalcidoid genus Perilampus, and
its relations to the problem of parasite introduction ”’, ‘U.S. Dep.

Agric. Tech. Series, Entom,’, no. 19, pt. iv.

478 MAUD D. HAVILAND

25. Timberlake. P. H. (1910).—-“‘ Observations on the early stages of two
Aphidiine parasites of Aphids ’’, ‘ Psyche ’, Mass.

26. —— (1913).—“‘ Preliminary Report on Parasites of Coccus
hesperidium in California”, ‘Journ. Econ, Entom.’, vol. vi,
pp. 293-303.

27. Wheeler, W. M. (1907).—‘ The Polymorphism of Ants, with an account
of some singular abnormalities due to Parasitism”’, ‘ Bull. Amer.
Mus. Nat. Hist.’, vol. 23, Art. I.

Notes on the Larval Skeleton of Spatangus
purpureus.

By
Hiroshi Ohshima,

Assistant Professor in the Department of Agriculture, Kyushiu
Imperial University, Fukuoka, Japan.

With Plate 21.

Autrnoucu ‘one of the very first Kehinoderms of which
artificial fertilization and rearing of the larvae were under-
taken’ (Mortensen, 6, p. 14), the development and
especially the structure of the larval skeleton of Spatangus
purpureus have been rather imperfectly known. Krohn’s
descriptions and figures (2, 8) are not quite satisfactory with
regard to the skeletal structure. and, moreover, the larvae
described in his second paper are doubtful as to their specific
identification (Mortensen, 6, p. 15). Through Mor-
tensen’s renewed observations on the artificially-reared
larvae of this species (6, pp. 14-17) the external features of
the larval development are now made clearer. As to the
larval skeleton, however, he was only able to give some brief
information owing to the unfortunately bad state of preserva-
tion of his specimens. Among other Spatangoids, Eehino-
cardium cordatum and Brissopsis lyrifera were
carefully studied by Macbride (4) and Mortensen
(7, pp. 144-8), and the larvae of these three species have been
shown to have such a striking resemblance to each other im
early stages that it is desirable to ascertaim some more minute
diagnostic characters for each species. In such circumstances

480 HIROSHI OHSHIMA

it seems not unnecessary to put on record detailed descriptions
of the skeletal structure of the larva of Spatangus pur-
pureus.

The material on which my work is based consists of a series
of larvae, reared and preserved by Mr. Elmbhirst at
Millport, and kindly handed over to me for study by Professor
EK. W. MacbBride.’ Although there are found several gaps
in developmental stages, the changes undergone by the larval
skeleton could be followed fairly satisfactorily. From the
labels which were found attached to the vials we obtain the
following chronological accounts. .

The earliest stage which is represented by segmenting eggs
is dated 16th May 1914. This is probably the day on which
the eggs were artificially fertilized. The further stages with
regard to the age in days are :

2nd day May 17th ; . Blastula.

SEG (i 55 ssa! 18th P . Gastrula.
Ain tes af 19¢h * . Young 2-armed pluteus.
bthis «; 5, 1e20bh ; . Fully-formed 2-armed
pluteus.
6th ~,; = a iealiat . . 4-armed pluteus.
7 bhih June Ist . , . 6-armed pluteus.
? ? : : . 8- or 10-armed pluteus.
DAS ,, June 8th : . 12-armed pluteus.

Thus in full accordance with the statements of Mortensen
(6, p. 15) the larva reaches its last stage in the course of
three weeks. It is to be regretted that those larvae whose
skeleton was best preserved had been kept together in one
vial, all the different stages being mixed up, and without any
label, so that it is not possible to give a chronological state-

! The present work was done partly in the zoological laboratory of the
Imperial College of Science and Technology and partly in the British
Museum (Natural History). My cordial thanks are due to Professor
E. W. MacBride of the College and to Sir Sidney F. Harmer of
the Museum, for help and encouragement in various ways and for the
privilege of the use of the laboratory and the libraries.

LARVAL SKELETON OF SPATANGUS 48]

ment in most cases as regards the first appearance of a new
calcification centre or its subsequent development, &e.

At the outset I may call attention to the fact that the
latticed rods, viz. the postoral, postero-dorsal, and posterior
unpaired (so-called aboral spike), are morphologically different
from the other simple, though often thorny, rods which serve
equally as the support of each corresponding arm. Théel
(11, pp. 40-1) described very clearly the early development of
the postoral rods of Echinocyamus pusillus as
follows: ‘they (the latticed rods) begin to arise during the
gastrula stage as three small processes, one on each rod of
the star close to its centre, Pl. ii, fig. 38. These processes
stretch in length, run parallel and become connected by
transverse beams’. ‘The same is exactly true for the corre-
sponding rods and also for the other latticed rods in Spatan-
gus. In all of these a three-rayed ‘star’ is first laid down
lying parallel to the surface of the body. From each of
the rays or arms, very close to the centre, is given out
a vertical process, directed towards the surface of the body.
The latter, three in number if, as in most cases, all developed,
give rise to a latticed rod. The postoral and postero-dorsal
rods of Echinocardium cordatum are both stated by
MacBride to be formed of only two parallel reds (4,
pp. 475, 477). As compared with the table-lke calcareous
body, which is commonly met with in all classes of Echino-
derms, the latticed rod corresponds to the spire, and the three- _
rayed portion to the base. Thus the above-named two-paired
and one unpaired latticed rods are morphologically com-
posite in structure and are from the beginning directed
vertically to the surface of the body. On the other hand,
those rods supporting the antero-lateral and postero-lateral
arms are morphologically simple, being produced either
as prolongations or branches of the three-rayed base, which
were lying originally parallel to the surface of the body. The
body-, recurrent, and horizontal rods are also either prolonga-
tions or branches of the basal part, which remained running
along the surface of the body without, however, pushing out

482 HIROSHI OHSHIMA

to support arms. The dorsal arch consists only of the three-
rayed portion, from which any vertical process fails to
develop. The pre-oral and antero-dorsal rods also belong,
according to this interpretation, to the simple type of the
rods. In a similar manner, it seems to me, in Arbacia,
Dorocidaris, Ecehinocyamus, &e., the posterior
unpaired star fails to produce vertical processes, which would
give rise to the aboral spike in the Spatangoid larva, the
laterally directed basal arms being only developed as the
postero-lateral rods. Prouho’s discovery of an abnormal
larva of Dorocidaris papillata which produced a
well-developed aboral spike (10, pp. 349-50, Pl. xxv, fig. 9;
ef. Mortensen, 5, p. 75) is exceedingly imteresting in this
respect. Mortensen (5, p. 71) maintains that the state-
ments of some authors, eg. Kolliker’s, who have
described plutei with six to ten latticed rods must be wrong.
My observations confirm this conclusion. Though in some
abnormal cases those morphologically simple rods may be
doubled or split, analogous to those I have observed, e.g. the
dorsal horizontal rod of the right side (Pl. 21, figs. 7 and 8, dh)
and the left recurrent rod (fig. 6, re), 1t is quite impossible that
they should assume latticed structure.

Late in the gastrula stage a pair of calcification centres
appear, which are bilaterally symmetrical in position. This
state of affairs is so well known in other Kchinoids that any
detailed description is quite unnecessary. I may, however,
point out that the body-rod represents one of the three-
rayed basal arms, not simply a posterior continuation of the
postoral rod, as might easily be wrongly inferred because they
both run in an almost straight line (fig. 1, br, po). The
other two arms of the base are represented respectively by
the ventral horizontal rod (vh) and the recurrent rod (re),
from which latter the antero-lateral rod (al) is given out later.

The third, unpaired calcification centre appears near the
posterior end of the body (ab). This may appear as early as
in the stage where the future postoral arms can as yet hardly
be recognized as arms, viz. when the larva has formed a slight

LARVAL SKELETON OF SPATANGUS 483

concavity at the oral field and has begun to assume roughly
a tetragonal shape. The star is situated in, such a position
that two of its arms lie bilaterally and the remaining one is
directed dorsally. The former two ultimately give rise to
the postero-lateral rods, while the third remains as a short
but distinct spur-like process all through the larval life
(figs. 3-8). From each of these arms a vertical process is
produced, directing posteriorly. These three vertical processes
form together the aboral spike (fig. 2, ab). Being robust in
structure the transverse beams extend rapidly so as to
obliterate the openings between them.

Hand in hand with the rapid growth in length of the post-
oral rods the arms of the basal portion develop to assume their
future position. The body-rods, which run straight postero-
medially, are the most rapid in growth among them, and their
posterior ends come to overlap each other (fig. 1, br.). In
the corresponding stage as well as later, as figured by Krohn
(2, Pl. vii, figs. 1-8, 6), the posterior ends of the body-rods
are shown standing fairly apart. Except in a later stage
where the rods begin to be absorbed at the posterior ends
(figs. 5 and 7), I have never met with such a state as shown
in his figures. The second arm, the recurrent rod (re), which
is at first directed dorsally, soon bends posteriorly. In the
meantime it produces a branch at its bent portion. This
branch, which is the future antero-lateral rod (al), proceeds
a little towards the median line, but soon bends anteriorly
to run almost parallel to its fellows of the other side, though
slightly approaching this as it runs. Its base is a little
broadened and bears a few minute processes, as shown in
Krohn’s figure (2, Pl. vii, fig. 5, ce) and confirmed by
Mortensen (6, p. 15). The remaining arm of the first
calcification centre runs along the ventral surface, almost
transversely towards the median line, but slightly deviating
anteriorly (fig. 1, vh). This is the ventral horizontal rod.
The end soon comes in contact with that of its fellow of the
other side, and they ultimately fuse, forming a characteristic
thickened joint (figs. 2 and 8, vh). This feature is constantly

NO. 259 Kk

484 HIROSHI OHSHIMA

seen and lasts for a fairly long period, and it seems to me that
this can be regarded as a specific character in identifying
Spatangoid larvae. In many other Spatangoid larvae this
is not the case; these rods either stand apart or pass across,
as in Echinocardium cordatum and its doubtful ally
(Miller, 8, p. 290, Pl. iii, fig. 2). In cases where both ends
come very close together, as in Echinopluteus fusus
(Miller, 9, Pl. vu, fig. 2), E. solidus (9, Pl. vi, fig. 9;
Pl. vu, fig. 1), and perhaps Brissopsis lyrifera also
(Mortensen, 7%, fig. 2), they do not form any thickened
joint. Only in Chadwick’s figures of an unidentified
form (1, Pl. ix, figs. 61 and 62) the similar state of the ventral
horizontal rods is very clearly shown.

By the time when the two-armed stage is fully developed,
when the post-oral arms have reached the length nearly equal
to the body proper, whereas neither the antero-lateral arms nor
the aboral process are as yet distinct, the following features
are to be noticed: the post-oral rods are usually solid and
three-ridged, and the margin of the ridges is not serrated.
Eixceptionally, however, some irregularly-scattered holes
may be met with even near the proximal end of the rod, but
owing to the very slight differences in the refractive indices
between the thin, filmy skeleton and the surrounding medium,
which consists of oil of cloves or Canada balsam, it is difficult
to demonstrate the holes clearly. Krohn (2, p. 256) observed
no fenestration in these rods in the corresponding stage.
Further, in his figure (Pl. vil, fig. 1) he showed only the antero-
lateral and body-rods besides the post-oral, whilst the ventral
horizontal and recurrent rods are not represented. The star
of the aboral spike should also have appeared in this stage.

The recurrent rod grows rapidly, and when its posterior end
comes in contact with that of its fellow of the other side
(fig. 2, re) fuses with it and increases in thickness, often being
beset with some irregular short processes near the end
(fig. 3, re). A little anterior to this end a branch is soon sent
out ventrally, while about the same time the body-rod
produces a branch dorsally, and these two branches meet

> tin) pee Os

LARVAL SKELETON OF SPATANGUS 485

and fuse midway between the body- and recurrent rods
(figs. 4 and 6, c). There are very often some irregular spines
or branches from the dorso-ventral connexion thus formed.
As the result of this connexion there is formed a rectangular
framework as seen from side (cf. Mortensen, 5, p. 75,
Pl. ix, fig. 9). From the poimt where the antero-lateral rod
diverges from the recurrent rod there is formed frequently
a short process directed anteriorly (figs. 2 and 6). This seems
to have no significance.

Lastly, at the end of the two-armed stage the body-rods
fuse at the point where they have been overlapping each other,
so as to form an oblique cross. Very often there is formed
an accessory connecting-span between the two body-rods.
This is a short transverse piece lying a short distance anterior
to the crossing-point. Now the caleareous framework
encircling the stomach has become fairly rigid. The body-,
recurrent, and ventral horizontal rods of both sides are fused
in the median plane with the respective fellow of the other
side, while, on the other hand, the body- and recurrent rods
are connected with each other near the posterior end on each
side of the body.

After having reached this state the body-rod increases no
longer in length, so that, as long as its posterior end remains
unabsorbed,-its length can be taken as unit in deseribing the
dimensions of other parts. The length of the body-rod can
easily be measured when the larva is laid with its ventral side
downwards, so that the rod is seen in its real length without
foreshortening. As expressed in terms of the ratios to the
body-rods, the post-oral rod reaches during the two-armed
stage a length more than twice as long as the body-rod, the
antero-lateral rod more than one-half, and the aboral spike
about one-third.

The aboral process and the antero-lateral arms become
discernible almost simultaneously. It may now be called the
four-armed stage (figs. 8-5). The change which takes place
during this stage is the enormous increase in lengths of the
post-oral and antero-lateral rods and of the aboral spike.

Kk2

486 HIROSHT OHSHIMA

The post-oral rods grow up to four or five times the length
of the body-rod, while the antero-lateral rod and the aboral
spike reach more or less twice the length of the same (fig. 5).

The post-oral rods seem in most cases to be devoid of
fenestration in their proximal half or one-third, whereas the
unfenestrated portion of the aboral spike is generally much
shorter. In an extreme case in the latter the fenestration
begins close to the proximal end (fig. 6, ab), exactly as the
feature seen by Krohn in an unidentified form (8, p. 210).
The distal parts of these rods are fairly regularly serrated.
The serration seems to begin roughly at the point where the
fenestration also begins (fig. 5, po, ab). The posterior ends of
both the body- and recurrent rods show towards the end of
this stage signs of degeneration, being gradually absorbed.
The dorsal arch makes its appearance near the end of this
stage, on the mid-dorsal line at the level where the oesophagus
opens to the stomach (fig. 5, da). The two arms of the star,
which he symmetrically and are directed antero-laterally,
increase rapidly in length, while the unpaired, posteriorly-
directed arm remains very short, sometimes even obliterated.

Krohn’s figure (2, Pl. vil, fig. 2) corresponds to the early
four-armed stage. It is the dorsal view, in which the ventral
horizontal rods and the body-rods are not shown, while the
descending rods, which J take as the recurrent, are not
coming to meet each other at the posterior ends. The post-
oral rods are shown as fenestrated on their distal three-fifths,
while the aboral spike remains unfenestrated. Both these
kinds of rods are, however, shown to have serrated edges
along their whole length.

The next, six-armed stage, is characterized by the appear-
ance of the postero-dorsal arms. Previous to the appearance
of these arms the supporting skeleton, which is called the
postero-dorsal rod, is formed underneath each of them
(fig. 6, pd). The rod develops in the manner similar to that
of the other latticed rods, and as deseribed and figured by
Théel in Echinocyamus pusillus (11, p. 44, Pl. vi,
fig. 88, y). The arms of the star lie in such a position that one

LARVAL SKELETON OF SPATANGUS 487

is directed anteriorly, another postero-laterally, and the
remaining one postero-medially. From the lack of adequate
material the fate of the former two arms cannot be stated with
certainty, though it seems probable that they do not develop
much farther. The postero-medially-directed arm in the later
stages continues to develop in a direction parallel to the dorsal
surface, reminding one of the body-rod on the ventral side
(figs. 7 and 8). Near the base of this arm a branch is sent out
in an antero-median direction, reminding one again of the
ventral horizontal rod. This is the dorsal horizontal rod (dh).
From each of the arms of the star, close to the centre, 1s
given out a vertical process, very often differing in the rate of
development, but ultimately the three in all give rise to the
latticed postero-dorsal rod.

Although from want of material, especially of the later
part of this stage, no definite statement can be made, yet,
judging from later specimens, it is highly probable that
the post-oral rod increases in length during the six-armed
stage up to nearly 6 times the length of the body-rod, the
antero-lateral rod 3 times, the aboral spike nearly 3-5 times,
and the postero-dorsal rod probably at least 1-5-2 times the
length of the same.

In Krohn’s figure (2, Pl. vu, fig. 3) is indicated the three-
rayed base of the postero-dorsal rod (e). The buds of the
pre-oral arms have already appeared (d), while the dorsal arch
is still in a rudimentary condition, of which, however, nothing
is mentioned. The fact that the pre-oral arms appear without
any mechanical influence of the underlying skeleton is also
seen in Echinocardium cordatum (4, p. 477, Pl. xxxu,
fig. 6). But both in MacBride’s case of Hchino-
cardium and my specimens of Spatangus the appear-
ance of the pre-oral arms takes place much later than the
stage as shown by Krohn, viz. even when the postero-
dorsal arms have attained a fair length, there was as yet no
sign of these arms found. Krohn gives some detailed
structures in a somewhat advanced six-armed stage (Pl. vil,
figs. 5 and 6). If the fig. 5 is really the dorsal view, as stated

488 HIROSHI OHSHIMA

by him, then the dorsal arch (d) should lie above the antero-
lateral rods (c). The two arms of the base of the postero-
dorsal rod (f, g) are shown very well developed, and that the
serrated recurrent rods meet each other at the broadened
posterior ends is also clearly drawn. His fig. 6, which is the
ventral view, 1s somewhat difficult to understand. ‘There
are two sets of rods which seem to correspond to the ventral
horizontal rods, both overlapping each other at the end.
Whether it is really an abnormal case, as in the right dorsal
horizontal rod in my oldest larva (figs. 7 and 8, dh), or due to
his misrepresentation cannot be decided at present.

The further advanced stages are represented by a small
number of eight- to ten-armed larvae with dissolved skeleton,
und a single specimen of the twelve-armed stage.

The fourth pair of arms to appear are the pre-oral, which are
supported respectively by the direct prolongations of each
end of the dorsal arch. The fifth pair are the postero-lateral,
supported by the lateral prolongations from the base of the
aboral spike. Irom want of material showing any adequate
stage I cannot decide whether the postero-lateral arms have
from the beginning a skeletal support, as e.g. in Kehino-
cardium cordatum (MacBride, 4, p. 479), or not,
as e.g. in Brissopsis lyrifera (Mortensen, ¥%,
pp. 147-8). Judging, however, from the fact that the arms
soon develop to assume their typical shape, instead of remain-
ing as ear-shaped lobes, I am strongly inclined to think that
the arms in question of Spatangus purpureus do
contain their skeletal support from their earliest stage.

Owing to the remarkable inerease in size of the stomach
during the eight to ten-armed stage, that skeletal framework
which formerly encircled the stomach must have undergone
corresponding changes. ‘his can be judged from the state
seen in the twelve-armed specimen (figs. 7 and 8). Both the
body- and recurrent rods are shortened at the posterior
ends, their side-by-side connexion being broken. ‘The ventral
horizontal rods of both sides are also separated from each other
at the joint. This broken framework does not now encirele the

LARVAL SKELETON OF SPATANGUS 489

stomach, but has gradually been pushed posteriorly, and the
angles between the body- and recurrent rods of one side and
their fellows of the other side are much widened.

The twelve-armed specimen (figs. 7 and 8) is much younger
than the larva figured by Mortensen (6, fig. 14), the total
length measuring only 2-1mm. The pre-oral and postero-
lateral arms are nearly equal in length, measuring 0-3 mm..
a little shorter than the antero-lateral, which measure 0-35 mm.
The antero-dorsal arms, which have appeared last, are only in
the form of buds. The other arms and process are remarkably
long, i.e. the posterior arms measuring | mm. in length, the
posterior process 0-9 mm., and the postero-dorsal arms 0-8 mm.

A short distance anterior to the point where the antero-
dorsal rod is sent out from the dorsal arch, the latter produces
a short lateral branch. The same is noticed by Miller
in Echinopluteus fusus (9, Pl. vii, fig. 3) and by
Mortensen in Kchinocardium cordatum (5, p. 103,
Pl. ix, fig. 6). In a Spatangoid larva, which has been doubt-
fully identified by Mortensen (5, pp. 102-8) with Echino-
ecardium cordatum, Miuller described and_ figured
a peculiar feature in that the median posterior branch of the
dorsal arch fused with the tips of the dorsal horizontal rods
(8, p. 290, Pl. i, figs. 1 and 4, d). So far as I know such a case
has never since been recorded by any other observers nor
have I noticed it in my specimens (figs. 7 and 8, da, dh).
The postero-lateral rod has no noticeable characteristics,
being of a uniform thickness throughout and rather smooth,
differing from the richly-serrated state as seen in Hehino-
cardium cordatum (Mortensen, 5, p. 103, Pl. ix,
figs. 7and8; MacBride, 4, Pl. xxxiu, fig. 11, pla).

The rectangle formed by the body- and recurrent rods as
seen in some younger stages (figs. 4 and 6) can no more be
found (fig. 8). The area roughly corresponding to the anterior
half of the rectangle is now occupied by an irregularly-per-
forated calcareous plate, which is developed more strongly
on the right side than on the left side. The bases of the post-
oral (po) and antero-lateral rods (al) are incorporated into this

490 HIROSHI OHSHIMA

calcareous plate, and the recurrent rod is now hardly distin-
guishable. Although it is difficult to make out clearly, it seems
highly probable that neither the bases of the postero-dorsal
(pd) nor of the postero-lateral rods (p/) are fused with that
plate. Similar features in the formation of calcareous plates
are frequently met with in other irregular sea-urchins, e.¢.
Echinopluteus fusus (Miller, 9, Pl. iv, fig. 7; Pl. vii,
figs. 3 and 11), Arbacia pustulosa (Miller, 9, Pl. im,
figs. 2 and 3), &c. Whether these plates have anything to
do with the definitive skeleton of the young sea-urchin is
still an open question, though it seems probable that they are
absorbed altogether at the time of metamorphosis.

SUMMARY.

1. The larva of Spatangus purpureus reaches its
last stage, which is characterized by its possession of six
pairs of arms, in the course of three weeks after fertilization.

2. The paired arms develop in the following order: post-
oral, antero-lateral, postero-dorsal, pre-oral, postero-lateral,
and antero-dorsal. The posterior process appears about the
same time as the antero-lateral arms become distinct.

3. These six pairs of arms and the unpaired process are
each supported by a caleareous rod. Of these calcareous rods
one can distinguish two classes which differ morphologically
from each other, viz. the simple and the composite.

4. To the class of simple rods belong the antero-lateral,
pre-oral, postero-lateral, and antero-dorsal rods. They are
either direct prolongations or branches of the three arms
produced from one of the calcification centres. They are
originally horizontal (parallel to the surface of the body)
in position, and are homologous with the body-, recurrent.
and horizontal rods.

5. The remaining rods, viz. the post-oral and postero-dorsal
rods and the aboral spike (posterior rod) are composite.
They are each composed of three parallel rods connected by
transverse beams so as to give a latticed appearance. Hach

LARVAL SKELETON OF SPATANGUS 49]

of the parallel rods is a branch given out vertically from an
arm of the calcification centre.

6. ‘The larval skeleton of Spatangus purpureus is
characterized chietly by (a) more or less considerable length
of the unfenestrated proximal portions in the latticed rods,
(b) fusion of the tips of the ventral horizontal rods forming
a thickened joint, (c) overlapping of the body-rods near their
posterior ends, and subsequent fusion of this part so as to
form an oblique cross, (d) rather simple appearance of the
postero-lateral rods, and (e) formation of a calcareous plate
on each side of the stomach in the oldest stage.

REFERENCES.

—_

Chadwick, H. C. (1914).—** Echinoderm Larvae of Port Erin ”’,
‘Proc. Trans. Biol. Soc, Liverpool’, vol. 28.
2. Krohn, A. (1853).—‘Uber die Larve von Spatangus pur-
pureus ”’, Miiller’s * Arch. Anat. Physiol.’, 1853.
3. —— (1854).—“* Beobachtungen iiber Echinodermenlarven ”’, Miiller’s
“Arch. Anat. Physiol.’, 1854. ;
MacBride, E. W. (1914).—** The Development of Echinocardium

4.
cordatum. Part 1. The External Features of the Develop-
ment ”’, ‘ Quart. Journ. Micro. Sci.’, vol. 59, Part IV.

5. Mortensen, Th. (1898).—‘* Die Echinodermenlarven der Plankton-
Expedition ’’, ‘ Ergebn. Pl.-Exped. Humboldt-Stift.’, Bd. II, J.

6. —— (1913).—* On the Development of some British Echinoderms ”’,
‘Plymouth Journ. Mar. Biol. Ass.’, vol. 10, no. 1.

7. —— (1920).—* Notes on the Development and the Larval Forms of

some Scandinavian Echinoderms’’, ‘Vid. Medd. nat. Foren,
Kj@benhavn ’, Bd. 71.

8. Miller, J. (1848).—-‘‘ Uber die Larven und die Metamorphose der
Ophiuren und Seeigel’’, ‘Phys. Abhandl. k. Akad. Wiss. Berlin’,

. 1846.

9, —— (1855).—‘* Uber die Gattungen der Seeigellarven. VII. Abhandl.
iiber die Metamorphose der Echinodermen ”’, ibid., 1854.

10. Prouho, H. (1887).-—‘‘ Recherches sur le Dorocidaris papil-
lata’”’, ‘Arch. Zool. Expér. Génér.’, 2° sér. 5.

11. Théel, H. (1892).—‘‘On the Development of Echinocyamus
pusillus (O. F. Miiller)”’, ‘Nova Acta Rg. Soc. Sci. Upsala ’.
vol. 15, fase. 1.

492 HIROSHI OHSHIMA

EXPLANATION OF PLATE 21.

(All figures were drawn by means of a camera lucida and magnified
200 times.)

Fig. 1.—Dorsal view of a young two-armed larva, in which the rudiment
of the aboral spike (ab) has just appeared. (An unusually long process
is seen arising from the base of the right post-oral rod.) ,

Fig. 2.—Dorsal view of an old two-armed larva to show the fusion at
the tips of the ventral horizontal rods (vh). (The posterior ends of the
body-rods, br, are not overlapping here as normally.)

Fig. 3.—Dorsal view of a young four-armed larva to show the fusion of
the posterior ends of the recurrent rods (re). The body-rods (br) are
also fused with each other (hidden behind the aboral spike, ab).

Fig. 4.—Left-side view of the same specimen as shown in fig. 3. A
rectangle is formed by the body- (br) and recurrent rods (re). (From
the base of the aboral spike an additional process is given out ventrally.)

Fig. 5.—Dorsal view of an old 4-armed larva, in which the rudiment
of the dorsal arch (da) has appeared and the posterior ends of the body-rods
have begun to degenerate.

Fig. 6.—Right-side view of a young six-armed larva to show the early
stage of the postero-dorsal rod (pd). (The left recurrent rod, re, is here
seen abnormally split into two.)

Fig. 7.—Dorsal view of a twelve-armed larva. The body- (br), recurrent
and ventral horizontal rods (vh) have all lost their connexion with the
fellows of the other side. (The right dorsal horizontal rod, dh, is abnormally
doubled. )

Fig. 8.—Right-side view of the same specimen as shown in fig. 7, to show
the calcareous plate formed between the body- (6r) and recurrent rods.

ABBREVIATIONS.

ab=aboral spike; ad=antero-dorsal rod; al=antero-lateral rod ;
br=body-rod ; c=dorso-ventral connexion between body- and recurrent
rods ; da=dorsal arch ; dh=dorsal horizontal rod ; pd= postero-dorsal
rod; pl=postero-lateral rod; po=post-oral rod; pr=pre-oral rod;
ye=recurrent rod; vh=ventral horizontal rod.

Quart. Journ. Micr. Sci. Vol. 65, N.S., Pl. 21.

On the Classification of Actiniaria.

Part I1I.—Consideration of the whole group and its relationships,
with special reference to forms not treated in Part I.
By

T. A. Stephenson, M.Sce.,
University College of Wales. Aberystwyth.

With 20 Text-figures.

CoNTENTS.

PAGE
1. InTRODUCTION ‘ 493
2. BrigeF HisToricaL SECON : : . 497
3. DISCUSSION OF CHARACTERS TO BE USED IN Cua scunea rae - 499
4. SpectaAL DiIscusSSIONS AND OUTLINE OF NEw SCHEME. P05
5. EVOLUTIONARY SUGGESTIONS . : : . ; : - BR
6. SUMMARY : 566
7. SHORT GLOSSARY . 572

1. INTRODUCTION.

Ir has been necessary, on account of the length of the present
paper, to confine Part IJ to discussions; the definitions of
families and genera involved, on the lines of those already
given in Part IJ, will be printed in another issue of this Journal
as Part III, which will also contain a lst of literature and an
_ Index to genera covering Parts ILand II]. The hist of literature
“will be additional to that printed in Part I, and any
numbers given in brackets in the following pages will refer
to the two lists as one whole.

Part I dealt with a relatively limited and compact group of

1 Part I was published in Vol. 64 of this Journal.

NO. 260 Ll

494 T. A. STEPHENSON

anemones in a fairly detailed way; the residue of forms is
much larger, and there will not be space available in Part II
for as much detail. I have not set apart a section of the
paper as a criticism of the classification I wish to modify, as
it has economized space to let objections emerge here and there
in connexion with the individual changes suggested. Part I
tried to clear the ground and discuss the method of attack,
so that the arguments there given need not be repeated, and
so that the general principle and method suggested there might
be taken for granted in Part II. I should like to record here
that in these papers on Classification there will be found points
in contradiction to certain remarks in earlier papers— Terra
Nova’ and ‘ Actiniaria collected off Ireland ’—but the point
of view is bound to become modified in some particulars as
further experience opens new vistas. That the view-point
should remain immovably fixed in the light of developing
knowledge would more need apology than that it should march
with necessity. Work on Part II has served only to strengthen
and confirm the plan suggested in Part I of this paper.

Definitions to be given in Part IIT are based as far as possible
on anatomically-described species, leaving the more doubtful
forms to fit themselves in as knowledge of them increases.
Consequently lists of species given include rather the better-
known forms on which the definition is founded, than exhaus-
tive enumerations. Even to identify an anemone from an old
figure or description is very risky ; to be sure of an old species
one must obtain and re-deseribe the type-specimens if such
exist. If there are none, it is guess-work—cf. Pax (75), p. 309,
and others.

One result of working through all the Actinian genera
(supported by a personal anatomical study of a large number
of them) is the recurrence of impressions connected with the
difficulty of species-identification of some of them from
preserved material—and the unfruitfulness of the pursuit.
It would seem that family and genus are fairly easily tracked
down when once a certain number of data are gained, and that
these are intelligible quantities. But when it becomes a matter

CLASSIFICATION OF ACTINIARIA 495

of species the variation of the different anatomical criteria
of distinction may be so wide, and the limits of specific varia-
tion so little known, that to go beyond the genus is little more
than guess-work ; especially when one thinks of the modifica-
tion caused to certain characters by mode of preservation,
degree of contraction or distension of the animal, age, reprodue-
tive condition, locality, and other things. ‘lwo paths there
are here which need following. Firstly, a large number of
anemones should be collected (some belonging to stable and
some to unstable species, and representative of various families)
in cases where it could be positively certified that all individuals
collected for any one species were undoubtedly the same.
These should be preserved in different ways and states, and
a study made which would reveal the limits of specific varia-
tion—or it might prove that sometimes there are no limits.
Even after this, many descriptions would need supplementing
before a revision of species within the group could be
attempted. The second path is the study of nematocysts ;
it may prove that measurements of these will provide
clear specific distinctions. 1 believe Professor Carlgren will
bring forward a good deal of evidence in this connexion.
I have not been able myself to give this point much attention,
but what I have done rather suggests that the size of the cells
is too variable and too similar in closely-related species to help
us. Pax has a note on this in his paper on the ‘ Family
Actiniidae’, pp. 80-2. At least it becomes evident that
species-identification from preserved material, with certainty,
is going to be extraordinarily laborious. It would probably
better repay effort to take more notice of the living animals,
for here one’s experience suggests that species-identification
from colour and habit in life would usually be easy and sure.
Experience is leading me to the view that among these low
and plastic forms a species may have its peculiarities of organic
constitution at an early stage of the development of their
expression, such expression having affected colour scheme and
general facies of the living animal but not necessarily to any
extent the internal anatomy which can be studied in preserved
L12

496 T. A. STEPHENSON

specimens. If this idea can influence the study of anemones,
it will turn the attention of some workers in the direction of
refuting it by minute research and measurement ; and others
towards ‘ leaving it at genera’ and looking into the matters
of living form and broader group-problems, in any case resulting
in better knowledge of the group. Special detailed studies
of individual families should yield good fruit. In some cases
at least further work would reveal interesting and instructive
similarities and variations running through all the members
of a given family, but of a kind beyond the scope of the short
definitions to which a paper like the present is limited. It
would also reveal which families are more and which less
homogeneous, and help to clear up ideas of relationships.
I have made a preliminary study of the Chondractinidae,
for instance, which promises to be interesting in this sense.

Once more I wish to record hearty thanks to several friends
who have given me their aid in one way or another, especially
to Professor H. J. Fleure for much kindness, and to Captaim
A. k. Totton, M.C., for kind help with literature and specimens
at South Kensington. J am also much indebted to Professor
Stanley Gardiner for the loan of a collection of specimens
without the aid of which it would have been very difficult to
complete the paper.

Some of the illustrations in this paper are copied from other
sources. ‘T'ext-fig. 14, K, is copied from Plate 22, No. 2, in
W. Saville-Kent’s ‘The Great Barrier Reef of Australia ’
(W. H. Allen & Co., Ltd., 100 Southwark Street, 5.H. 1) ;
Text-fig. 19 is from a photo by Saville-Kent in ‘ The Naturalist
in Australia’, p. 224 (Chapman & Hall, Ltd., 11 Henrietta
Street, W.C. 2), and later on printed in ‘ Marvels of the
Universe’, p. 1135 (Messrs. Hutchinson, Paternoster Row,
E.C.); Text-fig. 9 is copied from ‘ Journ. Mar. Biol. Soe.’,
N.S., vol. x, no. 1, 1913, p. 73; Text-fig. 8 is from ‘ Sei. Trans.
R. Dublin Soce.’, ser. 11, vol. iv, 1889, Pl. 35, fig. 1. I wish
to acknowledge with thanks permission to print my versions
of these figures, from Messrs. W. H. Allen, Chapman & Hall,
and Hutchinson, Dr. E. J. Allen, the Science Committee of the

CLASSIFICATION OF ACTINIARIA 497

Royal Dublin Society, and the executors of the late Mr. Saville-
Kent.

2. Brier HistoricaAt SECTION.

Unfortunately space forbids the inclusion here of even
outline histories of all the families dealt with in the paper
similar to those given for Sagartidae and Paractidae in Part I.
The number of families is far greater, and possibly the historical
interest is less than in the previous case. ‘The following
details, therefore, are limited to an outline of the more usual
classifications used up to date, and which it is the suggestion
of this paper to modify.

G. C. Bourne’s scheme is the following :

Class ANTHOZOA.

Sub-class I. Octactiniaria (Octocorallia, Carlgren).
Sub-elass II. Ceriantipatharia (Hexacorallia, Carlgren).
Sub-class III. Zoanthactiniaria (Dodecacorallia, Carlgren).
Order 1. Zoanthinaria.
Order 2. Edwardsiaria.
Order 3. Dodecactiniaria.
Sub-order A. Madreporaria.
Sub-order B. Actiniaria.

The principle of his three sub-classes is that of Carlgren,
Bronn’s Thierreich, 1908.

The position of the Zoanthinaria and Edwardsiaria varies
in different schemes. In Carlgren’s 1900 plan, for instance,
the Edwardsiaria go under his group Athenaria, and the
Zoanthinaria stand away separately and rank equal to the
Ceriantharia and Actiniaria. Bourne has recently shown (9)
that the Edwardsiids must be clearly separated from ordinary
Actinians, and it is his allocation of them which is to be accepted.

The subdivision of the sub-order Actiniaria will vary
accordingly as one follows Carlgren or not. Carlgren’s division,
as used by him in ‘Ostafrikanische Aktinien’ (1900), for
example, is as follows :

498 T. A. STEPHENSON

Sub-order ACTINIARIA.

Tribe 1. Protantheae.
Sub-tribe 1. Protactininae.
Sub-tribe 2. Protostichodactylinae.
Tribe 2. Nynantheae.
Sub-tribe 8. Actininae.
A. Athenaria.
B. Thenaria.
Sub-tribe 4. Stichodactylinae.

Other arrangements ignore the Protantheae and Nynantheae,
dividing at once into Actininae and Stichodactylinae, in
which case the Protactininae rank as Actininae, the Proto-
stichodactylinae as Stichodactylinae.

The Protantheae are separated from the Nynantheae by
the possession, usually, of an ectodermal muscle-sheet and
nerve-layer in the body-wall and generally in the actinopharynx
also; and in some of them by the absence of basilar muscles,
and ciliated tracts on the mesenterial filaments. The Actininae
and Stichodactylinae, and similarly the Protactinimae and
Protostichodactylinae, are marked off from each other by
the fact that in the Actininae (and Protactininae) only one
tentacle communicates with each exocoel and endocoel, at
most, whereas in the other groups two or more tentacles grow
out from at least the stronger endocoels.

This section may suitably contain a list of the more generally-
used families, which will be convenient for reference later,
assigned to their respective positions under Carlgren’s main
eroups.

1. PROTACTININAE: Gonactiniidae, Ptychodactidae,
Halcuriidae.
2 PROTOSBTICBHODACT Y LANA Ge Corallimor-
phidae.
3. ACTININAE:
ATHENARIA: Ilyanthidae, Halcampidae, Haleampo-
morphidae, Andvackiidae, Halcampactidae.

CLASSIFICATION OF ACTINIARIA 499

THENARIA: Sagartiidae, Paractidae, Boloceridae, Acti-
niidae, Bunodidae, Aliciidae, Phyllactidae, Dendro-
melidae, Minyadidae.

4. STICHODACTYLINAE: Discosomidae, Stoichac-
tidae, Heteranthidae, Homostichanthidae, Aureli-
anidae, Actinodendridae, Phymanthidae, Thalas-
sianthidae.

This is, of course, the list as it stands without taking any account
of the present paper, even Part I of it. The work of Part I
was chiefly devoted to a revision of the Sagartiidae and
Paractidae, taking those names in the old sense as used on this
page.

oa DIscUSSION OF CHARACTERS TO BE USED IN
CLASSIFICATION.

The characters already discussed in Part I, pp. 456-68, will
of course be used here again, where they come in, but a few
others remain to be mentioned.

In the families under discussion now, there are no mesogloeal
sphincters save in Halcampa, but it has to be decided
how far the character of the endodermal sphincter is to be
trusted as a family feature. All grades of it exist, from very
weak diffuse or very weak circumscribed to very strong
circumscribed, through various degrees of diffuseness and cir-
eumscribed diffuseness (cf. Text-figs. 11 and 12). It may
be quite absent. In some families the range is not more than
from absent to weak diffuse. But in other cases there are
so many grades that one can draw no line of demarcation
anywhere ; and it must be admitted that the form and grade
of development of the sphincter cannot be used as a family
character except where it is fairly stable. The same thing
really applies to mesogloeal sphincters, but here it has been
less noticed because no one happens to have suggested an
artificial distinction between diffuse and circumscribed meso-
gloeal sphincters. .

It has long ago been realized that presence or absence of

500 T. A. STEPHENSON

verrucae and acrorhagi? cannot be used in limiting
families, and this leads on to the question of vesicles.
A certain number of forms develop, either all over their bodies
or in certain parts only, various sorts of hollow vesicular out-
srowths of the coelenteron (see Text-figs. 2, a, and 18). These
may be slightly or very highly specialized. It may be argued
that they are only verrucae which have gone farther, but in
most cases they have gone a good deal farther, and really

TrextT-Fic. 1.

A. Small portion of the upper part of the body of Bunodactis
alfordi, somewhat enlarged, to show the vertical rows of
verrucae, three of them ending above in conical acrorhagi.

B. Half a transverse section of an acrorhagus of B. alfordi.
Mesogloea black, ectoderm and endoderm white, the black
strokes in the former representing nematocysts.

seem to constitute a definite and characteristic feature by
which forms possessing them may be separated from those
which do not. Since these forms also show an agreement among
themselves in other ways, falling naturally into sets, we may
fairly take * presence of vesicles ’ as a family character for use
among others.

The presence or absence of a definite base seems a valid

‘ In this paper the term ‘acrorhagi’ is used to cover ‘ marginal
spherules ’ of any sort, whether simple or compound, whether nematocyst
batteries or not. There seems to be too much variation in their structure
for it to be possible to maintain a serviceable distinction of them into
acrorhagi, pseudo-acrorhagi, &c. A sketch of typical acrorhagi from
Bunodactis alfordi is given in Text-fig. 1,

CLASSIFICATION OF ACTINIARIA 501

and useful distinction between the Ilyanthids and the (more
or less) adherent forms, even though in special instances the
Ilyanthid condition is partly retained or imitated by others.
Text-fig. 7 shows the contrast between the two states. The
conversion of the base into a definite float as in Minyas
provides a third useful type.

Among the forms without acontia or mesogloeal sphincters
one cannot make use of presence or absence of cinclides as
might have been hoped. They have here excited so little
interest that not much trouble has been taken to find them,
and the range of their distribution is not really known. They
are recorded in some forms such as Peachia and Haren-
actis, and I must record here that I have personally observed
them very clearly in a species of Phymanthus—quite an
unexpected find. It seems to me not unlikely, from noticing
the ways of living anemones, that there may be discovered
cinclides of some sort (even if only acrorhagial perforations)
in some or even many families. A study of Actinia equina,
Anemonia sulcata, Bunodactis gemmacea, and
Tealia crassicornis in this connexion might reveal
something quite interesting—and attention should be paid
to the thin region just near the edge of the base, as well as to
the rest of the body.

Among Stichodactylines we have to deal with characters
of quite a clear-cut sort affecting form and arrangement of
tentacles, and these provide simple and satisfactory family
distinctions. (See Text-figs. 2, B, 14, 15, 19.)

Taking these remarks, together with the similar ones in
Part I, we may list some of our more useful characters as
follows :

Presence or absence of (1) a definite base, (11) a float, (i) cin-
clides, (iv) a distinction of the body into regions, (v) vesicles,
(vi) a mesogloeal sphincter, (vii) acontia, (vill) mesogloeal
disc-and-tentacle muscles, (ix) a division of the mesenteries
into macro- and microcnemes, (x) macrocnemes over and
above six pairs, (x1) perfect mesenteries over and above six
pairs, (xii) more tentacles than one in connexion with some or

502 T. A. STEPHENSON

TExtT-FIG, 2.

A. Vertical section of a whole specimen of Phyllodiscus, toshow
two vesicles (v) and two tentacles (t) cut through. Mesenteries,
&c., are omitted for clearness. B. Vertical section of a portion
of the upper part of the body-wall and outer part of the oral dise
of Cryptodendron. The section passes through many short
tentacles (¢),and although all do not belong to the same mesenterial
chamber (mesenteries are omitted for clearness), there is not by any
means only one tentacle to each chamber as at A. 8s, sphincter;
b, body-wall.

CLASSIFICATION OF ACTINIARIA 5038

all of the endocoels, (xiii) more tentacles than one in connexion
with some or all of the exocoels, (xiv) permanent tentacle-
bearing arms of the oral disc.

This is of course an incomplete list, but other characters
not needing special mention here will reveal themselves
in their respective contexts. None of the characters can be
treated in an absolutely hard-and-fast way, and may need
special consideration in special cases. Of those listed, nos. iv
and viii affect genera more than families, but are interesting
even if their presence or absence does not in itself determine
the fate of a given form. No. vi has to be taken in connexion
with the fact that sphincterless forms have to be included
sometimes with forms which have a mesogloeal sphincter,
sometimes with those possessing an endodermal one, or else
alone, according to the sum of their other characters.
Characters such as presence or absence of brood-pouches are
not of much classificatory use.

There are many other things involved in classifying Anthozoa
which will be pointed out in due course, but a few need special
mention ; they affect most, on the whole, groups larger than
families. These may be taken one at a time.

(i) Presence or absence of ciliated tracts on
the mesenterial filaments. These ciated ‘ tracts’
or ‘pads’ (Flimmerstreifen of German authors) are
very definite structures, and their presence or absence seems
to be one of the soundest indications we have of difference of
tendency between one group and another. It forms also an
easily-made-out character and one to which there is hardly
any of the usual objection of intermediate conditions between
presence and absence. Their loss, as I conceive it (or their
non-development if it were that), by the corals and by certain
anemones seems to constitute a very distinct evolutionary
step, which may be seized upon for purposes of classification.
Its usefulness both as a clue and as a sound distinction has been
somewhat swamped by the amount of attention which the next
character has absorbed ; but I propose here to lay a good deal
of stress upon it as being more valuable than no. i. The

504 T. A. STEPHENSON

contrast between the kind of filament with ciliated tracts
and that without may be seen from Text-fig. 17, where three
of the four sorts of filament illustrated have the tracts (though
not all the same kind of tract, in detail), and the fourth has
none (Cc).

(ii) Presence or absence of ectodermal muscle
in body-wall.—I this case we are dealing with a universal
ancestral character which has been allowed to die out in most
forms. It persists in those retaining most primitiveness, and
is present, at least partially or as a vestige, here and there
among more advanced forms, physiological causes probably
accounting for its retention. It can therefore only be used
in a limited way in a classification—useful in defining primitive
groups, but not a criterion of relationship when it becomes
a question of forms some of which have retained it, in greater
or less degree, and others have shed it.

(iii) Presence or absence of spirocysts in ecto-
derm of body-wall.—tThis is another character about
which a similar view may be taken to that developed in con-
nexion with the last one.

(iv) Presence or absence of basilar muscles.—
These muscles are natural developments correlated with the
stabilizing of a well-marked basal disc. Their presence is
certainly a good characteristic of the higher forms in general,
but here again it may be misleading to think too much about
them in connexion with transitional forms or forms of doubtful
relationships. For purposes of family-definitions, it appears
that the presence or absence of the base itself is the first
consideration, basilar muscles or not.

(v) Presence or absence of any perfect meta-
enemes.—One set of forms (Gonactinia, Protanthea,
and Oractis) seem well distinguished from others by virtue
of the fact that they alone among <Actinians (excluding
Edwardsiids and odd individuals among Haleampas,
Aiptasias, &c.) have the four couples of protocnemes
(the eight ‘ Edwardsia-mesenteries ’) perfect, none of the
metacnemes being so, with the result that there are no perfect

CLASSIFICATION OF ACTINIARIA 505
pairs. This, taken among other things, seems to mark them
off pretty well from other primitives, and constitutes a character
upon which one is inclined to lay more weight than has been
done hitherto—it is another, though a less important one,
the value of which has been somewhat overshadowed as in
the case of the ‘ ciliated tracts’, by the discussion of ecto-
dermal musculature. A diagram showing this type of mesen-
terial arrangement for comparison with others may be found
in Text-fig. 16, B.

4. SprciaL Discussions AND OUTLINE oF NEW SCHEME.

§A. The Gonactiniidae.

This family has been made to include Protanthea,
Gonactinia, Oractis, and Boloceroides. For pur-
poses of this discussion we shall limit it to Gonactinia and
Protanthea, with Oractis as a probable but insuffi-
ciently-known member. Boloceroides requires separate
treatment. The Gonactiniidae, then, have in common a number
of characters, most of them primitive. The smooth unspecial-
ized body has a definite attachable basal end, but without any
basilar muscles. The animal is small and delicate, and has
both the inner and outer surfaces of the whole of its mesogloea
covered by a weak generalized muscle-layer, not specially
concentrated to form definite retractors or sphincters, and
present in ectoderm of body-wall and actinopharynx as well
as elsewhere. The body-wall ectoderm also shares the character
of that of the tentacles in that it possesses spirocysts. The
mesenterial filaments are without ciliated tracts, and only
the first eight mesenteries to appear (i.e. the protocnemes,
which arise as bilateral couples and not as pairs) are perfect
(see T'ext-fig. 16,8). These undifferentiated forms seem to
come nearer than any surviving thing to the probable ancestor
of the Zoanthactiniaria (‘Text-fig. 16, 4), which, whatever it
was, must surely have had in common with them the small
size and delicacy, the generalized musculature and generalized
distribution of spirocysts, and the eight perfect mesenteries

506 T. A. STEPHENSON

only. Not only have the Gonactiniidae a good deal approximat-
ing them to this ancestor, but also there are no other forms of
this grade which can fairly be placed in the same family with
them. It seems that the family must be looked upon as one
apart, and representative of past things; the remaining ques-
tion, which will receive attention later, beg the rank of the
group to which it must be allocated.

