st abe ates eo
Te ee

rie
» gly
ot GeO Nae emi

eS
eta s oo

chorea

Eo fonkm omit

me
etna

mieten de ine oy Md mth! pare:
Yr ne Mee ote Pwo or wre!

4 -yoyetoens Y=

Oe es ee land

se ptt tn
einer tm

ahve! apr tui ntfs fog
insge pa hehe Seal i ri porrehatory ae
ih aRie none prerret alae

”"

OA ae pe Rah ie gt Mee el aE ee PA OE PRO Re UO Te ee ee el aa

“a Nore eee: egy apna me remcsnnty

HARVARD UNIVERSITY

EEL

LIBRARY

MUSEUM OF COMPARATIVE ZOOLOGY

ait
fi eh

QUARTERLY JOURNAL

OF

MICROSCOPICAL SCIENCE:

EDITED BY

HE. RAY LANKESTER, M.A., LL.D., F.R.S.,
Linacre Professor of Comparative Anatomy, Fellow of Merton College, and
Honorary Fellow of Exeter College, Oxford.

WITH THE CO-OPERATION OF

ADAM SEDGWICK, M.A., F.RS.,
Fellow and Lecturer of Trinity College, Cambridge ;
AND
W. F. BR. WELDON, M.A... F.&:S.,

Jodrell Professor of Zoology and Comparative Anatomy in University College, London;
Fellow of St. John’s College, Cambridge.

VOLUME 37.—New Serrtzs.
With Mithographic Plates and Engrabings on Wood.

LONDON:
J. & A. CHURCHILL, 11, NEW BURLINGTON STREET.
1895,

EAT el PO, A

CHANTAL 7
mons (00%, MOO CUT “yy Se
FLY aa0tdaWAg |

P ' ery OP LT Tay
WAT) hy ae /. ¥ Wetec
: ae .
; i wake aoe Paty tat ied:
te 4 a MoT | yy j ieil ey by wil Plee inf La Hal i 4
, 7
: 1
- ; =
ul
e 7 -
i
{ : ty a .

ee Ce

Ase duce Habit paiue THA Eh SHUNT UTRD k

, z a ¥ Baie
fr | rn

CONTENTS.

CONTENTS OF No. 145, N.S., NOVEMBER, 1894.

MEMOIRS :
On Julinia; a New Genus of Compound Ascidians from the
Antarctic Ocean. By W. T. Caiman, reat NaN
Dundee. (With Plates 1—3) ; ;

Hermaphroditism in Mollusca. By Dr. Pav Petsenzer (Ghent).
(With Plates 4—6)

A Description of the Cerebral Convolutions of the Chimpanzee
known as “ Sally ;” with Notes on the Convolutions of other
Chimpanzees and of two Orangs. By W. Biaxtanp Benuaw,
D.Sc.Lond., Hon. M.A.Oxon., Aldrichian Demonstrator of Com-
parative Anatomy in the University of Oxford; Lecturer on
Biology at Bedford College, London. (With Plates 7—11)

On the Inadequacy of the Cellular Theory of Development, and on —

the Early Development of Nerves, particularly of the Third
Nerve and of the Sympathetic in Elasmobranchii. By Apam
SEDGWICK, F.R.S. : F

On Benhamia cecifera, n.sp., from the Gold Coast. By
W. Braxtanp Benuam, D.Se.Lond., Hon. M.A.Oxon., Aldrichian
Demonstrator in Comparative Anatomy in the University of
Oxford, &c. (With Plate 12)

CONTENTS OF No. 146, N.S., DECEMBER, 1894.

MEMOIRS :
A Re-investigation into the Early Stages of the Development of the
Rabbit. By Ricnarp Assueton, M.A. (With Plates 13—17)
On the Phenomenon of the Fusion of the Epiblastic Layers in the
Rabbit and in the Frog. By Rrcuarp AssHETON, M.A. (With
Plate 18) : : : : :
b

PAGE

19

47

87

103

113

165

iv CONTENTS.

On the Causes which lead to the Attachment of the Mammalian
Embryo to the Walls of the Uterus. By Richarp AssHETON,
M.A. (With Plate 19).