§ B. Boloceroides.

This is a genus of uncertain affinities and needs unusually
careful placing. Carlgren has thought of it as a Gonactinud,
and others as a Boloceroid. It certainly does not come within
the Gonactinudae as understood in Section A, nor even near it.
The characters by which it may be defined, those which most
affect us at the moment, are as follows. (i) There is a definite
base, but (11) no basilar muscles. The body is (i) smooth with
unspecialized margin. (iv) There is no sphincter. (v) There
is ectodermal muscle in the body-wall. (vi) Spirocysts are
present in the body-wall ectoderm. (vu) The tentacles are
deciduous. (vill) Six pairs of mesenteries are perfect. (ix) The
mesenteries are not divided into macro- and microcnemes.
(x) There are ciliated tracts on the filaments, but (xi) no true
siphonoglyphes.

Of these characters, the genus shares nos. 1 to vi and ix
and xi with the Gonactiniidae. Character vu turns up also
in Bolocera and Bunodeopsis, and need not trouble
us, because it is a special feature which may be taken as a
convergence—not necessarily a token of relationship with
Bolocera, and certainly not with Bunodeopsis.
Characters viii and x are the two of importance in which it
differs from the Gonactiniidae, but they are rather funda-
mental. Boloceroides_ represents a _ different stage
altogether, by its possession of ciliated tracts and its attain-
ment of pairs of perfect mesenteries, although at the same
time it retains several primitive traits. It shares five characters
(i, ll, vill, ix, x) with the genus Myonanthus (a form
which, as will be seen, requires special consideration), but

CLASSIFICATION OF ACTINIARIA 507

differs from it in six others. It becomes evident that if
we treat the sum-of-the-characters principle woodenly and
mechanically here, we shall run Boloceroides into the
Gonactinidae or near them; but that will not represent
the truth. It is a case for weighing individual points, and the
best we can do for the genus is to place it near Myonanthus.
Opinion will differ as to the relative value of the various
points, but taking the general line of this paper, nos. viii
and x will count more heavily for its relationship (not close)
with Myonanthus than all its points of similarity to the
Gonactiniids. For, after all, most of those points may be
summed up as aspects of one fact, the generalized nature of
the structure ; they are primitive features not shed, and these
are more numerous than usual outside the Gonactinudae.
There are other forms with much clearer relationships which
retain some of them, e.g. Bunodeopsis.

This means the inclusion of Boloceroides either in the
same family as Myonanthus, or in a family to itself near
the one containing the latter. Some of its differences from
Myonanthus are of generic importance only (deciduous
tentacles and lack of sphincter), and the question remains
whether its ectodermal muscles and spirocysts in the body-
wall, and its lack of basilar muscles and siphonoglyptes can
separate it. Considering the fact that in other coherent
families some at least of these things may be present or absent,
it leaves the separation a matter of doubt. In the present
paper, therefore, Boloceroides will be included in the
Myonanthidae (see pp. 524, 545, 564, &c.), with the reservation
that probably there would be no harm in having a separate
Boloceroididae (under Endomyaria and next to Myonanthidae)
if preferred. The genus is evidently a transitional one.

Any close relationship between Boloceroides and
Bolocera seems a matter of doubt. Bolocera may well
be a subsequent development of the same stock, which has
attained larger size and, with this, numerous perfect mesen-
teries, retiring to deeper water and losing the primitive condi-
tion of body-wall, &c. This, however, is no argument for

508 T. A. STEPHENSON

placing Bolocera with Boloceroides, but is additional
evidence for thinking of the former as an Actiniid, taking the
view that will be developed below, that the Actiniidae are
one of the next steps on from the Myonanthidae.

IT am conscious that the arguments used in this section are
rather dangerous, and that along some such line an attack
might be developed upon the whole system of classification
by summation of characters. But I feel that it is a special
case, like one or two others, and that, as suggested in Part I
(p. 470), the summation principle must not be used blindly
like an arithmetical measure; looking upon it as useful
typically, but needing modification here and there.

§C. The Ptychodactidae.

Carlgren (1911) has shown clearly that two curious genera,
very different in detail but similar in fundamentals (Pt ycho-
dactis and Dactylanthus), should be thought of together
as forming one family. The debatable ground here is as to
where the family fits into the general scheme. Carlgren includes
it in his Protantheae with the Gonactinudae. That the
Ptychodactidae must be kept apart from the ordinary Actinians
is pretty clear; also that they must come next to the Gonac-
tinids in a list. But apart from this general location, they
seem to have very little to do with the Gonactiniids, and
should be marked off from these by being placed in a group
of their own and of higher rank than a family.

Of primitive characters they share with Gonactiniuds the
following : absence of basilar muscles although there is a base ;
similarity of structure between tentacles and body-wall—
spirocysts and ectodermal muscle in both ; sphincter little or
none; mesenterial musculature weak, hardly forming retractors.
They have no ciliated tracts on the filaments. On the other
hand they have diverged from the Gonactinuds as regards
size—they can get quite large—and have attained not only
pairs of perfect mesenteries but often a good many of them.
Ptychodactis has become very broad and has almost lost
its actinopharynx (a unique case), and has numerous tentacles

CLASSIFICATION OF ACTINIARIA 509

and mesenteries. Dactylanthus has a good actinopharynx
but has tentaculiform outgrowths of the body, curious actino-
pharyngeal pouches, and a fusion of the lower ends of the
mesenteries into a columella-like network. Further, both
genera are unique in two ways: firstly, the upper extremities
of the filaments of the imperfect mesenteries are modified into
curious structures like bisected funnels, the analogy of which

>

TEXT-FIG. 3.

One-half of a specimen of Paradiscosoma. Note the cup-
shaped form, meuth on a cone at the bottom of the cup, tentacles
reduced to knobs lining the cup. Mesogloea, &c., black. The base
was injured, and is not fully shown. The tentacles have narrow
‘stems ’ running through the thick mesogloea of the disc.

among other forms it would be difficult to suggest; and,
secondly, the gonads and filaments are confined to different
parts 6f each mesentery, the free border of the latter (or what
corresponds to it in Dactylanthus) being occupied by
filament above and gonad below, quite an unusual state of
affairs.

From this one would judge that the Ptychodactids are
a collection of curiosities which have diverged along a little

NO, 260 Mm

510 T. A. STEPHENSON

line of their own. Since they are in some ways primitive
we may place them next to the Gonactinids for convenience ;
but because of their peculiarities they should be kept sufficiently
apart from those to represent a quite distinct evolutionary
line. The exact rank of the group Ptychodacteae which
I propose for their reception will be better discussed in other
sections (see pp. 540, 552, 554-6. &e.).

§D. The Corallimorphidae and Discosomidae,

There has been a growing feeling among those who have
worked at anemones that there is a good deal of inter-relation
between them and the corals, and that we can no longer insist
on a separation of them based on presence or absence of
a skeleton alone. This feeling has been best expressed by
Duerden (120) in a study of the Madreporarian relationships
of certain Stichodactylnes. Perhaps in this connexion too
little attention has been paid to the soft parts of corals. We
are undoubtedly justified in retamimg two groups, Actmiaria
and Madreporaria ; but the justification is to be found in the
sum-of-the-characters principle, and not im the presence or
absence of skeleton merely. ‘The reservation is, that if we
maintain these two groups we must include in the Madre-
poraria some forms without skeleton. I am not familiar
enough with Madrepores to generalize about them, but am
relying on the details given in Duerden’s paper—from which
I gather that there are certain aspects of their soft parts
which present a fair degree of uniformity through the group.
With the Actiniaria, as hitherto limited, this is not the case ;
but if certain forms were removed from among them it would
be so to a more reasonable extent. There are two families of
forms, hitherto called anemones, which have all the charac-
teristics of coral-polyps save a skeleton—in fact which are
corals but for that one thing. If these two families be removed
from the Actiniaria and placed under Madreporaria in some
way, the division into anemones and corals at once becomes
more intelligible, and various difficulties disappear. The
families in question are the Corallimorphidae and Disco-

~

CLASSIFICATION OF ACTINIARIA 511

somidae,' both *Stichodactyline’. One advantage of placing
these with the corals is that they are not like the remaining
true Stichodactylines, which apart from them form a har-
monious group (see p. 533),

Two further points arise: (i) are there any corals with the
Stichodactyline arrangement of tentacles? and (ii) to which
Madreporarian families do our forms most nearly approach ?
With regard to the first it does not much matter, for a Sticho-
dactyline condition of tentacles could arise as a convergence
anywhere, and has done so among the Ceriantharia. As to
the second it is for a coral expert to suggest, and pending
further investigation the families should simply go under
Madreporaria without closer allocation.

A vertical section of one of the animals in question is
shown in Text-fig. 8. It is a cup-shaped form in which the
tentacles have become reduced to mere knobs.

What are the pomts which make these forms like corals ?
A general statement about them might be made as follows:

They secrete no horny or limy skeleton. They may be
quite solitary, or quite gregarious, sometimes living in sheets
or carpets. Frequently they reproduce by fission, and often
compound individuals with several mouths, or individuals
connected by a basal coenosarce are found. The base is adherent.
The body is without verrucae, variable in form and consistency.
More than one tentacle connects with at least the older endo-
coels. The tentacles may be simple, or capitate (cf. Caryo-
phyllia and others among corals), or branched; or small
and wart-like, or even reduced to so little as to be invisible
externally. There are no siphonoglyphes (or rarely ?). The
mesenterial filaments have no ciliated tracts. Sphincters are
feeble or absent. Sting-cells of a size characteristic of Madre-
poraria, but not of Actinians in general, are usually found
somewhere in the body. There are usually a good many

1 The Discosomidae as referred to in this connexion means the family
taken in Carlgren’s sense, 1900, p. 58, and not in the wider sense of some
authors—including only the genera Discosoma, Paradiscosoma, Actinotryx,
Rhodactis, Orinia, and Ricordea.

M m 2

T, A. STEPHENSON

TEXT-FIG, 4.

Transverse sections of mesenteries, to show various types of muscu-
lature. Mesogloea black, endoderm white. A, Epiactis;
gp, Aureliania; oc, Cryptodendron; D, Actinotryx;
E, Phymanthus.

CLASSIFICATION OF ACTINIARIA 518

perfect mesenteries, and no distinction of mesenteries into
macro- and microcnemes. ‘The longitudinal mesenterial
muscle consists typically of a feeble layer, not forming the
sort of sheet or retractor characteristic of anemones. There
are no basilar muscles, and directives may be present or not.
The ectoderm of the body-wall may or may not contain a weak
muscle-layer. The mesogloea is Madreporarian rather than
Actinian.

Text-fig. 4 shows the contrast between various sorts of
Actinian mesenterial musculature and the sort of thing found
in these ‘soft corals’. In the former there may be seen
dendrites or processes projecting from the general mesogloea
for the support of the muscle-fibres. In the soft corals the
surface of the mesogloea is typically either straight or lobed
as at p, but has a weak fringe of muscle-fibres directly upon it,
not elevated on processes. The sort of thing is better seen in
Text-fig 5. Text-fig. 6 shows Discosomid sting-cells contrasted
with typical Actinian sting-cells from acontia and acrorhagi,
&e. The general difference in size between a and B (‘soft
corals ’) and the others is very marked. c is unusually large
for an Actinian cell, p and = providing more average examples.
A Discosomid filament, showing the absence of ciliated tracts,
is to be seen in Text-fig. 17, c.

A microscopical study of a few of these forms at once suggests
a difference from the anemone type running through the
histology and other things. Even when anemones have
weak musculature it has a different appearance. These are
things which one cannot well bring out in figures without an
extensive histological demonstration, but are easy to see in
actual sections. The curiously feeble mesenterial musculature,
the presence. of very large sting-cells, the absence of ciliated
tracts, the appearance of the mesogloea and cell-layers, the
lack of siphonoglyphes, the tendency towards compound
individuals and colonies, the weak or absent sphincters, and
sometimes the strong permanent actinopharyngeal ridges and
form of the tentacles, and so on, are points which, taken
together, suggest Madreporaria, of some or all of which they

514 T. A. STEPHENSON

appear, generally speaking, to be characteristic. One or
other of them may be found among anemones, but their com-

TExtT-Fic. 5.

Transverse section of a mesentery of Paradiscosoma treated
in the same way as those in Text-fig. 4. Note the heavy meso-
gloca (black) and absence of muscle-processes.

bination indicates coral affinities. Their distinctness from
anemones in general struck me decidedly, before I thought
of them as corals.

CLASSIFICATION OF ACTINIARIA ALS

The presence of ectodermal muscle-fibres in the body-wall
of Corallimorphus, &e., is doubtless a survival. Whether
the weak general musculature is primitive in this case it
would not be safe to say; there is much to suggest that it
is a well-established thing here. Some of the other characters

TEXT-FIG. 6.
\W2
| é

B
A

Sting-cells. All are drawn to same scale, as seen with } objective
and no. 3 ocular. A (Actinotryx) and B (Paradiscosoma)
show the size characteristic of many ‘soft coral’ sting-cells.
cis an unusually large Actinian cell from acontium of Artemi-
dactis, and D (acrorhagi of Bunodactis alfordi) and
E (Halcampa aspera, body-wall) show a more average
Actinian size.

suggest advancement—the tentacles and their specialization
of form and arrangement, the big sting-cells, numerous perfect
mesenteries, and the sometimes thick and rigid bodies. ‘The
condition of mesenterial filaments they share with all corals.
Taking them all in all suggestion of primitiveness here would

516 T. A. STEPHENSON

be much less safe than in the case of Gonactiniidae or even
Ptychodactidae.

The Actiniaria as freed from extraneous skeletonless corals
show general tendencies towards more complex individuality
rather than towards colonial development, towards a special
development of musculature in some way or another, towards
different histology and on the whole more activity. They go
in for expression of permutations and combinations of various
characters, leading to great diversity—this diversity affecting
differences among polyps, whereas it is perhaps more connected
with variation of skeleton and colony-form, among corals,
which may to some extent be compared with the Aleyonaria,
although of course the latter much surpass them both in
uniformity of the individual and diversity of the colony.

§ E.

The discussions so far have dealt with curious forms which,
whatever their fate, are special cases, coming outside the
main mass of anemones. ‘Those that follow are concerned
with forms the general position of which is fairly clear, i. e. they
all come under the main tribe (Nynantheae in the sense taken
on p. 540) of the sub-order Actiniaria, excluding Edwardsians,
Zoanthids, Gonactiniids, Ptychodactids and corals whether
hard or soft—or to put it another way, they are presumably
the descendants of a muscular Halcampa-like stage
(cf. Text-fig. 8) with ciliated tracts on its filaments. Among
these forms there seem to be four main sets which can be
followed, and in the following sections the exceptional sets
will be considered before the majority-forms. .

§F. The Ilyanthidae.

There has been a family Ilyanthidae in use for a long time
((Actinies pivotantes’), for the more or less vermiform
creatures with no adherent base. It has been subdivided
somewhat arbitrarily—that it needs subdivision is not in
question, but how to do it, Although, however, we are obliged

CLASSIFICATION OF ACTINIARIA 517

to have more than one family, it seems wise to retain the old
plan to the extent of having a group to cover them, the principle
of which is good. This group must be labelled by Carlgren’s
name Athenaria, with the Edwardsiids of course excluded.
The rank of this group will be discussed in a later section,
but here we may consider the general characters justifying it.

The Athenaria appear to be the representatives of those
forms which, being the outcome of a muscular Haleampa-
stage, have retained more similarity to their ancestor than
the majority of other forms, and have kept to a more or less
burrowing life. There is variation in size; the predominating
shape is vermiform, the relation of length to diameter varying
in different cases and different states of expansion, diameter
sometimes considerable. Text-fig. 7 shows the contrast between
some of these and one of the ordinary adherent anemones with
short wide form. In these Athenaria the aboral end is not
a definite base, but a rounded physa, which is sometimes able,
however, to adhere to small objects. There is little or no
sphincter. Often there are cinclides. The number of tentacles
is usually small, and at most does not pass about forty. The
number of mesenteries is similarly limited, and either these
all have the grade of macrocnemes, or else there is a division
into macro- and microcnemes—and in Peachia the state of
affairs is intermediate. The mesenterial filaments have ciliated
tracts.

The above may be taken as a sort of definition of the
burrowers or Ilyanthids. The subdivision of the group remains
to be discussed.

Of course, some of the forms formerly included here have
long since been removed, others more recently—the Cerian-
tharia and Edwardsiaria. Forms with no base but with
acontia are little known, but seem to fit in quite well with the
Phelliidae (see Part I, p. 524), though possibly a new family
may later on be needed for them. Carlgren has suggested
a Haleampactidae, but it is here treated as coming under
Phelliidae. Andvackiidae is not yet established. For Hal-
campactis see Part I, pp. 499, 509, 525.

518 T. A. STEPHENSON

The forms we are here concerned with are Halcampa,
Halcampoides, Pentactinia, Scytophorus, Har-
enactis, Kloactis, Peachia, Haloclava, Ilyanthus,
and Andresia.

If we go into detail about all these forms we shall find that

—

TEXT-FIG. 7.

A, Peachia hastata; B, Tealia crassicornis; © and D,
Halcampa chrysanthellum. To emphasize the contrast
between burrowing and adherent forms. All are natural size.

almost every one could claim distinction for one reason or
another ; because there is diversity in rather important ways.
But it would seem extravagant and hardly justifiable to give
a family to each, and failing that we have to do the best we

1 Andresia is a new name for Ilyanthus parthenopeus,
which is quite unlike the more typical British I. mitchelli, and has

to be separated as a distinct genus with new name. This will be formally
established in Part III.

CLASSIFICATION OF ACTINIARIA 519

can, allowing a fair latitude of definition. It is possible to
gather them into three fairly clear sets, which must be our
families. It seems impossible to be content with a subdivision
which has already been suggested, and based on the nature of
the sphincter only—-into Haleampidae, Haleampomorphidae,
and Ilyanthidae. ‘This, among other things, means that
Halcampa and Halcampoides go into different families,
and this seems to be straining things.

TEXT-FIG. 8.

ee
(aN

Transverse section of Halcampa chrysanthellum, showing
six pairs of macrocnemes and six pairs of microcnemes. 4a, acti-
nopharynx; 0, body-wall; m, microcneme; _ r, retractor.
(After Haddon. See acknowledgement on p. 496.)

Taking first the genera Halcampa, Halcampoides,
Pentactinia, and Scytophorus, we can make for
these a fairly precise definition, and call them Halcampidae.
They are Athenaria of more or less vermiform shape, with or
without suckers or papillae or cuticle or incrustation on the
body. ‘There may be cinclides in the physa. The tentacles
may be 8-12, 14, 20, or more, and their longitudinal muscula-
ture is ectodermal. The sphincter is absent, or weak endo-
dermal, or weak mesogloeal. The mesenteries have as their

520 T. A. STEPHENSON

main feature six pairs of macrocnemes; but there are varia-
tions ; the full six pairs may not be developed (Pentactinia
and some individuals of Halecampa), or there may be an
extra couple (Seytophorus). Microcnemes may be present
or not.

This idea regards the genera Halcampa and Halcam-
poides as constituting, jointly, types of the family, and no
separation of these on account of sphincter is wise. It brings
in Pentactinia and Scytophorus, the one as a slightly
under-developed, the other as a slightly over-developed,
Halcampa-form. Indeed, these two are very like Hal-
campas but for mesenterial oddities slightly deviating from
type. A case parallel to that of Pentactinia is that of
Decaphellia, a Phelliid with subnormal number of macro-
cnemes. Text-fig. 8 shows a transverse section of a Haleampid
for contrast with that of one of the Ilyanthidae in the strict
sense as described in the next paragraph and illustrated in
Text-fig. 9.1

If we now take the genera Ilyanthus (mitchelli),
Harenactis, Eloactis, Peachia, and Haloclaya,
we find a rather different type of structure. The mesenteries
are never fewer than ten pairs in adult animals, and vary
up to about eighteen pairs. They all have virtually the grade
of macrocnemes, even though there may be differentiation
among them—except that in Peachia some of them are
devoid of filament and gonad, but have strong retractors and
are not microcnemes. Tor the rest they often attain fair size
and may have stout bodies (capable of becoming vermiform)
or very long ones. Suckers present or absent. Cinclides may
occur. ‘Tentacles simple or capitate, eight, twelve, twenty, or
more, up to about forty. Little or no sphincter. There may
be only one siphonoglyphe, which in Peachia is specialized
into a conchula. In Peachia we have six perfect pairs of
mesenteries (or rarely fewer ?) and four secondary pairs; in

1 In this figure the gaps in the mesenteries are due to the fact that

the section passes through the region of mesenterial stomata—in most
regions the mesenteries would be continuous.

CLASSIFICATION OF ACTINIARIA 591

Kloactis and Haloclava ten pairs, all perfect; in
Harenactis twelve pairs in two cycles, all perfect; in
Ilyanthus the number of mesenteries varies, but is the same
as the number of tentacles, and all are perfect—but there are
some individual peculiarities as well.

Unless there is to be much multiplication of families the
above arrangement seems the best.

TEXT-FIG. 9,

Transverse section of Eloactis mazeli. The gaps in some of the
mesenteries are due to mesenterial stomata. Ten pairs of macro-
cnemes and no microcnemes. a, actinopharynx; 0, body-wall;
r, retractor. (After O. M. Rees. See acknowledgement on
p. 496.)

There remains the case of Jlyanthus parthenopeus—
or Andresia parthenopea as it must now be called.
This form does not seem to fall in well with the usual idea
of Ilyanthid structure, apart from its form and rounded
aboral end. It has long tentacles in four regularly-graded
cycles, and twenty-four pairs of mesenteries in three graded
cycles. The mesenterial musculature appears to form only a
weak layer, not rising into a thick (and typically circumscribed)

522, T., A. STEPHENSON

retractor or pad as in all other Ilyanthids. The body-margin
is notched in a way suggestive of acrorhagi. In fact, but for
its burrowing habit and rounded end, it would be a typical
member of the Actiniudae of the less muscular sort. Whether
it is an Ilyanthid which has passed the usual grade of develop-
ment and moved towards that of adherent forms, or whether it is
a retrograde adherent which has gone back to buried life and
lost its base, we cannot tell. But in classification it ought to
be separate, or probably go nearer the early Actinuds than the
Ilyanthids. I have in this paper made a family Andresidae
for it, placing this among the earlier Endomyaria (see Part IJ).

With regard to other forms without bases excluded from the
Athenaria (see Part I) these fit in better with the Mesomyaria
(see pp. 541, &c.) than with the Athenaria, because of their
acontia and mesogloeal sphincter, &c. In the case of some of
them (Phelliidae in part) we have our finger on the transition
from burrowing to adherence, and there are grades from
a physa to a well-marked base ; and as these seem to be getting
up to the attached stage it seems better to keep them out of
the Athenaria, especially since their acontia and mesogloeal
sphincter and other things show their relationship to be with
the Mesomyaria. Some of the Diadumenids are also almost
without base, but here it is obviously a case of retrogression or
arrested development; they are probably normally adherent
forms changing under special conditions.

§G. The Endocoelactids.

These forms start from a six-pairs-of-muscular-mesenteries
or Halecampa-stage basis, with ciliated tracts on the mesen-
terial filaments, but work onward from this in quite an unusual
way. ‘The secondary mesenteries appear in the endocoels
of the lateral primaries, and all of them have the character of
directives (i.e. the retractors of each pair face away from one
another). The usual plan is, of course, for the secondaries to
appear in the primary exocoels, and have their retractors
vis-a-vis. The contrast is indicated in Text-fig. 16, @ being
an Endocoelactid. Apart from this most fundamental structural

CLASSIFICATION OF ACTINIARIA 523

aberration, the Kndocoelactids are sphincterless, and nearly
always have spirocysts in the body-wall ectoderm. ‘There is
a definite base. The form of body and tentacles is variable,
and may be ordinary, but the wall may be thick and heavy,
the dise lobed, the tentacles often with aboral basal swellings.
In fact we find here a tendency not found on the main line of
Endomyaria (see p. 541, &e.) or * endodermal-sphinctered ’
anemones, towards a deep-water specialization similar to that
which we found earlier on in certain Paractids and Actino-
scyphids, &e. (see Part 1). Taking them as wholes, the Endo-
coelactids are a set very different from average forms, being
apparently a little line of evolution to themselves; and as
such they should have slightly higher rank than that of a family,
forming a group Endocoelactaria equal in level to the
Athenaria.

Carleren includes Endocoelactids in his Protantheae along
with Gonactiniidae and Ptychodactidae ; but since they seem
evidently derived from a muscular Halcampa-like ancestor
with ciliated tracts, and have no ectodermal muscle in their
body-walls, I cannot see the merit of that plan, or accept it.
(See also pp. 541-2, 560, &c.)

There are among the Endocoelactids two rather clearly
marked out groups, one of them containing Halecurias
and Carlgrenia, the other Actinernus and_ three
related genera. The two groups seem to have fairly good
claims to be regarded as families, and as such they are defined
later on in this paper (Part II). There is in one of the families
practically a division of the mesenteries into macro- and
microcnemes (macrocnemes six to ten pairs, with circum-
scribed retractors, gonads, and filaments; microcnemes
confined to upper part of body except for four pairs of them
in Carlgrenia—some of them may be perfect, but without
retractor, gonad, or filament), and also there is constantly
one siphonoglyphe only and no tendency to lobing of dise
or tentacles. These forms, especially Carlgrenia (Text-fig.
16, a), are nearer their Haleampid ancestor than the others.
In the other family we find the lobing tendency and charac-

524 T. A. STEPHENSON

teristically thick body-walls, two siphonoglyphes, and numerous
mesenteries, the older ones at least fertile and not much
marked off from the others, many being perfect and their
musculature not strong. The first family is the Haleurudae
sens. strict., the second the Actinernidae.

§.

The next five sections will deal with the ‘ Sea-Anemones ’
in the narrowest sense (i.e. such of them as were not dealt
with in Part I), the usual forms, the majority-forms, exclusive
of atypicals such as Athenaria and Endocoelactaria and the
pre-Halcampid groups.

§J. The family Actiniidae.

This family, containng our commonest and most familiar
anemones, has been the subject of a good deal of discussion
and fluctuation. As it is usually understood at the moment,
it is not much more homogeneous than the old group * Parac-
tidae ’, but contains three distinct types of mesenterial arrange-
ment. Any discussion of it involves also the families Bolo-
ceridae and Bunodidae, and these points will be dealt with
in order.

Firstly, the Actinudae. If we consider the aggregate of
genera usually included here—Actinia, Anemonia, Con-
dylactis,Gyrostoma, Actinioides, Condylanthus,
Myonanthus, Macrodactyla, and others, we find three
types of mesenterial formula, as follows :

(i) In Condylanthus the mesenteries are divided into
macro- and microcnemes, the macrocnemes numbering six
pairs (cf. Text-fig. 16, c).

(ii) In Myonanthus and Macrodactyla there is no
division of mesenteries into macro- and microcnemes, but
only six pairs are perfect (cf. Text-fig. 16, D).

(iii) In the others there is no division of mesenteries into
macro- and microcnemes, but there are numerous perfect
mesenteries as a rule, always more than six pairs in adults
(cf. Text-fig. 16, u, and Text-fig. 10).

CLASSIFICATION OF ACTINIARIA 525

Without going over the old arguments again, we take it
that if the ideas advocated in this paper be accepted at all, it
will have been made clear in Part I that forms exhibiting
these grades of mesenterial development need separation.
We have, therefore, at once three families, Condylanthidae,
Myonanthidae, and <Actiniidae sens. strict., and these will
be defined in Part II]. This gives a homogeneous and intel.
ligible Actiniidae, and has the advantage of providing two

TExtT-FIG. LO:

Diagram of a transverse section of Phymactis clematis, show-
ing numerous perfect mesenteries.

families as links between the Actiniidae and Haleampidae,
from near which they presumably arose. ‘The three families
might be compared with, for instance, the Diadumenidae,
Metridiidae, and Sagartiidae of Part I—in which we have
the same three grades of mesentery development, but acontia
and cinclides and mesogloeal sphincter in all. In our new trio
there is a common absence of acontia and mesogloeal sphincters
and also of vesicles—as to cinclides it is hardly safe to say
anything.

NO, 260 Nn

526 T. A. STEPHENSON

It has been so generally recognized that smooth-bodied and
verrucose forms, and forms with and without acrorhagi cannot
be separated into different families, that it seems hardly
necessary to discuss this here.

Secondly, there is the question of a separate Boloceridae.
Such a fanuly has been in use by some authors, and originally
I felt a need for it (see 1918 a, p. 19), but further work has
changed that feeling. It hardly seems that the deciduous
tentacles are a character giving the Boloceras any night
to separation, and otherwise they are exactly Actiniidae. This
is especially the case since Boloceroides and Buno-
deopsis have also the deciduous tentacles, and neither of
them could be imcluded in a Boloceridae im any case. One
has to think of the cases as convergences. Evenif Bolocera
and Bunodeopsis should be further stages, along different
lines, from a Boloceroides-like ancestor. this is no reason
for classing the three together.

Thirdly, the Bunodidae. It seems a pity to have to attack
an old-established family like this, but at the same time there
seems to be no valid way whatever of separating it from the
Actinudae (in the revised sense), with which it is continuous.

Originally the Bunodidae relied fer separation upon their
verrucae and their strong circumscribed sphincters. The
verrucal character was swept away by Epiactis and Iso-
tealia, which are without it. We must now tackle the
sphincter. In the first place the sphincter in Bunodactis
(Bunodes) itself is variable, and often not a strong one. In
the type-species, B. gemmacea, it may be half diffuse in
some cases (I have sections of a very typical specimen showing
this—see ‘Text-fig. 11,p), and poorly developed. It is in
Tealia and Epiactis (Text-fig. 12, a, B, c) that the really
strong sphincters are found, and even there the size varies with
species and individual. Further (this will be dealt with again
under Bunodactis in Part III), there are apparently no
criteria by which Bunodactis can be separated from
Anthopleura and Actinioides, even generically—the
three run right into each other and really form one large genus

CLASSIFICATION OF ACTINIARIA ESAT

varying as to sphincter from weak and diffuse to fairly strong
and circumscribed—with too many grades to draw a line
anywhere (a few are illustrated in Text-figs. 11 (p, Fr) and
12 (p, )). And Actinioides is one of the genera hitherto

Text-riac. 11.

i

Vertical sections of the sphincters of some Actiniidae, all drawn to
same scale; showing various grades. sa, Condylactis (here
the sphincter is ahsent or practically so, what is shown being
simply the upper part of the ordinary endodermal circular
muscle), B, Actinia equina; co, Bolocera tuediae;
Dp, Bunodactis gemmacea; F, Anemonia sulcata;
r, Bunodactis alfordi. B and © are typical diffuse
sphincters. D, FE, and F are more of the circumscribed-diffuse
grade. In all of them mesogloea is black, ectoderm and endo-
derm white.

classed as Actiniid. Proceeding still further towards the typical

Actiniidae, if a comparative study of, for instance, Buno-

dactis (Anthopleura) alfordi, a Condylactis,

and Anemonia sulcata be made, there is too much
Nn2

528 T. A. STEPHENSON

similarity between them for any separation greater than generic
to affect them.

In B. alfordiand A. suleata there is a definite base,
there are acrorhagi, long tentacles, lax habit of body only able
to retract with great difficulty, similar habitat, weak to
moderate circumscribed or circumscribed - diffuse sphincter
(Text-fig. 11, 5, F) (sometimes more diffuse in suleata),
numerous perfect mesenteries with fairly strong retractors,
gonads on most of the older mesenteries, and the longitudinal
musculature of the tentacles ectodermal. ‘The chief difference
is that B. alfordi has rows of verrucae which A. suleata
has not, and of course lesser species-differences. But obviously
the relation is too close for the two to be included in different
families, which has been done hitherto.

In B. alfordi and a Condylactis of which I have
specimens, there is a definite base, there are verrucae, good
tentacles, lax habit, numerous perfect mesenteries with fairly
strong retractors, gonads on most of the older or all the
mesenteries and ectodermal tentacular muscle. Here the
main differences are lack of acrorhagi and a sphincter in
the Condylactis. Between the pomts here given the
similarity of the three genera should be clear. It is not always
easy to distinguish them from each other if dealing with
preserved material.

These things being so, where is the line to be drawn between
Actinudae and Bunodidae ? Given a series of forms—such as
Anemonia, Condylactis, Bunodactis (incl. Actinio-
ides and Anthopleura), Tealia—-where is the division ?

Condylactis gives us verrucae but no acrorhagi and little
or no sphincter; Anemonia has the acrorhagi but no
verrucae, and a weakish circumscribed or diffuse sphincter ;
B. alfordi has both verrucae and acrorhagi and a moderate
sphincter, + circumscribed (its relations showing other grades) ;
and Tealia has verrucae (and rarely acrorhagi) and strong
circumscribed sphincter.

The conclusion seems to be, clearly, that Bunodidae must
be abandoned altogether. It should be noted that this does
not impair the homogeneity of the Actinudae, except as regards

CLASSIFICATION OF ACTINIARIA 529

TExtT-FIa. 12.

Further sphincters of Actiniidae for comparison with those in Text-
fig. 11. All are to same scale as Text-fig. 11, and treated in the
same way, but here we have various grades of circumscribed
sphincter. The difference in size between A and B, for instance,
is not due to a corresponding difference in size of the individuals
from which they were taken. a, Epiactis sp.; B, Tealia
crassicornis; 0, Epiactis novozealandica; D,
Bunodactis sp.; £, Bunodactis balli,

530 T. A. STEPHENSON

srade of sphincter ; and that, it is evident, is bound to vary
within the limits of some families.

A series of sphincters is illustrated in Text-figs. 11 and 12,
all of them being taken from Actiniudae in the new sense.
A more evenly graded set could, I think, be made, but 1 have
not material for it. But this one brings out the facts that
I have wished to emphasize fairly well.

§K.. The Forms with Vesicles.

The anemones provided with vesicles should (see p. 500)
be kept apart from those without them, but among themselves
there are two kinds at least.

Taking the vesicled genera together, one can list nine
clearly-distinguished ones—Alicia, Phyllodiscus, Thau-
mactis, Bunodeopsis, Phyllaetis, Phymactis,
Cradactis, Cystiactis, and Lebrunia.

There have been families in existence to cover these forms
(Aliciidae, Phyllactidae, Dendromeliidae, Thaumactidae), but
the definitions have been based chiefly on the form and situa-
tion of the vesicles, and this seems as unnatural as it used
tu be to separate Hormathia, Chitonactis, Chito-
nanthus, and Chondractinia on account of variation
in ridges and tubercles ; and it has not been a very intelligible
arrangement. So long as the vesicles are present, that is
the family-character ; their form and situation are more
questions of generic distinction. From this point of view the
families fall to the ground. ‘The Dendromeliidae lapses in any
case ; it was formed to cover Lebrunia and Ophiodiseus.
Ophiodiscus seems to be a typical Paractid (sce Part I,
p. 560), and in the present state of our knowledge it seems very
doubtful whether, although it is a distinct enough genus, there
is anything to keep Lebrunia out of the Phyllactidae.
The genus Thaumactis does not seem worthy, as we know
it, to have a family to itself either. The other two families
(Aliciidae and Phyllactidae) must be retained, but revised in
the light of mesenterial arrangement, &c.

(a) Alicia and Phyllodiscus are delicate creatures
with vesicled scapus and naked capitulum, or with the vesicles

cece at ORS sae

®
.
7

CLASSIFICATION OF ACTINIARIA 531

at or at and above the scapo-capitular junction. There is little
or no sphincter, and only six pairs of mesenteries are perfect.
For these the name Aliciidae should be kept, and for these only.
(b) The other genera are provided, usually, with numerous
perfect mesenteries, have various arrangements of vesicles, may
be less delicate, and have sometimes circumscribed sphincters.
Some of them have mesogloeal longitudinal musculature in
the tentacles. This collection has to be covered by the name
Phyllactidae, and at that it had better be left for the moment.
A fuller discussion of the family will fit better mto Part III,
where the generic definitions will be available for reference.

§L. The Actiniidae and Vesicled Forms
together.

If the old Actinudae and Bunodidae and _ Boloceridae
(re-sorted into the new Actiniudae and Condylanthidae and
Myonanthidae) and the Aliciidae and Phyllactidae be taken
together, a mass of forms is presented exactly comparable
to the set classified in Part I. It is worth while seeing whether
they will work into a companion table like that on p. 481 in
Part I. It is unfortunately necessary here to leave out cinclides,
as there are not enough data about them. And of course
absence of acontia and mesogloeal sphincter go right through.

No Vesicles. Vesicles present,
Mesenteries divided into Condylanthidae. |
macro- and microcnemes. Condylanthus. No known representatives.
Number of macrocnemes six |
pairs. 7
Mesenteries NoT divided Haconanthidas. Aliciidae.
into macro- and micro-| Myonanthus. | Alicia.
enemes. Number of perfect | Macrodactyla. Phyllodiscus.
mesenteries six pairs. | Boloceroides.
| Nevadne. 7 8
Mesenteries Not divided Actiniidae. Phyllactidae.
into macro- and micro- Actinia. Bolocera. | Phyllactis. Lebrunia.
enemes. Numerous perfect Anemonia. Leipsiceras. _ Phymactis. Thaumactis. |
mesenteries, or at least more | Gyrostoma. Ixalactis. | Cradactis. | Bunodeopsis |
than six pairs in normal | Condylactis. Pseudophellia. | Cystiactis. |
individuals. _ Bunodactis. Boloceropsis. |
| Tealia. Glyphostylum. |
| Epiactis. Parantheopsis.
| Isotealia. Dofleinia. 6 6)

In this table the number in the corner of each square is the number of characters which the
members of the family in that square have in common.

532 T. A. STEPHENSON

Here the same characters as before go down the side of the
table, but there are fewer to go above. And even without using
the full number of combinations possible there is an empty
square, no forms being yet described to fill it. Perhaps some
will turn up, or perhaps it indicates that vesicles are structures
not developed at stages of mesenterial evolution such as that
represented by the Condylanthidae.

The diagram representing evolution in this group, as far
as one may understand it, and for comparison with that given
on p. 504 in Part I, would be as below. More will be said
about it in the evolutionary section of the paper. An ancestral
Eoactiniia corresponding to the Kosagartia on the
other lme, may be imagined—a good deal like a Halcam-
poides.

oactinia

Condylanthidae

Myonanthidae Actiniidae

This diagram shows the Aliciudae and Phyllactidae as
parallels, and involves the assumption that they arose indepen-
dently from the main line, as some of them at any rate may
probably enough have done. If they had a common origin
among the pre-Actiniids, and the Phyllactids changed their
mesenterial arrangement afterwards, or if some Phyllactids
arose from early Actiniid forms and others from Aliciud forms,
that would modify the diagram, but it is all speculation.
Further details about it will be found under Phyllactidae in
Part ITI.

CLASSIFICATION OF ACTINIARIA 533

§M. The Minyadidae.

Probably unrelated forms have been placed here. There is
little evidence of their relationship to each other, and there
are few data altogether. It may later be found that there
is no need for a Minyadidae, and that the contained forms
may be allocated to different families as floating members.
One form at least, Nautactis olivacea, Les., seems to be
some sort of Stichodactyline. At the moment only Sticho-
phora torpedo, Bell, can be defined, so far as I am aware,
and that not fully ; but for this form there seems to be justi-
fication for a family Minyadidae, even if it is not very clear,
based on the definitely float-like character of the base taken
with other things. At the present time it seems inopportune
to say much about it, with the provision that so far there is no
evidence of its ability to sustain higher than family rank,
and it seems to fit in near the Actiniidae. If further details
come to light—if, for instance, $8. torpedo should have no
ciliated tracts on its filaments—the position of the family
will need reconsideration.

§N. The Stichodactylines.

Here I have no suggestions to offer (save that already made
about the Corallimorphidae and Discosomidae), but am _ pre-
pared to accept fully the families defined by Carlgren in his
‘ Ostafrikanische Aktinien’, 1900, and (Aurelianidae) in
a smaller paper on Stichodactylines, also in 1900. These
families seem to be excellently based and to represent relation-
ships very naturally. They are the Stoichactidae, Thalassian-
thidae, Actinodendridae, Phymanthidae, Aurelianidae, Heteran-
thidae, and Homostichanthidae. ‘They entirely supersede
other arrangements, including Duerden’s division of the group
into Homodactylinae and Heterodactylinae; they will be
defined in Part III.

There are a few points worth noting about the Stichodacty-
lines in general, excluding always the Corallimorphidae and
Discosomidae (this latter in the sense taken by Carlgren,

534 T. A. STEPHENSON

1900). In their main structural features they form a homo-
ceneous group. In all of them (with rare exceptional indivi-
duals) there is more than one tentacle situated over at least
some of the endocoels, and often over all the endocoels and
even exocoels as well. The contrast between a Stichodactyline

TErxt-FIG. 13,

Vertical sections of two sphincters of Stichodactylines. a, Actino-
porus; B, Aureliania.

hike Cryptodendron and an ordinary form (as regards
tentacles) like Phyllodisecus is brought out in Text-fig. 2.
A shows a vertical section of a whole specimen of the latter,
and passes through one tentacle on each side of the mouth.
Bis a vertical section through a corner of the oral dise and body-
wall of the former, and shows many short tentacles cut through
in the same section—they do not all belong to the same mesen-
terial space, but they have not by any means one space to

Text-Fic. 14.

Arrangement and form of tentacles. a, Aureliania, sector of
disc, with various sorts of tentacles, and two to each exocoel
and endocoel. B (surface view) and c (side view), of a dendritic
tentacle of Actinotryx. In c the stem is embedded in the
mesogloea of the disc. D is a sector of the disc of an Actiniine
form with plain tentacles, one to each endocoel and exocoel.
E, a radial group of dendritic tentacles and nematospheres from
Thalassianthus. All this belongs to one endocoel. ¥, sector
of dise of Antheopsis for contrast with D, showing some of the
tentacles in endocoelic radial rows; here some of the tentacles
have been cut off for the sake of clearness. G, knobbed tentacle
of Corallimorphus. 4, sector of disc of Phymanthus,
showing marginal pinnate tentacles in alternating cycles and
small disc-tentacles in radiating rows. The pinnate character is
not very clear, as the specimen was distorted. kK, arm or disc-
lobe of Megalactis, bearing tentacles. (After Saville-Kent.
See acknowledgement on p, 496.) In A, D, F, exocoels are white,
endocoels shaded,

536 T. A. STEPHENSON

each. The contrast is differently brought out in Text-fig. 14,
p and F, which represent two sectors of the oral disc of two
forms. One of these (bp) has the ordinary arrangement of
tentacles in alternating cycles, one to each mesenterial space ;
the other (F) is from an Antheopsis, and shows two of the
long radial rows situated over endocoels (which are shaded)
and also the arrangement of the marginal tentacles, one or
two to each endocoel, one to each exocoel (exocoels are not
shaded).

Text-Fiq. 15.

Diagrams of three types of tentacular arrangement. In each diagram
three cycles of mesenteries are shown, with their retractor
muscles as black thickening; endocoelic tentacles black, exo-
coelic white. In A there are cycles of tentacles, one tentacle
only to each exocoel and endocoel. In B there are radial rows
on the endocoels, but only one tentacle per exocoel (e. g. Stoichac-
tidae). In c the exocoels as well as the endocoels have more
than one tentacle (e.g. Homostichanthidae).

Stichodactylines have a definite base (occasionally reduced
and half like a physa). Cinclides are recorded in at least one
case (see p. 501). There is a complete absence of acontia and
mesogloeal sphincters, and almost complete absence of vesicles
(there is one case of somewhat vesicular verrucae). The
musculature is always reasonably well developed, at least in
the mesenteries. here is either no approach to a division of
the mesenteries into macro- and microcnemes, or if there is,
there are at least twelve pairs of the macrocnemes ; in the
first case there are usually numerous perfect mesenteries.

CLASSIFICATION OF ACTINIARIA 5a

Fundamental differences affect chiefly form and arrangement
of tentacles and strength of musculature, and details about
this will be found in Part III. Text-figs. 14 and 19 show
some of the variation in tentacle-form to be found among
Stichodactylines and skeletonless corals ; Text-figs. 14 and 15
show some of the modes of arrangement; and Text-figs. 4
and 13 give details of musculature.

Taking them all in all it may be said that the Stichodactylines
are the nearest analogue among anemones to the composites
among plants or the birds among vertebrates. A good deal
of fundamental structure is fixed, and variation more affects
details or additional features. They may be looked upon as
Endomyaria (see below) with, above all, tentacular specializa-
tions, often of a frilly nature. ;

§ O.
So far, taking this paper and Part I together, it has been
sought to establish the thirty-two families listed below. It

remains to discuss main subdivisions of the Anthozoa and
arrangement of families within groups of higher rank.

Corallimorphidae. Actinidae. Diadumenidae.
Discosomidae. Aliciidae. Phelliidae.
Gonactiniudae. Phyllactidae. Flosmarinidae.
Ptychodactidae. Minyadidae. Marsupiferidae.
Halcampidae. Aurelianidae. Metridudae.
Tlyanthidae. Stoichactidae. Chondractinudae.
Halcuriudae. Homostichanthidae. Actinoscyphiidae.
Actinernidae. Actinodendridae. Sagartiidae.
Condylanthidae. Heteranthidae. Choriactidae.
Myonanthidae. Phymanthidae. Paractidae.

Andresudae. Thalassianthidae.

538 T. A. STEPHENSON

§ P. The Groups larger than families, and the
Arrangement of the families within these
Groups.

At this poimt discussion becomes more difficult and more
dependent upon individual opinion. It may be simplest to
start with the class Anthozoa and work downwards. Neither
nomenclature nor main subdivisions are my special concern
here, but probably no one will object to one of the following
alternatives, whatever names be preferred.

Bourne’s division is

Sub-class 1. OcracTiNniaRiA.

Sub-class 2, CERIANTIPATHARIA.

Sub-class 3. ZOANTHACTINIARIA,
Or one could use

Sub-class 1. OCTACTINIARIA.
Sub-class 2, CERIANTHARIA.
Sub-class 3. ANTIPATHARIA,.
Sub-class 4. ZOANTHACTINIARIA.

Hither of these is a good arrangement, probably, leaving aside
the vexed question of Tetracorallia—it has recently been
suggested that these may have something to do with the
Endocoelactid type of structure.

The next step is the subdivision of the Zoanthactiniaria.
Few will object to having the Zoanthids as a separate set
among them, and although Edwardsiids are sometimes included
with ordinary anemones, Bourne has recently shown that
they must rank as a distinct group equal to that containing
the Zoanthids. So the Zoanthactiniaria may be divided into
three or four, with a number of common tendencies (see p. 551).

Bourne subdivides thus :

Order 1. ZOANTHINARIA.

Order 2. EDWARDSIARIA.
Order 3. DOoDECACTINIARIA.

His order Dodecactiniaria includes the sub-orders Actiniaria
and Madreporaria. Carlgren, however, divides — slightly
differently.

CLASSIFICATION OF ACTINIARIA 539

Order 1. ZOANTHARIA.
Order 2. ACTINIARIA.
Order 3. MADREPORARIA.

To this, now, Edwardsiaria would have to be added. In
_ this paper Bourne’s division will be used. It is when we come
to the subdivision of the sub-order Actiniaria that the main
divergence of opinion begins.
Carlgren divides into
Tribe 1. PROTANTHEAE.
Tribe 2. NYNANTHEAE.
Another division in use is

Tribe 1. AcTINIINAE.
Tribe 2. STICHODACTYLINAE.

In the following paragraphs I shall indicate the lines of
grouping which I wish to suggest, giving an outline only.
Further reasons, filling in this outline, will be found in various
parts of the paper, especially under the foregoing sections
dealing with sets of forms individually, and in the later evolu-
tionary discussions.

Much has been said about Carlgren’s division into Protantheae
and Nynantheae, and it has been rejected by some workers,
at any rate, in the sense in which Carlgren uses it. It is based
mainly upon the presence or absence of ectodermal muscle
and a nerve-layer in the ectoderm of body-wall and actino-
pharynx ; and this, as has been suggested before, is probably
a universal ancestral character surviving in more or less primi-
tive forms and, otherwise, in sporadic cases. I cannot accept
it as a good basis of distinction in itself, although it helps
to show relationships, in some cases, when taken with other
things. In this attitude I believe I am in agreement with
Haddon (1898, p. 411), Duerden (1900, p. 187, and 1902),
MeMurrich (1904) and Bourne. At the same time I accept
decidedly Carlgren’s Protantheae, but in a different and much
more restricted sense. I have tried to show that Carlgren’s
Protostichodactylines (a sub-tribe of his Protantheae) (and
also the Discosomidae) are corals (see p. 510), and this restricts
his Protantheae to Gonactinudae, Ptychodactidae, and the

540 T. A. STEPHENSON

Endoceelactids. As mentioned on p. 523, the Endocoelactids
seem to be definitely post-Haleampid and Nynanthean, and
will here be treated as such. This leaves us with the Gonactini
idae and Ptychodactidae (see pp. 508, &¢.); and I feel that
these represent two different side-lmes of evolution, not
necessarily very close together even though both have some
primitive features, and that in this case it is safer to
give each a group, the two equal in rank. I therefore
propose to limit the Protantheae to the Gonactiniidae (in
the sense taken on p. 505 and exclusive of Boloceroides),
and to erect a group Ptychodacteae for the Ptychodactids,
equal in rank to Protantheae and Nynantheae. The Nynan-
theae I accept as the main tribe, provided it include
Boloceroides (see p. 506) and the Endocoelactids (see
p. 522), and exclude the Edwardsids and ‘ soft’ corals (see
p- 510); and provided that not only it, but also the other
tribes, be re-defined on the sum of their main characters and
not on the presence or absence of ectodermal muscle in the
body-wall, simply.

My suggestion for subdividing Actiniaria is therefore this
one :

Tribe J. PRoTANTHEAE (including Gonactiniidae only, and not Bolo-

ceroides).
Tribe 2. PrycHopacTEaE (including Ptychodactidae).

Tribe 8. NyNANTHEAE (including Boloceroides and Endocoelactids,
and majority-forms, excluding Edwardsiids and skeletonless corals).

With regard to the other subdivision of Actiniaria into
Actinunae and Stichodactylinae—I used this, provisionally
only, in Part I, but am letting it lapse here in favour of the
above scheme. One feels that these groups have been useful
as a half-way house, but that in the light of developing know-
ledge of the group, it is now possible to go farther. The
‘ Actiniine’ condition is found in all Nynantheae save one
section; it prevails also in Protantheae, Ptychodacteae, and
most corals. ‘ Stichodactylinism’ occurs in Ceriantharia and
a few corals, and in one set of Nynantheae. There are, however,
among Nynantheae, four quite distinct sets, seemingly repre-

CLASSIFICATION OF ACTINIARIA 541

senting four lines of evolution; and the * Stichodactylines ’
form a compact group within one of these four sets. These
four groups can be defined by the sums of their main characters,
and clearly the Actiniine-Stichodactyline contrast must be
used simply in connexion with a subordinate division of that
one of the four groups in which it occurs—if it be used at all.
This is only making it one degree more subordinate than
Carlgren does in his scheme. It is evident that as primary
subdivisions of Actiniaria the two groups are no longer adequate
—they must be reduced in rank, at least, from tribes to less
than sub-tribes.
Carlgren’s scheme is :

Tribe NYNANTHEAE.
Sub-tribe 1. Actiniinea.
a. Athenaria.
b. Thenaria.
Sub-tribe 2. Stichodactylinae.

The grouping I wish to suggest, as expressive of the above-
mentioned four main lines of Nynanthean evolution, is :
Tribe NYNANTHEAE.

Sub-tribe 1. Athenaria.

Sub-tribe 2. Endocoelactaria.

Sub-tribe 3. Mesomyaria.

Sub-tribe 4. Endomyaria.
a, Actiniinae.
b. Stichodactylinae.

I have put in the Actiniinae and Stichodactylinae where they
must come, if used, in this scheme—as subordinate to Endo-
myaria.

The Athenaria of this plan is Carlgren’s Athenaria without
the Edwardsiids. I fully agree that it is a good group—but
it represents a line of evolution within Nynantheae, all of which
are derivatives of a Halcampa-like stage, and needs no
subordination to anything else. Nor is there any need for
a contrasting group Thenaria; the other three tribes are.
mostly ‘Thenaria’, but they represent three evolutionary
lines and are best kept independent (see p. 560 et seq.).

NO. 260 00

542, T. A. STEPHENSON

The Endocoelactaria form a decided small line apart,
and with very distinct characters (see p. 522), and it seems
inevitable to give them a group to themselves. Since they
seem Clearly post-Halcampid, this group must come under
Nynantheae, not outside it; and is distinct enough from
other Nynantheae to require no further subordination.

This leaves the Mesomyaria and Endomyaria, or
main mass of forms. It has been part of the purpose of this
paper to show that this main mass does fall into two chief
sets, following two great lmes of tendency, and these two
lines I propose to embody in the two sub-tribes named. The
Mesomyaria contains the forms classified in Part I, the creatures
with acontia and mesogloeal sphincters and so on; the Endo-
myaria contains those with ne mesogloeal (and typically an
endodermal) sphincter, no acontia, and often with vesicles,
frills, &e.—for more detail of Endomyarian and Mesomyarian
tendencies see pp. 560 et seq. The Endomyaria contains the
whole of the old Stichodactylinae (save soft corals) and part of
the Actininae, and if those names be still used it should be
only as subdivisions of this group.

For most of the matter supporting the various suggestions
made in this section, reference should be made to the sections
on evolution and on special sets of forms, and other parts,
both in this paper and in Part I.

It now remains to allocate the families listed on p. 587 to
their respective groups.

PROTANTHEAE. Gonactiniidae.
PTYCHODACTEAE. Ptychodactidae.
NYNANTHEAE.

A. ATHENARIA, Halcampidae, Ilyanthidae.

B. Enpocornactarta. Halcuriidae, Actinernidae.

C. Mersomyarta. Diadumenidae, Phelliidae, Flosmarinidae, Mar-
supiferidae, Metridiidae, Chondractiniidae, Actinoseyphiidae,
Sagartiidae, Choriactidae, Paractidae.

D. Enpomyarta, Condylanthidae, Myonanthidae, Andresiidae,
Actiniidae, Aliciidae, Phyllactidae, Minyadidae, Phyman-
thidae, Heteranthidae, Stoichactidae, Actinodendridae,
Thalassianthidae, Homostichanthidae, Aurelianidae.

Discosomidae and Corallimorphidae go to Madreporaria,

oo ho

CLASSIFICATION OF ACTINIARIA 543

§Q. Summation of Characters.

In case it should be felt that the foregoing sections are too
much of an outline and have too much connexion with evolu-
tionary speculation, it seems advisable to point out that the
conclusions have the backing of the sum-of-the-characters
principle. The following lists will show that the groups sug-
gested have a solid number of characters binding them together.
Only main features are included. In connexion with some of
the details given under the larger groups, Text-fig. 16 will be
found useful. I will take families first, then larger groups.

AMT ES:

GONACTINIIDAE. Genera: Gonactinia, Protanthea.
Common characters, 11.

1. Definite base. 2. No basilar muscles. 3. Ectodermal muscle
in body-wall and actinopharynx. 4. Spirocysts in ectoderm of body-
wall (and actinopharynx ?). 5. No developed sphincter. 6. Tenta-
cular longitudinal muscle ectodermal. 7. No true siphonoglyphes.
8. Only the eight protocnemes perfect. 9. Mesenterial musculature
weak, not forming true retractors. 10. Filaments without ciliated
tracts. 11. No acontia.

PrycHopactipAE. Genera: Ptychodactis, Dactylan-
thus. Common characters, 12.

1. Definite base. 2. No basilar muscles. 3. Ectodermal muscle
in body-wall and actinopharynx. 4. Spirocysts in ectoderm of
body-wall (and actinopharynx ?). 5. No developed sphincter.
6. Tentacular longitudinal muscle ectodermal. 7. At least six,
usually twelve or more, pairs of perfect mesenteries. 8. Weak
mesenterial musculature, hardly forming retractors, 9, Filaments
with no ciliated tracts. 10, Filaments of imperfect mesenteries with
curious half-funnels at upper termination. 11. Mesenteries with the
free edge (or its analogue) occupied by filament above, gonad below,
if present. 12. No acontia.

CoRALLIMORPHIDAE. Genera: Corallimorphus, Cory-

nactis, Isocorallion. Common characters, 16,

1. No horny or limy skeleton. 2. Definite base. 3. No basilar
muscles. 4, Ectodermal muscle in body-wall, at least sometimes,
perhaps always. 5. Large sting-cells typically present in some part

Oxo) 2

544

Dis

T. A. STEPHENSON

of body. 6. No developed sphincter. 7. Tentacular longitudinal
muscle ectodermal, 8. Tentacles not branched, but knobbed.
9, More than one tentacle on at least each of the strongest endocoels.
10. Not more than one tentacle per exocoel. 11. No true siphono-
glyphes.t 12. No division of mesenteries into macro- and micro-
cnemes. 13. Usually numerous perfect mesenteries. 14. Feeble
mesenterial musculature. 15. Filaments with no ciliated tracts.
16. No acontia.

COsOMIDAE. Genera: Discosoma, Paradiscosoma,
Orinia, Actinotryx, Ricordea, Rhodactis.
Common characters, 12.

1. No horny or limy skeleton. 2. Definite base. 3. No basilar
muscles. 4. No developed sphincter. 5. Tentacular longitudinal
muscle ectodermal, such as it is. 6. More than one tentacle on at
least each of the stronger endocoels. 7. No true siphonoglyphes.1
8. No division of mesenteries into macro- and microcnemes.
9. Usually numerous perfect mesenteries. 10. Feeble mesenterial
musculature, not forming true retractors. 11. Filaments without
ciliated tracts. 12. No acontia.

N.B.—In this family the tentacles may be reduced or practically
absent, and their form is variable ; sometimes there is more than one,
on exocoels as well as endocoels.

HaucamMpipag. Genera: Halcampa, Halcampoides,

Pentactinia, Scytophorus. Common charac-
ters, 8.

1. No base (correlated with more or less vermiform shape). 2. No
basilar muscles. 3. Sphincter absent or weak (if present may be
mesogloeal or endodermal). 4. Tentacular longitudinal muscle
ectodermal. 5. Mesenteries divided into macro- and microcnemes,
or all macrocnemes. 6. Six pairs of macrocnemes the average (may
be four or five to seven couples). 7. Few mesenteries and tentacles—
up to forty or so. 8. No acontia.

ILYANTHIDAE. Genera: Ilyanthus (mitchelli), Pea-

1

chia, Eloactis, Haloclava, Harenactis. Com-

mon characters, 8.

1. No base. 2. No basilar muscles. 3. No developed sphincter.
4. Tentacular longitudinal muscle ectodermal. 5. Mesenteries all

With regard to this statement, see definition in Part II], covering

Corallimorphidae and Discosomidae.

SEG at AS Oia

CLASSIFICATION OF ACTINIARIA 545

macrocnemes (in one case macrocnemes and some of an intermediate
sort). 6. Never fewer than ten pairs of mesenteries, mesenteries all
perfect (one exception). 7. Few mesenteries and tentacles—-up to
forty or so. 8. No acontia.

Hatcurnpar. Genera: Halcurias, Carlgrenia. Com-
mon characters, 8.

1. Definite base. 2. Spirocysts in ectoderm of body-wall nearly
always. 3. No sphincter. 4. Tentacular longitudinal muscle ecto-
dermal. 5. Only one siphonoglyphe. 6. After the first six pairs,
mesenteries develop as directives and in endocoels. 7. There is
a fairly sharp division between the first six or ten pairs and the rest,
the former being macrocnemes and the latter more or less micro-
cnemes ; retractors of macrocnemes circumscribed. 8. No acontia.

ACTINERNIDAE. Genera: Actinernus, Isactinernus,
Synactinernus, Synhaleurias. Common char-
acters, 10.

1. Definite base. 2. Spirocysts in ectoderm of body-wall. 3. No
sphincter. 4. Tentacular longitudinal muscle ectodermal (or with
mesogloeal tendency in part). 5. Two siphonoglyphes. 6. After the
first six pairs mesenteries develop as directives and in endocoels.
7. No division of mesenteries into macro- and microcnemes.
8. Numerous perfect mesenteries. 9. Mesenterial musculature not
strong. 10. No acontia.

CoNDYLANTHIDAR. Genus: Condylanthus. Main char-
acters, 7.
1. Definite base. 2. No vesicles. 3. No sphincter. 4. Tentacular
longitudinal muscle ectodermal. 5. Mesenteries divided into macro-
and microcnemes. 6. Macrocnemes six pairs. 7. No acontia.

Myonanruipar. Genera: Myonanthus, Macrodactyla,
Boloceroides, Nevadne. Common characters, 7.
1. Definite base. 2. No vesicles. 3. No mesogloeal_sphincter
(sphincter endodermal or absent). 4. Tentacular longitudinal muscle
ectodermal. 5. Mesenteries not divided into macro- and microcnemes.

6. Perfect mesenteries six pairs. 7. No acontia.

ANDRESIIDAE. Genus: Andresia. (One species only,
see p. 518.) Main characters, 7.

1. No base (correlated with very extensile body). 2. No vesicles.
3. Small circumscribed endodermal sphincter, 4. Tentacular longi-

546 T. A. STEPHENSON

tudinal muscle ectodermal. 5. Mesenteries not divided into macro-
and microcnemes. 6. All mesenteries perfect. 7. No acontia.
It has long tentacles in graded cycles.

AcTiInupAE. Genera: Actinia, Anemonia, Gyro-
stoma, Condylactis, Parantheopsis, Buno-
dactis, Tealia, Epiactis, Isotealia, Bolo-
cera, Leipsiceras, Boloceropsis, Dofleinia,
Glyphostylum, Pseudophellia, Ixalactis.
Common characters, 6.

1. Definite base. 2. No vesicles. 3. No mesogloeal sphincter
(sphincter absent or endodermal). 4. Mesenteries not divided into
macro- and microcnemes. 5. Numerous perfect mesenteries—at the
least more than six pairs in adults. 6. No acontia.

This is one of the few families in which the longitudinal muscle of the
tentacles is sometimes mesogloeal.

Atictipag. Genera: Alicia, Phyllodiscus. Common
characters, 8.

1. Definite base. 2. Vesicles present. 3. Body-wall delicate,
divided into scapus and capitulum, the vesicles occurring either on
the scapus or at and above its junction with the capitulum. 4. No
mesogloeal sphincter, no developed sphincter at all. 5, Tentacular
longitudinal muscle ectodermal. 6. Mesenteries not divided into
macro- and microcnemes. 7. Six pairs of perfect mesenteries. 8. No
acontia.

PHYLLACTIDAE. Genera: Phyllactis, Cradactis, Phy-
v 2 d
mactis, Cystiactis, Lebrunia, Bunodeopsis,
Thaumactis. Common characters, 6.

1. Definite base. 2. Vesicles present. 3. No mesogloeal sphincter
(sphincter endodermal or absent). 4. Mesenteries not divided into
macro- and microcnemes. 5. Numerous perfect mesenteries as a rule.
. No acontia.

Here again tentacular longitudinal muscle may be ectodermal or
mesogloeal.

MinyADIDAE. Genus: Stichophora. Chief characters, 7.

1. Base a float. 2. No vesicles. 3. Sphincter endodermal. 4. One
siphonoglyphe. 5. Mesenteries not divided into macro- and micro-
cnemes. 6. Ten pairs of perfect mesenteries. 7. No acontia.

CLASSIFICATION OF ACTINIARIA 547

AURELIANIDAE. Genera: Aureliania, Actinoporus.
Common characters, 9.

1. Definite base. 2. Sphincter strong endodermal circumscribed.
3. Tentacles have the form of small vesicles, and may be lobed.
4, More than one tentacle to each main endocoel. 5. More than one
tentacle to each main exocoel. 6. One siphonoglyphe. 7. All the
mesenteries, or all the older ones, perfect. 8. Mesenteries either
all with the grade of macrocnemes (and with unusually strong circum-
scribed retractors), or else more or less divided into macro- and
microcnemes. 9. No acontia.

Here the disc and tentacle radial muscle may be ectodermal or
mesogloeal, and there are vesicular verrucae in one genus.

PHYMANTHIDAE. Genus: Phymanthus. Main charac-
ters, 10.

1. Definite base, sometimes reduced. 2. No vesicles. 3. No
developed sphincter. 4. Disc and tentacle radial muscle ectodermal
or with a mesogloeal tendency. 5, Tentacles divided into marginal
and discal, the former tentaculiform and usually pinnate, the latter
more usually papilliform (rarely they are absent). 6. Marginal
tentacles not more than one per exo- and endocoel. 7. Discal tentacles
typically in radial rows—they may occur on exocoels as well as
endocoels. 8. Mesenteries not properly divided into macro- and
microcnemes as a rule, though coming very near it sometimes.
9. Numerous perfect mesenteries. 10. No acontia.

ACTINODENDRIDAE. Genera: Actinodendron, Mega-
lactis, Actinostephanus. Common characters,10.

1. Definite base. 2. No vesicles. 3. No developed sphincter.
4, Dise and tentacle radial muscle ectodermal. 5. Disc produced
into permanent arm-like lobes. 6. Numerous tentacles per endocoel.
7. Numerous tentacles per exococl. 8. Mesenteries not divided into
macro- and microcnemes. 9%. Numerous perfect mesenteries. 10. No

acontia.

HomostIcHANTHIDAE. Genus: Homostichanthus. Main
characters, 11.

1. Definite base. 2. No vesicles. 3. Sphincter endodermal, not

strong. 4. Oral disc not formed into arm-like permanent lobes.

5. Tentacles short and papilla-like, simple. 6. Numerous tentacles

per endocoel. 7. Numerous tentacles per exocoel. 8. Tentacular

longitudinal muscle ectodermal. 9. Mesenteries not divided into

548 T. A. STEPHENSON

macro- and microecnemes. 10. Numerous perfect mesenteries. 11. No
acontia.

THALASSIANTHIDAE. Genera: Thalassianthus, Crypto-
dendron, Actineria. Common characters, 10.

1. Definite base. 2. No vesicles. 38. Sphincter endodermal, not
very strong, may be circumscribed. 4. Disc and tentacle radial
muscle ectodermal. 5, Tentacles divided into dendrites and nemato-
spheres. 6. Not more than one dendritic tentacle per exocoel, and
no nematospheres. 7. Typically more than one dendrite, and nemato-
spheres, on endocoels. 8. Mesenteries not divided into macro- and
microcnemes. 9. Numerous mesenteries perfect. 10. No acontia.

STOICHACTIDAE. Genera: Stoichactis, Radianthus,
Antheopsis. Common characters, 11.

1. Definite base. 2. No vesicles. 3. Sphincter endodermal,
strong or not very strong, may be circumscribed. 4. Tentacular
longitudinal muscle ectodermal. 5, Oral disc not produced into
permanent arm-like lobes. 6. Tentacles simple, all of one sort (but
for sporadic cleft ones which are sometimes present). 7. Not more
than one tentacle per exocoel. 8. More than one tentacle on at least
each older endocoel, except in very rare cases; usually some or all
of the endocoels have several or many. 9. Mesenteries not divided
into macro- and microcnemes. 10. Numerous perfect mesenteries.
11. No acontia.

HETERANTHIDAE. Genus: Heteranthus. Chief char-

acters, 9.

1. Definite base. 2. No vesicles: 3. Sphincter endodermal, not
very strong, circumscribed. 4. Tentacles of two sorts, marginal and
discal. 5. Oral disc not produced into permanent arm-like lobes.
6. Marginal tentacles short conical, disc-tentacles wart-like. 7. Mesen-
teries not divided into macro- and microcnemes, 8. Numerous perfect
mesenteries. 9. No acontia.

The other ten families were listed and dealt with in Part I.
It will be seen from an inspection of the above lists, that at
the minimum each family has six common characters, and
most have 7 to 11, a few even more. It must also be remem-
bered that the lists are not exhaustive, and that most of them
could be added to and even some of the characters subdivided.
or instance, ‘ presence of ciliated tracts on the mesenterial

a

CLASSIFICATION OF ACTINIARIA 549

filaments ’, ‘ presence of basilar muscles’, ‘ absence of ecto-
dermal muscle in body-wall’, ‘ presence of not more than one
tentacle per exo- and endocoel’, and so on, could be added
where suitable ; but the addition of all these where not strictly
required for present purposes would be needlessly complicating
—it is mentioned only to show that the lists could be expanded
rather than otherwise.

I should like to repeat here a remark made in Part I, to
the effect that the arrangement suggested cannot be validly
criticized on the ground that in some cases there are only one
or two differences between two given families. Provided
that the differences are good ones, this is all right—if families
be fused up on that principle it is soon found that the whole
Actiniaria will go into one or two collections, and classification
breaks down altogether. The very fact that the families form
enough of a series to have few differences sometimes, supports
the idea that they represent relationships truly.

If sums-of-characters for groups of wider inclusion than
families be now taken, the difficulty of course arises that
they can be made less absolute, because in some cases there
are one or two exceptions to almost everything among large
series of anemones, and this is the same whatever Classification
be adopted. It must therefore suffice to make definitions
of tendency rather than of exclusive fact, in some cases.

GROUPS LARGER THAN FAMILIES.
Here the wider groups will be taken first.

Class ANTHOZOA. Sub-classes included: Cerianti-
patharia, Octactiniaria, Zoanthactiniaria.
Common characters or tendencies, 14.

Coelenterata with (i) no medusae, (ii) “ hydroid-generation * form,
(iii) nematocysts, (iv) characteristic muscularity as compared with
Hydrozoa, (v) bilateral symmetry typically, (vi) no primary cruciform
symmetry like that of Scyphozoa, (vii) mesenteries, (viii) no septal
funnels, (ix) no endodermal tentacles, (x) mesenterial filaments,
(xi) endodermal gonads borne on the mesenteries, (xii) an actino-
pharynx, (xiii) no canal-system comparable to that of a Scyphozoan,
(xiv) no specialized sense-organs in adults.

Text-r1. 16.

(
Ge
C 3

Diagrams of transverse sections showing various mesenterial for-
mulae. A, supposed ancestor of Zoanthactiniaria ; B, Gonac-
tinia; c, Haleampa; D, form with graded cycles of mesen-
teries but only six pairs perfect (e.g. Myonanthidac); 5, an
Edwardsia; r, Parazoanthus; G, Carlgrenia; H,
form with graded cycles of mesenteries, and sixteen pairs per-
fect, Compare this with a further stage shown in Text-fig. 10. -

CLASSIFICATION OF ACTINIARIA 551

Sub-class ZoANTHACTINIARIA. Orders included:
Edwardsiaria, Zoanthinaria, Dodecacti-
niaria. Common characters or tendencies, 9.

1. Directive mesenteries typically present, two pairs the standard
number. 2. The directive endocoels do not become subdivided in
most forms, but it may occur. 3. There are always more than eight
mesenteries, even if only eight strong ones, in adults. 4. The eight
protocnemes do not typically get pushed out of the way in the manner
typical of Ceriantharia. 5. In most cases the mesenteries form pairs,
not couples. 6. There are never just eight pinnate tentacles ; pinnate
tentacles at all are rare, and occur in a few forms of obvious relation-
ships. 7. There are no gular tentacles like those of Cerianthids.
8. There is no sheet of muscle in the body-wall ectoderm comparable
in strength to that of Cerianthids. 9. There is no horny axis like that
of Antipatharia.

Order Dopxcactiniaria. Sub-orders included: Madre-
poraria, Actiniaria. Common characters or
tendencies, 6.

1. More than eight mesenteries, but there may be only eight
perfect : but even so some imperfect ones pair with them: usually
at least six pairs perfect. 2. After the first six couples, typically
pairs in cycles are formed. 3. Both pairs of directives, if present,
are perfect, not one pair macro- and the other micromesenteries as
in Zoanthids. 4. The later mesenteries are not typically confined to
two lateral regions of growth only, as in Zoanthids, though they may
come in the directive endocoels. 5. Mesenteries not typically formed
in unequal pairs, one perfect and macromesenteric and the other
not, as in Zoanthids. 6. No canals in the body-wall save in the case
of some skeleton-building forms.

Sub-order Actrntarta. Tribes included: Protantheae,
Ptychodacteae, Nynantheae. Common char-
acters or tendencies, 6.

1. No horny or limy skeleton. 2. No colonies. 3. Sting-cells of
Madreporarian type do not occur much. 4. Tendency to muscularity
greater than in Madreporaria, but not found in the most primitive
forms and some others. 5. Siphonoglyphes present in the majority,
but not in certain primitive and other forms, 6. Save in the earlier
forms, the mesenterial filaments have ciliated tracts.

Tribe ProranrHEAr. 1 family. See Gonactinidae for char-
acters.

552 T. A. STEPHENSON

Tribe PrycHopacTEAE. 1 family. See Ptychodactidae for
characters.

Tribe NYNANTHEAE. Sub-tribes included: Athenaria,
Endocoelactaria, Mesomyaria, Endomyaria.
Common characters or tendencies, 7.

1. Ectodermal muscle in body-wall the exception and not the rule,
occurring only in sporadicaily-distributed cases. 2. Spirocysts in
body-wall ectoderm not the rule—only of regular occurrence in
Endocoelactids. 3. Siphonoglyphes present save in odd cases.
4. Mesenterial filaments with ciliated tracts. 5. Pairs of perfect
mesenteries present. 6. Mesenterial musculature does not very often
exhibit so low a grade of development as in the Gonactiniidae, Ptycho-
dactidae, and many Madreporaria, weakness being usually sporadic
and secondary rather than universal and inherent. 7. A fundamental
number for the arrangement of parts is six, but there are a good
many deviations.

Sub-tribe ATHENARTIA. 2 families. Common characters, 9.

1. No base (correlated with more or less vermiform shape). 2. No
basilar muscles. 3. No vesicles. 4. Sphincters weak or absent,
though if present they may be endodermal or mesogloeal. 5. Not
more than one tentacle per exo- and endocoel. 6. Tentacles and
mesenteries few, up to forty or so. 7. Secondary mesenteries exo-
coelic. 8. Mesenteries divided into macro- and microcnemes, or all
macrocnemes, with Peachia as an intermediate. 9. No acontia.

Sub-tribe ENpocomBLAcTARIA. 2 families. Common char-
acters, 9.

1. Definite base. 2. No genuine basilar muscles. 38. No vesicles.
4. Spirocysts nearly always in body-wall ectoderm. 5. Probably
no ectodermal muscle in body-wall. 6. No sphincter. 7. Secondary
mesenteries endocoelic and oriented as directives. 8. Not more than
one tentacle per exo- and endocoel. 9. No acontia.

Sub-tribe Mrsomyarta. 10 families. Common characters or

r

tendencies, 7.

1. Definite base with one or two exceptions. 2. Basilar muscles
usually present. 3. No vesicles. 4. Acontia OR a mesogloeal sphincter,
or both, present. 5. Not more than one tentacle per endo- and exo-
coel. 6. Secondary mesenteries exocoelic. 7. No acrorhagi or tenta-
cular complications of an Endomyarian sort—often there are basal
mesogloeal swellings to the tentacles, and thick body-walls, however,
and there are two cases of another sort of tentacular thickening.

Ee

CLASSIFICATION OF AGTINIARIA 553

Sub-tribe ENpomyarta. 14 families. Common characters or
tendencies, 6.

1. Definite base save in one case (it may be somewhat reduced, or
may form a float). 2. Basilar muscles usually present. 3. No
mesogloeal sphincters (sphincter endodermal if present). 4. No
acontia. 5, Secondary mesenteries exocoelic. 6. There may be
no external complications of the body or tentacles, but verrucae,
acrorhagi, vesicles, and complex tentacles are characteristic of different
members of the group, more than one of them sometimes occurring
in the same form; but there are no tentacles with basal mesogloeal
swellings.

Here there is often more than one tentacle on an endocoel, and there
may be a good many on each main endo- and exocoel; or, on the other
hand, there may be not more than one to each.

Fhe above lists show that even when one is dealing with
larger groups it is generally possible to base them on a fair
sum of characters or at least of tendencies. It should of course
be remembered that each family has not only its own special
family-features, as listed, in common, but also many of the
eroup-characters behind the family. To take a single example,
the Actinudae have in common 6 Actiniud characters + 6
Endomyarian features + 7 Nynanthean characters + 6 Acti-
marian characters + 6 Dodecactiniarian characters + 9 Zo-
anthactiniarian + 14 Anthozoan, not to-mention all their
Coelenterate and Metazoan points. So that they have, back
to Anthozoa, 54 common characters—the number has to be
reduced of course by any characters which may occur in more
than one of the lists involved, or which may be inapplicable
to the particular case in point, but even then the number will
be considerable.

5. EvoLutTioNARY SUGGESTIONS.

That the classification suggested here has a firm foundation
in character-summation will be evident from the above lists
and the definitions later on; but it allows a certain amount
of latitude for alternative ideas of evolutionary history, with
which it is necessarily a good deal mixed up, especially in cases
of large groups, where one is almost bound to think partly

554 T. A. STEPHENSON

in terms of evolution. The view here taken of the evolution
of the forms will now be further developed.

In Part I reasons were given for thinking of a Halecampa-
like form as more primitive than such a creature as Catadio-
mene (though of course more advanced than Gonactinia),
and if was concluded that whatever the detail, the main
direction of evolution would be in the direction Halecampa-
form ——->+Catadiomene and not the reverse, and that
this would generally apply. Without gomg into it all again
(see Part I, p. 487) it may be assumed that in dealing with
such a group as the Endomyaria, some Halcampoides-
like form is the end to start at, and Tealia or Phymactis,
or some Stichodactyline the antithesis, for much the same sort
of reason, with differences in detail. Before discussing the
Endomyaria further, however, it will be well to try to get
at the relationship of Endomyaria and Mesomyaria to other
groups.

If it is fairly clear that both these groups origmated some-
where near Halcampa, the same is still clearer of the
Athenaria—i.e. the Halcampids themselves and their burrowing
descendants. There is also a clear suggestion of origin from
a Halcampa-like ancestor in the Endocoelactaria, and they
must be thought of as Halecampa-stock diverging from the
main lines. The Stichodactylina (excluding the Corallimor-
phidae and Discosomidae) are to be thought of as specialized
Endomyaria. The first idea to establish then is that Endo-
myaria, Mesomyaria, Endocoelactaria, and Athenaria are the
outcome along different lines of a Halcampa-stage with
strong retractors and with ciliated tracts on the filaments.
That is, they are ‘ post-Halcampid’ and form a single class,
Nynantheae s.s. as defined on pp. 540 and 552, and in Part III.

Next, there are the Gonactiniidae, Ptychodactidae, and
Madreporaria to be considered. The idea I hope to work out
in connexion with these is that they origimated in an ancestor
earlier and less advanced than Halcampa (it would of course
also give rise to Haleampa itself), and in fact may be called
‘ pre-Halcampid ’.

CLASSIFICATION OF ACTINIARIA 5d5

What forms are more primitive than Haleampa? It
was suggested in Part [that Gonactinia and Protanthea
are survivals of something very early (see pp. 493, 496-7, &e.).
The grounds are these. The ‘Halcampa-stage’ in evolution
may be defined as a stage with six pairs of perfect mesenteries
(including two pairs of directives) bearing strong retractors,
gonads, and filaments with ciliated tracts; any mesenteries
beyond these six pairs would be rudimentary ; there would
probably be little or no base, a fairly narrow body, and little
or no sphincter (cf. Text-figs. 8 and 7, c, p). This is not the
Halcampa-stage sometimes used in an embryological sense,
but is the way im which the term is usually taken for purposes
of this paper. Now the Gonactiniids have paired mesenteries,
but not six pairs perfect—only the eight protocnemial couples
are fully developed. The filaments have no ciliated lobes,
and the mesenteries have very weak musculature, not forming
retractors as in the Halcampa-stage. Moreover, the body-
wall, tentacles, disc, and actinopharynx approximate to each
other in structure, at least as regards ectodermal muscle, and
mostly spirocysts. This gives something much nearer a possible
ancestor for the groups not specified as post-Halcampid than
anything else. ‘The consideration of Anthozoa generally,
suggests inevitably that mesenteries coupled before they
paired, and the Gonactiniids still keep a vestige of the coupling
which Halcampa has lost (see Text-fig. 16, Bk-and in
a case like this the generalized musculature may be taken to
indicate a stage before much differentiation of tentacles from
body-wall, and of good retractors, had set in.

There seems no reason to think that the Ptychodactidae or
Madreporaria ever passed through a Halcampa-stage in
the sense outlined above. They did not attain to much in the
retractor line, and the Ptychodactids did not differentiate the
parts of their ectoderm very markedly. They never have
ciliated tracts on the filaments, and their whole organization
and histology, especially of course in Madreporaria, suggests
a difference of direction in evolution from that of the post-
Halcampids.

556 T. A. STEPHENSON

Although these forms (Gonactinids, Ptychodactids, Madre-
pores) must be put down as pre-Halcampid, they have common
features establishing them as distinct from Edwardsiaria and
Zoanthinaria, and they form one group, Dodecactiniaria—
for instance, they have typically attained pairing of mesen-
teries and equality of directives, and the pairs are not usually
formed each of a macro- and a micromesenteric partner, nor
do they usually develop in two lateral zones of increase only,
after a certain point; there are no canals in the body-wall
save in some of the skeleton-making Madreporaria.

So that it may be said that the Dodecactiniaria present on
the one hand descendants of a Gonactinia-like form, and
these are poor in muscle and lack ciliated tracts ; and on the
other hand descendants of a Halecampa-like form (itself,
of course, the outcome of an earlier Gonactinia-like one),
with the ciliated tracts stabilized and a tendency to muscularity.

Is the ancestor of the Zoanthactiniaria, the group contaiming
the Dodecactiniaria as well as the Edwardsiaria and Zoan-
thinaria, simply the same sort of Gonactinia-like animal ?
The whole situation suggests that it must have had a good deal
in common with Gonactinia—it would surely be a small
form with weak muscle and generalized ectoderm and only
eight perfect mesenteries (see T'ext-fig. 16, A); the chief point
of debate is, had its filaments ciliated tracts ? At first glance
one would say No, the state without the tracts is more primi-
tive: but there are other things which do not suggest that it
was devoid of them. That the ancestor of all Anthozoa was
without them seems certain, but that is even farther back
than the one here visualized. Our Zoanthactiniarian ancestor
gave rise to Edwardsians and Zoanthids as well as to Dode-
cactiniaria, and both the former have ciliated tracts, even if
they are not quite the same as those of the Nynantheae. This
suggests that either (i) the Edwardsians and Zoanthids attained
them independently, or else that (ii) the Gonactinuds, Ptycho-
dactids, and Madreporaria lost them, while Haleampa and
its followers retained, stabilized, and developed them. (See
Text-fig. 17 for the main types of filament here mentioned.)

CLASSIFICATION OF ACTINIARIA 557

The first assumption, of independent acquisition, would not
be unreasonable, but at the same time it does seem lkelier
that the ancestor of all three had ciliated tracts, perhaps only
in a slightly differentiated form; and it is a simpler way of
putting things to think of some forms losing them than of

Trxt-rie. 17.

Olt
ay y

SSS.

Mesenterial filaments. A and £, a Zoanthid, with powerful ciliated
tracts (f). E passes twice through these, as it cuts through
a curved edge of mesentery. B, Edwardsia. Ciliated tracts
present but less marked than in the Zoanthid; here and in D
there are also reticular tracts (r). c, Paradiscosoma. Here
there are no ciliated tracts, but three large sting-cells are
shown. pb, Artemidactis. Typical Actinian filament,
with median cnido-glandular tract (c) and lateral ciliated and
reticular tracts.

three groups gaining them. ‘here seems no special reason

why such an ancestral form as that under consideration should

not have weak ciliated tracts, because although very distinct

structures. they would easily be differentiated early on, just

as acontia seem to have been at the Hosagartia stage in
NO, 260 Pp

558 T. A. STEPHENSON

the history of the Mesomyaria. It provides an idea parallel
to that of loss of acontia by various forms, advocated in
Part I.

I do not feel that the loss of ciliated tracts by some forms
can be very fully accounted for, but it is easier to explain than
their independent acquisition in three cases would be. The
suggestion I should like to offer in this connexion was made
to me by Professor Fleure, and does seem to make it intelligible.
In certain Gastropods where the gill-lamellae are not much
strengthened and kept apart skeletally, there is a device for
keeping open chinks between them, for the passage of water,
by means of pads of cilia. It is an attractive idea that part
of the function of the anemone’s ciliated tracts is something
of the same sort—a preservation of chinks allowing access of
water between the mesenteries, for respiratory purposes and
soon. In the hght of this several things may be noted. Among
the forms with no cihated tracts there is little or no sphincter,
which means not much tight closing-up of the body. The
forms with the tracts have above all developed strong retractors
or sphincter, or both (with fairly numerous exceptions), and
can often spend a good deal of time tightly shut up—in which
condition, of course, the pads would function very well. The
marked development of the tracts in Zoanthids fits in with
this idea. Among the tractless forms the only really successful
ones are the skeleton-making corals, and these have got over
any difficulty by keeping their mesenteries apart with septa ;
and the other groups are seemingly quaint survivors, and some
of them are so constituted that there is not much crowding
in the coelenteron. It is not impossible that certain appear-
ances in some of the filaments devoid of ciliated tracts represent
vestiges of them; similar appearances may be present, it is
true, in forms with the tracts—but even here they might be
vestiges of the weak tracts of the ancestor which were super-
seded by much better ones. On the other side of the question
it must not be forgotten that there are analogues of the ciliated
tracts in Ceriantharia, but here again the ancestor may not
have been far from that of the Zoanthactiniaria.

CLASSIFICATION OF ACTINIARIA 559

Summarizing so far, we get the suggestion of an evolutionary
course somewhat as follows :

From a small, delicate, bilateral ancestor, with eight feebly
muscular mesenteries, with some degree of differentiation of
ciliated tracts, and with generalized ectoderm, there arose

(i) Edwardsiaria,’ the mesenteries of which never
paired, but some of them attained muscularity (see
Text-fig. 16, &).

(i) Zoanthinaria, the mesenteries of which paired, but
which went in for various curiosities (see Text-fig. 16, F).

Gu) Dodecactiniaria, the mesenteries of which paired,
and which developed along the familiar ‘ Hexactinian’
lines.

There is just the possibility of an alternative view of the
Edwardsiaria to the one adopted in this paper—namely, that
they might somehow be Nynantheae in which certain mesen-
teries had been suppressed so that now there are only couples
and not pairs. It is their histology which rather suggests
Nynanthean affinities, but this idea is put forward very
tentatively and further work would be required to ascertain
how far it could be entertaimed as a possibility.

The Dodecactiniaria split on the rock of sluggishness versus
muscularity.1 The Gonactinia-like ancestors experimented
a little, and gave rise to the Gonactiniidae and Ptychodactidae,
perhaps trial-lmes, on the one hand, and to the corals on the
other; all these losing the ciliated tracts and never getting
very muscular, the majority-forms going in for strict sedentari-
ness and skeleton-building, often colonially. In a different
direction there arose from one of the Gonactinia-like
ancestors a muscular Halcampa-form; this, far from
losing the ciliated tracts, developed them further, and gave
rise to the individualized and typically muscular forms, which
fell into four sets—Athenaria, Endocoelactaria, Mesomyaria,
Endomyaria.

1 See in this connexion Chapter VIII in Thomson and Geddes, * Evolu-
tion’.

le p 2

560 T. A. STEPHENSON

From this pomt evolution among the Halcampa-descen-
dants or Nynantheae may be further considered.

About the Athenaria and Endocoelactaria little further need
be said beyond what may be found in the special sections on
those forms. The Athenaria are highly muscular as regards
their mesenteries, this bemg useful in a burrowing existence.
They have diverged among themselves in curious ways, and
some of them present rather interesting special features, such
as the immense siphonoglyphe and conchula of Peachia
(presumably a development connected with drawing in a water-
current when the animal is below the sand), and the knobbed
tentacles of Eloactis. Harenactis has become very
attenuated, with many cinclides—and indeed there are often
cinclides among these forms. The Endocoelactaria are obviously
divergent in another way. ‘The earlier ones, most nearly
represented by Carlgrenia, would be not far from the
Halcampa-stage, but with secondary mesenteries (micro-
cnemes at first) appearing in the lateral endocoels, and oriented
like directives—this modifying the whole plan of structure.
A stage further is represented by Halcurias, with ten pairs
of macrocnemes instead of six, and later in the Actinernids the
distinction into macro- and microcnemes has gone and numerous
mesenteries are perfect, and often there are lobed dises, swollen
tentacles, thick body-walls, and deep sea habitat. A sphincter
never appears.

This leaves the main mass of forms, the Meso- and Endo-
myaria (including Stichodactylines). With regard to the
justifiability of these two groups, if the work of this paper and
of Part I be taken into consideration it should emerge that
so far as we can know anything about these things, the Endo-
myaria did, as a bunch, follow a different line of tendency
from the Mesomyaria, and if that is established the grouping
follows. It is mainly a difference of tendency, there being, at
any rate low down in the two groups, probably no essential
histological difference—this might come in higher up, perhaps,
in comparing such formsas Actinoscyphia and Catadio-
mene with Thalassianthus.

CLASSIFICATION OF ACTINIARIA 561

Among the Endomyaria the sphincter, if present, is endo-
dermal. ‘There are never any acontia. After early evolutionary
stages are past, there are often vesicles, sometimes very complex
ones, on the body ; verrucae and acrorhagi are frequent ; and
in some cases the tentacles increase in number or become

f bn op:ca rca ete Ite

An enlarged view, from above, of a whole specimen of Phyllo-
discus indicus. The tentacles are not shaded, and form
the central part, and projecting beyond them is the corona
or ruff of compound vesicles. An example of complexity affecting
outgrowths of the body.

curiously modified in form-—vesicular or branched, sometimes
quite feathery in their subdivision. ‘There is little or no
tendency to thick body-walls of the sort found among Meso-
myaria, and never are there basal mesogloeal swellings to the
tentacles. The tentacular musculature rarely becomes meso-
gloeal. A definite base has been attained save in one case,
and typically there are basilar muscles. ‘The secondary

562 T. A. STEPHENSON

mesenteries appear in exocoels, and usually the musculature
of the body-wall ectoderm is lost. The habitat of the forms
with vesicles and elaborate tentacles is often tropical. ‘Text-
fig. 18 gives a good example of one of the forms with a frill
of vesicles. The crown of tentacles (umshaded) is seen to be
surrounded by a wider corena of compound vesicles, like a ruff,

TEXtT-FIG. 19.

Actinodendron plumosum, copied from a photograph of
a living specimen by W. Saville-Kent. See acknowledgement
on p. 496. An example of complexity affecting tentacles and disc.

projecting beyond it. A vertical section of the same form is
shown in Text-fig. 2, a. A case in which the tentacles are
dendritic and form a frill, bemg borne on permanent arm-like
projections of the disc, is shown in Text-fig. 19, and other
variations in Text-figs. 14 (tentacles) and 1 (acrorhagi).

In Mesomyaria, on the other hand, we get the sphincter,

CLASSIFICATION OF ACTINIARIA 563

when present, mesogloeal. Acontia are often present. Real
vesicles or frilled tentacles do not occur (the tentacles are
shghtly complicated in one or two cases), nor do acrorhagi ;
there is never more than one tentacle to an endocoel (often
there are more in Endomyaria, and it may be so on the exocoels
also), and the tentacles often have a thick basal mesogloeal
swelling aborally. Thick body-walls and knobs and crests
of mesogloea are fairly frequent (see Part I, Text-figs. 24, 25,
26, 27, 31). Tentacular musculature is more often mesogloeal
than in Endomyaria.

Possibly the acontia in the second group, the acrorhagi and
vesicles and complex tentacles in the first, are different expres-
sions of stinging tendencies along different lines, going with the
sphincter-difference and so on, the frills especially associated
with warmer seas, the curiosities of the Mesomyaria often
connected with deep water. One difference is that acontia
seem to have been ancestral in the Mesomyaria and to have
been lost in certain cases; whereas vesicles and such things
must be the attainments of certain individual sets of animals
at given points.

Lastly, evolution within Endomyaria may be a little more
closely thought of. [For Mesomyaria see Part I.

The general direction has been decided on (see p. 554).
The simplest way will be to put the route suggested by the facts
as narrative, as before, and it must have been something more
or less like the following :

From an Eoactinia (near to Kosagartia—see Part I)
or Halcampa-like form with little or no base, no sphincter,
and six pairs of macrocnemes and a few microcnemes, at
first one line of evolution only started.

An adherent base was gained, at first, and an increase in
the number of tentacles and microcnemes, but nothing else
(cf. Text-figs. 8 and 16, c). There are survivors of this stage
even now, the Condylanthidae.

Next, the distinction between macro- and microcnemes was
lost, but at first only the former macrocnemes remained perfect
(cf. Text-fig. 16, p). Some forms began to get an endodermal

564 T. A. STEPHENSON

sphincter, though not a very strong one; some developed
suckers on the body-wall, and one curious animal formed
special sphincters whereby it could cut off its tentacles at will—
it also retained some primitive features (Boloceroides).
The present-day forms which have gone no further than this
are the Myonanthidae.

A large number of forms, however, did go further, and attained
a larger number of perfect mesenteries (cf. Text-figs. 16, 4,
and 10). Often the endodermal sphincter developed and
sometimes became very strong, though some forms still remained
sphincterless, or with very little or a moderate sphincter.
Some of the advanced ones with strong sphincters have the
tentacular and discal musculature embedded in the mesogloea.
Among these forms the body either remained smooth, or
developed verrucae or acrorhagi or both, but never vesicles.
These are the Actiniidae s.s. in the sense taken on p. 546.

To go back a little, from somewhere near the Myonanthidae
arose a group of delicate forms which retained the six pairs
of perfect mesenteries, but the body became divided into
a scapus and capitulum, and either from the scapus or from
the region where scapus and capitulum join (and sometimes
above that region as well) there grew out hollow sac-like
diverticula, often compound—the vesicles. Little or no
sphincter was attained. These forms are the Alicidae.

There is another set of forms with these vesicles, but with
usually more numerous perfect mesenteries. They sometimes
have a less delicate body, and occasionally mesogloeal tentacle-
muscle. There is often a well-developed endodermal sphincter,
but it may be weak or absent. Perhaps these, or some of them,
arose, independently of the Aliciidae, from among the Actinudae,
or perhaps they arose from near the Aliciidae by a mesenterial
change. Whichever way it was, they represent onward steps.
They are the Phyllactidae—a somewhat heterogeneous group
to be further discussed in Part III. A section of one of them,
with many perfect mesenteries, is shown in T'ext-fig. 10.

A form, or perhaps several forms, which from our hitherto
incomplete knowledge of them would seem to have arisen near

Oe yt (oc Pia Oe ne

a ee

SE He eee a

ee

CLASSIFICATION OF ACTINIARIA 565

the Actinidae, took to a floating life, swimming upside down.
The base developed into a regular float, and certain anatomical
peculiarities appeared. These are the Minyadidae.

Other advanced stages are represented by the various
families of ‘ Stichodactylines ’. These arose from some Actiniid
or pre-Actinud ancestors, and they have usually the numerous
perfect mesenteries, and often endodermal sphincters, which
may be quite strong. The sphincter is endodermal or absent,
never mesogloeal. Among themselves they diverged into seven
families, easily distinguished from one another. The differ-
ences affected the arrangement of the tentacles on the endocoels
and exocoels, and their form—they might be simple, pinnate,
dendritic, sessile and vesicular, feathery, modified into special
stinging ‘ nematospheres ’, and so on. ‘The other part of the
structure chiefly affected by variation was the musculature—
there might be absence of sphincter in one case compensated
for by strong retractors ; or very strong sphincter and retractors
but poor tentacles ; and so on.

It will be seen from the above outline, and from that given
earlier for Mesomyaria, that there is one thing assumed as
having independently taken place in Endomyaria, Mesomyaria,
and Endocoelactaria—and possibly more than once in Endo-
myaria: that is, that i each of these cases a start was made
from the condition in which the mesenteries are divided into
macro- and microcnemes, this was lost, and in the end there
were graded mesenteries and numerous perfect pairs. This is,
however, a convergence quite to be expected among forms
making in a general way towards increase of size and diameter
of the individual, and correlated multiplication of organs.
Whatever arrangement be adopted, there is some convergence
cropping up, but when one thinks of the vertebrate and
cephalopod eye, or of the Marsupial and ordinary wolf, a
convergence like that assumed here seems very simple.