The Primitive Streak of the Rabbit ; the Causes which may deter-
mine its Shape, and the Part of the Embryo formed by its
Activity. By Ricnarp AssHEton, M.A. (With Plates 20—22)

On the Growth in Length of the Frog Embryo. By RicHarp
AssueTon, M.A. (With Plates 23 and 24)

CONTENTS OF No. 147, N.S., MARCH, 1895.

MEMOIRS :

On the Variation of the Tentaculocysts of Aurelia aurita. By
Epwarp T. Browne, B.A., University College, London. (With
Plate 25) : : :

On the Structure of Vermiculus pilosus. By E.8. Goopricu,
F.L.S., Assistant to the Linacre Professor, Oxford. (With
Plates 26—28) . 5 : :

On the Mouth-parts of the Cypris-stage of Balanus. By THExo.
T. Groom, F.Z.8., late Scholar of St. John’s College, Cam-
bridge. (With Plate 29) :

A Study of Coccidia met with in Mice. By J. Jackson CLARKE,
M.B. Lond. (With Plate 30)

Observations on Various Sporozoa. By J. Jackson CLARKE,
M.B.Lond. (With Plates 31—383)

A Revision of the Genera and Species of the Branchiostomide.
By J. W. Kirxatpy. (With Plates 34 and 35)

Sedgwick’s Theory of the Embryonic Phase of Ontogeny as an Aid
to Phylogenetic Theory. By E. W. MacBuripg, B.A., Fellow
of St. John’s College, Cambridge; Demonstrator in Animal
Morphology to the University of Cambridge

PAGE

173

191

223

245

253

269

277

287

303

325

CONTENTS.

CONTENTS OF No. 148, N.S., MAY, 1895.

MEMOIRS:

The Anatomy of Alcyonium digitatum. By Sypnny J.
Hickson, M.A., D.Sc., Beyer Professor of Zoology in the Owens
College, Manchester; Fellow of Downing College, Cambridge.
(With Plates 36—39) E : : ;

Note on the Chemical Constitution of the Mesogloea of Aleyonium
digitatum. By W. Lanepon Brown, B.A., late Hutchinson
Research Student, St. John’s College, Cambridge

A Study of Metamerism. By T. H. Morean, Ph.D., Associate
Professor of Biology, cs Mawr Nee U.S.A. (With Plates
40—43) . : : 5

On the Colom, Genital Ducts, and oe By Epwin $8.
Goopricu, F.L.8., Assistant to the Linacre Professor of Com-
parative Anatomy, Oxford. (With Plates 44 and 45)

PAGE

343

389

395

477

7 ey e/ j iWtAerrivtee =
HORE EA Ge eed vow AO. Oat) el '}
| ‘ enti

2 20’! :
j ‘
AP Fadtied 4 widedisth. aichio get A FO waite
yreew OF. OL) AV y ie Maite PP ose We Pe Ad ee Le
; . my
. Tae inet) a ond aot WeayW ; i if nye wae nepal ’
abe ‘ Re fit exalt ae
bite a ‘ °
aroine vila Wrad) z ead) &. SUA ed era (woth) hy oth BP
| eeinriiie Glee Sith A ae Hagel we Arie it
ul A it a > neil Ay eA. a) } | iy df
aa '
: rai) ] 1.UNG I | t av PPaAra i. Waites
) i woes ie ei? & d 7? ’ af : ’
‘ 4 ‘ } . Lio 4 twee
UT) ee i) !
We ; tte
ANGE 4 R : : .
thd ee Vie ee, et Gh aie) it ied 1p >
vita) ) ee ‘ l | thi mb be taly wr Al rpiree hil De
LP Ate 7 bay FI Pil ae Le) erie oe) one
i
. +
e
7 ‘ a
a 2 s 7
:
bee
_
i
io
)
«
J “
. «
1 ef
\ ‘
* ;
=
4 - i
eae, te
7 =. 7 ‘

On Julinia; a New Genus of Compound
Ascidians from the Antarctic Ocean.