In the two hypothetical ancestors of Hndomyaria and
Mesomyaria (Hoactinia and Hosagartia) there is no
harm in assuming for them ectodermal muscle in the body-wall,
and the same may probably be said for the Hoactinia-lke

566 T. A. STEPHENSON

aneestors of Athenaria and Endocoelactaria. This would
allow for the retention of the musculature, or of traces of it,
here and there, with a dominating loss of it along all the lines.

6. SUMMARY.

It is difficult to make a concise summary of a paper covering
a good many inter-related discussions, but the following is
an attempt to give some of the main points with redsonable
brevity.

1. There is difficulty in defining specificity among Actiniaria,
as in other lowly and plastic animals. Among British forms
species are well enough marked on the whole, if studied alive
so that colour and habit can be taken into account. When
preserved, however, too little is known of possible range of
specific variation in anatomy for much to be done. Foreign
forms are so often known in death only that species are some-
what in chaos and there is little firm ground. Experience
leads one to the view that among these low and plastic forms
a species may have its peculiarities of organic constitution
at an early stage of the development of their expression, such
expression having affected colour-scheme and general facies
of the living animal but not necessarily to any extent the
internal anatomy which can be studied in preserved specimens.
Much work needs doing by way of studying all forms alive,
and of killing and preserving numerous individuals which
belong certainly to the same species, in different ways, and
studying them so as to reveal effect of reagents, age, state
of contraction or distension, locality, reproductive maturity,
and so on, on the anatomy. When a better knowledge of the
limits of specific variation is gained (and they will be much
wider in some species than in others) a revision of species
might be attempted. Especially the value or otherwise of
measurements of nematocysts as specific characters should
be looked into.

2. Although species are in a poor way, genera and families
are on the whole much easier to understand and make use of, and
here there are enough data to start a methodical classification

CLASSIFICATION OF ACTINIARIA 567

with. Omitting for the sake of brevity any criticism of existing
classifications, and regarding Actiniaria as an unclassified
series, it may then be inqured what method can be applied
to them to find cut their inter-relations. Clearly unit characters
are not much help, since they may vary independently, and may
enter into combination in different genera with various sets
of others. It is therefore necessary to sum up the chief features
of each genus, and to see which genera have most in common
with which others; and those sharing most can be united in
families. The result is a natural grouping, and one which
expresses relationships of animals as wholes, and not analogies
of isolated parts of their bodies. The classifications of Lamelli-
branch Mollusca may be referred to as an example of several
overlapping schemes affecting the same group, founded on
few characters, and each expressing the relationships and evolu-
tion of one set of anatomical details (be it siphons and pallial
lines, hinge lines and teeth, adductor muscles, or gills), and not
expressing those of Lamellibranch animals as wholes.

It is found, however, that after applying the method of
summation of characters, families can be defined by
half a dozen or more common features, and may form so graded
a series that there are only unit-differences between some of
them. On the other basis there were sometimes only single
or few differential resemblances between the members of
a family, accompanied by important differences. To look at
it from another angle, it has been said that criticism is finding
out why one likes or does not like a given book or picture.
It seems fair to say that classification is findmg out why
a horse is more like a mule than like a wolf—we know instine-
tively that it is so, but if we can confirm that instinct by good
reasons we have a classification. Similarly, given enough
study of a group, and enough training of the relationship-
instinct, it is felt that from their whole organism and make-up
certain forms are more nearly related to some of their brethren
than to others. This may be of very great help, but of course
needs cautious exercise and confirmation. ‘The point is that
the principle of summation of chief characters gives this

568 T. A. STEPHENSON

confirmation in a way that an artificial system of unit-characters
cannot do—it justifies and bears out the instinct. The summa-
tion principle also enables the family to be used definitely
as the expression of a step in the evolution of any set of forms,
and the classification represents evolution of whole anemones,
not of their sphincters or tentacles only. It also provides
evolutionary hints which could not otherwise come to light,
and which, given a general idea of group-evolution, help to
confirm and enlarge it. The general idea itself can grow from
a comparison of early and advanced forms, embryology, and
so on. From working through a whole group in such a way it
does seem possible to get a glimpse of the rhythmical develop-
ment of the life in the creatures, expressing itself in the various
ways at its disposal and unfolding along various lines. It
should be noted that in dealing with a group as plastic as
Actiniaria, it is often necessary to define differentiation
of tendency without too much insistence on hard and fast
divisions without qualification or exception.

3. The classification worked out on the above lines, in this
paper, is as follows. For definitions of the groups and families,
and for limitation of the sense in which they are taken, refer-
ence should be made to the portions of the paper where these
things are dealt with. I have accepted the arrangements of

3ourne and Carlgren as regards sub-classes of Anthozoa ;
and that of Bourne for orders and sub-orders. The tribes,
sub-tribes, families, and genera have, however, been largely
revised in this paper. I have kept as near to Carlgren’s tribes
and sub-tribes as I felt possible, and have throughout used old
names where I could; but the sense of his groups has been
altered and they have been added to, and many of the families
more narrowly limited, so that the old names take on a new

meaning.
Crass ANTHOZOA.

ee oer 1. CERIANTIPATHARIA,
Sub-class 2. OcTAcCTINIARIA.
Sibetlass 3. ZOANTHACTINIARIA,

at hay PE 99s ey es ———

CLASSIFICATION OF ACTINIARIA 569

Order A. Edwardsiaria.
Order B. Zoanthinaria.
Order C. Dodecactiniaria.
(Sub-order (1) MApREPORARTA.
(Sub-order (ii) AcTrNtaRra.
Tribe a. Protantheae.
Tribe b. Ptychodacteae.
Tribe c. Nynantheae.
Sub-tribe a. Athenaria.
Sub-tribe 8. Endocoelactaria.
Sub-tribe y. Mesomyaria.
Sub-tribe 6. Endomyaria.

The families will be found listed under their respective
group on p. 542.

4, An idea of the evolutionary history of the group has been
worked out in connexion with the above classification, and may
be summarized as follows.

It is possible to guess at a small plankton swimmer with
eight tentacles and eight mesenteries, without much definiteness
of musculature, and with bilateral symmetry, and contrasting
with, not resembling, the cruciform Scyphistoma, which must
have been quite an independent outcome of a Hydrozoan.
This small creature would give rise to several types much like
itself but with differences of detail, each of which would give
rise to a main Anthozoan sub-class. Only the one which gave
origin to the Zoanthactiniaria need be followed here. ‘This
stock seemingly shed out curiosities at first; some of them
took to burrowing and life in cracks, and became vermiform,
but did not amount to much (Edwardsiaria) ; others went in
for colonialism and incrustation and had fair success in a coral-
like way (Zoanthinaria). The main line, however, divided fairly
early into two great groups, the split being upon the rock of
sluggishness and colonialism and skeleton-building versus
comparative activity, specialization of the individual, greater
muscularity, and no skeleton. The two groups are of course
corals (Madreporaria) and sea-anemones (Actiniaria). ‘There
are a few corals which developed no skeleton, or else lost their
skeleton, and which though often simple show colonial tendencies.

570 T. A, STEPHENSON

They have usually been classed with the anemones, but it
appears that they are almost identical in structure with
coral-polyps, but unlike anemones. Their lack of skeleton |
cannot keep them out of Madreporaria, and the transference
makes the division between the two groups, as regards soft
parts, more intelligible. They are the Corallimorphidae and
Discosomidae.

Returning to the sea-anemones proper, they seem first to
have experimented with further curiosities, which perhaps
diverged from the main stock about the same time as the
corals, or a little later. These experimental forms fall into two
sets, with a good deal that is primitive about them, one of
them resembling as nearly as any surviving form the supposed
ancestor of the whole Zoanthactiniaria. They are the Pro-
tantheae (Gonactiniidae) and Ptychodacteae. After this the
main. line attained a definitely muscular Halcampaz-like
stage with well-marked ciliated tracts on its mesenterial
filaments, and from this point two main lines of divergence
may be traced, and two lesser lines. Of the subsidiaries, one
sroup (Athenaria) took to, or simply remained in, a burrowing
life, and retained a good deal of simplicity ; the other (Endo-
coelactaria) went off in a curious direction, the reverse of that
taken by most forms, as regards some details of its mesenteries,
and possibly gives a clue to the origin of Tetracorallia. This
group shows one tendency in common with the two main lines
to be next dealt with—-a general move towards increase in size
of the individual, especially in diameter, and increase in the
number of effective organs; with musculature tending to
change from a few strong retractors on a few mesenteries to
a larger number of less specialized ones.

The two main lines both went in for development of a mar-
cinal sphincter, but otherwise their differences of tendency are
marked. The Mesomyaria developed mesogloeal sphincters,
and these, when they have special stinging organs, have
acontia, never or hardly ever acrorhagi or frills. And although
diverging among themselves, many of them tend after a time
to take to deep-water life. In correlation with this they may

CLASSIFICATION OF ACTINIARIA 571

lose their acontia and may lose mobility, and develop stiff
or thick body-walls, their metabolism slowing down and spare
energy sometimes being used up in the production of knobs
and crests and solid horn-like tentacular swellings. This is
a tendency towards fixity of character and possibly thence
towards ultimate extinction. It is interesting to note that
some of the above-mentioned Endocoelactaria have reached
a similar state, although along an entirely different line.

The other main line, Endomyaria, went in for endodermal
sphincters if any, and their special stinging organs are never
acontia, but they often have acrorhagi and other things.
Some of them develop vesicular blisters and compound acro-
rhagi which may reach wonderful complexity of structure ;
in others the tentacles increase in number and sometimes they,
not the outgrowths of the body, become complex, at their
finest with a frill-like effect. These forms, whether it be body
or tentacles that complexify, are more especially found in
the warmer seas, and here the tendency to fixity of character
does not seem much indicated. Along both lines various
forms halted by the way. of course.

This idea of the evolution of the group may be helped out
by the diagram printed below.

A more detailed outline of the history of Mesomyaria has
been worked out, and will be found in Part I, p. 498, &e. ;
a corresponding one for Endemyaria is given in this part,
p. 563, &e.

5. Apart from the above considerations, it has been the object
of the paper to revise and re-define all the families and genera,
sorting them out in such a way as to make them as homo-
geneous as possible, and to represent their relationships naturally,
with the idea of getting the definitions as precise as is feasible
in order to facilitate identification. It has the advantage of
collecting all the definitions together, but at the same time
is not meant to be an exhaustive compilation as regards
species-lists and so on. Only a minimum of synonymy is
included, and insufficiently known forms are left alone. The
classification worked out is, admittedly, complicated rather

572 T. A. STEPHENSON

than simple, but that is inevitable in a large and very old
croup.

6. It seems fair to suggest that the principles advocated
and put into practice here might with advantage be applied
to other animal groups (e.g. Gastropods and Lamellibranchs).
It is not for a moment implied that the classification of animals
as at present understood does not group them correctly, speaking
broadly and of the main groups; but that it needs revision
and supplementing on the plan suggested, especially in the
cases of some of the sub-groups, the classification of which
sometimes seems tentative and not very clear. It appears
that nearly encugh data are now collected about animals to
permit of entry on a new phase in the history of classification.
It is becoming evident, with regard to species for instance,
that some new system will shortly have to be devised which
will more adequately represent their inter-relations, and allow
for the idea of interlacing systems of concentric circles with
the characters of the central individual in each system as
those of the species, which has grown up. Some new conception
will probably work itself out about classification in general
also, and the revision of some groups in accordance with ideas
advocated here is suggested as a small beginning along the
road—a beginning which may possibly lead to further steps
in the realization of the new conception. If it prove to be
a blind alley, that conclusion should not take very long to
emerge.

7. SHORT GLOSSARY.

This is not in any way a complete glossary, but is meant for
use in connexion with a few terms which more than most seem
to require definition, for convenience in using the paper.

Aconria.—Slender white or coloured threads attached
to the borders of the mesenteries in some families of Actiniaria,
just below the mesenterial filaments. They are loaded with
nematocysts, and can be protruded through the mouth, and
in some cases also (accidentally) through pores (cinclides) in
the body-wall, for purposes of defence or to paralyse prey.
Histologically they differ from mesenterial filaments.

CLASSIFICATION OF AOTINIARIA 573
AcRORHAGI.—Marginal outgrowths of the body-wall
found in some genera of Actiniaria, and which may or may not
be specialized as nematocyst-batteries. They may be simple
(spherical, conical, &c.), slightly compound, or even frondose.

TExtT-FIG. 20.

--Z OANTHACTINIARIA

> --Dobe cacrin/ariA
bY

™“
EN

ae
-.

=

AnrHozoan ANCESTORS

Diagrammatic representation of the classification and evolution
of the Zoanthactiniaria.

CapritruLuM.—The bodies of some <Actiniaria show a
distinction into three regions: the main part of the body in

such cages is termed the scapus, and may be provided with

cuticle. The distal extremity, which bears the tentacles, is

NO, 260 Qq

574 T. A. STEPHENSON

termed the capitulum; it may or may not be very distinct
from the scapus ; usually it has no cuticle ; it may be delicate
and different in structure from the scapus, and introvertible
into the latter. The aboral end of the body if rounded and
able to become bladder-like is called a physa. Some adherent
forms possess scapus and capitulum, but ordinary base instead
of physa ; among these the capitulum may be delicate or may
be very thick-walled. There are grades between a physa and
a well-marked adherent base, and some bases may temporarily
become physa-like.

Cinratep Tracts (Flimmerstreifen) of mesenterial
filaments. In the filaments of Zoanthinaria, Edwardsiaria,
and Nynantheae, a transverse section cut at the right level
will show a trifoliate outline, portions of the lateral lobes of
the trefoil being composed of plain ciliated cells, these portions
forming, therefore, in the whole filament, lateral ciliated
tracts on either side of a median glandular or cnido-
glandular tract (Nesseldriisenstreif).

CINCLIDES.—Pores in an Actinian body-wall. Function
perhaps connected with water-currents ; in some cases they
seem to provide safety-valves against rupture of the wall on
sudden jerky contraction. Connexion with acontia secondary
and indirect.

Concnuta.—The specialized upper extremity of the
siphonoglyphe in the genus Peachia. Perhaps connected
with the entry or exit of a water-current when the animal
is embedded in sand up to the disc.

CouPpuLeE of mesenteries. See foot-note.

Enpvocoer.. The space between two mesenteries of the
same pair.!

' In this paper the word ‘ pair’ is used of two mesenteries, both on the
same side of the body, and adjacent to one another—and usually with
their retractor muscles vis-a-vis. The word ‘couple’ is applied to
two mesenteries arising at the same time and symmetrical about the
long axis of the actinopharynx, but one on one side of the latter, and one
on the other; their retractors facing the same way. Thus ordinary

directive mesenteries are strictly couples, though usually called
pairs for convenience.

CLASSIFICATION OF ACTINIARIA 575

Exocoru.—tThe space between two pairs of mesenteries.

Fossr.—Some anemones have the margin of the body
raised into a distinct rim or parapet, outside the bases of
the tentacles ; the circular groove between this parapet and
the tentacle-bases is known as a fosse.

MacROCNEME.—A typical macrocneme is a well-developed
mesentery which joins the actinopharynx as well as the body-
wall, has a strong and usually circumscribed retractor muscle,
a gonad, and a mesenterial filament. There are sometimes
variations in detail from this general plan.

MrtTACNEME.—Any mesentery formed after the earliest
eight mesenteries to appear (protocnemes).

MrircrocNeMeE.—Typically a narrow mesentery which
does not join the actinopharynx, has little or no muscle beyond
a * parietal muscle —no retractor therefore—no gonad, and
no filament. Variations from this typical scheme are found,
however.

NEMATOSPHERE.—A tentacle which has become con-
verted into a short structure rounded at the end, or into a
practically sessile sphere, and the ectoderm of at least part of
which is crowded with nematocysts.

Parr of mesenteries. See foot-note on previous page.

Perrect MrsentEeRyY.—In a form where there are
graded cycles of mesenteries (1.e. no division of the mesenteries
into macro- and microcnemes), any mesentery which joins
the actinopharynx as well as being inserted into body-wall
and oral disc, is termed ‘ perfect’. In a form where there
are macro- and microcnemes, the former are of course * perfect ’
as part of their macrocnemic nature ; but in some cases some
of the microcnemes may join the actinopharynx though
otherwise more or less rudimentary. They are then technically
‘perfect ’ mesenteries, but are by no means macrocnemes.,
In the forms with graded cycles, the perfect mesenteries have
filaments and retractors, but not always gonads, which in
such forms may appear on the ‘ imperfect ’ mesenteries only.
In such forms the older imperfect mesenteries, at least, may
have retractor, gonad, and filament, so that they are not

Qq?

576 T. A. STEPHENSON

microcnemes although less fully formed than the perfect
mesenteries.

Puysa.—sSee Capitulum.

ProrocnEeME.—The first four bilateral couples! of
mesenteries to be formed in a Zoanthactiniarian.

Scapus.—sSee Capitulum.

SpuinctHR.—The sphincter usually referred to in this
paper is the one running round within the upper margin of
the body, outside the tentacle-bases, in many anemones. It
may be embedded in the mesogloea of this region (meso-
gloeal), or its fibres may be supported on processes of meso-
gloea which project into the endoderm (endodermal). It
may be spread out a good deal (diffuse) or gathered up into
a definite sharply marked-off cord, which at its best forms
a marked projection from the body-wall into the coelenteron
(circumscribed). ‘There are various intermediate grades
between diffuse and circumscribed, and various degrees of
strength in sphincters.

STICHODACTYLINE condition of tentacles. This is the
term used to denote the state of affairs in which more than
one tentacle communicates with at least some of the endocoels,
sometimes with all endocoels, and with exoccels also.

VeRRUCAE.—These are local, slightly differentiated
sucker-like warts or slightly hollow outgrowths of the body-
wall, and often they attach foreign bodies to themselves.

VesrtcuEes.—tThese are truly hollow, bladder-like exten-
sions of the coelenteron into outgrowths of the body. They may
be delicate and thin-walled, simple or compound, and some
times are well provided with nematocysts.

1 See foot-note on p. 574

peut ISG Gs @ wie AE Oe INNS aS aha ee

The Development of the Sea Anemone
Bolocera Tuediae (Johnst.).

By

Prof. James F. Gemmill, University Coll., Dundee.

With Plate 22.

Bolocera tuediae was recorded and described from
deep water near Berwick by Johnston (11) in 1832. Gosse
described it more fully in 1860 (10, p. 185) and the following
is his summary of its characters: ‘Base adherent, not much
exceeding the column. Column _ pillar-like, the diameter
and height sub-equal ; surface generally very smooth, studded
with warts remotely scattered. Dise smooth, circular in
outline, not overlapping the column. Tentacles short,
thick, constricted at foot, obtusely pointed, loagitudinally
furrowed, flexuous and motile, easily separated, not retractile.
Mouth raised ona cone. Stomach capable of being greatly
protruded.’ The tentacles are, however, moderately long and
slender when fully extended during life.

Carlgren (8, pp. 34-6) adds that the genus Bolocera is char-
acterized by the presence of a relatively well-developed diffuse
or circumscribed endodermal sphincter, that the column has
no ectodermal longitudinal muscular layer, that the tentacles
have a well-marked endodermal sphincter at their bases, and
that probably all the mesenteries except the eight ‘ Kdwardsia ’
ones are fertile. Carlgren follows MeMurrich (16) in judging
that Bolocera must be placed in a separate Family, the Bolo-
ceridae. Its nearest allies are probably among the Antheinae
in which, however, the sphincter is extremely feeble if not
entirely absent (see Delage, 6, 1. 2, pp. 503-5).

In the Clyde Fauna List (Laurie, 18, p. 867) Bolocera

1 T am indebted to the Trustees of the Carnegie Trust for a grant
towards the expenses of this investigation.

578 JAMES F. GEMMILL

tuediae is put down as occurring at depths of from fifteen
to seventy-five fathoms. My own records lower the first limit
to thirty fathoms. While possessing an attaching base and
capable of adhering weakly to the sides or bottom of an
aquarium tank, Bolocera appears to live usually on muddy
bottoms, and is almost always brought up by itself when
taken with the dredge or on the long lines of fishermen. It
has great stinging powers, and one has to risk a somewhat
severe urticaria when handling it alive.

The sexes are separate and the gonads are at their largest
in the end of February and beginning of March. Unfortunately
the females very seldom spawn in captivity. The eggs are
retamed and undergo absorption after a time. Probably
want of the natural food is a contributing reason. The males
shed their sperm more freely.

Only a few eggs were obtained in 1916 and 1917, but in
March 1918 large numbers were extruded by a recently-taken
specimen. These after floating about in the Bolocera tank
were duly fertilized, although none of the males at the time
had emitted a noticeable amount of sperm.

Maturation must take place just prior to extrusion. Serial
sections of full-sized ovaries show the eggs with large-sized
germinal vesicles, but in similar sections of freshly-shed unfer-
tilized eggs the nucleus is so small and inconspicuous that
I could not detect it.

The eggs are spherical, 1-1 mm. in diameter, and pink or
flesh coloured, i.e. of much the same tint as the animal itself.
They tend to float, and when floating show no polarity as
regards upper and under sides. They are surrounded by
a membrane beset all round by small conical bunches of spines.
The interior is crowded with small granules faintly stainable
with haematoxylin, small yolk-spheres staining red with
eosin, and large clear spherules unaffected by re-agents, the
latter being relatively more numerous towards the centre of
the egg. In certain methods of preservation (e.g. corrosive
sublimate followed by graded alcohols) an inner core, about half
the diameter of the egg, tends to become separated from the

RT Me Pe

DEVELOPMENT OF BOLOCERA 579

outer zone. Just under the egg-membrane is a thin layer
where the first-named granules are very numerous, the clear
spherules absent, and the yolk-spheres few in number.

Bolocera has the largest eggs of all the Clyde Anemones
I have investigated. Their diameter, 1-1 mm., compares with
0-1 mm. for Metridium dianthus, 0-3 mm. for Anthea
cereus, 0-1 mm. for Sagartia, 0:25mm. for Adamsia
palliata, 06mm. for Urticina coriacea (the shore
Urticina), 0-7 mm. for Urticina crassicornis (the sub-
merged Urticina). Full-grown ovarian eggs of Gonactinia
prolifera and of Actinia equina measure respectively
0-07 and 0-15 mm. in diameter. The Bolocera egg-membrane
and its spines resemble but are hardly so strong as those of
Urticma. The egg-contents of the two are much the same.
Anthea (and Actinia equina, according to Lacaze Duthiers)
has spiny egg-membranes, but in Metridium, Sagartia, and
Adamsia the membranes in question are smooth.

In Bolocera, as in Urticina (Appellof, 1), the fertilized
nucleus gives rise to a number of daughter nuclei (sixteen in
Urticina) before the egg-mass undergoes cleavage. In particular
cases I have estimated the number as not less than twenty-
four. The fertilized nucleus probably lay at a point some-
where in the deeper layer of the outer zone, about a third
of the diameter of the egg inwards from the surface. The
daughter nuclei, as they increase in number, spread laterally
at this level from the point in question until they are more or
less equally distributed all round. In the egg illustrated by
fig. 1, eight nuclei were present, all of them in one hemisphere.

Slightly older eggs examined under reflected light begin to
show rounded bosses or humpings which appear first at one
side (no doubt the side towards which the fertilized nucleus
lay), and afterwards extend all over the egg-surface. They
soon become better defined and separated from one another
by linear furrows. Segmentation of the egg-mass is in progress,
and serial sections show that each hump is the outer end of
a large more or less conical cell the apex of which is directed
centrally. The whole egg increases slightly in size, and a small

580 JAMES F. GEMMILL

central cavity filled with coagulable fluid makes its appearance.
The egg-membrane is not separated off as a membrane of
fertilization, but is found to follow closely every surface
change of contour so long as it is recognizable. As segmentation
proceeds, non-nucleated portions separate offfrom the inner ends
of the cells, and, mixing with the blastocoelie fluid, form a central
diffuse trophenchyme. At this stage one or two whole cells
may share the same fate by migrating or getting pushed
inwards from the surface. Their nuclei proliferate; but,
soon losing control over the cell-contents which become
trophenchymal, are destined to degenerate along with the
other trophenchymal nuclei to be deseribed later.

A little later the Bolocera egg shows very markedly those
peculiar surface grooves and foldings which Masterman
first described in the case of Cribrella (17, p. 8), and which have
since been noted in many ova (8, p. 12). During this process there
is a tendency, better marked in some instances than in others,
for the egg to assume the form of a flattened dise the edges of
which become turned upwards like those of a saucer. The
surface grooves and the saucer cavity gradually fill out, so
that the egg becomes almost spherical again. The saucer
cavity is accordingly not the archenteron, though gastrulation,
which soon supervenes, affects the part of the egg-wall
which was formerly the hollow of the saucer. In the fully-
formed blastula this part often remains flat while the rest of
the blastula wall is spherical.

An important point to note is that as the surface folds smooth
out, many single cells and groups of cells are nipped off from
the recesses, and migrating inwards become included within
the trophenchyme. I thought at first that these cells were
going to form the endoderm of the larva. But this is not so.
Their cell outlines will disappear and their nuclei degenerate.

Gastrulation.—In typical cases (see e.g. figs. 7-9)
a relatively large portion of the blastula wall shows
flattening and sinks gradually downwards, the margins of
this portion closing in slowly to form the lip of the blasto-
poric opening. At the same time this hp becomes slightly

ated iieeeiet ee kee ee et ee EEE

DEVELOPMENT OF BOLOCERA 5$1

involuted giving rise to the rudiment of the stomodoeal
canal.

The invaginating area soon presses against the trophenchyme,
and we often find at this stage secondary flattening of the whole
egg and foldings of its walls, which are probably caused by
the resistance of the trophenchyme to the progress of invagina-
tion. However, in course of time, the trophenchyme finds its
way through the mpushing endcderm into the cavity of the
archenteron. Tirst, the fluid and fine granules begin to get
through, then the yolk-spheres, and lastly the clear spherules.
The process appears to be mechanical in the sense that the
trophic material passes through interstices between endoderm
cells, and is not first swallowed or assimilated and then excreted
into the archenteron.

As gastrulation proceeds most of the trophenchymal nuclei
disintegrate, but some pass with the trophenchyme into the
archenteron and are absorbed later.

It is of particular interest to note that in a few cases the
end-result of gastrulation is attained by a process which may
be described more accurately as unipolar immigration than as
invagination. In such cases the cells over a relatively small
area at one pole of the blastula begin to sink inwards through
the trophenchyme, at the same time proliferating and spreading
out so as almost to lose their continuity with one another.
This process continues until having passed through the whole
depth of the trophenchyme, they abut against the ectoderm
where they soon form a continuous sheet of endoderm lining
an archenteric cavity which now naturally contains all the
trophenchyme. Sometimes the process is intermediate between
that described above and open invagination. Similar differ-
ences occur among the eggs of different Crustacea, but not so
far as I know among the eggs of the same Crustacean species.
We may put down the variations in Bolocera as probably due
to differences in the character of the yolk, noting that those
blastulae which showed the fewest foldings and the least
deformation tend also to form their endoderm by unipolar
immigration.

BK

582. JAMES F. GEMMILL

A mesogloeal sheet only begins to form after the ectoderm
and endoderm have come in contact. Accordingly it appears
first at the oral end of the larva. Both layers seem to take
part in its formation.

Comparison with other Anemones as regards
the Stages up to the end of gastrulation.
Metridium dianthus.—Nuclear division and segmenta-

tion go together from the first ; blastula with a hollow central
cavity; endoderm formed by invagination (Gemmill, 9).
MeMurrich, however, stated (15) that the endoderm is formed
by delamination. Sagartia troglodytes.—As in Metri-
dium. Adamsia palliata.—Cleavage begins after the
second nuclear division; the preblastula is a wrinkled dise,
becoming saucer-shaped, and then smooth and spherical or
oval; the inner yolky ends of the cells separate off to form
a central trophenchyme normally without nuclei; gastrulation
is by imvagination (Gemunill, 9), and the trophenchyme passes
through the inpushing endoderm into the archenteron. Faurot

(7), however, stated that the endoderm is formed by delamina-

tion. Urticina crassicornis.—Development is much the
same as in Bolocera. Cleavage, however, begins when there
are sixteen nuclei in the egg, and the trophenchyme nuclei
are sparing or absent. The crumpling and folding of the wall
of the early blastula which I find to be very well marked in the
eggs of Urticina have not been described by Appelléf in his
otherwise excellent account of the development of this species
(1). Actinia bermudensis.—KHarly stages not deter-
mined; gastrulation by invagination (Cary, 5). Actinia
equina.—Harly stages not determined ; endoderm formation
by invagination according to Jourdan (12), but by immigra-
tion or delamination according to Appellof (1), who states that
the mouth opening is a secondary break-through.t Cerian-
thus and an Actinian allied to A. equina.—Endoderm
formation by invagination (Kowalevsky, 18).
Movements.—Cilia are acquired during the middle blastula
]

My own observations (Millport, 1920) are entirely in favour of the
open invagination method of endoderm formation in this species.

DEVELOPMENT OF BOLOCERA 583

stage and show activity before the egg-membrane spines have
disappeared. bBlastulae and early gastrulae move irregularly,
but late gastrulae and older larvae progress with the aboral
end in advance, rotating at the same time in the contra-solar
direction as viewed from this end. Meantime a change of
specific gravity has occurred and the larvae tend to remain
on or near the bottom. Elongation of the larva has also taken
place in the oral-aboral axis. The shape now varies according
to contraction but is usually pyriform, the aboral end being
the smaller. Over this end the ectodermal cells elongate,
becoming clear at their outer extremities. They are preparing
a cement in view of fixation. At no stage is there present
a specially elongated tuft of cilia such as is characteristic of
the larvae of Metridium and Sagartia and in a less degree of
Actinia equina.

Mesenteries.—The eight primary or Edwardsia mesen-
teries appear, first in the neighbourhood of the mouth, as folds
of the endoderm, each containing a thin mesogloeal sheet con-
tinuous with the general mesogloeal layer between ectoderm
and endoderm. The sulco-laterals (ventro-laterals) are the
first to develop. The remainder appear practically simultane-
ously, but I could sometimes make out that the sulculo-laterals
were a little ahead of the sulear directives, and the latter of
the sulcular directives. In the figures the mesenteries are
numbered 1, 2, 3, 4, corresponding to the above sequence.

All the primary mesenteries have appeared prior to fixation,
and at this stage the oral ends of the sulco-laterals are already
edged by a down-growth of stomodoeal ectoderm for the
mesenteric filaments, and project so far inwards as almost to
meet one another. The developing muscle banners on all the
mesenteries show the characteristic Edwardsia arrangement.

Fixation occurs about twenty-five days after shedding of
the eggs, and is at first by cement attachment, the larvae
adhering usually to the bottom but sometimes to the sides of
the hatching vessel. ‘The base, at first small and pointed,
soon becomes larger and dise-like. Shortening of the larva
takes place till the length of the column is less than its breadth ;
the oral surface flattens ; the mouth opens widely and elongates

584 JAMES F. GEMMILL

in the axis of the directive mesenteries. Then the young
anemones remain quiescent except in showing the following
changes.

1. Absorption of the trophenchyme within the archenteron.
It is partly used up and partly absorbed into the endoderm
layer, which becomes greatly thickened, as well as extended
by the fuller growth of the mesenteries.

2. Down-growth of stomodoeal ectoderm to form mesenteric
filaments on the sulculo-lateral mesenteries. This began prior
to fixation on the sulco-laterals.

3. Formation of a new mesentery in each lateral and sulco-
lateral Kdwardsia space. These mesenteries can be detected
near the middle of the column of the larva earlier than near
the mouth or on the base. In my oldest specimens their
developing muscle banners could with much difficulty be made
out, each bemg formed on the sulcular side of its mesentery
as in Urticina (1). They are thus suitably placed to form with
the Edwardsia sulco- and sulculo-laterals, the primary hexac-
tinian uleo- and sulculo-lateral mesenteric pairs on each side,
the r¢ maining pairs being of course the sulcar and sulcular
directives (fig. 15).

I tried to rear the young anemones further, but so far without
success, although I gave the larvae the chance of settling down
on shells, stones, glass, and mud, and of living after attachment
either in separate hatching vessels, or in a tank with sea-water
circulation. Those which settled on mud retained a rounded
base, but otherwise reached much the same stages as the
attached ones. None went the length of growing out tentacles.
The attached specimens were less firmly fixed, and yet crept
about less freely, than the corresponding stages in Urticina,
in which also, as was shown by serial sections, the mesogloeal
and muscular tissues were more strongly developed.

For further comparative details and a discussion of some
general problems connected with coelenterate development,
reference may be made to a recent paper by the author in
the ‘ Phil. Trans. Roy. Soc. Lond.’ (9) on the development
of Metridium dianthus and Adamsia palliata.

DEVELOPMENT OF BOLOCERA 585

SUMMARY AND CHRONOLOGY.

Egg large, floating ; maturation prior to shedding ; fertiliza-
tion external; at least twenty-four nuclei present before
cleavage of egg-mass takes place (fifteen hours); cleavage
total leaving a small central cavity ; the inner ends of the
cells separate off to form a central trophenchyme (twenty-four
hours); a greatly-folded ‘ preblastula’ stage during which
groups of cells are included in the trophenchyme (forty-eight
hours) ; the blastula becomes more or less smooth and spherical
(three to three and a half days) ; gastrulation begins (four to
five days) ; gastrulation complete and first mesogloea formed
(six and a half to seven and a half days), the trophenchyme
passing into the archenteron, and degeneration of its nuclei
taking place; blastopore narrows and virtually closes, involu-
tion of stomodaeum taking place; larva elongates (nine to
ten days); sulco-lateral mesenteries begin to form (fifteen
days) ; aboral end shows cement gland formation, and rudi-
ments of the other mesenteries appear (twenty days) ; fixation
and shortening of the larva (twenty-five days); formation
of four additional mesenteries (thirty-six days); complete
absorption of trophenchyme within archenteron (thirty-six
days). For cilia, movements, &c., see p. 582. At no stage is
there a tuft of specially elongated aboral cilia.

REFERENCES,
1. Appelléf, A.—‘‘Studien iiber Aktinien-Entwickelung”’, ‘ Bergens
Museum Aarbog’, 1900, pp. 1-99.
2. Carlgren, O.—‘‘ Beitrige zur Kenntniss der Aktiniengattung Bolo-
cera ’’, ‘ Oef. Vet. Akad. Forh. Stockholm ’, vol. 48, 1891, pp. 241-

50.

3. “Die Aktiniarien der Olga-Expedition ”’, “ Wiss. Meeres-Unter-
such. (Helgoland) ’, Bd. V, pp. 33-55.

4. “Studien iiber Nordische Aktinien”, ‘K. Svenska Vet.-Akad.

Handl.’ 25, no. 10, 1893.

5. Cary, L. R.—‘‘ The Formation of the Germ Layers in Actinia
bermudensis, Verr’’, ‘ Biol. Bull. Woods Hole’, xix, 1910,
p. 339-46,

6. Delage et Hérouard.—‘ Traité de Zoologie concréte ’, ii. 2, Coolenterés.

7. Faurot, L.—‘‘ Etudes sur l’anatomie, Vhistologie, et le développe-

586 JAMES F. GEMMILL

ment des Actinies’’, ‘Arch. Zool. expér. et gén.’, sér. iii, t. 3,
pp. 43-262.

8. Gemmill, J. F.—‘‘ The Development of the Starfish Solaster
endeca, Forbes”, ‘ Trans. Zool. Soc. Lond.’, 1912, pp. 1-71.

9. —— ‘‘ The Development of the Sea Anemones Metridium dian-
thus and Adamsia palliata”, ‘Phil. Trans. Roy. Soc.
Lond.’, Ser. B, vol. 209, pp. 351-75.

10. Gosse, P. H.—‘ A History of the British Sea Anemones and Corals ’,
London, Van Voorst, 1860.

11. Johnston, G.—“‘ Illustrations in British Zoology ”’, ‘Mag. Nat. Hist.’,
v, London, 1832, pp. 163-4.

12. Jourdan, E.—‘‘ Recherches zool. et histol. sur les Zoanthaires du
Golfe de Marseille ”, ‘ Ann. Sc. Nat.’, sér. vi, t. 10, 1878.

13. Kowalevsky, A.—See Hoffmann, W., *Schwalbe’s Jahresbericht f.
Anatomie *, 1873.

14. Lacaze Duthiers, H. de.—‘* Développement des Coralliaires ; Acti-
niaires sans Polypier”’, ‘ Arch. Zool. expér. et gén.’, t. 1, 1872.

15. MceMurrich, J. P.—‘‘On the Development of the Hexactiniae”’,
* Journ. Morph.’, vol. 4, pp. 303-30.

16. —— ‘ Report on the Actiniaria collected by the Bahama Expedition
of the State University of Iowa, 1897”, * Bull. Lab. Nat. Hist.
Univ. Iowa, 1898’ (quoted from Carlgren (8), p. 34).

17. Masterman, A. T.—‘‘The Early Development of Cribrella
oculata, Forbes, with remarks on Echinoderm Development ”’,
‘Trans. Roy. Soc. Edin.’, vol. 40, 1902, pp. 373-417.

18. Scott Eliot, Laurie, and Murdoch.— Fauna, Flora and Geology of the
Clyde Area’. Glasgow, MacLehose, 1901. (Laurie, p. 367.)

19. Wilson, H. V.—‘‘ The Development of Manicina areolata”,
‘ Journ. Morph. Boston’, vol. 2, pp. 191-252.

EXPLANATION OF FIGURES.

bl.c., blastocoele cavity ; end., endoderm ; g., commencement of gastru-
lation; n.f., furrows on blastula from which trophenchyme nuclei
are nipped off; x.tr., these nuclei degenerating within the trophen-
chyme; st., stomodaeum; ?r., trophenchyme; ¢r.a., trophenchyme
within the archenteron; 1-6, the primary hexactinian mesenteries
numbered according to their order of development (see explanation in
text).

Fig. 1.—Section of Bolocera egg, 10 hours after fertilization, showing
five nuclei; all the nuclei are in one hemisphere of the egg.

Fig. 2.—Section of egg about 18 hours after fertilization, showing the
characteristic complete segmentation (see p. 579).

|
|

a

DEVELOPMENT OF BOLOCERA 587

Fig. 3.—Similar section about 28 hours after fertilization. Note the
passage of one of the cells inwards from the surface, and commencement
of trophenchyme formation.

Fig. 4.—Similar section about 36 hours after fertilization.

Fig. 5.—Similar section about 48 hours after fertilization. Note the
extremely folded and crumpled surface (see p. 580).

Fig. 6.—Similar section about 34 days after fertilization. The folds
have mostly straightened out leaving behind numerous groups of cells
nipped off from their recesses and enclosed within the trophenchyme.
The outlines of these cells disappear and the nuclei degenerate now or
later.

Fig. 7.—Similar section about 4} days after fertilization. Commence-
ment of gastrulation.

Fig. 8.—Similar section about 5} days after fertilization.

Fig. 9.—Similar section about 7 days after fertilization.

Fig. 10.—Similar section about 8 days after fertilization, showing
(a) the progress of gastrulation, (b) the passage of the trophenchyme through
the inpushing endoderm into the archenteron, and (c) the involution of
the lips of the blastopore to form the stomodaeum.

Fig. 11.—Longitudinal section of larva 12 days old. The shape is now
pyriform and the cells at the aboral end are becoming elongated and
glandular.

Fig. 12.—Transverse section across larva 15 days old near its oral
extremity showing the two first mesenteries—the sulco-laterals.

Fig. 13.—Transverse section through larva 20 days old showing forma-
tion of all the Edwardsia mesenteries, viz. (1) the sulco-laterals ; (2) the
sulculo-laterals ; (3) the sulcar directives, and (4) the sulcular directives.
In this specimen the last named are the smallest and were no doubt
the latest to appear (see p. 583). The sulco-laterals are now edged by
a down-growth of epiblast for the mesenteric filaments.

Fig. 14.—Diagram of transverse section of attached specimen (25 days
old) to illustrate the arrangement of the eight Edwardsia mesenteries
which are numbered as in the previous figure, and on which the rudiments
of muscle banners can now be made out.

Fig. 15.—Similar transverse section of attached specimen (36 days old)
in which a new mesentery (numbered 5 and 6 respectively) has developed
in each sulco-lateral and lateral Edwardsia space. Muscle banners are
beginning to develop on their sulcular sides. The six primary hexactinian
mesenteric pairs will consist of the sulcar directives, the suleular directives,
two pairs made up of two and five on either side, and two pairs made up
of one and six on either side.

a ? ve... ,
= 7
Jj lo, ?, on
etka mel
f iy 7 *Z
m2 3 4

|
;

li del.

Gemini

. . a f 2a) YA €
Luan Fourn, Mier Sti. Vt. C5 NS PAL

Huth London

Observations on the Shape of the Nucleus
and its Determination.
By
Christian Champy,
Professeur agrégé & la Faculté de Médecine de Paris,
and

H. M. Carleton,

Demonstrator in Histology, University of Oxford.

With Plates 23 and 24 and 11 Text-figures.

CONTENTS.

PAGE
1. INTRODUCTORY , 589

2. THe RELATION BETWEEN ements Ree AND SURFACE
TENSION . A A : F ‘ 590
3. MECHANICAL DEFORMATION OF THE NUCLEUS 592
4, NUCLEAR SHAPE AND THE CENTROSOME . 595
5. THE RELATION BETWEEN CELL SHAPE AND Nugind AR Sin: APE 596
6. CANALICULI IN THE NUCLEAR MEMBRANE : 597
7. Fotps AnD INcISIONS IN THE NUCLEAR MEMBRANE 600
8. THE UnrotpinG oF NUCLEAR INVAGINATIONS 602
9. INTRANUCLEAR RODLETS, ETC. ‘ P 602
10. THE RELATION BETWEEN NUCLEOLI AND | Nackaas SHAPE 604
11. Ceti Diviston anp NucLEAR DIFFERENTIATION 605
12. SumMaRyY 606
13. BIBLIOGRAPHY 608
14, EXPLANATION OF PLATES 608

1. INTRODUCTORY.

Tat the nucleus is extraordinarily variable in shape, not

only in different animal cells but also in the same cell during the
different phases of its ontogeny and metabolism, is a notorious

fact.

In the following notes, which embody a brief description
NO, 260 RY

590 CHRISTIAN CHAMPY AND H. M. CARLETON

of nuclear shape, we have also attempted to analyse, when
possible, the factors responsible for this.

In the present stage of cytology the interpretation of cell-
function is largely based on purely descriptive methods. There-
fore such reasons as we have been able to put forward in explana-
tion of the diversity of nuclear shape are to be regarded more
as reasonable suppositions than as proven statements. We
are of opinion that it is better to run the risk of assigning false
causes to the phenomena which we have observed, than to
explain nothing by confining ourselves to purely morphological
considerations.

Only when cytology has acquired experimental methods will
it be possible rigorously to determine the factors responsible
for nuclear form and function.

Although the details of the structure of the nucleus—and
particularly those concerning the disposition of the chromatin
and the alleged ‘ linin’ network—are controversial, observa-
tions on nuclear shape are easily verified. For not only are
the appearances similar with widely different methods of
fixation and staming, but they can be controlled by observa-
tions on living material. And, finally, corroborative evidence
can sometimes be obtaimed by experimental methods such as
tissue culture.

2. Tur RELATION BETWEEN NUCLEAR SHAPE AND
SURFACE TENSION.

A spherical nucleus is found in hepatic and most
other gland-cells, also in many nerve-cells and spermatocytes.
Its shape may often be attributed to surface tension, being
the result of a relatively fluid (nuclear) mass that is immiscible
in the surrounding cytoplasm. Such nuclei are relatively
rare in the animal body, for the spheroidal condition is not
uncommonly associated with mechanical factors, e.g. furrows
or canaliculi in the nuclear membrane. Such structures,
which occur more often than is generally supposed, sometimes
make it difficult to say whether a spherical nucleus is the result
of surface tension alone or of accompanying mechanical causes.

;

1 te

SHAPE OF THE NUCLEUS 591

It is a curious fact that such nuclei usually contain a single
nucleolus only, and that this body tends to be in the centre of
the nucleus or somewhat deviated towards that pole of it which
is farthest from the centrosome. The latter seems to exercise
a repellant action on the nucleus also—a fact which can be
verified in many cells with large nuclei, e.g. spermatocytes.
Lobulation of the nucleus can sometimes be attri-
buted to variations in surface tension at the interfaces of

TExtT-FiG. 1.

Phagocytic cell (amoebocyte) from a larva of Phryganea sp.—
a caddis fly. Extreme polymorphism of the nucleus probably
due to variations in surface tension over the nuclear membrane.
Nu, nucleolus; va, vacuole. Technique: Bouin and iron
haematoxylin.

nucleus and cytoplasm. <A striking example of this is furmshed
by the large cells accompanying histolysis during metamor-
phosis in insects. Here, as is well known, the larval tissues
are destroyed by large phagocytic cells known as Amoebocytes.
Text-fig. 1 shows such an element from a larva of Phryganea
sp. Here the polymorphism of the nucleus is extreme, while
the nucleolus, which is single and central, does not appear to
be involved in the lobulation. Of the latter, every degree
Rr2

592 CHRISTIAN GCHAMPY AND H. M. CARLETON

may be observed in such cells, and it seems definitely to be
related to variations in surface tension caused by exchanges
between nucleus and cytoplasm, as has been suggested by
various authors (e.g. Prenant, 10).

In other instances, however, the shape of the nucleus,
notwithstanding its extreme lobulation, is too definite to permit
of its being attributed to surface tension alone. Examples of
this are the spermatogonia of some Amphibia, in which the
shape of the nucleus is constant ina given species (PI. 28, fig. 3).
Here the nuclear polymorphism is apparently due to the
intervention of other factors (to be considered later), and
only such variations from the normal as occur during periods
of intensive cell-activity—such as growth, differentiation, &¢.—
can be ascribed to the surface-tension changes that accompany
such phenomena. Somewhat similar are the modifications
which occur in many oocytes during development, as shown
in Pl. 23, fig. 2. In the early stages of differentiation the nucleus
in such elements is oval, containing one large central nucleolus
and many smaller peripheral ones. But subsequently the
nucleus becomes polymorphic, while around it is established
a clear (endoplasmic) zone in the cytoplasm. Here again do
we find extreme nuclear lobulation coinciding with enhanced
metabolism of the cell.

3. MECHANICAL DEFORMATION OF THE NUCLEUS.

The study of our material has convinced us that nuclear
shape is often due to pressure exerted on it by various cell
inclusions. An obvious example of this is furnished by the
thin and crescentic nucleus entirely pressed to the periphery
of the fully developed adipose cell. Somewhat similar is the
deformation of the nucleus in the duct-cells from the pronephros
of Triton (see Pl. 23, fig. 4). This is due to the centre of the
cell being occupied by the lumen of the duct. Again, in the
early segmentation stages of ova containing much yolk, the
nuclei are indented by the large, mert yolk-dises. Text-fig. 6
shows such a nucleus from a blastula of the Amphibian
Triton alpestris. On one side there are deep indentations

SHAPE OF THE NUCLEUS 593

between the centrosome and the nuclear membrane: these
are due to other causes and will be referred to later. And,
finally, similar appearances can be seen in the nuclei of the
interstitial cells of the testicle of Rana esculenta, the
nuclear membrane here being pitted by the lecithin globules
in the cytoplasm.

Sometimes the inclusions are localized in a particular area
of the cytoplasm. This may give rise to a peculiar deformation
of the nucleus such as is depicted in PI. 28, fig. 6, which illustrates
a cell from the hepato-pancreas of the isopod crustacean
Oniscus. Here the nucleus at the basal, i.e. attached,

TEXxtT-Figc. 2.

Cell from pronephros of a 3mm. larva of Triton alpestris.
Note constriction of middle of nucleus due to pressure from
Tonofibrillae, TN.

pole of the cell is strikingly indented by large cytoplasmic
globules of a lipoid nature. It follows from this that nuclear
deformation can be produced by relatively fluid bodies.

Another instance of nuclear shape being modified by ecyto-
plasmic structures is afforded by the intestinal epithelial cells
of the same species. By appropriate stainmg methods (see
Pl. 24, fig. 3) fine fibrils lying in the cytoplasm around the
nucleus can be distinguished. They run from the basement
membrane to the cuticle, apparently function as an intra-
cellular skeleton, and may be termed Tonofibrillae after
the French ‘ Tonofibrilles ’.

We have also observed a similar condition in cells from the
excretory tubules of larvae of Triton as shown in Text-
fig. 2.

In those muscles which are characterized by cross-striation

594 CHRISTIAN CHAMPY AND H. M. CARLETON

of their fibres, 1.e. ordinary striated and cardiac muscle, the
influence of cytoplasmic structures on nuclear shape is very
marked. ‘Thus, in striated muscle there is obvious flattening
of the nuclei against the sarcolemma due to pressure from the
areas of Cohnheim (i.e. groups of fibrils) of which the fibre is
composed.