By

WwW. T. Calman,
University College, Dundee.

With Plates 1—3.

Tuer Ascidian described in the present paper was collected
in the Antarctic Ocean by Dr. C. M. Donald, of the whaler
“ Active,”’ during the Dundee Whaling Expedition of 1892-93,
and was placed along with other specimens in Professor
D’Arcy W. Thompson’s hands. Its large size and remarkable
appearance attracted attention, and closer examination sug-
gested its possible identity with the species described by
Professor Herdman in his report on the Compound Ascidians
of the Challenger Expedition (‘Chall. Rep.,’ vol. xiv, pp 250)
as (?)ignotus. The specimens on which this species
was then tentatively founded, two of them in the Challenger
collection and the third in the British Museum, were in an
unsatisfactory state of preservation, so that Professor Herdman
was unable to determine the genus, or even with certainty the
family, to which the species belonged. Our specimen was pre-
served simply in methylated spirit, but on sectioning it was
found to be in a better state than might have been anticipated,
giving good results with the usual stains and making possible
a fairly complete study of its anatomy. In the following
description, Maurice’s monograph on Fragaroides (‘ Arch.
de Biol., viii, pp. 205—495) has been used for comparison

VoL. 37, PART 1.—NEW SER. A

2 W. T. CALMAN.

throughout, as being the most complete account we possess of
the structure of a Compound Ascidian.

Our species, whose relation to Professor Herdman’s (?)
ignotus I shall again return to, evidently forms the type of a
new genus, which at the suggestion of Professor D’Arcy
Thompson (under whose guidance this paper has been pre-
pared), I propose to dedicate, under the name of Julinia,
to the learned naturalist whose researches on the subneural
gland and other anatomical features form classical contributions
to our knowledge of the Ascidians.

The specimen was found floating on the surface of the sea
in the north of Erebus and Terror Gulf, and Dr. Donald tells
me that considerable quantities were seen. As there is nothing
in its structure to suggest a pelagic mode of life, the colonies
may have been torn from their submarine attachment in shallow
water. At the same time the remarkable genus Celocormus
must be remembered as showing the possibility of a Compound
Ascidian leading an unattached life though destitute of any
special means of locomotion; moreover in the present case
the extreme elongation of the colony renders it improbable that
it could have been supported by a narrow base of attachment.

External Appearance and General Features of the
Colony.

The colony is irregularly cylindrical in shape (Pl. 1, fig. 1),
measuring 78'5 cm. in length and from 1°5 to 2°5 cm. in dia-
meter. One end is torn and ragged, while the other is slightly
tapered and smoothly rounded. No attaching fibres such as
Herdman describes are preserved. Over large areas, particu-
larly along one side, the surface presents a decayed appearance,
the zooids being absent or scattered here and there and
partially macerated. Over the rest of the surface the zooids
are arranged in fairly regular round or oval systems (PI. 1,
fig. 2), each comprising from about six to twelve zooids. The
colour of the whole is yellowish with a tint as though in life it
had been orange, the zooids appearing as lighter spots. The
buccal orifices are small and six-lobed. In the centre of

ON JULINIA. 3

each system is a large and irregularly-lobed common cloacal
opening.

In a transverse section through the colony (Pl. 1, fig. 9)
the zooids are seen arranged round the circumference. From
their proximal ends vascular processes pass inwards, branching
aud interlacing in the interior of the colony. In the further
structure of the framework of the colony we have to deal with
a number of elements more or less obscure, which may be
briefly enumerated but which remain here as in other cases
imperfectly understood. Just below the ascidiozooids large
numbers of peculiar vesicles occur scattered among the vessels.
These vesicles are spherical, from ‘3 to °7 mm. in diameter, and
consist of a denser layer of the test-substance surrounding a gra-
nular mass of variable size and appearance, in which numerous
nuclei are occasionally found. It is possible that these masses
may represent degenerating zooids like those described by
Maurice in Fragaroides, though their great abundance is
rather against this view. Many buds in different stages of
development occur scattered among these vesicles. The centre
or core of the colony is composed of the spongy test-substance,
traversed by the vascular processes, which here run more or less
in a longitudinal direction.