Often, however, other causes intervene, chief amongst which
is the influence of the Membrane of Krause (‘Strie Z’
of the French and ‘ Zwischenscheibe ’ of the German authors).

TEXT-FIG. 3.

A, After Cajal, showing intranuclear rodlet, NR, in pyramidal cell
from cerebral cortex of rabbit. Technique: Cajal method for
Golgi apparatus, Ga. B, After Retzius, depicting peri-nuclear
structure (xX) in spermatozoon of the Gasteropod Cypraea.

This structure is segmentally disposed along the muscle fibril
and appears in the middle of the dark bands as a clear and
narrow line. It is best studied in the large fibres of insects
(see Text-fig. 10), where it can be seen to constrict the nucleus
at regular intervals by its projection out of the fibrils into the
surrounding sarcoplasm.

A similar appearance of the nuclei can be seen in human
cardiae and other vertebrate muscle. This is shown in Pl. 28,
fig. 7. But here we have not been able to follow the membrane
of Krause as far as the nuclear membrane. Nevertheless, the
nuclei bear definite constrictions corresponding to the mem-
branes of Krause of adjacent fibrils, while the curious blunt-

SHAPE OF THE NUCLEUS 595

ended nuclei—so characteristic of human heart muscle—can
only be explained by assuming the presence of these membranes
lying invisible in the sarcoplasm at each end of the nucleus.

TExt-rie. 4.

_ Oocyte of Esox lucius—a pike. Note the pouches in nuclear mem-
brane usually in relation with the nucleoli. Nv, nucleolus; yp, yolk-
discs. Technique: Bouin and iron haematoxylin.

4. NUCLEAR SHAPE AND THE CENTROSOME.

We deliberately confine ourselves to the consideration of
the centrosome and nucleus in the resting cell, as the question
of the spindle fibres, amphiaster and chromosome formation
is beyond the scope of these observations. In the resting
cell the centrosome often lies very close to the nuclear mem-
brane and opposite an indentation in it. And since this body
often does not touch the nucleus, one must surmise that the
depression is due not to mechanical causes but to repulsion
between nuclear membrane and centrosome. When an
amphiaster is present, its influence upon the nucleus is still
more marked, as is shown in Text-fig. 6, which depicts a blasto-
mere from an egg of Triton. It will be seen that here
nuclear shape is due partly to pressure from the yolk-dises
(as already pointed out), partly to invaginations in the nuclear
membrane in the vicinity of the centrosome. The astral rays
in fact deeply indent the nucleus wherever they come into
contact with it—a point possibly in favour of the view that the
cytoplasmic radiations around the centrosome are of a relatively
solid nature,

596 CHRISTIAN CHAMPY AND H.M. CARLETON

5. Tur RELATION BETWEEN CELL SHAPE AND NUCLEAR
SHAPE.

It is notorious that the longer a cell, the longer (usually)
is its nucleus. Muscle, columnar epithelium, and connective-
tissue cells are familiar examples of this (see Pl. 23, figs. 1, 8,
and 10; Pl. 24, fig. 2). This elongation of the nucleus is often
due to mechanical causes. ‘Thus, in epithelia it is sometimes
due to mutual cell-pressure, while the long nucleus of the
smooth muscle-fibre must be ascribed to pressure from the
myofibrillae. Further, the nucleus shortens or lengthens as
the fibre contracts or extends. Again, in preparations of
amphibian intestine fixed in different degrees of distension,
there are marked differences in the height of the epithelial
cells and their nuclei—the two varying in length in a parallel
ratio between certain limits. Exceptions, however, exist to
this general rule. For instance, in the intestinal epithelial
cells of the dragon-fly Libellula (see Pl. 24, fig. 5) the
small oval nucleus is quite disproportionate to the elongated
cell.

As claimed by Martin Heidenhain (8), we must surmise the
existence of a force which tends to push the nucleus towards
the centre of the cell. And in view of the plasticity of the
nucleus there can be no doubt but that this force must influ-
ence its shape also.

It is a fact of no small significance that the nucleus
never comes into contact with the cell mem-
brane, except in a few instances due to powerful
mechanical factors, e.g. pressure from bulky cyto-
plasmic inclusions forcing the nucleus against the cell mem-
brane. ‘Two possibilities suggest themselves in explanation
of this :

(1) That the position of the nucleus is due to
forces exerted on it by the surrounding cyto-
plasm, forces which might conceivably be proportional to
the mass of the cytoplasm around the nucleus. Were this so,
nuclear shape in a cell of greater length than breadth would

nS ea

SHAPE OF THE NUCLEUS 597

be as in Text-fig. 9, B on p. 600, which is never the case
in nature.

(2) That there is mutual repulsion between
cell membrane and nuclear membrane. Such a
force, acting in an inverse ratio to the distances between the
two membranes is indicated in Diagram C, p. 600. This
supposition explains :

(a) Why nuclear and cell membranes practically never come
into contact with one another.

(b) Why the nucleus tends to elongate concurrently with
the cytoplasm.

(c) Why the nucleus is never round so long as the length of
a cell is greater than its breadth, although there is often ample
room in the cytoplasm for it to become spherical.

Of the nature of such a force responsible for the antagonism
between cell membrane and nuclear membrane we know
nothing.

6. CANALICULI IN THE NUCLEAR MEMBRANE.

Intranuclear canaliculi are more common in spherical and
oval nuclei than is usually thought. They have been described
in the spermatogonia of Amphibia by Champy (4), and are
easy to demonstrate in Rana esculenta andthe Axolotl.
Canaliculi in the nuclear membrane occur in many types of
cell; we have observed them in the epithelium lning the
Wolffian duct in the salamander, and in pyramidal cells of
the cerebral cortex in the guinea-pig. ‘These structures are
illustrated in PI. 24, fig. 4, and in Text-fig. 8.

The intranuclear canaliculus is essentially a narrow invagima-
tion of the nuclear membrane. Its blind extremity, which
may be bifid, often ends in the vicinity of the nucleolus. That
this structure is a definite tube and not a deep furrow in the
nuclear membrane, is shown in transverse sections of it. In
many spermatogonia there seems to be some relation between
the canaliculus and the centrosome ; at the prophase the latter
comes to lie very close to the former, often exactly opposite
its aperture in the nuclear membrane.

8 CHRISTIAN CHAMPY AND H. M. CARLETON

TERxXT-1rIG. 5.

A, Skin of sucker of Lepadogaster guannii-—a ‘suck-fish’.
B, Supporting tissue of the same organ with intercellular
stroma of cartilage. Both a and B show intranuclear canaliculi
in all the cells. Probably mechanical in origin, e.g. mutual cell-
pressure. cs, Intercellular cartilaginous stroma; Ni, Nuclear
incision.

TEXT-FIG. 6.

Blastomere from blastula of Triton alpestris. Note deforma-
tion on one side by yolk-dises, yp, and on the other by astral rays.
c,Centrosome, Technique : Champy’s fluid and iron haematoxylin,

os 2m 4. £64 @6 7 68

SHAPE OF THE NUCLEUS 599

The intranuclear canaliculus of nerve-cells (see Text-figs. 7
and 8) is sometimes demonstrated by the Cajal method for the
Golgi apparatus, and has apparently been observed by Cajal
himself. With standard cytological stains—such as iron
haematoxylin—it appears as a single invagination of the
nuclear membrane. Its aperture is often opposite the point

-~I

TEXT-FIG.

Binucleated sympathetic ganglion cell from rabbit.
Intranuclear canal in one of nuclei.

TEXxtT-FIG. 8.

Pyramidal cell from cerebral cortex of guinea-pig.
AD, Apical dendrite; Nc, Intranuclear canal; ns, Nissl substance.

of insertion of the apical dendrite in the case of pyramidal
cells (see Text-fig. 8). In these elements the relation of the
canaliculus to the centrosome is obscure, largely owing to the
uncertainty of the existence of this structure in adult nerve-
cells.

Intranuclear canaliculi are also readily observed in the cells
lining the Wolffian ducts in Amphibia, while apparently similar
structures can sometimes be seen in the tissues of the higher
Vertebrates, though here, except in the case mentioned above,
the small size of the cells renders observation difficult,

600 CHRISTIAN CHAMPY AND H. M. CARLETON

7. Foups AND INcISIONS IN THE NucLEAR MEMBRANE.

Such modifications of the nucleus are common, though care
is required in their observation. This is easiest after fixation
in fluids which do not precipitate the nuclear contents in too
coarse a manner. Fixatives such as Gatenby’s Flemming
without acetic (6) and Champy’s carbol-formalin (4) give the
best results.

TEXT-FIG. 9.

Fig. a.—Diagram showing the relation of intranuclear incisions
to the nuclear membrane and nucleoli.

Figs. B and c showing that the shape of the nucleus is governed
rather by its distance from the cell membrane than by any mass
action of the cytoplasm. Were nuclear shape the product of
repulsion between the nuclear membrane and the mass of the
cytoplasm, the shape of the nucleus in an elongated cell would
be as shown in B. But this is never the case. In nature the
long axis of the nucleus is always in the long axis of the cell,
as indicated in c. The explanation of this seems to be that the
nuclear membrane is repelled by the cell membrane, and that
the nearer it is to the latter, the greater the degree of repulsion.

Gastric epithelium of Amphibia, e.g. of Bombinator or
Alytes, in which the cells are very large, shows clearly the
folds in the nuclear membrane. In longitudinal sections of the
nuclei there may be several of these structures, which may or
may not traverse its entire length. They are illustrated in
PI. 23, figs. 1, 7a, and 8. That we are dealing with folds and not
with canaliculi is made clear by transverse sections of such
nuclei, which are depicted in PI. 23, figs. 7 B and c, and 8.

SHAPE OF THE NUCLEUS 601

Folds in the nuclear membrane are found in a great variety
of cells in addition to gastric epithelium in Amphibia. They
occur in cardiac muscle in Man and Astacus (the crayfish),
and also in the connective-tissue cells of the Testis in the
latter species. In germinal epithelium they are especially
common, not only in that of the Axolotl (PI. 23, fig. 10)
but also in some mammalian tissues. But in the latter it is
usually difficult to make sure that the structures one can see

TEXT-FIG. 10.

Portions of muscle-fibres from nymph of Phryganaea sp.—
a caddis fly. In a the membranes of Krause can be seen running
across the sarcoplasm and constricting the nucleus at regular
intervals. In B only the nuclear constrictions are visible, the
section passing outside the zone of myofibrillae. ui, Hensen’s
line; MK, Membrane of Krause.

in germinal epithelium are truly intranuclear folds, though it
is interesting to note that undoubted incisions exist in the
pathological cysts—Cystadenomata—which are derived
from this epithelium.

The nuclei of smooth muscle-fibres, after impregnation by
the Cajal method for the Golgi-apparatus (Cajal, 1; Carleton,
2), show a peculiar spiral peri-nuclear band which has been
observed by Rio Hortega (11). After careful differentiation,
iron haematoxylin sections show that this structure is not
a thickened portion of the nuclear membrane but a. series of
usually rather irregularly arranged spiral folds. ‘Transverse

602 CHRISTIAN CHAMPY AND H. M. CARLETON

sections of such nuclei confirm the existence of these incisions,
which we have observed in non-striated muscle from the
intestine in Amphibia (see Pl. 24, fig. 2), in Mammals (muscle
layers of intestine of cat), and in certain invertebrate muscle-
fibres, e.g. heart of Helix as shown in PI. 24, fig. 6.
Finally, we have noted similar folds in the nuclear mem-
brane of developing oocytes (already described in Section 3), while
a peri-nuclear reticulum—possibly comparable to that found in
smooth muscle-cells—has been described by Retzius in the
spermatozoa of certain Gasteropods as shown in Text-fig. 3, B.

8. THE UNFOLDING OF INVAGINATIONS IN THE NUCLEAR
MEMBRANE.

Tt seems certain that nuclear folds and incisions expand
under certain conditions, thus altermg both volume and shape
of the nucleus. That such a phenomenon occurs during
differentiation of some cells is shown by the following example :

In Urodele Amphibia there exists a layer of lymphoid tissue
surrounding the liver. Study of the lymphocytes in this layer
(see Pl. 23, fig. 5) show that their nuclei, though round or oval,
bear a large number of narrow incisions. The latter can be
observed in various degrees of ‘ deployment’ in these cells,
and there is no doubt that polymorphonuclear leucocytes can
be formed in this manner from lymphocytes in some Amphibia
—a point in favour of the ‘ Unicist ’ theory of blood-formation.
The persistence of some of the nuclear folds gives rise to the
lobulation characteristic of the polymorphonuclear leucocyte.

Mutual cell-pressure may apparently in certain cases
inhibit expansion of the nuclear membrane. We have observed
an instance attributable to this in cells from the epidermal
and sub-epidermal tissues of the sucker of the fish Lepado-
gaster guannii. This is illustrated in Text-fig. 5, a and B.

9. INTRANUCLEAR RODLETS, ETC.’

Intranuclear rodlets and allied structures, which are only
found in highly specialized cells such as spermatids or certain
red blood corpuscles, are responsible for the shape of the

(on) om

SHAPE OF THE NUCLEUS 603

nucleus in such elements. The peculiar shape of the head of
the spermatozoon is doubtless an adaptation enabling it
rapidly to move in fluids and to penetrate the ovum. In
some instances, which have been described by Champy (4),
the changes in the shape of the nucleus durmg the stages
termed ‘Spermateleosis’ by Gatenby (7) are due te the
influence of a special intranuclear apparatus.

Trext-Fic, 11.

Fig. a.—Normal red blood corpuscle of bird with intranuclear
rodlet faintly indicated.

Figs. B and c.—Avian red cells after four days’ culture (pigeon’s red
cells in chicken plasma). The nuclei have become swollen and the
chromatin reduced in amount; consequently the intranuclear
rodlet is clearly visible.

The latter is best studied in Amphibia such as Bom-
binator, the Salamander, and the Axolotl. In these it
can be seen within the spermatid as a thin and usually
refringent rod, lying in the long axis of the nucleus. It appears
to be developed from the centrosomes, originating from either
the posterior or the anterior of these structures. Or sometimes
it may be developed from both simultaneously. When the
intranuclear rodlet does not extend the whole length of the
nucleus, its free extremity, which may be bifid, is sometimes
in relation with the nucleolus. All this is indicated in Pl. 24,
fig. 1, which depicts spermatid nuclei from Bombinator.
That this structure is not a fold in the nuclear membrane
is seen in the figures of transverse sections of these nuclei.
But it often co-exists with intranuclear canaliculi, from
which, however, it can be further distinguished by its greater
refringency.

It is well known that the red blood corpuscles of birds have

604 CHRISTIAN CHAMPY AND H.M. CARLETON

oval nuclei of a very definite aspect. These nuclei are remark-
ably stable, for they often retain their shape after the rest of
the corpuscle has been haemolysed. Now, observation of the
normal Avian red cell reveals little beyond a rather dark
central portion and, often, a small invagination at both poles
of the nuclear membrane. The general appearance is such
(see Text-fig. 11) as to suggest the presence of some supporting
structure , within the nucleus, though the density of the
chromatin makes its observation difficult. But when a bird’s
red blood corpuscles are aseptically cultivated in their own
plasma, the nucleus slowly swells up before actual death of the
cell occurs. As the nucleus becomes spherical, the chromatin
becomes condensed into a single nucleolus, and an axial
rodlet can frequently be seen under such conditions.

10. Tue RELATION BETWEEN NUCLEOLI AND NUCLEAR
SHAPE.

The nucleolus remains one of the most enigmatical of the
cell components, in spite of the attention devoted to it by
many biologists, and by Montgomery (9) and Vigier (12) in
particular. The nucleolus is of interest in that it often shows
amoeboid movements and undergoes independent fission
during the life of the cell. In these observations the term
‘nucleolus’ is used in its widest sense, as signifying any
condensation of nuclear material within the nucleus. Con-
sequently, the word as employed in this paper applies to both
karyosomes (or chromatin nucleoli) and plasmosomes (i.e. con-
densations of the oxyphil substance called plastin). Not only
do both chromatin and plastin often occur within the same
nucleolus, but karyosomes or plasmosomes sometimes contain
one or more granules of unknown composition, which have
been shown by the aid of special impregnation methods to
divide by fission during mitosis (Carleton, 8).

Clearly the nucleolus is often a complex structure of doubtful
significance, and it is impossible at present to dogmatize on
the relation of this element to nuclear shape. At the most,
certain deceptive appearances may be cleared up.

SHAPE OF THE NUCLEUS 605

Nuclear polymorphism is often—though by no means
always—associated with the presence of multiple nucleoli as
shown in Pl. 23, figs. 3 and 5. In elongated nuclei the nucleoli
usually lie parallel to the main axis of the nucleus as depicted
in Pl. 24, figs. 2 and 6, and in Text-fig. 10.

But it is in developing oocytes that the relation between
nucleoli and nucleus is particularly deceptive. In the earlier
stages of development the nucleoli come to lie at the periphery
of the nucleus, and when invaginations subsequently appear
in the nuclear membrane, they do so opposite the nucleoli.
Pl. 23, fig. 2, and Text-fig. 4 illustrate this, and they suggest
the possibility of nuclear incisions being formed under the
influence of the nucleoli. On the other hand, it must be
observed that in the case of some nuclei, the indentations in
which are obviously due to certain of the mechanical causes
already considered, the nucleoli are yet often in relation to
the blind ends of the pouches in the nuclear membrane. In
muscle, too, infolded portions of the latter often come imto
contact with the nucleoli, though here again nuclear incisions
are primarily mechanical in origin. And finally, there are
cells the nuclei of which contain nucleoli and yet have a nuclear
membrane of regular contour, as shown in PI. 23, fig. 11.

The main outcome of all this is that the relations so often
seen between nucleoli and nuclear invaginations are usually
secondary, and that the position of the nucleoli in such instances
is rather an effect than a cause.

11. Cent Division AND NucLEAR DIFFERENTIATION,

It is not without significance that mitoses are extremely
rare—if not altogether absent—in cells the nuclei of which
contain well-developed canaliculi or incisions. Such, at any
rate, is the case with the following tissues in adult mammals :

Non-striated muscle.
The various segments of the urinary tubule in the kidney.
The epithelium lining the vesicles of the thyroid gland.
Nerve-cells.
Our observations suggest that while highly developed nuclear
NO, 260 Ss

606 CHRISTIAN CHAMPY AND H. M. CARLETON

canaliculi or incisions seem to be incompatible with mitosis,
direct division may occur in cells—other than those enumerated
above—which contain such structures. Thus, amitosis has
been observed in nuclei of the cells of the Wolffian ducts and
germinal epithelium and Sertoli cells; also possibly in the
gastric mucosa of some animals,

The behaviour of smooth muscle when cultured in plasma
confirms this idea. It has been shown (Champy, 5) that the
nuclei of this tissue, when removed from the inhibitory influ-
ences of the organism, multiply actively. At first they do so
amitotically, and only when the typical structure of these
nuclei has disappeared by a progressive ‘ de-differentiation ’
do they multiply by mitosis. Cultures of ovarian germinal
epithelium behave in a similar manner. Again, the fundus
glands of the human uterine mucosa have nuclei without
incisions, while the cervical glands possess them. The former
divide mitotically, the latter amitotically. And further, even
in Adenomata (i.e. benign tumours) derived from the cervical
glands does direct division persist. Only when such growths
become carcinomatous do mitoses appear.

We would here point out that incisions or lobulations of
nuclei have only too often been mistaken as evidence of direct
division. In our experience such appearances are only of value
when an actual increase of the number of nuclei can be
established.

In conclusion, then, there is evidence that well-developed
intranuclear canaliculi and incisions are incompatible with
mitosis, a fact which possibly explains the tendency towards
direct division in certain cells with specialized nuclei.

12. SumMaRY.

Variations in the shape of the nucleus have been described
in different animal cells. In addition, the following factors
have been shown to be responsible for nuclear shape :

(1) Surface tension: when this is equal over the surface
of the nuclear membrane, the nucleus tends towards the spheri-
cal condition. When surface tension varies over the interface

SHAPE OF THE NUCLEUS 607

between nucleus and cytoplasm, nuclear polymorphism may
result.

(2) Mechanical deformation of the nucleus is com-
mon and may be due to various causes, chief amongst which
are: (a) Pressure from cytoplasmic inclusions,
e.g. fat, lecithin, and yolk; (b) Tonofibrillae; (c) in
striated muscle, the influence of the Membranes of
Krause which constrict the nucleus along its length—and
limit its ends—by their prolongation from the myofibrillae
into the sarcoplasm.

(3) The centrosome, which has been shown (in the
resting cell) often to repel that part of the nuclear membrane
which is nearest to it.

(4) The relation between cell shape and nuclear
shape has been briefly discussed. It has been noted that
the nucleus never comes into contact with the
cell membrane, except in the rarest instances due to the
intervention of mechanical factors. Evidence has been brought
forward in favour of our view that there is a mutual
repulsion between cell membrane and nuclear
membrane.

(5) Canaliculi and incisions in the nuclear membrane
have been described in various cells.

(6) The unfolding of such incisions during development and
differentiation of some such cells has been described.

(7) Intranuclear rodlets and their importance in
the maintenance or the modifying of nuclear shape have been
discussed. .

(8) Mitotic division and a certain degree of nuclear
differentiation have been shown often to be incompatible—
thereby accounting for amitosis in certain highly specialized
nuclei.

(9) The need for care in distinguishing between nuclear
incisions and genuine amitotic division of nuclei has been
emphasized.

June 1921.

608 CHRISTIAN CHAMPY AND H. M. CARLETON

13. BrBLioGRAPHY.

1. Cajal.— Algunas variaciones fisiolgicas y patolégicas del aparato
reticolar de Golgi’’, ‘ Trab. del Lab. Invest. Biol. Madrid’, vol. 10,
1912.

2. Carleton.— Note on Cajal’s Formalin-Silver nitrate method for the
Golgi-apparatus ”’, ‘ Journ. Roy. Micr. Soc.’, 1919.

3. —— “ Observations on an intranucleolar body in columnar epithelium
cells of the Intestine ’’, ‘ Quart. Journ. Micr. Sci.’, vol. 64, 1920.

4, Champy.—“ Spermatogenése des Batraciens”, ‘Arch. de Zool.
Exp.’, vol. 52, 1913.

5. —— “ Résultats de la méthode de culture des tissus en dehors de
Yorganisme ’’, ‘ Presse médicale’, no. 9, 1914.

6. Gatenby.—‘** Cytoplasmic Inclusions of the Germ Cells ’’, Pt. I, « Quart.
Journ. Micr. Sci.’, vol. 62, 1916-17.

7. —— “Cytoplasmic Inclusions of the Germ Cells ”, Pt. III, ibid.,
vol. 63, 1918-19.

8. Heidenhain.—‘‘ Neue Untersuchungen iiber die Centralkérper und
ihre Beziehungen zum Kern v. Zellenprotoplasma”’, ‘ Arch. f. Mikr.
Anat.’, vol. 43, 1894.

9. Montgomery (Jr.).—‘“ Comparative Cytological Studies with special
regard to the Nucleolus ’’, ‘ Journ. of Morph.’, vol. 15, 1899.

10. Prenant, Bouin et Maillard.—‘ Traité d’Histologie ’, t. 1.

11. Rio Hortega.—“ Investigations sur le tissu musculaire lisse ’’, ‘ Trab.
del Lab. Invest. Biol. Madrid ’, vol. 11. 1913.

12. Vigier.—‘‘ Les Pyrénosomes dans les cellules de la glande digestive
de l’Ecrevisse”’, ‘Comptes rendus de l’Assoc. des Anatomistes ’,
nos. 3-4, 1901.

14. EXPLANATION OF PLATES 28 AND 24.

Illustrating Champy and Carleton’s paper on ‘ Observations on the
Shape of the Nucleus and its Determination ’.

High-power figures drawn at various magnifications.

Arrows point towards the distal (i.e. unattached) ends of the cells.

LETTERING,

BC., bile canaliculus; BM., basement membrane; C., centrosome ;
C POST., posterior centrosome; H CAN., Holmgren canaliculi; LG.,
lipoid granules; MJZ., mitochondria; MK., membrane of Krause ;
NC., nuclear canal; NJ., nuclear incision; NR&., nuclear rodlet ;
NU., nucleolus; PNF., peri-nuclear fold; SB., striated border;
7'N., tonofibrillae; 'S., transverse section: X., invagination of
nuclear membrane.

SHAPE OF THE NUCLEUS 609

PLATE 23.

Fig. 1.—Showing nuclear incisions in a connective-tissue cell from the
Testis of Astacus.

Fig. 2.—Oocytes of the fish Silurus sp., showing how the nucleus
becomes polymorphic at a later stage of development.

Fig. 3.—Spermatogonium of Bombinator igneus, illustrating
that the relation of nucleoli to nuclear folds is not constant. Here the
nucleus has many incisions and yet the nucleoli bear but little relation
to them.

Fig. 4.—Tubule cell from a nephridium of Aulostomum—a leech.
Folds in nuclear membrane orientated in relation to flattening out of
nucleus.

Fig. 5.—Two leucocytes from the lymphoid layer of the liver of the
Axolotl. In a the nucleus is oval and its membrane highly pleated.
B shows a polymorphonuclear white cell derived from a by the partial
unfolding of the nuclear incisions. Technique: carbol-formalin and
ferric Brazilin.

Fig. 6.—Cell from the hepato-pancreas of Oniscus (an Isopod Crusta-
cean) showing deformation of the nucleus by large lipoidal granules in
the cytoplasm. Technique: Benda fixation and iron haematoxylin.

Fig. 7—Human cardiac muscle cells. A is a longitudinal section
showing (i) the pleating of the nuclear membrane, each incision correspond-
ing to a membrane of Krause, and (ii) the square ends of the nucleus.
B illustrates the arrangement of the nuclear incisions in transverse section
at a higher magnification. At the blind end of each incision there is usually
a nucleolus. c is a longitudinal and somewhat oblique section of the
nucleus, showing the relation of its shape to the fibrils. Technique :
carbol-formalin and iron haematoxylin.

Fig. 8.—Cells from gastric epithelium of the Axolotl, The nuclear
membrane shows deep longitudinal incisions. 7’S.=a transverse section
of the nucleus, the relation of the nuclear incisions to the nuclear membrane
being clearly shown.

Fig. 9.—Nucleus of cardiac muscle of Astacus, showing relation
between nuclear incisions and nucleoli.

Fig. 10.—Longitudinal nuclear folds in germinal epithelium cell of
Axolotl,

Fig. 11.—Spermatocyte of Lithobius forficatus—a Myriapod.
An example of a nuclear membrane of regular contour in spite of
multiple nucleoli

PLATE 24.

Fig. 1.—Spermatid nuclei of Bombinator—a toad. Showing the
fully formed axial rodlet in a. 8B and © are different stages in its

610 CHRISTIAN CHAMPY AND H. M. CARLETON

formation. pD, E, F, and G show its appearance in transverse section.
Technique: Bouin and iron haematoxylin.

Fig. 2.—Nuclei of smooth muscle from the intestine of the Axolotl.
A and B are longitudinal sections of nuclei, while c is transverse. All
show the spiral circular incisions in the nuclear membrane. Technique :
carbol-formalin and iron haematoxylin.

Fig. 3.—Cell from intestine of Oniscus. Note deformation of nucleus
by * Tonofibrillae ’.

Fig. 4.—Cell from Wolffian duct of Salamander showing the intra-
nuclear canaliculus and centrosomes opposite its aperture.

Fig. 5.—Intestinal epithelial cell from Libellula sp.—a dragon fly.
Note that here the length of nucleus is not proportional to that of the cell.

Fig. 6.—Nuclei in longitudinal and transverse section from heart of
Helix pomatia (snail). Incisions in nuclear membrane. Technique :
Flemming and iron haematoxylin.

Fig. 7.—Hepatic cells of Salamander. At x the nucleolus is in
contact with the nuclear membrane, which is slightly invaginated at this
point. Technique: Bouin and iron haematoxylin.

Champy & Carleton

Quart. Journ. Micr Sci. Vol. 65, NS., PL. 23.

Champy & Carleton

On the calcium carbonate and the calcospherites
in the Malpighian tubes and the fat body of
Dipterous larvae and the ecdysial elimination
of these products of excretion.

By
D. Keilin, Se.D.,
Beit Memorial Research Fellow.

(From the Quick Laboratory, University of Cambridge.)

With 5 Text-figures.

CoNTENT3.

PAGE
1. THE PRESENCE OF CALCIUM CARBONATE IN THE MALPIGHIAN
TUBES : s ; : : On
2. CALCOSPHERITES IN THE ore Bone ‘ ; : 5 Sale
3. CALCOSPHERITES IN THE MALPIGHIAN TUBES . ; 617
4, Ecopystat ELIMINATION OF CALCIUM CARBONATE DURING Meee
MORPHOSIS : : : ‘ : Ole,
5. HypoTHESES AS TO THE Gores AND FUNCTION oF CaLcIuM
CARBONATE IN THE LARVAL Bopy . : : : » 621
6. CONCLUSIONS ; ; ; : i ; : : ; 623
7. REFERENCES : : ! : : f 5 ; . 624

1. THe PRESENCE OF CALCIUM CARBONATE IN THE
MALPIGHIAN T'UBES.

Lyonst (11, 1832) was the first to notice in the larva of
Ptychoptera two milky-white vessels running throughout
the length of the body. Similar vessels have been discovered
in the larva of Eristalis by Batelli (1, 1879), who has
rightly described them as saccate dilatations of the anterior
pair of Malpighian tubes filled with calcium carbonate. Quite

612 D. KEILIN

independently Valery Mayet (18, 1896) has shown that in
Cerambyx larvae, of the six Malpighian tubes, four are
larger and are filled with calcium carbonate. The excretion
of this product, which was described by Valery Mayet as a new
function of the Malpighian tubes, was denied by Kinckel
d’Herculais (10, 1896), who at a meeting of the Entomological
Society of Paris made an observation that Valery Mayet
probably misunderstood the anatomy of the larva, and that
the organs containing calcium carbonate were not Malpighian
tubes but the intestinal caeca. Later, Valery Mayet (14, 1896)
succeeded in demonstrating that the tubes in question were
actually the Malpighian tubes; Kiunckel d’Herculais then
suggested that the calcium carbonate of Cerambyx larvae
is probably formed in other special glandular cells, and that
the Malpighian tubes were eliminating only the excess of this
product. P. Marchall (12, 1896), who took part in this discus-
sion, observed that the excretion of CaCO, by the Malpighian
tubes has nothing surprising in it; he thought, however, that
the excretory function in insects is not localized in one particular
organ: uric acid, for instance, can be found not only in the
Malpighian tubes but in the intestine and the fat body.
Calcium carbonate has been found also by Vaney (19, 1900 ;
20, 1902) in the anterior pair of the Malpighian tubes of the
Stratiomys larva, and by Pantel (16, 1898) in the parasitic
larvae of Tachinidae and in the larvae of Ptychoptera
(17, 1914). In the latter, two of the five Malpighian tubes
are transformed into large saes filled with calcium carbonate.
I myself have found the excreted calcium carbonate in the
Malpighian tubes of many Dipterous larvae: Hristalis
tenax, L., Myiatropa florea, L., Mallota erista-
loides, Lw., Merodon equestris, F., Syritta
pipiens, L, Eumerus strigatus, Fin, Ptycho-
ptera contaminata, L., several species of Stratiomyidae
belonging to the genera Stratiomys, Sargus, and
Odontomyia, and among the Trypetidae in Anastrepha
striata, Schiner. In all of these larvae the carbonate-
containing Malpighian tubes differ from the rest by being

CALCIUM CARBONATE IN DIPTEROUS LARVAE

TEXtT-FIG. 1.

i,

»

\
>

scveein putreceeasaceOCyy, 2
amg os

pair of Malpighian tubes ;

Myiatropa florea, dissection of a full-grown larva. a.m., anterior

a.8., anterior spiracles; c.p., cal-
careous or terminal portion of the Malpighian tubes; h.g., hind-
gut; v., mid-gut; %.c.,

intestinal caeca
System ; 0., oesophagus

n., central nervous
3 p., pharynx; p.m., posterior pair
of Malpighian tubes ; s., salivary glands ;

‘r., tracheal trunks.

6138

614 D. KEILIN

more developed and of a milky colour. In the larva of
Ptychoptera contaminata and of a few Eristalids,
these tubes, at least in their terminal portions, are excep-
tionally well developed and can be easily seen by transparency
with the naked eye. Text-fig. 1, which represents a complete
dissection of the larva of Myiatropa florea, L., shows
to what extent the calcareous portion of the Malpighian tube
can be developed in a full-grown larva. In this example the
posterior pair of Malpighian tubes (p.m.) is composed of two
short branches of normal structure ; the anterior pair (a.m.),
on the contrary, is very long, its two branches in their proximal
portion are of normal structure and diameter and extend to
the anterior portion of the body, where they suddenly pass
into two enormous sacs (¢.p.) with milky contents, which run
backwards and reach posteriorly the anal segment. These two
sacs are even thicker than the intestine of the larva; they are
very fragile, and the slightest puncture causes their milky
contents to flood out. The milky fluid is composed of a thick
suspension of very small calcareous granules which are almost
completely soluble in dilute acid, only a small central particle,
probably of an organic nature, remaining.

2. CALCOSPHERITES IN THE Fat Bopy.

In all of the above-mentioned larvae the calcium carbonate
of the Malpighian tubes appears in the form of crowded small
sranules suspended in the fluid which fills the lumen of these
tubes. There are, however, other larvae which contain the
calcium carbonate in form of calcospherites. The
latter are enclosed either in the anterior pair of the Malpighian
tubes or in special cells connected with the fat body.

The term caleospherite we owe to Harting (7, 1873), who
was the first to prepare, artificially, calcareous corpuscles:
composed of two substances, mineral and organic.
He obtained these bodies by precipitating calcium carbonate
(CaCle+ K2CO3 = CaCo3 + 2KCI) in a liquid containing organic
matter (albumen, for instance). The calcareous corpuscles
thus obtained were elongated or spherical, highly refractive,

a ee ee

CALCIUM CARBONATE IN DIPTEROUS LARVAE 615

composed of numerous concentric layers surrounding a
central or excentric granulated body and bearing some resem-
blance to starch grains. When the ealcospherites are dissolved
in dilute acetic acid there remains an albuminoid stroma
consisting of calcoglobulin. Examined in polarized light, the
ealcospherites show a black cross. The calcospherites, or
Harting’s corpuscles, have been well described by Nathusius
(15, 1890), who found them in numerous animals and plants,
and by Pettit (18, 1897) in cases of pathological ossification in
mammals.t

In insects the calcospherites were discovered simultaneously
by Henneguy (8, 1897) and Giard (unpublished observations
quoted by Henneguy). Henneguy found them in the larvae
of Phytomyza chrysanthemi, Kowarz. According to
this author each calcospherite of this larva is enclosed in
a special hypertrophied cell of the fat body. The fat of these
cells disappears completely, and all that remains of the cell is
reduced to a thin protoplasmic layer and a small degenerated
nucleus. The calcospherites still appear in the pupa, but they
are absent in the adult flies, and Henneguy thought that the
imagines which he examined were probably obtained from the
‘normal’ larvae, i.e. ‘larvae devoid of calcospherites ’.
Giard has observed similar calcospherites in the larvae of
Phytomyza lateralis, Fall., which attacks the inflores-
cence of Matricaria inodora.

Personally I have found the calcospherites in the fat body
of many Phytomyzine and Agromyzine larvae (Text-fig. 2).
In all the species where the calcospherites are present they are
to be found in every individual larva throughout its life.
In this my observations differ from those of Henneguy and
Giard, who considered the presence of calcospherites as
abnormal and probably only seasonal. The cells which contain
the calcospherites are always connected with the fat body,
although they never contain droplets of fat. As a rule they
lie in alveolar spaces formed among the fat cells (Text-fig. 3).

1 To these two papers the reader is referred for numerous observations
and references concerning this subject.

616 D. KEILIN
The calcospherites are already present in very young larvae,
but not in the embryo or in those just hatched ; they seem to

TEXxtT-FIG. 2.

Agromyza sp., full-grown larva, slightly compressed, showing
by transparency 120 calcospherites disseminated throughout the

body.

TEXT-FIG. 3.

OG00 O°
cee .
Mess

° N M
6 oC

7 $8 860 0
Pp Lo",

oR

Le

38
’

02mm.
Agromyza larva, a portion of the fat body, /., showing the
calcospherites, c.

appear only after a short period of feeding. The existence of
calcospherites in larvae belonging to the families Phytomy-
zinae and Agromyzinae seems to be so general that this character

CALCIUM CARBONATE IN DIPTEROUS LARVAE 617

assumes a taxonomic importance and helps one to recognize
these larvae and to differentiate them from the phytophagous
larvae belonging to other families like Anthomyidae and
Trypetidae, the fat body of which is devoid of calcospherite
cells.

3. CALCOSPHERITES IN THE MALPIGHIAN TUBES.

The only case of the existence of calcospherites in the
Malpighian tubes is that of the larva of Acidia heraclei,
the celery-fly larva. On examining a living larva of Acidia
gently compressed between the slide and coverglass, I have
noticed that its body contains a number of large calcospherites
similar to those of Agromyzine larva. J thought at first
that the calecospherites of Acidia larvae were also formed
in special cells connected with the fat body. The dissection
of these larvae revealed that such was not the case; all the
calecospherites were lying free in the lumen of the Malpighian
tubes and especially in their terminal portions (Text-fig. 4,
I and II, ¢.). The calcospherites of various sizes, from 8 pu
to 140, in diameter, distend these tubes, which have the
appearance of being composed of highly refractive beads.
The calcospherites when small are very often double, i.e. with
two or more central granules (Text-fig. 5, b, c, and d). The
occurrence of the calcospherites in the Malpighian tubes
_(Acidia heraclei) and in the fat body (Agromyzinae)
of the phytophagous Dipterous larvae demonstrates once
more the similarity in the excretory function of these two larval
organs.

4, EcpysiaL ELIMINATION OF CALCIUM CARBONATE
DURING METAMORPHOSIS.

All the foregoing shows that the larvae of a great number
of Diptera contain in their Malpighian tubes, or in the cells
connected with the fat body, a large quantity of calcium
carbonate stored in the form of minute granules or large
calcospherites.

618 D. KEILIN

A question now arises: What becomes of the stored calcium
carbonate during the ultimate stages of the life of the insect ?

TEXT-FIG. 4.

Acidia heraclei. I. a.m., anterior pair of the Malpighian
tubes; c., terminal portion filled with the calcospherites ;
g.. gut; p.m., posterior pair of Malpighian tubes. II. portion a
of the anterior pair of Malpighian tubes showing the calcospherites
free in the lumen of the tube.

According to Pantel (17, 1914) the calcium carbonate of
the Ptychoptera larva disappears before the metamor-
phosis takes place. He considered that it does not dissolve

CALCIUM CARBONATE IN DIPTEROUS LARVAE 619

in the body of the larva, but passes from the Malpighian tubes
into the hind-gut, whence it is expelled from the body just
before the larva begins to pupate. He admits, however, that
he never actually saw the process of expulsion of this product

TEXT-FIG. <6

Caleospherites of Acidia heraclei, ato e. a, a caleospherite
examined by polarized light, showing the black cross; 6 and ¢,
double calcospherites ; d, very small simple, double or multiple
calcospherites ; ¢, calcospherite in diluted acetic acid showing
collapsing stroma ; /, intracellular caleospherite of Agromyza
larva.

of excretion. In several cases he found the calcareous sub-
stance retained in the pupae of Ptychoptera.

Henneguy (8, 1897) found that the calcospherites of Agro-
myza larvae persist in the pupae, but he did not find them
in the adult flies. He considered that the existence of calco-
spherites was not general, and was very probably abnormal,

620 D. KEILIN

assuming that the adult flies which he examined were derived
from normal larvae devoid of calcospherites. Personally,
I have found that in all Diptera, the larvae of which contain
stored calcium carbonate, this substance disappears during
the pupal phase, and the adult flies are completely devoid
of this product of excretion. I must say, however, that the
disappearance of the calcium carbonate seems to be a more
complicated process than that suggested by Pantel (17, 1914).
In the case of the Ptychoptera larva, where calcium car-
bonate, in form of a thick suspension of small granules, is
enclosed in the distended portion of the Malpighian tubes,
it is possible that the milky contents of these tubes are emptied
into the hind-gut and are thus expelled from the body. It is,
however, difficult or even impossible to suppose that the large
calcospherites can follow the same channel in the larva
of Acidia heraclei, for the Malpighian tubes of this
larva, as in other Dipterous larvae, are completely devoid of
peristaltic movement.

In the case of the Acidia larva and the larvae of Agromy-
zinae, which attack Cirsium lanceolatum, I was able
to follow, step by step, the disappearance of the calcospherites
during the metamorphosis of these insects. Each of these
larvae, as is the case in all the Cyclorhaphous Diptera, trans-
forms into a pupa which remains enclosed in the puparium
formed by the contracted and hardened last larval cuticle.
During the first day of the metamorphosis the calcospherites
of these larvae can be easily seen either by transparency or
by dissection. When the pupa is completely formed and
separated from the last larval cuticle or puparium it loses its
calcospherites, which are gradually dissolved. At the same
time the puparium becomes very brittle and presents a white
opaque ‘ fossilized’ appearance. After the emergence of the
adults the empty puparia become so fragile that it is difficult
to detach them from the plant, for they pulverize under the
shghtest pressure. On treating such an empty puparium with
dilute hydrochloric acid a very active effervescence takes
place, with the evolution of carbon dioxide, and all that

i. Agios," ty a a tinal ee oe ee ee ee ee ee ee

CALCIUM CARBONATE IN DIPTEROUS LARVAE 621

remains is reduced to a thin transparent larval cuticle. All
this shows that the puparia of Acidia heraeclei and
of Agromyza are composed of a thin larval cuticle, the
internal surface of which is lined and strengthened with a layer
of calcium carbonate. As to the Dipterous larvae, like M yla-
tropa florea, EHristalis tenax, hL., Syritta pipiens,
L., and others, the Malpighian tubes of which contain caleium
carbonate in form of a granular suspension, I could not follow
with the necessary precision the course of their calcareous
excretion. It is possible that a certain portion is mechanically
expelled from the body before the metamorphosis takes
place. On the other hand, it is certain that a good part of the
stored calcium carbonate remains in the pupa and undergoes
a similar process of dissolution which we have seen to occur
in the Acidia and Agromyza larva. In fact, the empty
puparia of these flies are also internally lined with calcium
carbonate and effervesce when immersed in dilute acid. It
is evident then that in Dipterous larvae the calcium carbonate
(in the form of small granules or calcospherites stored in the
Malpighian tubes or in the cells connected with the fat body)
remains wholly or partly within the body of the larva until
the latter pupates. During the first day of the metamorphosis,
when the last ecdysis takes place, this product of excretion
(CaCO,) dissolves gradually in the perivisceral fluid of the
insect. It then passes through the newly formed cuticle of
the pupa into the ecdysial fluid which fills the space between
the pupal and the last larval cuticle. Finally, when the eedysial
fluid is absorbed, the calcium carbonate remains as a deposit
upon the internal surface of the puparium.

This mode of elimination of an excretory product from the
body of an insect, being connected with the process of moulting,
may well be named ecdysial elimination.

5. HypotrHEsEs AS TO THE ORIGIN AND IfUNCTION OF
Cancium CARBONATE IN THE LARVAL Bopy.
According to Valery Mayet (18, 1896) the calcium carbonate
stored in the Malpighian tubes of Cerambyx larvae forms
NO, 260 pT ft

622 D. KEILIN

a real reserve substance which has an important function
during the metamorphosis. Just before the last moult the
larvae, which form galleries in the wood of pine-trees, disgorge
the calcium carbonate, with which they cover the walls of
the galleries, thus protecting the pupae from the sap of the tree
and preventing the invasion of the galleries by fungi. The
opercula which close the galleries are also formed from calcium
carbonate of the same origin.

Vaney (19, 1900; 20, 1902) also speaks of calcium carbonate
stored in the Malpighian tubes of Stratiomys larvae as a reserve
substance.

According to Henneguy (8, 1897) the calcospherites of A gro-
my za larvae, which he wrongly supposed to be only seasonal,
are probably attributable to the special conditions of feeding
of these larvae during the autumn.

Pantel (17, 1914) considers the calcium carbonate as an
ordinary product of excretion, which is probably due to an
excess of calcareous substances present in the food of the lar-
vae. The formation of calcium carbonate in the larva of
Hristalis, Ptychoptera, and Stratiomys, which
live in putrefying organic substances, and in the parasitic
Dipterous larvae, reminds one somewhat of the calcareous
excretion observed in several other organisms. Calcospherites
are known, for instance, to exist in the parenchyma of Cestodes
and in the excretory tubules of Trematodes, and, according to
Burian (4, 1912, pp. 401-5), these calcospherites are derived
from the neutralization of carbon dioxide. He explains thus
how the parasitic worms, which live in a medium which already
has a high CO, content, get rid of the CO, derived from thew
respiration.

According to Combault (5, 1909), the erystals of calcium
carbonate, which fill the calciferous or Morren glands of
Oligochaetes, are also products of the neutralization of the
carbon dioxide which passes from the blood circulating in the
lamellae of these organs.

The neutralization of the CO, of respiration and the forma-
tion of calcium carbonate was also shown by Bohn (2, 1898) -

CALCIUM CARBONATE IN DIPTEROUS LARVAE 625

to exist in several Crustacea, e.g. Gonoplax rhomboides
and others. According to this author, several Crustacea,
Collapa or Ebalia, for instance, which live upon the
Red algae, in a medium rich in ammoniacal alkali, do not
eliminate the CO, derived from their respiration; they retain
it to neutralize the ammonia which reaches their blood and
tissues.

We know, on the other hand, that in insects the respiratory
function differs markedly from that of other groups of animals.
In insects oxygen is supplied to the tissues by means of
a highly developed ramified system of tubules—the tracheae,
while the carbon dioxide is given up by the same tissues to
the perivisceral fluid and thence eliminated through the
whole surface of the body.

It is possible that a part of this CO, is neutralized in the
blood or perivisceral fluid; but at present this is purely
hypothetical and needs verification by proper experimental
inquiry. It indicates, however, that it would be of great
interest to determine correctly the respiratory quotient of an
insect larva which, like Eristalis, Ptychoptera, and
others living in putrefying media, contains within its body
a large quantity of calcium carbonate.

6. CONCLUSIONS.

1. The larvae of a great number of Diptera, parasitic, phyto-
phagous, or living in putrefying substances, contain in their
bodies a large quantity of stored calcium carbonate.

2. The latter is present in the Malpighian tubes or in special
cells connected with the fat body.

3. Calcium carbonate is stored either in form of a thick sus-
pension of small granules (in the Malpighian tubes) or in the
form of calcospherites (in the Malpighian tubes or the fat
body).

4. Calcium carbonate remains wholly or partly in the body
of the larva when the latter passes into the pupal stage, but
disappears by the time the adult stage is reached.

5. During the first days of metamorphosis the caleium

Tt2

624 D. KEILIN

carbonate dissolves in the perivisceral fluid (haemolymph or
blood of imsects) and then passes through the newly formed
pupal cuticle into the ecedysial fluid. When the latter is
absorbed, the calcium carbonate remains as a deposit upon
the internal surface of the puparium.

6. This mode of elimination of calcium carbonate from the
body of an insect may be termed ecdysial elimination.

7. The excretion of calcium carbonate in Dipterous larvae
is comparable with calcareous excretion as observed in other
organisms like Cestodes, Trematodes, Oligochaetes, and
Crustacea, where this product of excretion is supposed to
be derived from the neutralization of the carbon dioxide
of respiration. This explanation has not yet been proved
experimentally.

7. REFERENCES.

1. Batelli, A. (1879)—‘‘ Contribuzione all’ anatomia ed alla fisiologia
della larva dell’ Eristalis tenax”, ‘Bull. Soc. Ent. Ital.’,
xi. 77-120, Pls. I-V.

2. Bohn, G. (1898).—-“‘ De l’absorption de anhydride carbonique par
les Crustacés décapodes ”’, ‘C. R. de la Soc. Biol. France ’, série 10,
vol. v, pp. 1008-10.

3. —— (1898).—* Variations des échanges gazeux chez les Crustacés
décapodes suivant la saison, Vhabitat, la taille des animaux ’”’,
ibid., pp. 1011-13.