In its minute structure the test resembles that described by
Herdman in Colella (1. ¢., p. 78). In section (PI. 3, fig. 32)
it presents a reticulate appearance, the cavities occupied by the
large vesicular or vacuolated cells known as “ bladder-cells”
reducing the matrix to a mere interstitial network. These
“ bladder-cells”? are of common, but not universal occurrence
among Compound Ascidians.

In the immediate neighbourhood of the zooids and of the
vascular processes the tissue is denser, forming tracts of homo-
geneous matrix-substance which show, besides a few bladder-
cells, small test-cells of the more usual type, with scanty
protoplasm and irregular outline. Near the surface of the
colony numerous groups of yellowish pigment-cells occur.
These are irregularly rounded (probably shrivelled by the
spirit), highly refracting, and apparently homogeneous. In

4 W. T. CALMAN.

the deeper parts of the colony are found scattered here and
there a few small crystalline masses (Pl. 1, fig. 11). These
bodies dissolved somewhat slowly and without effervescence in
hydrochloric acid, but on account of their small number their
chemical nature could not be definitely ascertained, and they
cannot be certainly assumed to represent the stellate calcareous
spicules, e.g. of the Didemnide.

Scattered here and there in the test-substance are groups or
nests of rather large cells (Pl. 1, fig. 10) containing rounded
deeply staining masses of varying size and appearance. The
protoplasm is often very granular, and no nucleus is visible.
It is possible that these obscure bodies represent the test-
phagocytes of Maurice, to which he ascribes the function of
absorbing the dead zooids, but it is to be noted that here they
do not occur in relation to the masses described above as
possibly representing degenerated zooids.

General Conformation of the Ascidiozooids.

The body (Pl. 1, fig. 3) is divided into two regions, the
. thorax and abdomen, united by a narrow neck. The thoracic
region contains the branchial sac, and varies in shape accord-
ing to the state of contraction. In the most expanded speci-
mens it is roughly square-shaped when seen from the side, and
broadly oval in transverse section (PI. 2, fig. 14). A number
of longitudinal muscle-bands run in the mantle, about four to
six on each side. Along the ventral edge of the thorax runs
the large undulating endostyle. On the dorsal surface of the
animal is the wide atrial orifice, behind which is attached the
vesicle described below as probably representing an incubatory
pouch. In front of the atrial orifice is the very large atrial
languet. The languets belonging to the several zooids of a
system meet at their edges to form a composite roof to the
common cloacal chamber; seen from the outside this roof is
hidden by a continuation of the common test, in the centre of
which is the common cloacal opening.

As a result of this arrangement, the individual languets are
somewhat irregular in shape. There appears to be no such

ON JULINIA. 5

continuation of the test-substance on the lower surface of the
languets as is seen in Fragaroides. In the abdomen the
loop of the gut lies transversely to the dorso-ventral axis, the
oval stomach being on the right side. On the dorsal side of
the loop is the ovary, containing one or two large orange-
coloured ova. In front of this the opaque white vas deferens
leads forwards to the cloaca. The heart in its pericardium
lies on the ventral side of the abdomen.

Branchial Cavity.—Tke buccal siphon is short, ending in
front in six simple rounded lobes (Pl. 1, fig. 8). It is lined as
usual by a reflected portion of the test, and a strong sphincter
muscle is developed in its walls (Pl. 2, fig. 21). The oral
tentacles (Pl. 1, fig. 8) are twelve in number, situated on a
hexagonal ridge surrounding the base of the buccal siphon.
Six of the tentacles, situated at the angles of the hexagon, are
larger than the others which alternate with them. Separated
from the tentacles by the “ peribranchial zone” is the “ peri-
pharyngeal band,” here apparently a simple ridge, not grooved
as in Fragaroides. On the ventral side it is continuous with
the anterior end of the endostyle, while on the dorsal side it
appears to fuse, as in Fragaroides, with the small dorsal
tubercle which bears the simple rounded opening of the sub-
neural gland.