4. Burian, R. (1912).—** Die Exkretion ’’, in ‘ Handbuch der Vergleichen-
den Physiologie’, Jena. See pp. 401-5.

5. Combault, A. (1909).—*‘ Contribution a l’étude de la respiration et
de la circulation des Lombriciens’’, ‘ Journ. de l’anat. et de la
physiol.’, xlv, pp. 358-99 and 446-9, 1 PI.

6. Giard, A. (1897).—MS. note; see Henneguy (1897).

7. Harting, P. (1873).—‘* Recherches de morphologie synthétique sur
la production artificielle de quelques formations calcaires
organiques ”’, ‘ Verhand]l. d. kon, Akad. van Wetensch, Amster-
dam’, xiv, 84 pp., 4 Pls.

8. Henneguy, L. F. (1897).—‘‘ Note sur l’existence de calcosphérites
dans le corps graisseux de larves de Diptéres”’, ‘ Arch, d’Anat.
Micr.’, i. 125-8.

9. Kiinckel d’Herculais, J, (1896).— Discussion in ‘ Bull. de la Soc. Entom.
de France’, pp. 126-7.

10. —— (1896).—‘ Sur les fonctions des tubes de Maipighi. Réponse
a M. Valery-Mayet ”’, ibid., p. 206.

ad.

12.

13

14.
15.

16.

A7.

18.

19.

20.

CALCIUM CARBONATE IN DIPTEROUS LARVAE 625

Lyonet, P. de (1832).—“* Recherches sur l’anatomie et les métamor-
phoses de différentes espéces d’Insectes”’, ‘Ouvrage posthume
publié par M. W. de Haan’, Paris, J. Balliére, 580 pp., 54 Planches.

Marchall, P. (1896).—‘‘ Remarque sur la fonction et Vorigine des
tubes de Malpighi ”’, ‘ Bull. Soc. Ent. de France’, p. 257.

Mayet, V. (1896).—‘‘ Une nouvelle fonction des tubes de Malpighi ”’,
ibid., pp. 122-6.

—— (1896).—* Encore sur les tubes de Malpighi”’, ibid., pp. 207-8.

Nathusius, W. (1890).—‘‘ Untersuchungen iiber Hartingische Kor-
perchen ”’, ‘ Zeitschr. f. wiss. Zool.’, xlix. 602-48, Pl. XXVIII.

Pantel, J. (1898)—‘“Le Thrixion Halidayanum, Rond.,
essai monographique sur les caractéres extérieurs, la biologie et
Panatomie d’une larve parasite du groupe des Tachinaires”’, ‘La
Cellule ’, xv.

(1914).—* Signification des ‘glandes annexes’ intestinales
des larves des ‘ Ptychopteridae’ et observations sur les tubes
de Malpighi de ces Nématocéres (larves et adultes) ”, “ La Cellule ’,
xxix, 393-429, 1 Pl.

Pettit, A. (1897).—“‘ Sur le réle des calcosphérites dans la calcification
& état pathologique”’, ‘Arch. d’Anat. Micr.’, i, pp. 107-24,
FL. VIL.

Vaney, C. (1900).—‘* Note sur les tubes de Malpighi des larves de
Stratiomys”’, ‘ Bull. Soc. Ent. de France’, p. 360.

—— (1902).—‘‘ Contributions 4 l’étude des larves et des métamor-
phoses des diptéres”’, ‘Ann. de Univ. de Lyon’, nouv. série,
i, 178 pp., 4 Pls.

a

ae ;atey rf,

ag p D r 7 sioee 8 a r i
ee i) s

(4&4 Cs sunt k ete, iy ae

mr iar t sia ws fae ik ae .

i 7 tees) prt ai oe <a nie ify —(

¥ : 4) ae Ree ie je inal

: 7 i \ = aT = wi |

rT. ;
Waris

7 i
;
r
-
;
‘
Sm
; =
:
i
-
¢ ’
-
.
ry
j

The Early Development of the summer ege of
a Cladoceran (Simocephalus vetulus).

By

d

H. Graham Cannon, B.A.,

Demonstrator in Zoology, Imperial College of Science, South Kensington.

With Plate 25 and 1 Text-figure.

CONTENTS.

PAGE
1. INTRODUCTION : é : : : : i : . 627
2. METHOD ‘ , : ‘ , : : : ‘ 1K628
3. EGG-LAYING . P ” : ; : : : : - 1630
4, CLEAVAGE . : E : : : . ; : paGol
5. FORMATION OF THE GERM-LAYERS . : : : 3 a aliRY
6. Discussion . : : , 3 : : ; : . 636
7. SUMMARY : : . : : : : ; : . 640
8. BIBLIOGRAPHY : : : : : : : : 4) G41
9. EXPLANATION OF PLATE : : a : : ‘ . 642

1. INTRODUCTION.

Tis paper deals chietly with the method of germ-layer
formation in the parthenogenetically-produced summer eggs
of Simocephalus vetulus. A considerable amount of
work has been done on the development of the eggs of various
Cladocera, and a complete summary of this work is to be
found in Vellmer’s paper (14) on the development of winter
eggs of Cladocera. In the same year that this paper appeared
Kiihn (8) described very fully the development of the summer
ege of Polyphemus pediculus, and here again is given
a résumé of the work that has been done on Cladoceran
development. ‘This paper also reviews, in this connexion,

628 H. GRAHAM CANNON

all the work done on determinate development in the Crustacea.
Since then no further work has appeared on this subject, so
that it would be mere repetition if that work were to be again
summarized here.

Vollmer (14), in his summary, states! ‘that these results
peint to the fact that we must differentiate between two
categories of eggs which possess different modes of develop-
ment in relation to ther yolk-content ; eggs poor in yolk
show a determinate development with practically total seg-
mentation, eggs rich in yolk show an indeterminate superficial
type of development ’. Referring to his own work, he suggests
that the developmental processes that he describes demonstrate
an intermediate form between these two methods and make
the transition less abrupt. From the work recorded in this
paper it would seem that the development of the summer
eggs of S$. vetulus shows, perhaps more markedly, an inter-
mediate stage between the determinate and indeterminate
methods of development of Cladoceran eggs.

This work was carried out in Professor MacBride’s laboratory
at the Imperial College of Science, and I must thank Professor
Macbride for valuable suggestions and for kindly reading the
manuscript.

2. MernHop.

In all cases the embryos were dissected out of the brood-
pouch into the smallest amount of water possible before being
fixed. Iixing the whole Daphnid with the embryos still in
the brood-pouch gave unsatisfactory results.

It was found necessary to employ different fixatives for the
various stages of development. For the early stages no reliable
method was found. Carnoy’s fluid (Ae. Ale. Chloroform)
gave good results, but the difficulty experienced was the
unreliability of the fixative. The egg is surrounded by a tough
inembrane and it is this that causes the trouble. It is never
possible to say whether it will burst or not under the action
of the fixing agent. In the segmenting egg, if the membrane

' My translation.

DEVELOPMENT OF SIMOCEPHALUS 629

bursts, the fixation is not good, while after the blastula stage
the reverse is the case. If the membrane bursts, it produces
a certain amount of distortion, and with all fixatives except
Carnoy this was so bad as to make the material useless. With
Carnoy a variable amount of swelling was produced, but with
the other fixatives the whole egg usually burst. If the mem-
brane remains intact, the egg becomes very difficult to embed
owing to the embedding material not penetrating the mem-
brane, and so it was found extremely difficult to obtain sections
of the segmenting egg. One method used for the earliest
stages was to employ hot water as the fixative. The eggs
were dissected out of the brood-pouch and flooded with
boiling water. After thirty seconds they were transferred to
70 per cent. alcohol. This gave fairly good results, but with
later stages the nuclei were not well preserved and so the
method was not of much use. Fixing in bichromate-formol
and subsequent treatment with 5 per cent. formalin gave good
results, but here again it was unreliable and gave results no
better than those obtained with Carnoy’s fluid. Gilson’s
mixture (Subl. Ac. Ale. Chloroform) gave very good fixation
when it succeeded in fixing the embryo without producing
excessive distortion.

For later stages Carnoy was again used, but better results
were obtained with hot Flemming. The embryos were placed
in Flemming’s strong solution at 56° C. for ten minutes and
then washed out in water. Strong picro-sulphuric gave fair
fixation.

The embryos were stained with alcoholic eosin before
clearing in clove oil and embedding in clove-oil ‘ celloidine’.
This made them more conspicuous and hence easier to manipu-
late. After hardening the celloidine they were embedded in
paraffin at 56° C.

Sections were cut 6 » and 7» thick and stained on the slide.
The best stain was Ehrlich’s haematoxylin. Iron haematoxylin
was used after Flemming fixation. Haemalum, picro-indigo-
carmine and thionin were among other stains used which
proved satisfactory.

630 H. GRAHAM CANNON

3. EGG-LAYING.

On several occasions the actual laying of the egg into the
brood-pouch was observed. Hach egg is laid separately as
a continuous stream of foam. The foam appears to consist
of more or less opaque drops—probably yolk-spheres and
transparent colourless globules—presumably oil in a continuous
mass of protoplasm. Immediately after laying, the egg is of
an irregularly elongated shape tapering at the end nearest to
the opening of the oviduct. In a few minutes it has rounded
itself off and become regularly shaped and almost spherical.
The oil-drops now commence to coalesce to form one large
oil-globule. About two hours after laying this large oil-drop
is most distinct. It is excentrically placed in the egg and at
this time has a diameter very slightly greater than half that
of the egg.

As stated above, the only fixative that was found satisfactory
for the earliest stages of the egg was Carnoy’s fluid. In sections
of eggs fixed in this liquid it was possible to recognize, according
to Gatenby’s diagnosis (5), the following structures: (1) one large
oil-globule excentrically placed and surrounded by a few much
smaller globules—these appeared as sections of empty vacuoles ;
(2) a mass of protoplasm placed almost centrally and on the
edge of the large oil-globule; (3) a large number of yolk-
discs staining very faintly with thionin and pervading the
remainder of the egg; (4) a less number of smaller bodies
scattered among the yolk-discs and staining deeply with
thionin—presumably the remains of mitochondria or Golgi
bodies.

Lebedinski (9) describes a similar arrangement of materials
in the egg of Daphnia similis, but does not mention the
mitochondria.

An egg-membrane is clearly distinguishable soon after the
egg has been laid, and it would appear very probable, from the
fact that the egg is laid as a fluid mass which subsequently
rounds itself off, that this egg-membrane is produced by the
egg itself after this rounding-off has taken place. It is not

DEVELOPMENT OF SIMOCEPHALUS 631

a vitelline membrane if this term is restricted, as McMurrich
(10) maintains, to a membrane which is connected with the
process of fertilization, but must be termed a primary ege-
membrane or * Dotterhaut’ as defined by Korschelt and
Heider (7).

4. CLEAVAGE.

Cleavage is completely superficial. At first the separate
blastomeres remain deep in the egg as apparently amoeboid
masses of protoplasm. After five hours they begin to appear
on the surface, and soon after, each blastomere becomes
separated from its neighbours by furrows extending a short
distance into the yolk. Hight hours after the egg has been
laid cleavage is complete and results in a uniform blastoderm
enclosing the yolk-mass. No yolk-cells were found in the
interior of the blastula. In Daphnia similis Lebedinski (9)
found that certain blastomeres remained behind in the centre
of the egg while the remainder migrated towards its surface
to form the blastoderm, and that the former blastomeres
functioned in absorbing the fat or yolk-drops. Vollmer (14)
in the winter eggs of Cladocera describes the formation of
a blastula with greatly reduced blastocoele by total cleavage,
and states that cells are budded off from the blastomeres into
the interior of the egg which function as yolk-cells. In
Leptodora hyalina Samter (12) found that yolk-cells
were budded off from the blastoderm at the same time that
the endoderm plate commenced to immigrate into the egg.
Agar (1) in Holopedium gibberum states that ‘ fairly
late stages show occasional very flat nuclei on the separate
yolk-masses, as figured by Samassa (11). Doubtless each
yolk-mass is contained in a single cell. The origin of these
yolk-cells has not been observed, but it may be safely assumed
that they arise in the same way as that described by Semassa,
i.e. by budding off from the mesendoderm’. Similarly, in
S. vetulus embryos in which the endoderm ‘has already
separated from the mesoderm often contain yolk-masses
against which flattened cells are seen to be lying. ‘Their origin

632 H. GRAHAM CANNON

cannot be stated with certainty, but it is thought very probable
that they arise from cells originally lying round the genital
rudiment which pass inwards on the inside of the blastoderm,
as will be described below.

5. FoRMATION OF THE GERM-LAYERS.

The first sign of differentiation of the blastoderm is the
appearance of a group of cells—more vacuolated than the
rest—on one side of the embryo which subsequently proves
to be the ventral side. These cells contain a large amount of
yolk, and in their earliest stages their nuclei are very obscure.
They will be called collectively the ‘ Ventral Mass’ (fig. 1).

When cleavage is complete each blastomere consists of an
inner yolky part and an outer non-yolky part. In their very
earliest stages the cells of the ventral mass are completely
pervaded by yolk and so are conspicuous by not showimg the
outer non-yolky zone. Soon a few of these cells pass inwards,
so that the ventral mass becomes a small heap of vacuolated
yolky cells on one side of the embryo, but as yet shows no
further sign of differentiation.

The cells of the ventral mass on one side, which is seen later
to be the anterior side, now proliferate and form a mass of
yolky cells whose protoplasm stains comparatively deeply
(fig. 2). The compactness of these cells and the distinct manner
in which they are marked off from each other indicate that
their protoplasm has a greater surface-tension than that of
the cells of the remainder of the ventral mass. The nuclei
of these cells, which are now becoming distinct, are large
compared with those of the blastoderm cells—approximately
twice as large. Their nucleoli are distinct and stain deeply.

Behind these cells, that is at the posterior part of the ventral
mass, are a few cells which still form a single layer. They are
very much vacuolated and contain a large amount of yolk.
Their protoplasm does not stain at all deeply and the cells are
not at all compact. At first their nuclei are not distinct, as
with the remainder of the cells of the ventral mass, but soon
these become quite clear and show very marked characteristics.

DEVELOPMENT OF SIMOCEPHALUS 633

They are several times as large as those of the blastoderm
cells, as will be seen from fig. 8. The chromatin in them is
either very scattered or very scanty. Each nucleus contains
several nucleoli which stain to varying degrees, but none stain
at all deeply. These cells are the primordium of the gonads.

Commencing at the earliest stages when the nuclei of the
cells of the ventral mass are still obscure, cells can be seen
round the posterior periphery which are passing inwards and
dorsally up the inside of the blastoderm. These cells are
apparently formed by proliferation of the blastoderm cells
round the edges of the ventral mass and then migrate inwards
at its periphery. They are mesoderm cells and will be spoken
of as the Ectomesoderm. Later their nuclei become more
distinct and are seen to be larger than those of the blastoderm
and to contain distinct deeply staining nucleoli.

Soon after the genital rudiment becomes distinct there
appear on the dorsal side of the embryo the primordia of the
nervous system—the ‘Scheitelplatten’. These consist of
two groups of tall columnar cells symmetrically placed about
the median plane, in which the nuclei are large and oval,
approximately twice as long as the nuclei of the neighbouring
blastoderm cells. The nucleoli are deeply stainmg and very
conspicuous, and there is a marked absence of chromatin in
the remainder of the nucleus. They agree with those described
by other workers on Cladocera, and their further development
will not be treated here.

A very conspicuous change is now brought about in the
embryo by the invagination of the genital rudiment. An
early indication of this inward migration can be seen in fig. 3,
where the surrounding cells are seen to be pushing their way
over the primordial germ-cells. The primitive germ-cells sink
into the egg, a variable but sometimes considerable distance.
The pit caused by this sinking in has been seen to stretch
a third of the way across the embryo. The lips of this pit are
formed of the ectomescdermal cells which are continually
pushing their way under the edge of the genital rudiment
to lie on the inside of the blastoderm (fig. 4), and as the

634 H. GRAHAM CANNON

invagination proceeds these lips gradually approach one another
(figs. 7 and 8) and thereby tend to enclose a space which
sometimes persists for a short time as a small cavity (fig. 5).
The lips ultimately fuse (fig. 6), so that the primordium of the
gonads comes to lie completely internally. With the closure
of this invagination the passage of the ectomesodermal cells
into the interior stops in this region.

At the time of invagination the number of cells constituting
the genital rudiment is about ten, but there seems to be no
constant number. Cell divisions among these cells were found
but rarely. Vollmer (14) states: ‘ Teilungsfiguren habe ich
aber niemals in der Gonadenanlage nachweisen kénnen’,
but in 8. vetulus the number of cells in the genital rudi-
ment most certainly increases by cell division from about four
at its earliest apparent differentiation to about ten at its
invagination.

While these changes have been taking place at the posterior
end of the ventral mass the formation of the ‘ mesendoderm °
has commenced at the anterior end. The original compact
yolky cells at this end apparently separate into two parts—
an inner mass of cells which spread themselves as mesodermal
cells over the anterior part of the blastoderm conspicuously
in the region of the ‘ Scheitelplatten ’, and an outer region
which remains as part of the blastoderm. In the centre of this
region, that is about midway between the genital rudiment
and the level of the ‘ Scheitelplatten ’, the mesendoderm makes
its first appearance as a group of tall, compact, comparatively
non-yolky cells in the blastoderm (fig. 7). The nuclei of these
mesendoderm cells show at first no difference from those of the
blastoderm cells, but later, as the mesendoderm mass grows,
the nuclei are seen to be nearly double as large as the blastoderm
nuclei, with conspicuous nucleo. This enlargement can be
seen to take place as the cells pass inwards. The mass enlarges
and its posterior end pushes its way backward in the median
plane (fig. 8). The area of origin, which may be termed the
blastozone, is marked a little later by a depression from wine”
later grows the stomodaeum,

DEVELOPMENT OF SIMOCEPHALUS 635

In its backward growth the mesendoderm comes up against
the primordium of the gonads. There is no strict relation
between the times of mesendoderm formation and of the
invagination of the genital rudiment--—sometimes the latter
is completely internal before the mesendoderm begins to
grow posteriorly. The mesendoderm pushes its way under-
neath the genital rudiment between this and the blastoderm,
which may now be called ectoderm, so that the genital rudi-
ment comes to lie on the dorsal side of the mesendoderm
(fig. 9).

During the formaticn of the mesendoderm, mesoderm cells
are formed at the periphery of the blastozone, most con-
spicuously at the anterior and lateral borders, the posterior
border being obscured by the backwardly-growing mesendo-
derm. The mesoderm at this stage is grouped, in the posterior
portion of the embryo, ventro-laterally, while in the anterior
part it extends dorsally, covering the ‘ Scheitelplatten’.

When the mesendoderm has finished its backward growth
it is a very clearly defined mass, and is sharply separated from
the lateral mesoderm, as can be seen in fig. 11. It now begins
to flatten out, and its lateral borders cease to be sharply cut
off from the neighbouring mesoderm. Ultimately the whole
of the mesoderm and mesendoderm form one flat plate of cells
lining the inside of the ventral ectoderm. While this fusion
is taking place the nuclei of the mesoderm and mesendoderm
cells become smaller, so as to be indistinguishable from those
of the ectoderm. From this plate of cells in the median plane
a solid rod of cells separates off, which is the endoderm (fig. 10).
At this stage the rudiments of the second antennae are already
showing. Much later, when the large stomodaeum and the
smaller proctodaeum have grown in from the ectoderm, this
solid rod acquires a lumen.

At the time of separation of the endoderm the genital rudi-
ment still exists in the ventral part of the embryo lying on the
gut, as a mass of yolky cells with very large nuclei showing
the game characteristics as the original primordial germ-cells
of the ventral mass.

636 H. GRAHAM CANNON

6. Discussion.

Kxiihn has shown in his paper dealing with the development of
the summer eggs of Polyphemus (8) that the early develop-
ment of Moina as described by Grobben (6) is very similar,
in fact almost identical, to that of Polyphemus, and
Samassa’s description (11) of a totally indeterminate method
of development of the eggs of Moina receives no support
from his work. These are the only two Cladocera in which
a determinate type of development has as yet been described.

Vollmer in his work on the resting eggs of the Cladocera
(14) states that when the blastoderm consists of about two
hundred cells a migration inwards takes place of about eight
to ten of these from the future ventral side of the embryo.
These multiply and form the genital rudiment. <A similar
proliferation of cells from a ventral blastozone later forms the
‘untere Blatt’, and from this is subsequently separated a
solid rod of cells which forms the gut. Because of the early
separation of the genital rudiment Vollmer states that this
method of development is intermediate between the deter-
minate development of Polyphemus and Moina and
the indeterminate type of development as described by Agar
(1) in Holopedium and Lebedinski (9) in Daphnia.

A comparison of the genital rudiment as deseribed by
Vollmer for Daphnia with that of Simocephalus
described in this paper shows certain differences. Jirstly.
the mode by which it passes into the interior of the embryo
is different in the two cases. In Daphnia this is brought
about by a few cells that wander from the blastoderm into
the interior of the egg, presumably by the action of an inwardly
directed cytotaxis. In Simocephalus, on the other hand,
the group of cells forming the genital rudiment passes into
the interior by an invagination and only becomes internal
when the edges of the pit caused by this invagination have
srown together and fused. Here again the invagination may
be brought about by a similar force. However, the extent of the
invagination yaries considerably, sometimes the pit is very

DEVELOPMENT OF SIMOCEPHALUS 637

shallow, while at other times, as stated above, it has been
seen to stretch one-third of the way across the egg. This fact,
when it is also remembered that the surrounding cells are
actively proliferating and producing cells which push their
way inwards between the edge of the genital rudiment and the
blastoderm, suggests that the invagination may be brought
about by the ectomesoderm cells pushing the genital rudiment

TrExt-Fic. 1.

Mesendoderm al Cells Endoderm

Genital |
Rudiment

™~ Eclomesoderm-*

(2) (6)

(a) Diagram of the ventral view of embryo of Simocephalus
vetulus, showing the ventral mass before the formation of the
mesendoderm.

(6) Ventral view of embryo of Polyphemus pediculus in
thirty-two-cell stage (from Kiihn).

in front of them as they themselves pass into the embryo.
A second difference lies in the fact that in Daphnia the
primordial germ-cells when they have passed into the interior
lose their yolk. Vollmer states (14): ‘auch in den Blastoderm-
zellen schreitet die Dotterresorption fort, wenn auch nicht in
demselben Grade wie in der Gonadenanlage’. In $8. vetulus
the cells of the genital rudiment always consist of large yolky
cells which retain their yolk all through the development.
Their protoplasm also stains very faintly, not as im Daphnia,
where Vollmer states that these cells show an increased
affinity for stains. However, from the position of origin of the
NO. 260 UU

638 H. GRAHAM CANNON

genital rudiment in the two forms, and from its relation to
the mesendoderm and ultimate fate, it would seem that the
differences are of small significance and that the two structures
described as genital rudiment are really homologous.

A comparison of the mode of development of S. vetulus
with that of Polyphemus as described by Kiihn reveals
some very close analogies. Text-fig. 1 (a) shows a diagram
of the ventral mass of S$. vetulus before the formation
of the mesendoderm. In the posterior region are the large
primordial germ-cells bordered laterally and posteriorly by
ectomesodermal cells. In front is the group of yolky cells
which are mesendodermal. ‘The mner layers of this latter
cell-mass spread out over the anterior part of the blastoderm
as mesodermal cells, and from the outer layer is developed the
very definite mesendoderm. While this is growmg backwards
mesoderm cells are still being proliferated inwards at the anterior
and lateral edges of this group and possibly at the posterior
edge. The fact that these latter cells origmate by proliferation
of cells at the edge of this mesendodermal group, together
with the fact that they form mesoderm distinct from the
mesoderm included in the backwardly growmg mesendoderm,
suggests that possibly they are a separate source of mesoderm,
that they are ectomesodermal cells—a continuation forwards
of the ectomesodermal cells which are formed at the periphery
of the genital rudiment. If this were so, an analogy might
be drawn with the development of Cyclops as described by
Urbanowicz (18), where he states that larval mesenchyme
arises from cells surrounding the primitive endoderm cell
while the secondary mesoderm arises from the gut. The more
recent work of Fuchs (4) on Cyclops has, however, failed
to confirm the findings of Urbanowiez, and has, on the contrary,
demonstrated an extraordinary resemblance between the
development of Cyclops onthe one hand and Polyphemus
and Moina on the other, in neither of which is there any
larval mesenchyme as distinct from secondary mesoderm.
But in $. vetulus when the mesendoderm is growing
backwards, although its hinder end is very sharply separated

DEVELOPMENT OF SIMOCEPHALUS 639

from the laterally lying mesoderm (fig. 11), at the anterior
end no such clearness exists and at the blastozone the mesendo-
derm merges into the plate of mesoderm lining the anterior
part of the embryo. But both this anterior mesoderm and the
mesendoderm clearly arise from a sharply defined group of
cells at the blastozone, and it is suggested that there is no
distinction between the mesoderm of the mesendoderm and
the other mesoderm formed in this anterior region. If this is
so, a very complete analogy can be found with Polyphemus.
Text-fig. 1 (b) shows a view of the vegetative pole of a Poly-
phemus embryo in the thirty-two-cell stage. ‘wo central
primordial germ-cells forming the genital rudiment are placed
posteriorly to two cells which give rise to the whole of the
endoderm. Laterally and posteriorly to the genital rudiment
are six cells which give rise to both ectoderm and mesoderm.
Each of these six cells divides into two cells, one of which
becomes an ectoderm cell and the other gives rise to mesoderm
cells. In the comparison of these two figures it is seen in the
two cases that the germ-cells are completely segregated in
the genital rudiment as two cells in Polyphemus and as
a group of about four cells in S. vetulus. Forming a
crescent posteriorly round this primordium in both cases are
mesectodermal cells, but anteriorly in Polyphemus are
two endoderm cells, while in $. vetulus are a group of
mesendoderm cells. The chief difference between the two
forms is thus that the endoderm is segregated very late in
S.vetulus, while it separates very early in Poly phemus—
in the sixteen-cell stage. Similarly the mesoderm is segregated
later than the endoderm, but still very early in Polyphemus
compared with 8. vetulus where the separation of mesoderm
is only complete with the separation of the endoderm.

In Moina and Polyphemus Weismann (15) has proved
that the parents nourish the young in their brood-pouch,
and it is probably due to this fact that the yolk im the eggs
of these two forms has diminished so considerably, and in
correlation with this disappearance of yolk is the appearance
of the teloblastic type of development. In S$. vetulus

uu2

640 H. GRAHAM CANNON

and also in Daphnia Agar (2) has shown that while the
embryo is in the brood-pouch it does not receive nourishment
from its parent. And yet §. vetulus shows a type
of development which differs considerably from that of
Daphnia in that there is a very early segregation of the
genital rudiment but shows such obvious similarities to the
development of Moina and Polyphemus. ;

The fact that has been pointed out by Fuchs (4), that among
sroups so far apart as the Copepoda and the Cladocera, in
forms where there has been loss of yolk owing to the develop-
ment of other modes of nutrition of the embryo, there is such
an extraordinary similarity in the cell lineages, suggests firstly
that the arrangement of the ‘ Anlagen’ in the eggs of these
forms is a very archaic character, and secondly that in cell
lineage there is a representation of the arrangement of the
‘Anlagen’ in the yolky eggs that do not show a teloblastic
mode of development. ‘This view is upheld by the similarity
between the cell lineages of the Cladoceran eggs that contain
little yolk and of the egg of the Cirripede Lepas, where
although there is abundant yolk, yet there is determinate
cleavage (Bigelow, 3). In the development of 8. vetulus
there is further support of this view in that in this apparently
indeterminate method of development the earlest arrangement
of the germ-layer ‘ Anlagen’ shows such a close resemblance
to the arrangement of the teloblasts in the non-yolky eggs of
the Cladocera.

7. SUMMARY.

1. Kach egg is laid as a yolky mass of a foam and later forms
a primary egg-membrane.

2. Cleavage is completely superficial and apparently indeter-
minate.

3. The first differentiation of the blastoderm is the appear-
ance of a group of vacuolated yolky cells on the ventral side
of the embryo which are called the ventral mass.

4. This subsequently differentiates into a few large cells
with very large nuclei which form the genital rudiment, sur-
rounded laterally and posteriorly by ectomesodermal cells,

DEVELOPMENT OF SIMOCEPHALUS 641

and anteriorly to this a mesendodermal mass of cells from which

arises the mesendoderm.

5. The genital rudiment surrounded laterally and posteriorly
by inwardly growing ectomesodermal cells invaginates and
becomes internal by the lips of the invagination growing
together and fusing.
6. The mesendoderm grows backwards as a solid mass of
cells, which later spreads out flat and becomes indistinguishable
from the laterally-lying mesoderm, and from this layer the
endoderm separates as a solid rod in the median plane.
SoutH KENSINGTON,

June 1921.

$. BrBLioGRAPHY.

. Agar, W. E. (1908).—‘** Note on the early development of a Clado-

ceran (Holopedium gibberum) ”, * Zool. Anz.’, 33.

: (1913).—‘* Transmission of environmental effects from parent to
offspring in S. vetulus”, ‘ Phil. Trans. Roy. Soc. Lond.’, 203.
. Bigelow. M. A. (1902).—** The early development of Lepas”’, * Bull.

Mus. Harvard Coll.’, 40.

. Fuchs, K. (1913).—** Die Zellfolge der Copepoden ”’, * Zool. Anz.’, 42.

. Gatenby, J. B. (1919).—* The Identification of Intracellular Struc-
tures’, “Journ. Roy. Micro. Soc.’

. Grobben, C. (1879).—‘* Die Entwicklungsgeschichte der Moina

- rectirostris”, ‘ Arb. a. d. Zool. Inst. Wien’, 2.

. Korschelt and Heider (1903).—‘ Lehrbuch der vergleichenden Ent-
wicklungsgeschichte der wirbellosen Thiere.’

. Kithn, A. (1913).—‘* Die Sonderung der Keimbezirke in der Ent-
wicklung der Sommereier von Polyphemus _ pediculus
(de Geer)”, ‘ Zool. Jahrb., Abt. f. Anat.’, 35.

. Lebedinski, J. (1891).—“* Die Entwicklung der Daphnia aus dem
Sommerei ”’, ‘ Zool. Anz.’, 14.

. MeMurrich, J. P. (1895).—‘‘ Embryology of the Isopod Crustacea ”’,
‘ Journ. Morph.’, 11.

. Samassa, P. (1893).—‘‘ Die Keimblatterbildung bei den Cladoceren: I.
Moina rectirostris (Baird)”’, ‘ Arch. Mikr. Anat.’, 41.

. Samter, M. (1900)—‘‘ Studien zur Entwicklungsgeschichte der
Leptodora hyalina”, ‘ Zeit. f. wiss. Zool.’, 68.

13. Urbanowicz, F. (1884).—‘‘ Zur Entwicklungsgeschichte der Cyclo-
piden ”’, ‘ Zool. Anz.’, 7.

. Vollmer, C. (1912).—‘‘ Zur Entwicklung der Cladoceren aus dem
Dauerei ”’, ‘ Zeit. f. wiss. Zool.’, 102.

. Weismann, A. (1877).—‘‘ Beitriige zur Naturgeschichte der Daph-
noiden ’’, ibid,, 28,

642 H. GRAHAM CANNON

EXPLANATION OF PLATE 25.

List oF ABBREVIATIONS.

bl, blastozone; ect, ectoderm; em, ectomesoderm; end, endoderm ;
ga, genital rudiment ; gac, cavity of genital rudiment; 7, pit produced
by invagination of genital rudiment ; me, mesendoderm ; mes, mesoderm ;
mm, mesendodermal mass ; v.m, ventral mass; ¥.c, yolk-cells.

Figs. 1, 2, 3, 4, 8, and 9, are from material fixed in Carnoy’s fluid, The
remainder are from Gilson material.

Fig. 1.—Section through an embryo showing the earliest sign of differen-
tiation of the blastoderm. The ventral mass is marked off from the rest
of the blastoderm as a group of cells completely pervaded by yolk.

Fig. 2.—Median section through embryo showing differentiation of
ventral mass into (1) genital rudiment, (2) anteriorly, the comparatively
deeply staining mesendodermal mass, and (3) posteriorly, the ectomesoderm
cells which are passing inwards. The nuclei at this stage are not at all
distinct. .

Fig. 3.—Slightly oblique section—almost median—of an embryo slightly
older than that figured in fig. 2. Shows the same as in fig. 2, but nuclei
are now distinct. The cells surrounding the genital rudiment are seen to
be pushing their way over the latter.

Fig. 4.—Transverse section of the genital rudiment showing how the
lips of the pit caused by its invagination are formed of inwardly migrating
ectomesoderm cells.

Fig. 5.—Transverse section through the genital rudiment after it has
become completely internal, showing its cavity.

Fig. 6.—Transverse section through the invaginating genital rudiment
showing the fusion of the lips of the invagination pit.

Fig. 7.—Median section showing commencement of mesendoderm, The
genital rudiment is not yet completely internal.

Fig. 8.—Median section showing mesendoderm growing backwards from
the blastozone which is marked by a small depression,

Fig. 9.—Median section. The mesendoderm has grown backwards
underneath the genital rudiment which is now completely internal.-

Fig. 10.—Transverse section showing endoderm separated as a solid
rod from the laterally lying mesoderm. The genital rudiment is imme-
diately dorsal to the endoderm. Yolk-cells are seen in this figure enclosing
yolk and oil-drops.

Fig. 11.—Transverse section through the posterior region of the blasto-
zone showing mesoderm formation at the lateral borders of the blastozone,

Fig.2.

Cannon del.

L, )
J 4
~w

§

~
sy
XQ
RS
3
R
N
©
S
»
NY
:

c/
Huth London

Studies in Dedifferentiation.
II. Dedifferentiation and resorption in
Perophora,

By

Julian S. Huxley,
New College, Oxford.

With Plates 26-28 and 1 Text-figure.

CONTENTS.

PAGE

1. INTRODUCTION é ‘ ‘ . ‘ ; : . . 643
2. DEDIFFERENTIATION ; é : ; : . : . 645
(a) General : : 2 ; : : ; . 645

(6) Simple Dedifferentiation (Clavellina type) é < . 647

(c) Dedifferentiation with Resorption . : ; : . 648

3. EXPERIMENTS WITH PotTasstuM CYANIDE : ; : a (as
4, EXPERIMENTS ON ANIMALS WITHOUT CIRCULATION. ; . 656
5. EXPERIMENTS WITH Low TEMPERATURE . ; P : . 660
6. MiscELLANEOUS NOTES ON PEROPHORA . : : : . 660
7. EXPERIMENTS ON OTHER SPECIES . ‘ ‘ ; j . 662
8. Discussion . : ; ; : : , . ; ~ 1067
9, SUMMARY . : : : ; : : F : . 692

1. INTRODUCTION.

THE observations of Driesch (1906) and E. Schultz (1907) and
myself (unpublished) upon the reduction or dedifferentiation of
the social Ascidian Clavellina have been mainly morphological.
Accordingly I decided, while in the United States, to take up
the problem from the physiological aspect. ‘The work was
carried out at Wood’s Hole. Clavellina itself is not found
there, but another social Ascidian, Perophora viridis,
is common, and proved to be a useful form for experimental]

644 JULIAN 8. HUXLEY

work. As is well known, the social Ascidians reproduce
asexually by means of buds given off at intervals from creeping
branched stolons; but whilst in Clavellina the zooids may
reach two inches, in Perophora the maximum length is only
about one-quarter of an inch, and the span of life is probably
limited in proportion. The branching and budding of Pero-
phora is also much easier to follow, the stolons often growing
in a straight line for a considerable distance, giving off buds
at regular intervals. It is thus easy to trace a sequence from
young to old individuals in Perophora, but hard in Clavellina.
In Perophora it is also possible to isolate single zooids of any
age by cutting the stolon midway between the neighbouring
zooids on either side; and in such preparations the piece of
stolon is of the same order of magnitude as the zooid, while in
Clavellina the volume of the stolon is quite negligible in pro-
portion to that of an adult, a half-grown, or even a quarter-
erown zooid.

Such preparations we may call stolon-zooid systems.
They are composed of two very distinct parts. The stolon
is very simple : it consists of a thin external test-layer surround-
ing a single-layered tube of flattened ectodermal epithelium,
which in its turn is divided into two by a horizontal partition
composed of two very thin endodermic epithelia flattened
together to form a single sheet ; the space between ectoderm
and endoderm contains blood, with numerous cells of several
different kinds. At either end of the stolon the partition
stops short, so that the blood can circulate from one half-tube
to the other. It is normally kept in motion by the heart-beat
of the zooid, which, as in all Ascidians, undergoes a periodic
reversal of direction. The cut surface of test and ectoderm
soon heals over. In a healed preparation the ectoderm at
either (cut) end of the stolon is more or less cuboidal, and
presents the appearance of undifferentiated tissue.

The zooid, on the other hand, is of high organization, con-
taining, as it does, heart, stomach and intestine, elaborate
branchial apparatus, nervous, muscular, and excretory systems,
and hermaphrodite reproductive organs. It is also highly

DEDIFFERENTIATION IN PEROPHORA 645

sensitive in the region of the two siphons. It is connected with
the stolon by a narrow tube of less diameter than the stolon,
separating above into two tubes; this is generally longer in
proportion in older individuals. The stolon may grow in
length and form buds at the proximal end or the distal end
or both.

Suitable food for Perophora has not yet been discovered ;
but in spite of this stolon-zooid preparations may be kept alive
in the laboratory for a considerable length of time.

2. DEDIFFERENTIATION.

(a) General.—Processes may occur in living matter
whereby whole organisms or parts of them become visibly
simpler. This occurs, for instance, in Clavellina when kept in
unfavourable conditions, in Hydra when starved (Schultz, 1906),
and in various other Coelenterates, in encysting protozoa and
in other protozoa in the ordinary course of the life-cycle, with-
out encystment (Lund, 1917), in sponges (Maas, 1910 ; Miiller,
1911), &e. Such a process is the reverse of differentiation, and
is best called dedifferentiation. It has also been termed
involution and reduction. The latter word will here
occasionally be used as a convenient synonym for the more
accurate but clumsier term.

In Clavellina the original observations of Driesch and the
later work of Schultz was carried out on half-animals,
the individuals being cut in two and the half containing
the branchial sac (pharynx) used for the experiments. ‘This
portion proved capable of regenerating the whole organism.
Sometimes it remained intact and produced a restitution-bud
in which the missing organs were formed; at other times it
dedifferentiated completely to form an opaque spheroid which
later redifferentiated into a normal whole individual; or it
might show a combination of the two processes. Here, when
dedifferentation occurred, it was as the result of the shock
of the operation and of the changes produced by it.

646 JULIAN 8S. HUXLEY

However, Driesch also mentions in one of his papers (1906)
that he had been able to secure dedifferentiation in whole
individuals. In my work I used whole individuals only. With
them I found that the simplest method of obtaining dedifferen-
tiation was to leave unchanged the water in which the organisms
were kept, the accumulation of toxic waste products probably
initiating the process. It was also found that only young
individuals underwent dedifferentiation easily, mature and
half-grown zooids speedily dying.

When full dedifferentiation, whether of half or whole zooids,
occurs in Clavellina, a spheroidal white mass results, in which
all the organs are very much simplified, both morphologically
and histologically, becoming reduced to a series of separate
sacs, some simple, others compound, of roughly spherical
shape with walls of embryonic-looking cuboidal cells. On
being replaced in clean water the opaque mass usually grows
out to form a new perfect zooid, quite normal but smaller than
the original; and this alternation of differentiation and
dedifferentiation may be repeated several times. It 1s obvious
that the term dedifferentiation may be applied equally to all
retrogressive changes resulting in simplification of visible
structure, provided that the reduced tissues remain alive,
whether or no redifferentiation from the reduced condition
is possible or not. When it is possible, an added interest
attaches to the whole phenomenon ; but dedifferentiation is
essentially similar whether subsequent redifferentiation can
occur or not, just as differentiation is essentially similar in
all cases whether subsequent dedifferentiation can occur
or not.

In Perophora similar methods were at first adopted, the
animals being kept in watch-glasses containing approximately
either 5 or 7-5 ¢.c. of water.

Zooids that were adult or more than half-grown never
achieved successful reduction. They all died after a few days,
but always after a preliminary attempt at dedifferentiation.
The siphons were closed, all appearance of vigour and tone was
lost, the body became contracted and opaque. The appearance

DEDIFFERENTIATION IN PEROPHORA 647

was very similar to that presented by an early stage of dedif-
ferentiation. After this, however, a brownish colour appeared
in the animals, and this heralded true degenerative changes
leading to death. Adult individuals are often found in nature
in a similar state, and these, too, always appear to die without
full dedifferentiation ; in fact it would appear that natural
death occurs in Perophora through this means, the conditions
in old zooids being such that they cannot maintain themselves
in full tone, and thus undergo incipient dedifferentiation,
which, in these old zooids, is not able to complete itself, and
so leads on to degeneration and death. Similar failure of old
individuals to adjust themselves to changed conditions is of
course well known in the case of regeneration ; a discussion
of the whole subject will be found in Child’s book, ‘ Senescence
and Rejuvenescence ’ (1915 a).

When smaller zooids were taken, however, quite different
results were obtained.

(b) Simple Dedifferentiation (Clavellina type).

If the stolon be cut very close to the zooid on either side,
the zooid will usually dedifferentiate as in Clavellina. That is
to say, the siphons contract, the zooid shrinks, becomes
increasingly opaque, and eventually draws right away from the
tunic. ‘The final stages of this process were represented by
opaque spheroidal masses with a diameter of one-third to one-
half that of the original zooid, and often with no or extremely
slight trace of siphons. The heart usually continued to beat
even in this condition. Examples are shown in fig. 1. Here
the shortness of the stolon is noticeable.

In most examples of this process the stolon was either very
short, or underwent dedifferentiation concomitantly with the
zooid, or both. In all such cases the system, with its relatively
small proportion of stclon, was similar to a stolon-zooid
system in Clavellina, and behaved in an essentially similar
way.

In one point there was a difference. I never observed such
complete reduction in Perophora as in Clavellina. l’urther,
[ was not able to obtain redifferentiation by replacing the

648 JULIAN S. HUXLEY

spheroids in clean sea-water. This, however, is probably due
simply to a greater susceptibility of Perophora to laboratory
conditions, in the same way as one species may develop well
after artificial insemination in the laboratory, while a closely-
related species cannot be got beyond early segmentation
stages.

(c) Dedifferentiation with Resorption.

(1) Stolon Resorption.—In systems with healthy
young or moderate-sized zooids which were changed to fresh
sea-water daily, the interesting fact soon came to hght that so
long as the full tone of the zooid was maintained and its siphons
continued wide open, it did not decrease in size at all, but
maintained itself at the expense of the stolon. This would
also occur sometimes when the zooid was in the form of
a partially-differentiated bud (e.g. fig. 4, cf). The bud
remained of the same size and at the same stage of development
for over seven days, while the stolon was almost completely
resorbed.

Later it was found that in other systems in which the zooid
portion was represented by similar developing buds, these
might pot merely maintain themselves but actually develop
further into perfect zooids at the expense of the stolon,
e.g. fig. 2, where in the course of three days a very great change
in the relation of zooid and stolon has taken place.

Tt is thus clear that in certain circumstances the zooid may
be physiologically dominant over the stolon, and may either
develop or maintain itself at the latter’s expense.

(2) Zooid Resorption.—In other cases, however,
a change in the opposite direction takes place. In most
systems, after the lapse of a few days without change of
water (and in some even when the water is changed), the
premonitory signs of dedifferentiation become visible: the
siphons close, the general tone decreases, and the whole animal
shrinks slightly. But the sequel is quite different. Instead
of becoming more and more opaque, on account of the cells
of the various organs and epithelia becoming cuboidal and so
bringing about a marked decrease in the size of all the cavities

DEDIFFERENTIATION IN PEROPHORA 649

in the organism, the zooid remains transparent. At the same
time, however, it decreases in size. It is, in fact, being
resorbed into the stolon. Appearances indicating the
occurrence of this process are also found in nature, though
not commonly. Successive stages of the process are shown in
figs. 4, a—b, 5, 6, 9, 12, and isolated stages in figs. 7, 8, 10, 11,
13-15.

After a very short time the siphons disappear entirely,
and a spheroidal mass of two-thirds or one-half the zooid’s
original diameter is left. In this, the ovoid heart, very little
diminished in size, can always be seen pulsating steadily.
A steady diminution of size continues, the heart too decreasing
absolutely, although becoming relatively larger. A certain
degree of opacity may appear, but it is never striking.

At a certain moment the pulsation of the heart slows down
and ceases. Soon after this the heart becomes invisible
altogether. ‘Traces of other orgens are visible. At first they
are somewhat masked by the slight opacity caused by accumu-
lation of blood-cells in the shrunken zooid, but later, as the
zooid becomes smaller and smaller, they become increasingly
clear. At about the stage when the heart disappears they are
seen as two or three translucent rounded bodies, some colourless,
some faintly yellowish.

The shrinkage continues after the disappearance of the heart,
and soon the zooid comes to appear as a minute knob, scarcely
bigger than the stalk connecting it with the stolon. This stalk
represents the stolon-connexion of the original zooid, and has
itself decreased in size, although but slightly. At this stage
a single clear refractive area, which I take to be the vestige
of the stomach, is usually the only structure to be seen in the
knob. Finally the knob all but disappears, and a mere trace
of the clear area remains visible. Presumably the stalk itself
would also eventually become resorbed into the stolon, but
resorption is much retarded after the cessation of the heart's
action, and becomes progressively slower and slower as the
size of the zooid decreases, so that J have never actually
observed this ultimate step in the resorption of fully-formed

650 JULIAN S. HUXLEY

zooids. Complete resorption of very young buds has, however,
been noted. When dedifferentiation is rapid, and especially
in larger zocids, the connexion between zooid and stolon may
be severed, and a spheroidal mass left isclated in the old
tissue. This, of course, precludes further resorption.

Two further points of interest should be mentioned. The
first is that the tunic of the zooid undergoes considerable
decrease in size, presumably by means of some form of resorp-
tion. ‘This reduction, as shown in the figures, is usually
irregular, but I have seen cases of reduction in buds where
the test remamed closely apposed and of firm outline.

The second is that the stolon, especially during the late
stages of the process, performs spontaneous movements of
contraction, thereby causing a rudimentary and irregular
form of circulation through the system. This may be called
stolon-circulation. The contraction is effected by the
ectoderm cells becoming cuboidal in one place and later
extending again to become flattened ‘ pavement ’-epithelium
(fig. 24). Corresponding with these circulatory movements
back and forth, the now minute zooid could be seen now to
contract, now to expand slightly, cells moving from it imto
the stolon or vice versa. A similar contractibility of the
ectoderm J have also observed in the stolon of Clavellina,
and in the coenosare of Hydroids (Campanularia and Obelia).

During the resorption of the zooid the stolon usually grows
in length, at least during the earlier stages (figs. 5, 6a). Later
on the stolon often remains constant in size, or decreases
slightly. It then becomes more or less opaque, owing to the
accumulation in it of cells from the zooid. Such packed opaque
stolons, however, may send out transparent slender new
srowths at one or both ends. Quite often the final length
may be greater than the original length, and buds may even
be formed. The process of resorption may take a considerable
time. The zooid in fig. 9 took seven days in all, four days to
the cessation of the heart-beat and three days more until
only a stalk was left, but in other specimens it was much
more rapid. For convenience the process may be divided into

DEDIFFERENTIATION IN PEROPHORA 651

stages as follows: (1) shrinkage alone, (2) siphons closed,
(3) siphons withdrawn from test, (4) spheroidal form assumed,
(5) cessation of heart-beat, (6) reduction to stalked knob.

It will be seen that this process is the reverse of that pre-
viously described as stolon-resorption. In both eases, however,
the equilibrium of the stolon-zooid system is altered, the altera-
tion results in the resorption of one or other of its members,
and this resorption may be total.

Resorption of an organ like the stclon cannot be considered
a very unusual phenomenon. It is paralleled, for instance,
by the resorption of various larval organs at metamorphosis,
such as the gills and tail of a frog-tadpele. Resorption of whole
individual organisms, however, is much more unusual. So
far as I am aware, it has only been noted at all adequately by
Loeb (1900), who found it to occur in the Calyptoblast Hydroid
Campanularia. I have re-investigated the phenomenon in
Campanularia and also in Obelia, and can confirm the facts
entirely. Something rather similar occurs in those Echino-
derms where almost the whole of the larva is absorbed into
the growing rudiment of the adult, but there remains an
essential difference, namely, that resorption in such a case is
determined as part of a normal development, whereas in
Perophora and Campanularia it does not occur except as the
result of circumstances which must be called abnormal. This
is also true for the interesting observation made by Child
(1904), who found in the chain-forming Turbellarian Steno-
stomum that, if a cut be made through one of the zooids,
the posterior half of such a zooid is completely resorbed by
the zooid behind it. Resorption of whole zooids is also recorded
(see later, p. 675). The case of Perophora is more remarkable
than any yet recorded, partly owing to zooids being resorbed
by subordinate systems, and partly owing to the great com-
plexity of the zooids, which is very much greater than in
Hydroids or Turbellaria.