The branchial sac is large and comparatively simple. There
are four rows of long and narrow stigmata, each row contain-
ing about twenty-four slits on each side (Pl. 2, fig. 12).
Between the four rows three shelf-like membranes project into
the interior of the sac. These are the “ lames intersériales ”
of Maurice, and are continued round the circumference of the
sac, being only interrupted by the endostyle. Along the middle
of each row of stigmata runs a transverse vessel connected with
each of the interstigmatic bars, but not otherwise interrupting
the stigmata. A similar arrangement to this occurs in the
genus Distaplia. On the dorsal side, but to the left of the
middle line, each interserial lamella is produced into a large
ciliated languet. Of these the two anterior are placed some-
what nearer the middle line than the third. A similar un-

6 W. T. CALMAN.

symmetrical position of the dorsal languets to the left of
the middle line occurs according to Herdman in Distaplia
rosea, and Maurice describes the same arrangement in
Fragaroides.

The branchial sac is connected with the wall of the atrium
by small trabecule placed at the level of the interserial
lamelle (Pl. 1, fig. 8). These are few in number, apparently
only one on each side for each lamella, though I could not
ascertain this with certainty. This condition presents the
opposite extreme to that seen in Fragaroides, where an
almost continuous membrane connects the branchial sac with
the body-wall between each row of stigmata. The wall of the
branchial sac, in contrast to that of Fragaroides, is here
entirely without muscles, except for the longitudinal bands
running in the lips of the endostyle and those at the sides of
the dorsal sinus.

The minute structure of the interstigmatic bars agrees
perfectly with that described by Maurice for Fragaroides.
Along each side facing the cleft is the specialised stigmatic
epithelium (Pl. 2, figs. 17 and 18), formed by six regular rows
of ciliated cells. Each cell forms at its free end a narrow
longitudinal ridge along which the cilia stand in a single row.
The surfaces of the bar turned towards the branchial and
atrial cavities are covered by simple flattened epithelium. In
the centre is a blood-sinus bounded by the basement membrane
of the epithelium.

The endostyle is large aud much undulated. In transverse
section (Pl. 2, fig. 20) it shows essentially the same structure
as that of Fragaroides. The everted lips are covered by
ciliated epithelium. At the bottom of the groove a narrow
band of columnar cells bears the enormously long cilia charac-
teristic of the Ascidian endostyle. On each side of the groove
is the “ glandular ” epithelium, consisting of elongated curved
cells tapering towards their free ends, and this side wall of
“slandular” epithelium is divided into three parts by two
small bands of ciliated epithelium. These two bands are here
identical in structure, while in Maurice’s account of Fraga-

ON JULINIA. 7

roides the inner or more median of the two appears as a
“troisiéme épithélium vibratile de caractéres tout particuliéres.”
Just beneath each of the recurved lips of the endostyle a
muscle-band runs along its entire length. At the posterior
extremity the right lip is continued backwards as a ciliated
ridge (“raphé postérieur”’?) which becomes continuous with
the esophageal funnel.

Digestive Tract.—The funnel-shaped opening of the
cesophagus into the branchial chamber is large and convoluted
(Pl. 2, fig. 13), and the columnar ciliated cesophageal epi-
thelium which covers it is sharply marked off at its edge from
the flattened epithelium of the branchial sac. Close to the
free surface of the cesophageal epithelium is a layer of deeply
staining granules of unknown nature. The csophagus has
thick and convoluted walls (Pl. 2, fig. 19), and its hinder
end projects for a little distance into the stomach, where it
forms a sort of valvular opening (PI. 2, fig. 23).

The stomach (Pl. 1, fig. 4) has the form of an oval sac,
smooth on the outside but having the lining epithelium thrown
into oblique longitudinal folds. This epithelium is very badly
preserved in all our specimens, but it can be seen that the cells
are columnar and stain deeply.