In all three cases, however—Ascidian, Flatworm, and
Hydroid alike—the mechanism of resorption appears to be
the same, namely, that the organs all decrease in bulk by the

652 JULIAN S. HUXLEY

actual migration of single cells out of their union in the tissues
into the cavities of the body (in Hydroids into the coelenteron,
in Stenostomum into the parenchyma, in Perophora into the
haemocoel). In no other way can we explain the rapid decrease
in size of the zooid, or the marked increase in the number of
cells in the cavities. The stolon in Perophora always becomes
crowded with cells during the later stages of resorption. I have
seen no sign of the cells disintegrating on release, there being
no increase in the number of granules, &c., in the plasma ;
and the process can certainly not be explained as due to the
using up of cells as nutriment in situ.

We have thus the singular spectacle of the organs and
tissues unbuilding themselves. It is as if a house were to
become smaller and smaller through individual bricks leaving
their places here and there in the walls and accumulating in
the passages and garden, the rooms meanwhile closing the gaps
in their walls and progressively diminishing in size.

During the process it appears that dedifferentiation also is
going on. Tl*or one thing, the ectodermic epithelium becomes
more and more cuboidal, and then also all cells that appear
in the blood-stream are of a simple, irregularly-rounded type,
and not visibly specialized in any way.

The long persistence of the heart as a functional organ, and
its final sudden disappearance are closely paralleled in simple
dedifferentiation in Clavellina.

Presumably what occurs when the stolon is resorbed into the
zooid is similar, the cells of the ectodermic epithelium and of
the endodermic partition also becoming dedifferentiated and
migrating out of the tissues into the blood-stream. The
process is merely not so remarkable here, owing to the less
differentiation of the tissues involved, and the subordinate
status of the stolon as an organ. To sum up, we find that in
Perophora (and in Campanularia) adverse conditions lead to
a form of reduction in which dedifferentiated cells migrate
out of their fixed position in the tissues into the general cavity
of the body, and the whole differentiated zooid finally dis-
appears by resorption. This combination of deditferentiation

DEDIFFERENTIATION IN PEROPHORA 653

and resorption will probably be found to occur also in other
colonial organisms, the zooids of which are united by relatively
undifferentiated portions.

When the stolon is resorbed in Perophora a similar process
appears to be at work. It is further probable that in many
other cases of resorption of subordinate organs, and of grafted
tissues, a combination of dedifferentiation and resorption is
also taking place, although in many higher organisms the factor
of phagocytosis also enters, but probably often as a secondary
phenomenon.

3. EXPERIMENTS witH PorTassium CYANIDE.

The next step was to find out something as to the factors
involved in the reversal of dominance and the initiation of
resorption. With this end in view some experiments with
dilute solutions of KCN were made. J have to thank Professor
Child for advice.

As a preliminary the effect of an n/250 solution of KON
in sea-water was tested. It was found that this affected the
whole system, zooid and stolon alike. Shrinkage of all parts
took place, and death-changes were in progress after twenty-four
hours. A series of solutions was therefore prepared as follows :
n/250, n/500, n/1,000, and so on to n/64,000, tegether
with a control vessel. All vessels were protected as far as
possible from evaporation, and the solutions changed every
twenty-four hours.

The detailed results are to be found in Table I. ‘They may
be summarized as follows: Solutions of n/1,000 and higher
concentration affect both stolon and zooid very adversely,
and lead to death in about forty-eight hours. The ciliary
action of the gills is much slowed down, and the action of the
heart badly affected. Almost always the stolons become
contracted and opaque. The zooids were never drained com-
pletely by resorption; they usually shrank slightly, became
opaque, and then died. In one or two cases the appearances
were very similar to those seen in the dedifferentiation of
Clavellina. In solutions from n/2,000 to n/8,000 inclusive

NO, 260 X X

654 JULIAN S. HUXLEY

there was no growth of the stolons (except a very slight growth
in one case). In n/8,000 the appearance of the stolons was
nearly normal, but in the two higher concentrations they were
adversely affected and showed contraction. As regards the
zooids, the circulation was in all subnormal. <A considerable
degree of drainimg (resorption) tock place, but was never
complete. Several became opaque and spheroidal without
appreciable draiming (Clavellma type of dedifferentiation).
The zooids mostly still showed normal tone after twenty-four
hours, while in higher concentrations all had begun to shrink
by this time. A shght effect on the stolon was indicated by
opacity and clubbing of the ends.

In solutions from n/16,000 to n/64,000 inclusive, a consider-
able proportion of the stolons showed new growth. In no case
was the stolon adversely affected, but it always remained of
normal appearance with flat cells. Of those zooids which did
not die the large majority had begun to be resorbed in the
typical way before forty-eight hours, and some of them became
completely drained. The n/32,000 solution seemed to be the
most effective in causing this draining, but this may have
been an accident, although it is perfectly possible that the
n/64,000 solution is less effective because too weak.

The controls, apart from a small proportion which started to
drain early (an occurrence which takes place in all collections
of stolon-zooid systems chosen at random, and presumably
depends on the internal condition of particular zooids),
remained normal, the zooids completely expanded, for forty-
eight hours and most of them for seventy-two hours. Most of
them showed slight new growth of the stolons, as is customary
in the early stages of stolon-zooid systems, but they were not
kept long enough to see whether stolon-resorption, which only
occurs after several days, would supervene.

We can classify the effects broadly as follows. High con-
centrations lall the whole organism speedily. The next
lower degree of concentration causes contraction (dedifferen-
tiation) of both stolon and zooid. No resorption is possible
in this case, whether of the zooid or of the stolon. The next

DEDIFFERENTIATION IN PEROPHORA 655

lower grades of strength adversely affect the zooid, but only
affect the stolon sufficiently to inhibit its growth, not to cause
its dedifferentiation. Partial resorption may take place in
these circumstances.

Still lower concentrations have no appreciable effect upon the
stolon, but yet adversely influence the more sensitive zooid.
The stolon is thus able not only to maintain its form, but to
grow. The zooid starts deditferentiation, and this is followed
by resorption, which, typically, is complete. Finally, we get
dilutions beyond which no effect is produced on the zooid or
the stolon, with the result that the normal dominance of the
zooid is maintained, and it is the stolon which is resorbed.

We thus see that these processes occurring in nature can
be experimentally controlled to a considerable degree. Other
toxic agencies were not tried on Perophora; but from what
we know of the reactions of other organisms we should expect
that the results of KCN treatment are non-specific, and that
essentially the same phenomena would occur in other toxic
solutions.

Our results of observation are therefore to be thought of as
due to the following causes :

(1) In Perophora, in the absence of food, there is a competi-
tion for nutriment among the parts of the colony.

(2) In normal conditions, in the absence of food, the most
active and differentiated parts (the zooids) are dominant in
this competition over the less active and differentiated parts
(the stolons), which are used up as nutriment by the zooids.

(3) Correlated with this difference of success in competition
there is a difference of susceptibility, the more highly-organized
zooids being more susceptible than the stolon to unfavourable
agencies.

(4) The result of unfavourable agencies on Perophora is to
cause dedifferentiation.

(5) Once dedifferentiation has started the zooid ceases to be
more active than the stolon, and so ceases to be domimant
in the intra-organismal struggle.

(6) In Perophora dedifferentiation may be followed by

xx 2

656 JULIAN S. HUXLEY

resorption due to the migration of cells from the tissues into
the blood-stream ; when the stolon is little affected, therefore,
zooid-resorption, or the reverse of (2), occurs.

In the most general terms we have a system the two parts
of which are in equilibrium. This equilibrium may alter in
either of two opposed directions. ‘There is differential activity
of the two parts; the one which is more active is capable
of causing the reduction of the other and utilizing it as food.
But differential activity is correlated with differential sus-
ceptibility, which results, in certain unfavourable conditions,
in a reversal of the direction of change; for these induce
dedifferentiation of the zooid, and in this condition it is less
active than the stolon.

Similar conditions, viz. (1) a balance in an organic system ;
(2) differential activity of the parts of the system leading
to physiological dominance of the most active part ; (8) con-
sequent differential susceptibility of the parts leading to
a possible reversal of dominance; and (4) the resultant
reversibility of the reactions of the system—play an important
part in general physiology. Often they are not easy to investi-
gate; but in Perophora we are fortunately provided with an
organism in which they appear in a striking form, and are
readily accessible to study.

Tt should be added that in all but the weakest KCN solu-
tions a grey tinge, not seen in dedifferentiating individuals
in sea-water, was observed in the zooids during resorption.

4, EXPERIMENTS ON REDUCTION IN ANIMALS
WITHOUT CIRCULATION.

At Professor Loeb’s suggestion, to whom I here tender my
thanks, experiments were undertaken to see whether the
action of the heart im Perophora was stopped by potassium
chloride, and if so whether zooids without an active circulation
would show typical reduction.

The experiment was carried out as follows. A large and
a small stolon-zooid system were placed together in finger-bowl

DEDIFFERENTIATION IN PEROPHORA 657

TABLE I

EXPERIMENTS WITH KCN,

Series A, Young and medium individuals,
Series B, Very young individuals and almost complete buds,

A and B, four stolon-zooid systems in each vessel,

Dediff, = opaque, Clavellina type of reduction.

resorption,
Strength of Twenty-four
KCN, hours.
Control A Normal

Control B

n/64,000, A

n/32,000, A

n/32,000, B

n/16,000, A
n/16,000, B

n/8,000, A

n/8,000, B

n/4,000, A

n/4,000, B
n/2,000, A

n/2,000, B

n/1,000, A
n/500, A
n/250, A

3 normal,
1 dediff.

3 normal,
1 stage 4
Normal

2 normal,
2 stage 3

3 normal,

1 dediff,
2 normal,

1 stage 3,

1 stage 4
Expanded, cir-
culation af-
fected

2 normal,
2 stage 3

Expanded, cir-

culation poor,

2 normal,
2 stage 3
Expanded, cir-
culation very
poor
2 subnormal,
1 stage 3,
1 stage 5
All subnormal

1 subnormal,
3 dediff.
All abnormal,

dediff,

Forty-eight
hours,
Normal

2 normal,
1 dediff.
1 draining,
stage 4

2 dead, 2 stage
5

All unhealthy,
1 stage 4

All draining,
stages 3-5

2 dead, 1 stage
5, 1 stage 3
dead,

1 dediff.,

2 stages 5-6
dead,

3 dediff,

—

—

_—

stage 1,
1 dediff.,

2 stages 4—5

bo

dying,
2 stage 4
dediff. ,
3 stages 4-5
1 dying,
2 dediff.,
1 stage 4
2 dying,
1 dediff.,
1 stage 6
All dead or
dying
All dead or
dying
2 dead, 1 dying,
1 dediff,, un-
healthy,

—_

Seventy-two
hours,
3 normal,
1 draining,
stage 3
As at 48

hrs.

2 dead,
1 stage 5,
1 stage 6

1 dead,
1 dediff,,

2 stage 5

All dead

Stages 1-6 refer to stages of

Remarks,

New growth
stolons,

on all

Slight new growth on all
stolons.

Stolons healthy,
growth on 2,
Stolons healthy, slight-
ly turgid, new grewth
on 2.

Stolons healthy, no new
growth.

new

Stolons healthy, new
growth on 1,
Stolons healthy, new

growth on 2,

Stolons nearly normal,
no new growth.

Some stolons healthy,
some opaque, | with
new growth (very
slight).

Most stolons contracted;
no new growth.

All stolons opaque ; no
new growth,

All stolons contracted.
Cilia slow.

All stolons opaque,
clubbed,
All stolons early af-
fected,
All stolons early af-
fected,
All stolons early af-
fected.

io 2)

65 JULIAN 8. HUXLEY

containing 50 ¢.c. sea-water together with a certain amount

of n/2 KCL. ‘The results are summarized in Table II.

TABLE II

+ denotes active heart-beat ; (+) slow; (—) slow and intermittent ;
— no heart-beat, The upper sign in each compartment denotes the larger
zooid, the lower the smaller,

No. of c.c. n/2 Minutes.
KCl added, 15 20 30 35 40 50 70 160
0 (control) == =F =F SF af =e se +
37 at Sa =5 oF +P = de
2 + =F =e ae at == oF =
ae a5 str Te + ot i =
4 =e = ap EW eh ai DY ied oe
= = =
8 eae A ieEOt nok | ae
3h ar 35 (—) noted —
10 “F is Se lily ba eet
+ oF a
15 te “7 a aS) ary =
aa sae Gad a Gt fee
20 ae)
+. (-) -
40 a =

KCl thus exercises a very marked effect upon the Ascidian
heart. The stronger action of the salt on small zooids is to
be noted. The organisms were left m the solutions to see what
type, if any, of dedifferentiation they showed.

Those in the two highest concentrations died in under
twenty-four hours without reduction ; their stolons also were
killed or damaged. Those to which 10 and 15 ¢.c. KCI had
been added were searcely affected after twenty-four hours,
but were dead by forty-eight hours, having previously shrunk
very considerably and become opaque.

In the solution with 8 ¢.c. one had died; the other had
started to dedifferentiate. Both stolon and zooid
were affected (fig. 20). The zooid showed a characteristic
sign of KCl reduction in the cellular strands extending from the
retracted siphons to the test. Also characteristic, and directly
dependent on the absence of circulation, was the congestion

DEDIFFERENTIATION IN PEROPHORA 659

of the network of small blood-vessels close to the surface
with the green blood-corpuscles. This gives a premature
green opacity to zooids dedifferentiating in KCl. This animal
was dead on the succeeding day.

In the solution with 4 ¢.c. one died, after only slight reduction,
after three days. The other exhibited dedifferentiation of a type
very similar to that just considered, but this time accompanied
by a little growth in the stolon, which remained healthy and
tonic. Although reduction had started, resorption never
ensued, and after five days the zooid had died and was repre-
sented by a blackish spheroidal mass about half its original
diameter, while the stolon was still healthy. (Fig. 19.)

In the solution with 2 ¢.c. matters were very similar. The
stolons remained healthy, though distended with blood-cells
(and possibly others) from the zooids, for over five days. The
zooids withdrew their siphons from the test, shrank, and
became opaque (i.e. started to dedifferentiate), but died with
change of colour to brown or blackish before any marked
resorption had occurred.

It will thus be evident that there are at least two factors
concerned in resorption in Perophora. The first is the shrinkage
of the whole organism and reversion of its cells to a cuboidal
type which we may call. simple dedifferentiation, the second
is the migration of cells out of the tissues, which does not take
place, or takes place only to a negligible degree, in the absence
of the circulation.

Thus in the presence of KCl, with consequent cessation of
heart-beat, the aspect of the process is altered in many parti-
culars. High concentrations of KCl damage both zooids and
stolon, and both contract. The cessation of the circulation in
lower concentrations leads to a very speedy dedifferentiation of
the zooid; but this never goes very far before death supervenes,
and is unaccompanied by resorption.

The experiments were repeated, with variations, with
forty-five more specimens; essentially similar results were
obtained. Twenty of these showed dedifferentiation without
resorption. In addition one showed a slight, one a moderate,

660 JULIAN 8. HUXLEY

degree of resorption. Twelve formed new stolon outgrowths of
fair length. The solutions used were 2 ¢.c. and 4 ¢c.c. n/2 KCl
in 50 ¢.c. sea-water.

5. EXPERIMENTS witH Low TEMPERATURE.

Hight vessels, each containing several individuals, were put
in an ice-chest, with a temperature of 3° to 8° C.

Several points were noted when these were examined eight
days later. Over half had turned brown or blackish, and were
dead or dying. No cases of extreme or even considerable
resorption were found. Most healthy-looking individuals had
shrunk and become opaque, i.e. had dedifferentiated. The
opacity was more marked than usual. Usually, however, the
siphons were left open and attached to the test at a stage when
at room-temperature they would have been closed and with-
drawn. The heart-beat was very slow or absent, though the
heart was usually visible. Sometimes the heart-beat began
again soon after transference to room-temperature for examina-
tion. Very young individuals were less dedifferentiated than
older ones.

The stolon seemed to be unaffected, and often remained
of normal appearance even when the zooid was dead or dying ;
no new growth, however, was ever seen. Recovery did not
occur at room-temperature.

Here again it is clear that the roti has been much more
affected than the stolon, and that the slowing or cessation of
circulation has, as in KCl, prevented resorption.

In one system a new bud was produced on return to room-
temperature, and grew to a normal zooid after six days.

6. MisceLLANEous NorTEs.

(a) Tone of Stolon.—The turgescence of the stolon
appears to depend on two quite different causes—first the
physiological condition of the ectoderm cells, and secondly
the pressure of the blood. Observation on a stolon which was
undergoing retraction showed that the ectoderm cells were
capable of great passive extension. At intervals the tip of

DEDIFFERENTIATION IN PEROPHORA 661

the stclon was dilated by the blood-pressure, the flattened
ectoderm cells becoming still more flattened. The test also
underwent passive dilatation.

However, even when the heart has ceased to beat, the stolon
may be quite turgescent, and the ectoderm cells flattened, not
cuboidal. Tullest turgescence, however, is thus only to be
expected when the circulation is active and when the ecto-
derm cells are healthy.

It may be mentioned that the first step in dedifferentiation
may be regarded usually as a diminution of tone (turgescence).

(b) Growing-points of Stolon.—At the tips of
srowing stolons the ectoderm is usually columnar (fig. 26)
and the lumen generally filled with a dense mass of cells,
into which the circulation does not penetrate. Sometimes,
as in fig. 27, there is an increase in the number of green cells
as we pass away from the tip. Often a layer of blood-cells will
become attached to the walls of the stclon over a considerable
distance, giving it an opaque appearance, though circulation
continues internally.

(c) Lateral Outgrowths of Stolon.—Some lateral
outgrowths, as in fig. 25, were occasionally seen. They did
not represent rudimentary branches. Their meaning and origin
is obscure. -

(2) Attachment of Stolons.—The stolons will usually
attach themselves to the substratum, This I] have seen
accomplished within three and a half hours.

(ce) Bud -formation.— When medium-sized zooids
attached to stolons of fairly large size were employed, buds
were often formed from the stolon when dedifferentiation
began in the zooid. Sometimes two buds or more might
form. Buds may form at either or both ends of a piece of
stolon. Resorption might occur at any stage in the develop-
ment of the zooid from the earliest bud up to half-grown
individuals.

(f) Penetration of Zooids by Stolon Branches .—
An individual was seen in which apparently a branch of the
stolon had grown up inside the test of the stolon-connexion and

662 JULIAN S. HUXLEY

encircled the zooid. The actual origin of the branch could not
be traced in vivo. When old zooids die, stolon branches
will frequently grow into the test previously oceupied by the
zooid.

(g) Death-changes.—Death-changes in Perophora usually
involve a change of the green colour to a hard brown or black.

(hk) Change of Position of Stolon.—When a stolon-
zooid system is isolated, and new growth of the stolon with
subsequent bud-formation takes place at one end, not only
may the original zooid be completely resorbed, but the stolon
tissue may abandon the original region and become con-
centrated in the region of the new bud. ‘This ‘ moving-on’
of the stolon is common in regeneration in Hydroids.

(i) Segmentation of Stolon.—In not very dilute
solutions of KCl and KCN in which the stolons were affected,
the stolon-tissue sometimes contracted into a series of separate
ellipsoid portions giving the appearance of a necklace without
a string.

7. EXPERIMENTS ON OTHER SPECIES.

(a2) On Amaroucium.—Some experiments were also
made on a form of compound Ascidian very abundant at
Wood’s Hole—Amaroucium pellucidum, var. con-
stellatum. For information and advice as to this form
I have to thank Professor Caswell Grave.

Twenty small pieces of Amaroucium colonies, consisting
each of from two to twelve or fifteen individuals, were cut
out and placed in separate dishes in a small volume of water.
The experiment was started on July 11 and was terminated
after twenty-nine days. Controls were kept in the circulation-
tanks.

Those kept in the unchanged small volumes of water showed
alterations as follows. The larger pieces remained normal
longer than the smaller. The larger individuals, however,
usually showed reductional changes sooner than the smaller,
ceteris paribus; but they did not usually remain as
healthy as the small ones during reduction. Often they

DEDIFFERENTIATION IN PEROPHORA 663

exhibited a phenomenon characteristic of Amaroucium—the
protrusion of the pharynx from the test and its subsequent
decay, the abdomen and post-abdomen remaining and dedif-
ferentiating. The small individuals underwent a process
obviously analogous to the dedifferentiation of Clavellina.
They shrank in size and decreased in transparency. The
siphons at first remained attached to the test (unlike Clavellina),
but later became completely detached. The pharyngeal
region, as in all other reducing Ascidians, shrank much more
than the rest, and finally a stage was reached in which the
two main portions of the body were still distinguishable,
separated by a shght constriction ; the general shape was thus
that of a constricted sausage ; the organism was completely
opaque, the colour being white with patches of red. (Certain
organs of the normal zooid show this same red colour.) A
curious feature was the frequent formation of clear projections
of the test. These were generally stalked, and spheroidal or
ellipsoidal, like bubbles or bladders. Healthy-looking test-
cells could be seen in them. Very frequently new buds would
be formed from the dedifferentiating zooids during the process
of reduction. These would attain a certain degree of organiza-
tion, but would not usually reach full development unless the
piece were replaced in clean and regularly-changed water.
This replacement in clean water, however, did not lead to the
redifferentiation of the reduced original zooids.

After seven to twenty days, when it had become evident that
it was not possible to obtain the extreme stages of dedifferentia-
tion seen in Clavellina, the surviving pieces were all placed under
gauze in the circulation. When examined twenty-nine days
after the inception cf the experiment it was found that a few
had remained in approximately the same condition im which
they had been placed in the circulation. More than halt,
however, while the original zooids had not redifferentiated,
had given rise to now zooids, usually in one or two clusters of
four to six zooids each.

It thus becomes clear that Amaroucium shows yet a third
type of dedifferentiation. The specialized method of forming

2

664 JULIAN 8S. HUXLEY

a large number of buds practically simultaneously by segmenta-
tion of the very long post-abdomen, with subsequent differentia-
tion of each segment to form a whole zooid, is apparently
responsible for this. After dedifferentiation of the primary
zooid has proceeded a certain way, either death supervenes
or else the post-abdomen, released from subordination now that
the dominant region is thus adversely affected, manifests its
independence by producing new individuals. Once these new
individuals start to develop they become dominant. The non-
recovery of the partially-dedifferentiated original zooids may
be ascribed to this, or to greater susceptibility. In spite of
this absence of the power to redifferentiate the process of dedif-
ferentiation is very similar to the early stages of the same
process in Clavellina. J*or such behaviour there is ample
evidence as regards numerous forms reproducing asexually in
the work of Child and his pupils (Child, 19156). We may thus
say that, under the conditions which prevail in the colony,
or in pieces of it, in Amaroucium, complete dedifferentiation
of single zooids is not possible. The colony or piece regarded
as a whole, however, may be said to undergo dedifferentiation
followed by redifferentiation.

Oozoites.—These had the advantage over blastozoites
that they could be obtained singly. They were got by allowing
larvae to metamorphose in the laboratory. They could be
induced to dedifferentiate either by lack of change of water,
or, after a longer period, by starvation. The process was very
similar to that in the blastozoites, with the exception that the
formation of buds was never observed. ‘This latter fact is
undoubtedly to be correlated with the small relative size
of the post-abdomen and the small absolute size of the whole
organism.

Here, too, dedifferentiation never got beyond a stage im
which a sausage-shape was assumed (fig. 27, Text-fig. 1).
The complete opacity and the spheroidal shape of the final
stages of the process in Clavellina were not observed ; neither
did I succeed in obtaining redifferentiation.

On the whole, dedifferentiation in oozoites went a little further

DEDIFFERENTIATION IN PEROPHORA 665

than in blastozoites, and appeared to be a healthier process
unaccompanied by so many abnormal swellings of the test,
extrusions of parts of zooids, phenomena of local decay, &e.
Treatment with Alcohol.—A few experiments were
made to test the effect of a 2 per cent. solution of alcohol
on the process. Jt appeared that under its influence, dedif-
ferentiation, both in oozoites and blastozoites, started sooner

Text-Fia. 1.

Reduction in oozites of Amaroucium.

A. Zooid in stage 3, test spherical, test of tail degenerating,

B. As A, except that the zooid shows detached cell-masses, and
lies in a spherical portion of test detached from the rest.

than in the controls, but that it did not progress in a normal
way. Opacity might be attained, but the loss of form, especially
in the pharynx, was not as great as usual, e.g. the siphons
remained visible relatively much longer after their retraction
from the test than in normally-reducing specimens (fig. 28).
This appears to indicate that there are two distinct pro-
cesses at work in normal dedifferentiation, the first bemg a mere
shrinking as a result of exposure to an unfavourable environ-
ment, the second a real despecialization of the cells, resulting
in loss of typical form. This latter then is due to active positive

666 JULIAN S. HUXLEY

changes in the cells, changes which are partially restrained
by the action of a narcotic like alcohol.

It should perhaps be mentioned that not only oozoites which
had lived some time in the circulation, but also those which had
only just metamorphosed, could be induced to dedifferentiate.
Larvae were allowed to fix on slides on July 28. After seven
to nine days in the laboratory they showed the first signs of
reduction. On the tenth day their water was changed, but
without effect on the result, for on the eleventh and twelfth
days all were markedly reduced. A sausage-shaped mass,
sometimes showing a slight constriction between pharynx and
abdomen, lay in a much-swollen, but healthy, test, which was
usually attached to the substratum in the form of a flattened
sphere. The remains of the test of the larval tail could be seen
attached to one point of the main test. In some examples
an interesting modification was observed—a small portion of
test surrcunding the reduced zooid became constricted off
from the main portion, which, though thus empty, remained
healthy (‘Text-fig. 1, B).

In one or two specimens, detached, rounded masses of cells
were to be seen outside the limits of the reduced zooid ; these
were also occasionally seen in reduced blastozoites. I believe
them to have been derived from the organism itself, and not
to have been merely collections of cells of the test. Such
collections were also seen, but never had the compact appear-
ance of the first-mentioned masses.

Swellings of the Test.—These have been already
referred to. In connexion with experiments on dissociated
sponges which were proceeding at the same time, it was decided
to see whether portions of test were capable of re-organization
or of regeneration in sea-water, or of growth in a nutrient
solution. Accordingly a number of these ‘ test-bladders ’
were snipped off and isclated. After cutting the pieces were
always tern and quite flabby. Some were placed in sea-water,
others in weak solutions of peptone made up either in tap-
or sea-water. In all cases a marked reorganization had taken
place within twenty-four hours. The wound was completely

DEDIFFERENTIATION IN PEROPHORA 667

healed, and the piece was irregularly lobed, as if swollen
out from two or more centres. It would appear that there was
an actual accumulation of fluid in the interior, as in the spheres
produced by sponge choanocytes (Huxley, 1921). Usually the
test-cells were nearly absent in some regions, rather densely
aggregated in others. No regeneration, however, took place,
and death occurred quicker in the peptone than in the sea-
water. Death took place in one to three days in peptone,
two to four in water. An interesting point was that, before
death, many of the test-cells always left the matrix of the
test, and crawled out on to the bottom of the dish. They still
preserved their characteristic shapes at first, but eventually
all rounded off preparatory to dying.

(b) Botryllus.—tThis genus is unsuitable for experiment
owing to the small size of its zooids and their intimate connexion.
One system, however, was seen in which all the zocids had
become reduced te shapeless but healthy-looking lumps ;
the test round them had degenerated save for a thin layer.
Some form of dedifferentiation had cbviously occurred.

8. Discussion.

Perephora happens to be an organism in which dedifferentia-
tion and resorption affect the whole individual in a very striking
way. In higher forms, thanks to thei self-regulating
mechanisms, their size, the bulk of their skeletons, and other
factors, the processes do not affect the individual as a whole.
None the less, similar processes play a large part in many
phenomena, both normal and abnormal, throughout the animal
kingdom.

In the first place it is important to realize that the ‘ struggle
of the parts’, to which Roux (1881) first drew attention, is
a very real struggle ; that the organism is in one aspect simply
an equilibriam between a number of parts, some in a relation
of simple competition, some in a relation of control of or
subordination to others; and that the relative success or
failure of any one part, the degree to which it is developed,
depend or have depended upon its success in this struggle.

668 JULIAN S. HUXLEY

Secondly, we must realize that success in the struggle, i.e. time
and degree of development, may depend largely on rate of
metabolic activity. It is not for a moment suggested that this
is the only factor at work, nor that it is the most important
factor (in the higher organisms the relationship of the nervous
system to the tissues of course masks it to a considerable
extent), but that it is an important factor.

Child (19156) has drawn attention to its importance for
problems of regeneration and asexual reproduction; he finds that
the most actively-working portion of the organism (or, in higher
forms, the portion containing the higher centres of the nervous
system) is net only formed first in regeneration, but exerts
some sort of controlling effect upon the rest of the organization
of the body. For instance, once a head is formed in the
regeneration of a Planarian or an Oligochete the old organs
are remodelled, some being broken down, others built up,
until what exists stands in normal relation to the new head.
But if, for some reason or other, a head is not formed (in
Planaria it can be experimentally prevented from forming),
then this remodelling does not occur. The production of a new
pharynx, for instance, in a pharynxless posterior half of
a Planarian, will not take place unless a head is formed at the
anterior end.

However, this controlling effect of the head is only exerted
up to a certain distance. Once this distance is overpast
the tissues of the bedy are free to react in the way characteristic
for them when not under any control, ie. by the formation
of a new head. In other words this control or dominance of
the head or oral end (or apical bud in plants) is what regulates
the important temporal and spatial relations of asexual
reproduction. As is to be expected, it varies with external
circumstances, and Child has performed some pretty experi-
ments on the experimental control of dominance.

It would appear, especially from some of his recent work
upon plants, that this dominance exerts an effect analogous —
to that of the nervous system by means of some form of
conduction, and that it is not, as might at first be expected,

DEDIFFERENTIATION IN PEROPHORA 669

sunply dependent on nutritional relations. Child has naturally
stressed this important point (Child, 1919). However, the
nutritional aspect is also important, and does as a matter of
fact determine many relations of dominance and subordination
of parts in organisms ; and it is to some of the implications of
this aspect that I wish to draw attention.

If one reaction or associated set of reactions is proceeding
faster than another, in the same system, it will occur to
a correspondingly greater extent; cf. Mellor, ‘ Chemical
Statics and Dynamics’, p. 70: ‘In any system of parallel
chemical reactions which consume the same substrate and
are proceeding simultaneously in a mixture, the extent to
which each reaction will occur is proportional to its velocity.’
This means that if two sets of reactions are going on in an
organism at an equal rate and that subsequently one of them
is stimulated to a 10 per cent. increase, then the end-products
of those reactions (the amounts of two different types of tissues,
let us say) will, if the available food remains constant, change
from the proportion 1000:1000 to 1048:952. <A similar
result will occur if the other reaction’s intensity is correspond-
ingly lowered. This is important in explaining many changes
resulting from a change in environment acting upon the tissues
which respond to the change at different rates (see Child, 1916 ;
Robertson and Ray, 1920; Lillie and Knowlton, 1902, &e.).?

One of the best examples is the relation of head-size to
body-size in a regenerating piece of Planaria. Apparently
the temperature-coefficient of the processes of the head-region
is greater than those of the body, for the relative development
of head increases with temperature. It is also decreased by
increase in concentration of narcotics.

There is, however, another aspect of the question which it
is rather more difficult to understand. That is the fact that if
in an organism two sets of reactions are going on at different

1 This will, of course, only occur up to a certain limit. A condition of

tion, or by excess of nervous stimulation in certain forms of neurasthenia,
which results in a wasting of the tissues concerned,
NO. 260 yy

670 JULIAN S. HUXLEY

rates, in two different regions, then, if the food-supply is
reduced, the one which, ceteris paribus, has the higher
speed will be able to maintain itself in its nermal state and at
its normal level This may be due to the fact that the
assimilatory processes are reversible; this would imply that
not merely the dissolved food-substances in the bedy-fluid
are to be regarded as the ‘ substrate’ from which the various
reactions draw their materials, but that this substrate must be
taken as including the tissues themselves. Jf, therefore, two
reversible reactions A and B were proceeding simultaneously
in two regions of an organism while the organism was starved,
we should have each reaction making demands upon the end-
products of the other, i.e. upon the tissues of the two regions.
Four processes would therefore be involved—first and secondly,
the reactions A and B proceeding in their normal direction ;
thirdly, B proceeding in reversed direction in response to the
demands of A; and fourthly, A proceeding in reverse direction
in response to the demands of B. Since A’s speed is greater
than Bb’s, the end-product of A will contmue to increase,
while that of B progressively diminishes. We can represent
such a state of affairs symbolically thus: P =, X 27 Q,
where P is the end-product of A, Q of B, and X the common
substances utilized by both. If the rate of formation of P is
ereater than that of Q, the reaction will proceed until no Q
remains.”

However that may be, we are confronted with the fact that
if two reaction-systems are competing in the organism for an
amount of nutriment which is not sufficient for both, then
the more rapid, or the one which subserves the more highly-
differentiated region, will not only get first call on the available
nutriment, but will actually nourish itself at the expense of the
other.

1 This again is masked in higher animals by the fact that the nervous
system, apparently owing to its controlling and co-ordinating function,
has come to be the system least affected by starvation.

* Similar ideas are put forward by Runnstrém (1917) in his important
paper on dediffsrentiation in Echinoid larvae, to which unfortunately
(owing to the war) I have only just had access.

DEDIFFERENTIATION IN PEROPHORA 671

This is particularly well seen in malignant tumours, which
will continue to grow at the expense of the rest of the body,
even when this is in a condition of relative starvation. In
tumours derived from adipose tissue, the tumour-cells may be
full of fat after all vestige of fat has disappeared from the
normal tissues. The fact that regeneration will proceed
actively and normally in starving Planarians exemplifies the
same state of affairs.

We saw above that we should expect the reaction to proceed
to a limit, all the product of the slower process being utilized
by the faster. As a matter of fact this limit can often not be
reached, since life is not possible when the ‘ subordinate ’ region
is absent, or else its reduction brings about subsidiary changes.

It is also complicated by the supervening of dedifferentiation.
That starvation can produce dedifferentiation has been shown by
Sehultz (1906), by Runnstrém (1917), &e. It therefore follows
that the tissues of the less active region will usually, as a result
of the starvation induced, reach a stage at which they are
unable to maintain themselves, and will start to dedifferentiate.
In the dedifferentiated state they will possess a still lower
rate of metabolic activity, and so the resorption-process will
be accentuated.

Next we meet with the fact of differential susceptibility.
This is a corollary of difference in rates of reaction. The more
highly-differentiated region and system, or the one with
higher metabolism, will be, ceteris paribus, more sus-
ceptible to unfavourable conditions. If it is placed in a toxie
solution, for instance, it will enter into reaction with more of
it in a given time than will a slower system. There are a number
of complicating factors (such as acclimatization) which enter
into the problem, but, broadly speaking, we may say that
a more highly-differentiated and more active system will
be relatively more interfered with than a less highly-differen-
tiated and less active system.

After a certain point of interference is reached, dedifferentia-
tion will set in. Dedifferentiaticn is the primitive reaction of
organisms to unfavourable circumstances. More energy is

yy 2

672 JULIAN 8S. HUXLEY

necessary to maintain a cell in a differentiated than in a dedif-
ferentiated condition. This is especially clear when, as in
most instances, differentiation involves an increase in the
surface of the cell relative to its bulk; here the maintenance
of differentiated form alone involves the expenditure of more
energy. Thus when the processes of life are interfered with
by unfavourable agencies, the cell is unable to continue to
produce the energy necessary for the maintenance of its
ditferentiated state, and must either die or dedifferentiate.

The main characteristics of dedifferentiation are the follow-
ing :

(a) Cells revert, if isolated to a spheroidal, if in epithelia to
a cuboidal form.

(b) Cytoplasmic differentiation is lost.

(c) Organs containing cavities revert to simple spheroidal
sacs. Junctions between organs are often broken.

(d) Apertures usually disappear altogether.

(ec) The whole organism diminishes in size and reverts to
a spheroidal form, owing to the form-changes in its constituent
cells. This has the effect of increasing the opacity and density
of the organism.

Once dedifferentiation has started in any region or system
the previous level of metabolic activity m that system is in-
evitably much reduced. ‘Thus, if dedifferentiation occurs in
a dominant and not in a subordinate system, this dominant
system will lose its dominance and become subordinate.
Such alteration of equilibrium by unfavourable agencies we
may call differential inhibition; it is a corollary of
differential susceptibility. Differential inhibition need not,
however, involve dedifferentiation, nor reversal of dominance.
In a growing organism unfavourable agencies will depress the
srowth of the dominant or more active regions relatively
more than that of the rest, and we shall, as outlined above
(p. 669), get a decrease in size of the former, an increase in the
latter—a decrease and an increase which will be absolute as
well as relative. This is illustrated in some of Child’s experi-
ments (see later).

DEDIFFERENTIATION IN PEROPHORA 673

A chemical analogy, for which I am indebted to Mr. H. R.
Raikes, of Exeter College, Oxford, may help illuminate the
point. If one equivalent each of hydrochloric acid, boric acid,
.and ammonia are mixed, a negligible amount of boric acid will
react with the ammonia owing te its small degree of dissocia-
tion. We may say that the hydrochloric acid is completely
‘dominant ’ in the system, owing to a greater speed of reaction.
If, however, the mixture is heated, the more volatile hydrochloric
acid will be driven off, and the less volatile borie acid left to
react with the ammonia. This we may call ‘ differential sus-
ceptibility’ (to rise of temperature) involving ‘ differential
inhibition’ of one portion of the system, and consequent
‘reversal of dominance’. If the mixture were contained in
a very large closed space, cooling after heating would restore
the original ‘dominance’ of the hydrochlorie acid, giving
a parallel to reversible dedifferentiation.

The emergence of the cells from the tissues in dedifferentia-
tion is a phenomenon which deserves further study. Though
probably by no means universal it is doubtless commoner
than is generally assumed. It occurs not only in Perophora
but also in Hydroids, in Turbellarians, and in Echinoderm
larvae, and in many cases of actual poisoning, e. g. by mercury
salts (Child 1917, Huxley 1921b) and other agencies (Gray 1920).

Once the cells start to emerge they may collect close to their
place of origin, or if space and means of transport are available,
be removed to regions at a distance. When the stolon portion
is large in a Perophora stolon-zooid system, and the heart is
beating normally, the latter is the case ; it is also the case in
Hydroids when the coenosareal portion is large in comparison
with the hydranth. ‘The difference between the two possi-
bilities appears to be similar to that between a reversible
chemical reaction when the end-products are not removed,
and the same when they are removed. Why, in the first case,
the tissues should not simply resolve themselves into their
constituent cells in situ is difficult to see, but the fact
remains that they do not (e.g. Clavellina ; Perophora with
very small stolon attached, or with circulation stopped by KCI).

674 JULIAN 8. HUXLEY

In ordinary organic systems, therefore, we must recognize
that we may have to deal with any of the following pheno-
mena :

(1) Physiological dominance and subordination of parts,
manifesting itself first as regards conduction and control of
asexual reproduction, secondly as regards nutrition.

(2) Differential susceptibility.

(3) Dedifferentiation.

(4) Differential inhibiticn.

(5) Resorption.

(6) Reversal of dominance.

Parallel phenomena occur in other organic systems in which
parts are related in equilibrium. Thus dominance, subordina-
tion, differential susceptibility and inhibition, a form of dediffer-
entiation, and reversal of dominance, also occur in psycho-
physical systems, both in some where consciousness 1s involved
and in some where it is not, as will be dealt with more fully
later. Here the dominance may be called neurological and
psychological, and the dedifferentiation is of course unaccom-
panied by physical dedifferentiaticn of nerve-tissues.

Many of the phenomena of inhibition, e. g. of buds by growing
tips in plants, ard within the central nervous system, ebviously
depend upon relations of dominance and subordination. A few
examples will perhaps serve to illustrate some of these general
statements.

We may start with the example already referred to, of
Stenostoma (Child, 1904), since here dominance and resorption
are very Clearly shown. When a solitary Turbellarian is divided,
regeneration of a head usually occurs from the anterior cut
surface. In Stenostoma, however, which is a chain-forming
organism, if a cut is made across the body of one of
the central zooids, such regeneration from the anterior cut
surface does not take place. Instead, the half-zooid which is
attached to the anterior end of the posterior half-chain will
shrink, assume a more rounded form, and eventually disappear
altogether.

Not only this, but the relative age of zooids determines

DEDIFFERENTIATION IN PEROPHORA 675

dominance. In Stenostoma, fission occurs according to a regular
system, so that the relative age of each head-region in a chain
can be determined. If now a cut is made so that a younger
zooid is left in front of an older zooid at the anterior end of
a piece, this younger zooid, though morphologically complete,
will be resorbed by the posterior. If completely isolated from
the posterior zooid the younger one would have been capable
of leading an independent and ncrmal existence, so that
the age-relation of zooids clearly determines dominance.

During the process, ‘ disintegration ’ (presumably migration
of the cells from the tissues) of the subdermal structures
occurs, and the pseudoceel becomes filled with cells and
granules. The posterior undestroyed zooid grows more rapidly
than usual, apparently because of the excess of nutriment thus
provided (although this nutriment is in the pseudocoel and not
in the gut). Child did not undertake a histological examina-
tion. From his observations in vivo, however, it is clear
that the cells migrate out of the tissues, as in Perophora.
The most highly-differentiated organ, the pharynx, disinte-
grates very early. The intestine, however, does not do so until
late. From the ectoderm of the resorbed portion a very
gradual migration probably occurs. The portions undergoing
resorption are wrinkled and collapsed.

The reversibility of the process is shown by the following
observations. If an older posterior zooid has in front of it
another almost as old, resorption will begin, but fission will
oceur before it has finished, and the two zooids will separate ;
after this the anterior zooid redifferentiates. The converse
of this is seen when a long anterior fragment is present. In
this case the beginning of regeneration occurs, but reduction
finally takes the upper hand, and the whole fragment is
resorbed. The rapidity of the change is noteworthy, complete
resorption usually occurring in twenty to thirty hours. Pro-
vided that the brain-region of an anterior fragment is absent.
resorption will occur; even when a system consists of a very
long but brainless anterior fragment, and only the brain-region
of the posterior zooid, resorption happens.

676 JULIAN §. HUXLEY

To sum up, whenever in Stenostoma a system is artificially
produced in which a posterior brain-region is older than any
brain-region anterior to itself, or has a brainless region anterior
to it, resorption of such anterior regions will start; it will
be completed unless fission of the system occurs during the
process, which only happens when an anterior zooid is far-
developed.

A brain-region is physiologically dominant over all
other tissues of the same and other zooids, and over all younger
brain-regions than itself. When a region comes to lie anteriorly
to a physiologically dominant region, it cannct maintain
itself, and is resorbed. Antagonistic to resorption is the
process of regeneration (morphallaxis). Both processes often
start simultaneously in a fragment ; which of the two even-
tually gains the upper hand is determined by the age of the
fragment. The systems resemble the stolon-zooid systems of
Perophora, except that the different members of the system
are all similar to each other except in age. Further, reversal
of the effects by altering the environment has not been
attempted. This would provide an interesting field for
experiment.

As the facts stand, the dominance is caused entirely by the
internal factor of physiological state due to (1) presence and
(2) age of brain or brain-region, and resorption is produced when
a part is caused to lie in an abnormal position relatively
to a deminant region. As Child points out, similar resorption
of parts in abnormal positions is frequently seen in grafting
experiments in Hydra and Planarians. Subordinate portions
in a normal position relative to a dominant region do not
of course become resorbed.

Once more the essential fact is that, in certain conditions,
parts of a system are unable to maintain themselves
of their normal size or their normal form, and, once they start
dedifferentiating, become subordinate in the system, and can
be used as food for the remaining dominant part.

I suspect that investigation would show that the first change,
here as in Perophora, is the loss of the normal cell-form of the

DEDIFFERENTIATION IN PEROPHORA 677

differentiated organs of the subordinate region, and _ that
resorption follows upon this.

Numerous other cases of tissues, regions, and whole organisms
being unable to maintain themselves as such in changed
circumstances are known. Of these may be mentioned the
degeneration of muscle-fibres when the nerves supplying them
are cut. Here the ‘ normal environment ’ apparently includes
constant nervous stimulation, and in the absence of this
the elaborate structure of voluntary muscles cannot be main-
tained in equilibrium. Similar dedifferentiation of muscle-
fibres takes place in the stump of an amphibian limb which has
been cut off preparatory to regeneration (Towle, 1901).

In the interesting studies of Child on differential imhibition
during development we do not get the total disappearance of
one part of the system, but merely a change in the proportions
of the various parts. The simplest example studied was the
effect of dilute poisons upon the development of the marine
Polychaet worm Chaetopterus (Child, 1917).

He found that during the earliest stages of development
the apical region of the egg and blastula is the most susceptible
to various poisons, in certain concentrations a regular death-
gradient being obtained from the animal to the vegetative pole.
By the time the early trochophore larvae has been produced,
however, a new development occurs ; the posterior (previously
vegetative) region suddenly becomes highly susceptible, its
metabolic rate being raised apparently in preparation for the
active growth-processes that are about to occur in this region ;
for the formation of the permanent growth-zone, from which
all the body-segments of the adult worm will be produced,
takes place here.

The death-gradient will now advance from the two ends of
the larvae to meet in the middle region, which, with its lower
metabolic activity, survives the effects of the poison longer
than the rest. In the later larva the anterior region is differen-
tiated as a head with ciliated band and apical tuft; and
posteriorly there is a well-defined growing-region, with a small
posterior prolongation,

678 JULIAN 8. HUXLEY

Immersions of the fertilized egg in solutions of poisons so
dilute as to allow development to proceed, while yet exerting
an influence on the more susceptible parts of the organism,
give the following results. (Essentially similar facts were
discovered for other Polychaetes (Nereis and Arenicola).)

Immersion continuously up to the late larval stage gives
a form with both anterior and posterior regions smaller and
less differentiated than the normal. The middle region is either
almost as large, and of the same form as the normal, or else
considerably distended. ‘This latter condition implies possibly
that the cells of this region have been able to develop practi-
cally normally. The anterior and posterior regions are not so
active as normally, and hence are not able to make use of so
much of the yolk; there is thus more for the middle region,
which is capable of utilizing it, and secretes an excess of fluid.
If immersed for eleven hours only, and then replaced in sea-
water, the apical region is small, but the growing region as well
as the middle region is nearly normal. If, on the other hand,
the development is allowed to proceed in sea-water for twelve
or twenty-four hours, and the larvae are then placed in the
solution, the apical region, having been completed before
immersion in the toxi¢ solution, is normal and the posterior
end is much affected.

In another paper, giving an account of similar experiments
on Hehinoderms, he makes an interesting suggestion to account
for the great over-development of the skeleton often found in
larvae which have grown in dilute solutions of toxic agents.
‘The mesenchyme cells appear to be least susceptible, and thus
when the other cells of the organism are inhibited, can obtein
a greater quantity of food, which results in a multiplication
not only of themselves but of the products of their activity,
i.e. the skeleton (Child, 1916).

A recent important attempt to apply similar principles has
been made by Robertson and Ray (1920, where reference to
earlier papers are given).

tobertson found that mice to whose diet had been added
tethelin from the anterior lobe of the pituitary, showed first

DEDIFFERENTIATION IN PEROPHORA 679

a retardation of growth in weight, then an acceleration,
and finally lived about 12 per cent. longer than normal
controls. Other experiments had led him to conclude that
tethelin (cr pituitary extract) caused increased growth in
cellular tissues, a conclusion strengthened by the recent
grafting experiments of Allen (1920) on tadpoles. His explana-
tion of the facts is as follows. Tethelin causes at first an
absolute increase in the growth-rate of the cellular tissues of
the body ; this involves, as we have seen, a relative decrease
in the weight of the supporting tissues. Since these latter
are the heavy tissues, this involves an absolute decrease in
total weight. Eventually, however, the characteristic relation
between the amounts cf cellular and supporting tissues is
established, but later than normal. WHelative increase of the
supporting tissues characterizes old age; and the onset of
senility is delayed by that period by which the establishment
of the cellular-supporting balance was postponed. ‘The
reason for the more rapid growth of the cellular tissues at the
beginning is that the tethelin stimulates them to greater
activity, and that consequently they obtain first call on the
available foodstuffs.

This view-point, it will be seen, is very similar to that of
Child.

A beautiful example of differential inhibition depending only
on the two quantitative factors of size and distance is given in
the interesting paper of Detwiler (1920; see especially pp. 149—
51). Detwiler transplanted the limb-rudiments of Ambly-
stoma autoplastically, cutting the rudiments out and trans-
planting them a varying number of segments posteriorly
from their normal position. The experiments were undertaken
at a stage when the rudiments were represented only by
circular thickenings of somatopleuric mesoderm in segments
3-5. He found, as had previous workers such as Harrison,
that in many cases the rudiment was not completely excised,
a few of its cells being left in the normal position. When this
was so, these cells usually begin to regenerate on their own
account. It is of interest to note that this regeneration is

680 JULIAN S. HUXLEY

greater when the wound is not covered—a result presumably
due to the greater stimulation which the unexcised limb-cells
then receive (Harrison, 1915).

After a short time a small nodule of cells begins to protrude
from the bedy in this region. If the main limb-rudiment is
completely removed the nodule may grow into a perfect limb.
When, however, the main limb-rudiment is transplanted
less than four segments back on the same side, these nodules,
after growing a longer or shorter time, begin to shrink, and
eventually disappear altogether. When the limb-rudiment was
only transplanted one segment back the nodules appeared after
about four days, but very speedily began to decrease and had
disappeared after eight days. When the limb was transplanted
two segments back the nodules continued to increase till the
fifth or sixth day, and had disappeared by the eleventh day ;
when the distance of transplantation was three segments,
nodule-growth continued until the tenth or eleventh day,
when the ‘nodule’ was almost as large as the transplanted
limb; but after this, decrease set in, and all nodules eventually
disappeared, although not until the eighteenth to twentieth
day. Finally, when the main limb-bud was removed more
than three segments from its original site, the regenerating
nodules always developed into a normal appendage, so that
two limbs were produced from the one original rudiment.

The cells of the imb-bud constitute an equipotential system,
as Harrison has shown. It is therefore clear that the inhibiting
effect exerted by the main transplanted rudiment on the cells
left at the original site must be due simply to the greater size
of the former. The strength of this ‘ dominance ’, however,
also depends upon the distance of the two systems ; and when
this distance is increased beyond a certain limit, there is
no longer any inhibitory effect. If we like, we may say that
the reason why the cells constituting the normal limb-rudiment
of Amblystoma do not usually form more than one limb is
that they occupy such a small area that any one rudiment
growing within that area inhibits the growth of any other.

Detwiler did not investigate the actual mechanism by which

of mci <7e%

DEDIFFERENTIATION IN PEROPHORA 681

the “nodules” decreased in size, and leaves it open as to
whether the cells composing them are actually translocated
into the main limb-bud, or are simply resorbed into the body.
The former view is less probable on general grounds, and the
latter is supported by the facts of resorption in Perophora.
The lmitation of physiological dominance by distance has
already been brought out by Child (1915 a, chap. 5), but is here
particularly well illustrated. The relation of dominance to
simple size-difference between two portions of otherwise
identical tissue hag not, however, so far as I am aware, received
any special attention, but is obviously of considerable theoretical
importance. Further, in no other case with which I am familiar,
is the importance of purely quantitative relations so well
brought out. It is perfectly clear that inhibition and conse-
quent resorption can take place at any stage of growth of the
‘nodule’ (regenerating limb-rudiment), and that it is not due
to anything in the nodule itself, but entirely to its relations
with a second developing system.

We now pass to the very different field of neurology and
psychology.

In recent years the phenomenon known as mental regression
has been carefully studied. Patients suffermg from this
return to an earlier stage of mental existence. Grown men
may show the behaviour and the mental processes of boys of
ten or five or even younger. A review of our knowledge of
this condition is given by Nichol (1920).

When properly analysed this state of affairs would seem
definitely to be due to the presence, in individuals affected by
it, of two competing systems of mental organization, i.e. of
two possible main channels for the flow of ‘ nervous energy ’.
(I purposely use this latter somewhat vague but non-committal
term to emphasize the fact that the existence of competing
systems and of some form of activity transmissible along their
paths is all that we need to assume for a preliminary discussion
of the problem.) In normal conditions the adult system is
dominant, the main flow of nervous energy is along its paths,
and the childish system or systems are dormant, existing for the

682 JULIAN S. HUXLEY

most part only as potentialities of ection. Under severe
stress (e.g. modern warfare, prolonged worry, &c.), the adult
system becomes in some way affected. It is no longer so easy
for the nervous energy to flow aleng its paths. Under these
conditions there is more nervous energy available for the other,
juvenile, system, which has remained undamaged. Finally,
there will come a moment at which the balance is so altered
that the adult system ceases to be deminant, and the poten-
tialitv of the juvenile system is transformed into actuality.
The juvenile system now becomes dominant in its turn, and
the adult system retreats into potentiality. During recovery
a remarkable picture is presented: the two systems are almost
equally balanced, and we get—not a blending of the effects of
both—but a rapid alternation, first one and then the other,
the two never co-existing. A somewhat similar state of affairs
exists in Perophora ; once absorption of either portion has
started it proceeds rapidly. Alternation, however, is not
possible, simce in Perophora it is structure, and not merely
possibility of function, that is being destroyed.

In the neurological cases structure is not destroyed. Further,
the rapidity of change from the dominance of one system to that
of the other is enormously more rapid, since this is apparently
accomplished simply by the passing of a threshold-value.
Once this is passed a sluice is opened, and a different neural
system flooded so as to permit of function. For this sudden
appearance of one or the other sub-system some~ psycho-
therapeutic writers use the expressive term ‘ puffing-up ’.
It is a well-known phenomenon of convalescence in such cases.

Such occurrences are one aspect of the general principle
laid down by Hughlings Jackson, that, as the result of lesion,
‘ dissolution occurs first in the most highly-organized products
of neural or mental activity, leaving the more lowly at liberty
to express themselves freely in the resulting symptoms’. This,
however, only stresses the aspect of differential inhibition, not
that, of equal importance, of intra-organismal struggle.

Part of this latter aspect of the question is expressed, how-
ever, by Head (1918), who lays down as one of his general
principles of neurology that * Integration of function within

DEDIFFERENTIATION IN PEROPHORA 683

the nervous system is based on a struggle for expression between
many potentially-different activities ’. Integration of function,
however, is not all. A number of integrated minor systems
may exist, one in actuality, the rest in potentiality, in the
developed human psycho-neural system as a whole ; and there
is also a form of struggle between them. The particular type
of mental disorder known as regression is only one special
case of the results of differential susceptibility among two or
more such minor systems. In other so-called neurasthenic
eases the second, normally-suppressed system may not be a
system of childish memories, but an imaginary ‘ ideal’ world
of thought along whose paths consciousness flows instead of
along those necessary to maintain adaptation to everyday
life; or else it may be the system of ‘ negative’ emotions,
leading to depression and possibly to suicidal attempts.
Dissociation of personality and subsequent alternation of the
sub-personalities may also, though less directly, be included
under the same rubric. Rivers, in a recent work (1920), has
emphasized the same point of view ; he points out for psycho-
logical systems what I have drawn attention to in this paper
for physical systems—that reversal of dominance in a balanced
system may occur either through the action of unfavourable
agencies on the dominant system (differential inhibition) or
of favourable agencies on the subordinate system (differential
stimulation).

In a case of regression mentioned by Dr. W. MacDougall and
Dr. Hadfield in their lectures and confirmed to me in conversa-
tion by Lt.-Col. Good, of Ashhurst Hospital, a young man
actually regressed to the condition of an infant.1_ He was
unable to talk or walk, and could tolerate no food except milk.
(By some freak of the nervous mechanism two associations
and two only remained from adult life: if a cigarette were
offered him he would light and smoke it ; when shown a horse
or a picture of a horse, he would get astride of some object
and ‘ tchk’ as if encouraging a horse. It turned out that he

1 Since the above was written, I find that an account of this and

similar cases has been published by MacDougall in ‘Journ, Abn, Psych.’
15, 1920, p, 136,

684 JULIAN S. HUXLEY

had been a jockey.) His recovery was interesting for various
reasons. The intolerance for all diets save milk he lost earlier
than the other infantile symptoms. As regards purely mental
symptoms his growth or redifferentiation was gradual and
progressive, though with considerable rapid oscillations.
It is therefore clear that the picture is not quite as simple
as I have drawn it above. Each stage is really in some ways
dominant to the one below, subordinate to the one above,
and if there has been a considerable degree of regression, the
redifferentiation must apparently be by steps (although the re«
gression itself is a sudden instantaneous process). In the normal
adult each lower stage is kept inits proper place in the hierarchy,
and most of the associations and types of reactions connected
with it exist in posse only. When it is released from the
inhibitory contrel of the processes associated with higher stages
it becomes dominant, and then these potential associations,
memories, and reactions become actual and functional again.
Normally, since each stage of growth represents a necessary step
towards the next stage, some of the reactions of each stage are
functional even in the adult, as foundations for normal adult
activity; but they are altered by the dominant higher processes
to a form different from that which they would have if released
from control. This is parallel, though not identical, with the
behaviour of dominant and subordinate regions in regenera-
tion (see later). Regression takes place suddenly to that
stage whose system has been encouraged ; if the patient has
dwelt upon a particular time of childhood, to the system
associated with that time; if he has dwelt on mere release
from control, to an infantile stage. But recovery must be by
gradual building-up, as in physical development.

Individual mental development is thus an epigenetic pro-
cess ; and the different stages of this development are arranged
in a functional hierarchy or series in which each stage 1s
dominant to the one below, subordinate to the one above.

' The alternation of dominance seen in dual and multiple personality

(Prince, 1908, 1920) is presumably based upon essentially the same principles,
the difference being that typically the two systems are very evenly balanced,

DEDIFFERENTIATION IN PEROPHORA 685

We shall now see that similar relations may exist in non-
conscious neural processes, of which the lower have never
been fully dominant in ontogeny (though possibly in phylogeny).

This is well shown by the observations of Head and Riddoch
(1917) on the activities of ‘spinal man’. ,They found that when
the spinal cord was completely divided, the reflex activities
which manifested themselves after the initial shock-period
were very different from those occurring in the uninjured
individual. In the normal person the activities of the spinal
cord are modified by influences reaching it from pre-spinal
levels. The isolated spinal cord, however, responds to stimula-
tion predominantly by a type of ‘ mass-reflex’ not normally
seen in man. In‘ spinal man’ any form of nocuous stimulation
to a hind-limb causes not merely flexion of the limb stimulated,
but violent flexion of both limbs, abdominal contraction,
voiding of the contents of the bladder if the contained fluid
is above a certain very small volume, and sweating. Con-
versely, injection of the bladder with fluid induces a flexor
spasm of the lower limbs, combined with sweating. (The
reaction may be called an excessive and non-discriminate
reaction to harmful stimuli, resembling in many ways that
seen in certain lower animals, e.g. the toad, in which voiding
of the bladder accompanies limb-flexion when the animal is
alarmed by handling.) The same mass-reflex also appears in
higher forms and in man himself when the higher centres are
put out of action under the influence of an excessive degree of
an emotion such as fear (differential inhibition). ‘The mass-
reflex may be looked on as a very primitive response of the
organism to nocuous stimuli.

In higher forms the mass-reflex has become subordinated to
the influence of other types of reaction ; among these are the
postural reactions and the conscious direction of movements
of escape. Head and Riddoch found that so long as any

and both adapted (though incompletely) to adult life. The emergence of
the juvenile personality ‘Sally’ in Morton Prince’s case is especially
interesting as it only occurred when the normal control was impaired
through the dissociation of the adult personality into two.

NO. 260 “4 Z

686 JULIAN 8. HUXLEY

remains of postural control were present in their patients—
which indicated that some connexion was still present with
pre-spinal centres—the mass-reflex did not appear. In other
words, in the course of phylogenetic evolution, a compound
mechanism has been evolved, the parts of which stand to each
other in a relation of dominance and subordination. But here
the dominance appears to be only slightly reversible, as opposed
to the cases of Perophora and of mental regression. Here the
subordinate system is so thoroughly under the control of the
other (presumably owing to certain structural relations and to
innate physico-chemical peculiarities inherent in synapses
concerned with inhibition), that it is apparently impossible
to tilt the balance so as to make the subordinate system the
dominant one for long together, so long as both are in organic
connexion. It is only when the two systems are separated
from each other that the real nature of the subordinate system
can be studied as it exists apart from controlling influence from
without. As indicated above, differential inhibition through
fear may induce a short temporary reversal of dominance.t

Child has pointed out that a somewhat similar (and also
simpler) relation subsists between the dominant and the
subordinate regions in many low forms of animals, such for
example as Planarians. Here, so long as the head region is
exerting its dominant or controlling influence, other portions
of the organism cannot form a head. But when this influence
is removed, either by the amputation of the head or by the
‘physiological isolation’ of parts of the organism (by their
removal, through growth, beyond the radius of influence of
the head), then the most anterior part of the isolated region
at once reacts by producing a head (Child, 1915), p. 96 et seq.).
Ip Head’s spinal case, however, after isolation the subordinate
system does not take on the characters of the dominant
system, but assumes a form which is peculiar to itself.

! The views of Head and Riddoch have been recently criticized (e.g.
‘Medical Science ’, vol. 4, 1921, pp. 141, 480). The fact of decerebrate
rigidity, however, would, among others, equally well serve to illustrate

the principle of neurological dominance and subordination, although here
we remain without phylogenetic analogies.

DEDIFFERENTIATION IN PEROPHORA 687

We may now leave the nervous system and return to physi-
ology. As an example in mammals, and one concerned only
with the parts of one organ, the following will serve.

As is common knowledge, the testis in mammals consists of
several functionally-distinct parts. Apart from blood-vessels
and nerves there are (1) the germ-cells (spermatogonia, sperma-
tocytes, spermatids, and spermatozoa), (2) the cells of Sertoli,
(3) the interstitial cells or cells of Leydig, (4) connective-tissue
cells. In the normal testis these exist in proportions which
do not vary beyond narrow limits. Various agencies, however,
will upset this balance. The germ-cells are the most suscep-
tible. Exposure of the testis region to X-rays or to Meso-
thorium; or ligature or section of the vas deferens; or
abnormal position in the organism, which can come about
spontaneously as in natural cryptorchism or can be produced
experimentally as in artificial cryptorchism or by transplanta-
tion, will bring about some degree of degeneration of the germ-
cells. This is accompanied in every case by a hypertrophy
of the interstitial cells. The cells of Sertoli are usually
unaffected. It would appear that these latter are not cells
capable of rapid multiplication. The chief competition is
therefore between the germ-cells and the interstitial cells.
The former are in some way dominant ; when they are damaged,
a check on the latter is removed, and their active increase
results. Whereas removal of the testis to an abnormal environ-
ment usually results in the permanent disappearance of the
germ-cells, X-ray treatment, if not very intense, only damages
them temporarily. Later they regenerate, and finally come
to have their old proportion once more. The increase in the
number of interstitial cells only lasts until this regeneration
starts, and is followed by a decrease. Finally, the normal
equilibrium is re-attained.t

1 See also R. Goldschmidt, ‘ Biol. Centralbl.’, 36, 1916, p. 160. In
Lepidopteran testes cultivated in tissue-culture, normal spermatogenesis
occurs. But the germ-cells always die before the cells of the follicle.
When this happens, the follicle-cells, which have till then remained
normal, start at once to multiply at a rapid rate.

ZZ2

688 JULIAN §. HUXLEY

The germ-cells are thus, in normal circumstances, partially
dominant over the interstitial cells, and are also more susceptible
than they are. This is the same relation that we found to hold
good between the zooid and stolon of Perophora. Furthermore,
it appears that in the testis a similar relation is to be found
between the interstitial cells in their turn and the connective
tissue (and Sertoli cells). Transplanted testes, as we have
said, first lose their germ-cells and show increase of inter-
stitial tissue. Within a few months the Sertoli cells also
degenerate and disappear (Stemach, Sand). We may take
this to mean that these cells, while not increasing after the
loss of the germ-cells because they are not a multiplicative
type of cell, are slightly less resistant than the interstitial
cells. Even these, however, are less resistant to unfavourable
conditions than the connective tissue. After a longer or
shorter period (usually several months) in the abnormal
situation, the interstitial cells in their turn start to decrease in
number, and now it is the connective-tissue cells which show
a corresponding increase. Finally, the ‘ testis ’ comes to consist
of nothing but connective tissue and blood-vessels. This is
also seen in some few cases of cryptorchism.

We have thus a system in which there enter four variable
sub-systems. One of these, for a reason which we can conjecture
but not prove, does not increase when others decrease. The
other three, however, are all in that state of dynamic equilibrium
which we have seen in its simplest manifestation in Perophora.
But this time they are arranged in a series, A being physio-
logically dominant over B, and B in its turn over C. Normally,
therefore, the relative proportions of the three tissues are
regulated according to the activity of A. When A is adversely
affected B increases, but not C. OC, however, increases when
both A and B have been affected.

If such a type of system were to exist, it should follow that
in some (abnormal) circumstances somewhat different condi-
tions should obtain, and that a slightly different end-result
should be brought about. As a matter of fact, in some of the
transplantations of Sand, this did occur. In three cases both

DEDIFFERENTIATION IN PEROPHORA 689

germ-cells and interstitial cells disappeared, leaving only
Sertoli cells and connective tissue. In one other case the
germ-cells and Sertoli cells were much less affected than the
interstitial tissue. This recalls the varying behaviour of
the stolon-zooid system in Perophora according to the internal
condition of the zooid.. (See Lipschitz, 1919, Chap. IV, where
full references are given.) Another view of an almost identical
problem is given by the varying response of the mammalian
ovary to different imtensities of X-ray treatment (Lipschitz,
1919, Chap. V, p. 205).

The conclusions we reached in discussing Detwiler’s results
(pp. 679-687) are of importance when we come to apply the
principles of dominance, differential inhibition, and resorption to
an explanation of the phenomena of metamorphosis. In meta-
morphosis, as [ have pointed out elsewhere (Huxley, 1921 b),
we have to think of the full-grown larva as consisting of two
minor systems in competition with each other—the differen-
tiated system of larval organs, and the developing system of
adult organs. The two enter into a state of balance. This
balance may be tilted in favour of the adult, or kept at the
existing tilt which favours the larval system. It has often
been maintained that the time of metamorphosis was deter-
mined by the production of a given relative quantity of some
definite substance within the organism, e.g. thyroid secretion
in the larvae of Amphibia. Such a concentration of a particular
substance is often the effective agent in tilting the balance, but
it is not the essential cause of metamorphosis. The essential
cause of metamorphosis is that two mutually meompatible
systems are in a state of dynamic physiological equilibrium
within the same organism.

In Echinoderm metamorphosis the mechanism for upsetting
the balance appears to be simpler than in Amphibia. Experi-
ments of Runnstrém (1917) and of my own, an account of which
is now in the press, indicate that exposure of the pluteus tissues
to unfavourable agencies of various descriptions will lead to
their dedifferentiation and partial resorption. In nature the
actual chain of events leading to this result appears to be as

690 JULIAN 8. HUXLEY

follows: the Echinus rudiment at the start grows concomi-
tantly with the Pluteus. After a certain time, however, it
becomes so large that its weight drags the larva to the bottom.
Here the conditions, as regards both food and general environ-
ment, are unfavourable to the pluteus tissues; these begin
to dedifferentiate, and as soon as they have passed a certain
critical stage in the process the Echinus tissues become dominant
and are able to develop further at the expense of the larval
organization. In the broadest terms the balance in Amphibia
is regulated mainly from within, in Echinoids mainly from
without ; but in both cases the possibility of the sudden
change which we call metamorphosis depends on the co-existence
of two systems in the same organism which are very closely
balanced as regards physiological dominance.

To sum up, we may say that the facts of physiological
dominance of inhibition of growth, of resorption, and of the
state of balance which exists among the parts of any organism
and is the dynamic expression of Roux’s ‘ Kampf der Teile’,
are all intimately connected. As a matter of fact physiological
dominance is rendered most obvious when it can be reversed,
as in Perophora or in metamorphosis—and that is when the
balance between sub-systems is very close.

The various examples discussed may perhaps be made
clearer by the use of symbols. In every case let A =a dominant
system; B a system normally subordinate to A; C one
normally subordinate to B and also to A. An arrow |, indicates
dominance, pointing towards the subordinate system. Brackets
( ) indicate subordinate condition. Dashes (A’, B, &e.)
indicate alteration of the system from its original condition
to another. Erasure (®, “B, &c.) indicates disappearance
of a system by resorption. Suffixes (A,, B,, &¢.) indicate
homologous systems in order of age or size. Enclosure

|

|
A

indicates passage to a non-functional state.

A
sal

Plus sign (A+, B+, &c.) indicates increase of the system,

DEDIFFERENTIATION IN PEROPHORA 691

1. Clavellina. A=zooid, B=stolon.

(a) Normal. (6) Reduced,
A A’
v V
(B) (B’)
2. Perophora. A=zooid, B=stolon.
(a) Normal. (6) Starved, water (c) Starved, water
changed, not changed.
A A AL
V M t
(B) TEL B+

3. Planaria or single Stenostoma zooid. A=brain-region, B=
pharynx-region, C = tail-region.
(a) Normal. (6) After decapitation. (c) Subsequent regeneration.
B’

de

=

| B |
(B) J (B)

| C) q
(C) (C)

4. Stenostoma chain. A=brain-region, B=rest of zooid.
A, B,=oldest, A, B,= youngest zooid.
(6) Transection giving brain-region

behind posterior brainless region (c) End-result
(a) Normal. of another zooid, from (0),
Bi A3 ( ae )
ae ((B 3 ))
( B - (d) Transection giving older
\( Oe \\ zooid posterior to younger, (e) Result of (d).
Baz
( A4 (v he Ay
(< a1) Bs.
( Ae ) § ai: )
\(B'2) J UiB2))
5. ‘Spinal man.’ A=cerebral centres, B=mass-reflex.
(a) Normal, (b) After transection of cord.
A
{I B
(B’)

6. Mental regression. A=adult system, B=juvenile system,
C=infantile system. (Only three systems given for simplicity’s sake.)

(a) Normal, (b) After regression. (c) During recovery.
A ama i
(B’) J
V (B’) t
(C’) iE B

692 JULIAN §. HUXLEY

7. Testis. A=germ-cells, B=interstitial tissue, C=Connective
tissue.

(a) Normal, (6) Transplanted, after (c) Ditto, after longer
short period, period,
A ” AL
v B+ BR
e) y C++
v (C+)
(C)

8. Amblystoma limb-buds. A,=transplanted limb-bud, A, =
regenerated remains of limb-bud in original position.
(a) After trans- (b) End-result = (c) After trans- (d) End-result

plantation to a from (a). plantation to a from (c),
distance of more distance of less
than four seg- than four seg-
ments, ments,
Ag A’, A, A
V
Ay AY (A,) AK

9. Metamorphosis of Echinoids. A=larval tissues, B= adult
tissues.
(a) At the time when (6) Shortly after. (c) End-result,
larva sinks to bottom.

A (A’) AA
ty
B B B’

9. SUMMARY.

1. The social Ascidian Perophora viridis may dedif-
ferentiate in either of two distinct ways, or by a mixed method :
(a) by reduction to a spheroidal mass, as in Clavellina ; (b) by
incipient reduction as in (a), but followed by total resorption
into the stolon, which may grow during the process.

2. Resorption is due to the migration of the individual cells
out of the tissues into the haemocoel.

3. In certain conditions the zooid maintains itself, in spite
of food not being provided, at its original size and in perfect
health. This it does by resorbing the stolon.

4, Experiments with dilute solutions of KCN show that
resorption of the zooid occurs in slightly unfavourable condi-
tions, which affect the sensitive zooid more than the less
highly-organized stolon.

DEDIFFERENTIATION IN PEROPHORA 693

5. The results are to be explained as follows: (a) In the
competition between zooid and stolon the zooid normally
is dominant because metabolic processes take place at a greater
rate in it than in the stolon. The stolon is therefore starved
at the expense of the zooid. (b) The zooid is more susceptible
than the stolon to toxic agencies. (c) In low concentrations of
such agencies it is therefore affected while the stolon is not.
(d) As a result it begins to dedifferentiate. Dedifferentiation
is here accompanied by the migration of the cells out of the
tissues. (e) The speed of its metabolic processes is now no longer
greater than that of the stolon’s. It is therefore now starved
at the expense of the stolon. (f) Any cells migrating out of the
tissues are removed by the normal circulation, by the stolon-
circulation (irregular pulsation of the stolon), or by utilization
as food by the stolon. As in chemical reactions where the
end-products are removed, the reaction thus runs to its limit,
i.e. to complete resorption of the zooid.

6. Stopping the circulation by means of KCl results in dedif-
ferentiation accompanied by a much smaller degree of resorption.

7. At low temperatures (about 5° C.) some dedifferentiation
oceurs ; but there is very little resorption, apparently owing
to the cessation or slowing of the heart-beat.

8. Partial dedifferentiation is recorded in Amaroucium and
Botryllus.

9. The significance for general biological problems of domi-
nance due to high rate of metabolism, of differential suscepti-
bility and of dedifferentiation, is discussed.

10. The similarity of certain psychological and neurological
phenomena is noted (mental regression, alteration of spinal
reflexes when freed from cerebral control, &c.).

LITERATURE LIsT.
Allen (1920).—‘ Science ’, 52, 1920, p. 274.
Child, C. M. (1904).—* Arch. Ent. Mech.’, 17, 1904, p. 1.
——- (1915 a).—‘ Senescence and Rejuvenescence ’, Chicago, 1915,
—— (1915b).—‘ Individuality in Organisms ’, Chicago, 1910.
—-— (1916).—‘ Journ. Morph.’, 27, 1916,
—— (1917).—Ibid., 30, 1917.

694 JULIAN S. HUXLEY

Child, C. M. (1919).-—‘ Science ’, N.S., 50, October 17, 1919.

Detwiler (1920).—‘ Journ. Exp. Zool.’, 31, 1920, p. 117 (see, especially,
pp. 149-51).

Driesch, H. (1906).—‘*Skizzen zur Restitutionslehre”’, ‘Arch. Ent.
Mech.’, 20, 1906.

Gray, J. (1920),—* Quart. Journ. Micro. Sci.’, 64, 1920, p, 345.

Harrison, R. G. (1915).—* Proe. Nat. Ac. Sci.’, 1, 1915.

Head (1918).—‘ Brain’, 41, 1918, p. 344.

Head and Riddoch (1917).—‘ Brain’, 40, 1917.

Huxley, J. S. (1921 a).—‘* Quart. Journ. Micro, Sci.’, 65, 1921, p. 293.

—— (1921 6).—‘ Studies in Dedifferentiation ’, i (in the press).

Lillie and Knowlton (1902).—‘ Biol. Bull.’, 1, 1902.

Lipschiitz (1919).—-‘ Die Pubertiitsdriise u. ihre Wirkungen ’, Bern, 1919.

Loeb, J. (1900).—‘ Amer. Journ. Physiol.’, 4, 1900, p. 60.

Lund (1917),—‘ Journ. Exp. Zool.’, 24, 1917, p. 1.

Maas (1910).—‘* Festschrift f. R. Hertwig’, 1910, 3.

Miiller. K. (1911).—‘ Arch. Ent. Mech.’, 32, 1911.

Nichol (1920) in ‘ Functional Nerve Disease’, ed. H. Crichton Miller,
Oxford, 1920,

Prince, Morton (1908),—‘ The Dissociation of a Personality ’, 1908.

—— (1920).—‘J. Abn. Psych.’, 15, 1920.

Rivers (1920).—‘ Instinct and the Unconscious ’, Cambridg>, 1920.

Robertson and Ray (1920),—‘ J, Biol. Chem.’, 42, 1920, p. 71.

Roux (1881).-—‘ Der Kampf der Teile im Organismus ’, 1881.

Runnstrom, J. (1917).—‘ Arch. Ent. Mech.’, 43, 1917, p. 223.

Schultz, E. (1906).—Ihbid., 21, 1906.

—— (1907).—Ibid., 24, 1907.

Towle (1901).—‘ Biol. Bull.’, 2, 1901, p. 289.

ILLUSTRATIONS.

All figures are drawn to scale with the Abbé Camera lucida.
Otherwise they are semi-diagrammatic. All were drawn at
table level; the magnification is indicated for each figure.

Fig. 1.—Clavellina type of reduction (x25). Two zooids, a and B,
isolated without stolons. a. Day of operation, B has a trace of stolon-
connexion. b. After forty-eight hours. Reduction started earlier in
B. Both have formed short stolons, but that of B remains within the test.
A has only just started to reduce. c. Advanced reduction (three days for B,
four days for 4), The stolon of Bis large and lobulated, but has not emerged
from the test. B is spheroidal and opaque, in stage 4-5. A is in stage 3,
which it did not reach till after three days. Its stolon has grown, and is
distended with cells. ’s heart was beating slowly, B’s had almost stopped.

Figs. 2 and 3,—Growth or maintenance of zooid at expense of stolon.

DEDIFFERENTIATION IN PEROPHORA 695

Fig. 2 (x 80).—a. Immature zooid on day of operation. b. The same,
perfect, after three days. The stolon has been much reduced both in
length and breadth. A small bud had formed and been absorbed. c. After
five days. Further reduction of the right end of stolon; zooid in first
stage of reduction, which has led to a slight dilatation of the left end of
the stolon. d. Stolon-tip from a similar system after three days, showing
shrunken appearance.

Fig. 3 (x25).—a. A system on the day of operation. 6. The same
three days later. Zooid actively functional, stolon much drained in all
dimensions.

Fig. 4.—Maintenance of bud following resorption of first zooid.
a. (x80). Original zooid in stage 4 of reduction, after two days. Stolon
healthy. 6. (x80). After three days. Stage 4, but smaller; meanwhile
a bud had formed to the left of the zooid, and by now was 50 per cent.
larger in diameter than the zooid. c. (x40). After five days the zooid
had disappeared. The remains of its test is seen. The right part of the
stolon, to the right of the bud, has also been resorbed, and resorption is
beginning in the other portion, as shown by the extent of its test.
d., e. (x40). After seven days. In d. the tip of the stolon is shown
contracted, in e. expanded with blood. /. After twelve days. The stolon
has almost disappeared. The zooid is practically unchanged in size or
development.

Figs. 5-15.—Resorption of zooids and growth of stolons.

Fig. 5 (x 25).—Stolon-growth. a. A system on the day of operation.
b. The same, but stolon only ; four days later. The zooid was in stage 2
of reduction. A bud is seen on one stoion-branch. Note the test bridging
concavities of the stolon.

Fig. 6 (x 40).—Early stages of resorption. a. After two days. Zooid
just reaching stage 3. The exhalant siphon is still slightly attached to
the test. Faint traces of gills visible. The stolon-branch B and the tip
of a represent new growth. 6. Eight and a half hours later. The zooid is
in stage 3-4, and has shrunk considerably ; B has grown.

Fig. 7 ( x 64).—Zooid in stage 3-4 of reduction, showing heart and traces
of inhalant siphon and stomach ; note the double stolon-connexion.

Fig. 8 ( x 64).—Zooid in stage 4 of reduction. The heart is seen end on.

Fig. 9 (x 64).—Later stages of reduction. «a. Zooid in stage 4, after
two days ; ectoderm in places cubical. A stolon outgrowth had occurred.
b. The same, ten hours later (test omitted) ; further shrinkage. Ectoderm
all cuboidal. c. Fourteen hours later (test omitted) ; further shrinkage.
A new stolon outgrowth has occurred. d. Forty-eight hours later (five
days in all). It is now in stage 6 (after four days it had reached stage 5).
The pale ovoid is probably the remains of the stomach. Note the slight
reduction of the test. e. Twenty-four hours later (six days). Zooid por-
tion smaller than stolon-connexion,

696 JULIAN 8. HUXLEY

Fig. 10 (x 64).—Zooid reaching stage 5. The stolon was attached to
another zooid, and showed active circulation. It was hard to be sure
whether the heart was beating.

Fig. 11 (x 100).—Zooid in stage 5. Stolon as in 10. Zooid ectoderm
cuboidal. Solid organ-remains fill most of the zooid.

Fig. 12 (x64).—a. Zooid in stage 4, after two days. b. The same,
nine and a half hours later (from a different aspect). Stolon as in 10.

Fig. 13 (x 64).—Zooid in stage 5 of reduction. Zooid of the same
opacity as the stolon.

Fig. 14 (x 64).—Zooid in stage 5-6. Opacity as in 13.

Fig. 15 (x 64).—Zooid in stage 6. Some remains of organs visible.

Figs. 16-18.—Reduction at low temperature.

Fig. 16 (x 32).—After eight days; early stage of reduction. Note
considerable opacity combined with open siphons. Heart beating, but
circulation only in right half of the stolon-connexion. Note a new stolonic
outgrowth into the test of the zooid.

Fig. 17 ( x 64).—A similar zooid after eight days. Débris on siphons,
which are open. Heart not beating, but visible,

Fig. 18 (x 64)—Similar; but a slightly later stage of reduction.
Depressions still mark the siphons. Heart beating, but very faintly.

Figs. 19-20.—Reduction in KC] solutions.

Fig. 19 (x25) (50 c.c. sea-water+ 4 c.c. m/2 KCl).—a. Early stage of
reduction, after one day. Inhalant siphon-lobes of test separate from
siphons. Outgrowths at the end of stolon. b. The same, twenty-four
hours later. Cell-strands attach siphon-regions to test. No sign of
internal organs. Stolon healthy.

Fig. 20 (x25) (50 c.c sea-water+8 c.c. m/2 KCl). Similar to 19, 6.,
except that the stolon as well as the zooid has been adversely affected
(shrinkage, cuboidal epithelium).

Figs. 21-3.— Reduction in KCN solutions.

Fig. 21 (x25).—In m/2,000 KCN. a. Before treatment. The zooid
is a not quite developed bud. b. After twenty-four hours. Zooid in stage 3
of reduction. Stolon slightly shrunk, but crowded with cells, and with
attempts at new growth.

Fig. 22 (x25)—In m/4,000 KCN. a. Before treatment. 6. After
forty-eight hours. Zooid much reduced. Stolon crowded with cells, but
shrunken ; no new growth.

Fig. 23 (x 25).—In m/32,000 KCN. a. Before treatment. Zooid a not
quite developed bud. 6. After twenty-four hours. Zooid considerably
reduced, stolon with clubbed ends and with new growth (within test only).
c. In reversed position after forty-eight hours. Zooid much reduced.
Stolon crowded with cells, and with new growth outside test.

Fig. 24 (x 340).—To show pulsation of stolon. The same stolon-tip
(a) expanded ; (6) (less than a minute later), contracted. The position of

DEDIFFERENTIATION IN PEROPHORA 697

the test (2) did not change, and a space was left between it and the ecto-
derm when contraction occurred. Note the thickened epithelium in con-
traction, with irregular outline externally. Note the small outgrowth
in the expanded state; this was not observed after contraction. The
stolon remained for a few minutes contracted, then expanded in under
a minute; and vice versa. The blood-cells are not figured.

Fig. 25 (x 340).—A large lateral outgrowth, on the same stolon as that
shown in fig. 24. The blood-cells are shown in the outgrowth itself, but
only a few indicated elsewhere.

Fig. 26 (x340).—A normal growing stolon-tip. Note the columnar
epithelium at the extreme tip. Close to the tip there are very few green
blood-cells, the majority being white. Then comes a zone where a con-
siderable proportion are green, and then one where they are in the majority.
The circulation, though active, did not extend into the densely-packed
region drawn.

Figs. 27 and 28.—Dedifferentiation in Amaroucium.

Fig. 27 (80).—A young oozite in stage 2-3 of reduction. The dense
anterior mass was orange-red. Portions of the intestine are seen below.
Muscular contraction of the whole organism still took place at intervals.

Fig. 28 ( x 80).—A blastozooite dedifferentiated in weak alcohol, stage 3.
Note the cell-masses outside the main body of the organism.

ek

genese
Tie

T a
7 _

{

a ~oe
“eal

a

we

Huzley, del.

gt"
Quart. Journ. Mier. Se. Vol. Go; N.S. Pee

.
-
7

7 eee

"ss -

ae ate Vt

o
2

19a

Huxley, del.

5, N.S., Ph. 26:

Vol. 6

26
CF i

Quart. Journ. Micr. Scr.
23a
23b

23a¢

Review

The Microtomst’s Vade-Mecum, by Arruur Boies Leer.
8th edition, edited by Professor J. Bronr& GaATENBY.
London, J. A. Churchill, 1921. Price 28s. net.

Tuts eighth edition of Mr. Bolles Lee’s well-known Micro-
tomist's Vade-Mecum has been edited and entirely revised
by Professer J. Bronté Gatenby with the assistance of five
collaborators.

Readers of this Journal will not be surprised to find that
Professor Gatenby has himself written special sections dealing
with chromatin, chromosomes, and cytoplasmic inclusions in
which he gives us the full benefit of his thorough practical
experience. Moreover, he has re-written the part on Mammalian
Embryological methods. An innovation is the inclusion of two
new methods for staining bacteria in tissues which will doubtless
be very useful to biologists not versed in bacteriological
technique. In the next edition we may hope to see mentioned
the important ‘ Carmine Claudius’ method for differentiating
yeasts, as well as Gram-positive bacteria in tissues.

Professor Gatenby is also personally responsible for a very
valuable chapter on the cultivation of tissues ‘in vitro’.

Zoologists will be grateful to Professor Bayliss for re-writing
the chapter on Staining. Here we find an authoritative
general account of the principles involved in the staining
of living as well as dead cells. A careful reading of this lucid
summary will save histologists from many pitfalls.

The chapters on Neurological Techniques have been to a great
extent re-written by Dr. Da Fano. The additional directions
given for the carrying out of the Bielschowsky and other
complicated impregnation methods will be much appreciated.

For an important section embodying some of the new work
in micro-chemistry on the lipoids and true fats and their
differentiation we are indebted to Dr. W. Cramer, while
Mr. J. T. Carter has revised the section on bone and teeth.
In addition to the short account of the Protozoa given by

700 REVIEW

Mr. Bolles Lee in former editions, Dr. A. Drew includes many
useful notes on the culture and staining of Amoeba, in which
he has had much experience. The general arrangement of this
section leaves something to be desired; for example, there
are two paragraphs headed ‘ Flagellata’ (1003 and 1038),
and the fixation of Coccidia is considered in paragraph 1001,
and again under Sporozoa in § 1031.

A new departure is the inclusion for beginners of a final
chapter tabulating general procedure in the making of micro-
scopical preparations.

It will be gathered that though the bulk of the volume
remains little changed there has been considerable rearrange-
ment of the contents. Some have been eliminated and much
useful new matter has been added.

Professor Gatenby is to be congratulated on the success
of his editorship. The appearance of this new edition of
the familiar and indispensable vade-mecum will be heartily
welcomed by all working zoologists.

INDEX TO VOL. 65
NEW SERIES

Acrosome, Gatenby, 265-292.

Actiniaria ; classification, Stephen- _

son, 493-576.

Anopheles ; development of ovary,
Nicholson, 395-450.

Aphides ; hyperparasites, Haviland,
101-128 and 451-478.

Aphidius ; parasites of, Haviland,
101-128 and 451-478.

Chromosomes ;

Archotermopsis ; protozoa parasites |

of, Cutler, 247-264.
Arenicola ;
Goodrich, 157-162.
Autotomy; tails of Gecko, &c.,
Woodland, 63-100.

Bahl, K.N. On the Blood-Vascular

System of the Karthworm.
Pheretima, and the course

of the Circulation in Earthworms,
349-394.

Bionomics of Lygocerus, Haviland,
101-128.

parasites of Aphides, Haviland,
451478.
Bolocera, Gemmill, 577-588.

Calcospherites of Dipterous larvae,
Keilin, 611-626.

Cannon, H. G. Early Development
of the summer egg of a Cladoceran
(Simocephalus vetulus),
627-642, pl. 25.

Carleton, H. M. See Champy, 589-
610.

Cavia ; sperm, Gatenby, 265-292,

. . |
Gregarine in egg of, |

of certain Cynipid hyper- |

Champy, C., and Carleton, H. M.
Observations on the Shape of the .
Nucleus and its Determination,
589-610, pls. 23, 24.

structure
division, Lee, 1—32.

Cucumaria ; development, Ohshima,
173-246.

Cutler, D. Ward. Observations on
the Protozoa parasitic in Archo-
termopsis wroughtoni Desn. Part
III. Pseudo-trichonympha._pris-
tina, 247-264, pl. 10.

Cytoplasmic inclusions, Gatenby,
265-292.

and

Dakin, W.J. The Eye of Peripatus,
163-172, pl. 7.
Dedifferentiation
Huxley, 643-697.
Development :
Lygocerus, Haviland, 101-128.
Cucumaria, Ohshima, 173-246.
Cynipid hyperparasites of
Aphides, Haviland, 451-478.
Bolocera, Gemmill, 577-588.
Ovary of Anopheles, Nicholson,
395-450.
Summer egg of a Cladoceran,
Cannon, 627-642,
Diptera ; larval excretion, Keilin,
611-626.

in Perophora,

Karthworm :
circulation, Bahl, 349-394.
pharyngeal or salivary gland,
Keilin, 33-62.

702
Eye of Peripatus, Dakin, 163-172.
Gatenby, J. B. Lee’s Microtomist’s

Vade-Mecum, eighth
(review), 699.

Gatenby, J. B., and Woodger, J. H. |

The Cytoplasmic inclusions of the
Germ-Cells, IX. On the Origin of
the Golgi Apparatus on the
Middle-piece of the Ripe Sperm
of Cavia and the Development of
the Acrosome, 265-292, pls. 11, 12.

Gecko (Hemidactylus); caudal
autotomy and _ regeneration,
Woodland, 63-100.

Gemmill, J. F. The Life-history of
Melicertidium
tum (Sars),a Leptomedusan with

a theca-less Hydroid Stage, 339-

348, pl. 16.

Anemone Bolocera tuediae
(Johnst.), 577-588, pl. 22.

Golgi apparatus, Gatenby, 265-292.

Gonospora minchinii, Good-
rich, 157-162.

Goodrich, E. S., and Pixell Goodrich,
H. L. M. Gonospora min-
chinii, n.sp., a Gregarine in-
habiting the egg of Arenicola,
157-162, pls. 5, 6.

Proboscis of the
Structure, 323-

Haswell, W. A.
Syllidea, Part 1.
338, pl. 15.

Haviland, M. D. Bionomics and
Development of Lygocerus
testaceimanus, Kieffer, and
Lygocerus cameroni, Kieffer
(Proctotrypoidea - Ceraphronidae)
parasites of Aphidius (Bra-
conidae), 101-128.

Bionomics and Post-Embry-

onic Development of certain

Cynipid Hyperparasites of

Aphides, 451-478.

edition |

octocosta- |

The Development of the Sea |

INDEX

Huxley, J. S. Further Studies on
Restitution-bodies and freeTissue-
culture in Sycon, 293-322, pls.
13, 14.

| —— Studies in Dedifferentiation.

II. Dedifferentiation and resorp-

tion in Perophora, 643-697, pls.

26-28.

Kaburaki,T. Terrestrial Planarians
from the Islands of Mauritius and
Rodrigues ; with a Note upon the
Canal connecting the Female
Genital Organ with the Intestine,
129-156, pl. 4.

Keilin, D. Pharyngeal or Salivary
Gland of the Earthworm, 33-62,
pl. 3.

On the calcium carbonate and

the calcospherites in the Mal-

pighian tubes and the fat body of

Dipterous larvae and the ecdysial

elimination of these products of

excretion, 611-626.

Lee, A. Bolles. The Structure of
certain Chromosomes and the
Mechanism of their Division, 1-
32, pls. I, 2.

Microtomist’s Vade-Mecum,
eighth edition (review), 699.

Lygocerus, parasite of Aphidius,
Haviland, 101-128.

Melicertidium, life-history, Gemmill,
339-348.

Microtomist’s Vade-Mecum, A.
Bolles Lee, eighth edition, edited
J. B. Gatenby (review), 699.

Nicholson, A. J. The Development
of the Ovary and Ovarian Egg of
the Mosquito, Anopheles
maculipennis, Meig, 395-450,
pls. 17-20.

INDEX

Nucleus; shape, Champy, 589-610.

Ohshima, H.
Cucumaria
Marenzeller,
So:

Larval

tangus purpureus, 479-492,

pl. 21.

Development of
echinata v.
173-246, pls.

Parasites :
Pseudo-trichonympha of Archo-
termopsis, Cutler, 247-264.
of Aphidius, Haviland, 101-128
and 451-478.
Gonospora of Arenicola egg,
Goodrich, 157-162.
Peripatus, eye of, Dakin, 163-172.
Perophora; dedifferentiation and
resorption, Huxley, 643-697.

Pheretima ; blood-vascular system, | Syllidae ; proboscis, Haswell, 323-
Sy ; scis. swell, 32

Bahl, 349-394.
Pixell Goodrich, see Goodrich, 157-
162.

Planarians from Mauritius and
Rodrigues, Kaburaki, 129-156.
intestinal connexion with

genital organ, Kaburaki, 129-156.

Skeleton of Spa-

Proboscis of Syllidae, Haswell, 323- |

338.

Pseudo-trichonympha, Cutler, 247—
264.

Pygopus; caudal autotomy and
regeneration, Woodland, 63-100.

Regeneration ; tails of Gecko, &c.,
Woodland. 63-100.

703

Resorption in Perophora, Huxley,
643-697.

Restitution-bodies,
322.

Huxley, 293-

Simocephalus; summer egg. Cannon,
627-642.

Spatangus; larval skeleton, Ohshi-
ma, 479-492.

Sperm of Cavia; Golgi apparatus,
Gatenby, 265-292.

Sphenodon ; caudal autotomy and
regeneration, Woodland, 63-100.

Stephenson, T. A. On the Classifi-
cation of the Actiniaria. Part IT.
Consideration of the whole group
and its relationships, with special
reference to forms not treated in
Part I, 493-576.

Sycon; tissue culture, Huxley, 293-
322,

338.

Tails ; autotomy and regeneration
in Gecko, &c., Woodland, 63-100.

Tissue-culture in Sycon, Huxley,
293-322.

Woodger, J. H. See Gatenby, 265--
292.

| Woodland, W.N. F. Observations

on Caudal Autotomy and Re-
generation in the Gecko (He mi-
dactylus' flaviviridis,
Riippel), with Notes on the Tails
of Sphenodon and Pygopus, 63-
100.

Oe
o

? = 7 oye

4
VAAL sen
vane ey = fg ae Seating
r */) 7 i - oie val

elie ii “ Sivinnyol eat. H
tas Alva PA

uo 4 ADAM MTE ' C
iy me eT: At q
ar : 5 : v
oT «
« *eitees Maral
?

? ! iL} 3 $3 a> ‘t-<- baa

ig

weet.

ip! v
"A ] ent OUT ¢
refta ye
va Shee
. tat
jepuit al
y 7 4h
} 7}
i ee,
ot :
. ta a bl
2 . cs Oe
— : =
; ‘A ou ‘7 wn ti yy Ai
' a ,

Ue

ara 7) Oat

eres
a he wis

Cri;

pai fag,

Lv ®
iy
A,

>a. ae

\ Marcerry ‘ ae 4 ry

of

QH Quarterly journal of Micro-
201 scopical science

Q2
NoSe
v.65
COPe2

saainey

BIOLOGICAL &
MEDICAL SERIALS

PLEASE DO NOT REMOVE
CARDS OR SLIPS FROM THIS POCKET

a renee
UNIVERSITY OF TORONTO LIBRARY

EEE

ae

feq=:

ay

ci vesnepener

>

Speneg str)

2S:

ine

. : ret ty ; Fito
“Site Srereenpeane pi se erdier ae cgheseahitits

satesiessboe Sp AS Sc: etsy

PRPS OTH FN:

Fe tel: Sa ee a)

ree tthe

POO:

~ darth ae shabe
eran ty bet wearer

ayy
(haste

Ft steiyy
diasatt

Ss

=the
eee
SPIRE
ow
Sessa

==

teh
tuk

feseces

wae

v
Oe

rarest. ft i
esse te ae
ant
+

Ri

ae uaa}
ible. ¥, |
ese

Bd 4 '
Reet

2ehe44 9b tn

Nanirayi
/ Ser

1)
i
ey