Embedded in the lining epithelium of the stomach, or lying
free in its cavity, are certain large oval or pyriform cells
(Pl. 2, fig. 25). These have sharp outlines and very granu-
lar protoplasm, stained only faintly by hematoxylin and not
at all by carmine. The nucleus is large and transparent, and
contains a large deeply-staining nucleolus. In one or two
cases a small segment, marked off from one end of the cell as
shown in fig. 25 a, confirms a belief suggested by the nuclear
and other characters that these are parasitic Gregarines. _

A slight constriction separates the stomach from the
intestine (Pl. 2, fig. 23). The latter is of nearly uniform
diameter throughout its length, there being no division into
duodenum, chylific ventricle, and rectum. Its walls are
slightly folded, and it is lined throughout by columnar epithe-
lium. The terminal part projects slightly into the cloacal

8 W. T. CALMAN.

chamber, forming what Maurice calls a “ pavillon anale,” but
I have failed to find an anal sphincter, though such in all
probability exists.

The intestinal gland is well developed. A series of tubules
ramifies over the surface of the intestine (Pl. 2, fig. 24);
these are lined by somewhat irregular cells, and unite to form
a duct which crosses the intestinal loop to open into the
stomach at about its middle (Pl. 2, fig. 23).

Nervous System.—The nerve-ganglion is oval in form,
and gives off numerous nerves. Two large nerves in front
and a large median nerve (the cloacal), flanked by two smaller
lateral ones behind, are easily seen; but it was found impossible
to determine the number or arrangement of the smaller lateral
nerves.

On the dorsal wall of the dorsal blood-sinus a slight thicken-
ing composed of a few small cells (Pl. 2, fig. 22) represents
the “cordon viscéral ganglionnaire” of van Beneden and
Julin, the remains of the central nervous axis of the larva. In
a fortunate series of sections this cord can be traced into con-
nection with the cerebral ganglion, which it enters at the root
of the cloacal nerve.

In minute structure the ganglion corresponds exactly with
that of Fragaroides (Pl. 2, fig. 21). Outside is a layer
of large nerve-cells, finely granular and with conspicuous
nuclei, sending processes inwards—the “grey matter’’ of Julin,—
while the interior is composed of the usual fibrillated “ punkt-
substanz”’—the “ white matter” of Julin.

Subneural Gland.—The subneural gland (Pl. 2, fig. 21)
is very feebly developed, consisting merely of a thin layer of
cells lying on the posterior face of the nerve-ganglion, separated
from it in the middle line by a space representing the back-
ward continuation of the duct. The tissue of the gland con-
sists of ill-defined rounded cells, many of them containing
vacuoles.

The duct, on the other hand, is well developed and normal
in structure. As in Fragaroides, its anterior wall is con-
tinued back over the posterior face of the ganglion for some

ON JULINIA. 9

distance, while the ventral wall stops short of this and becomes
continuous with the band of glandular tissue. The columnar
epithelium lining the duct is reflected at its mouth over the
‘dorsal tubercle,” and in the interior of the duct bears the
usual long cilia, one to each cell.

It is interesting to note that the extreme degeneracy of the
gland itself (carried to a greater degree, I think, than has been
described in any other Ascidian) is not associated with any
corresponding reduction in the size of the duct. Indeed, even
in other cases where the gland is much better developed (e. g.
Fragaroides), the duct, both in respect to size and to the
richness of its ciliation, seems out of all proportion to any
secretory function we can ascribe to the gland.

Circulatory System.—The heart in its pericardium lies
on the ventral side of the intestinal loop (Pl. 1, fig. 4;
Pl. 2, fig. 16). It is a thin-walled, somewhat fusiform tube,
but I was unable to trace any vessels connected with it.
A space lined by endothelium and lying to the left of the peri-
cardium may possibly represent a vestige of the epicardial
system. A large blood-sinus runs along the dorsal edge of the
branchial sac (Pl. 2, figs. 14 and 22); in its lateral walls
run strongly developed longitudinal muscle-bands, while in the
middle of its dorsal wall runs the ganglionic nerve-cord above
alluded to.

The “vascular process” arises near the hinder end of the
abdomen, a little to the left side. In the substance of the
test these processes branch and probably anastomose, and
finally end in club-shaped dilatations (Pl. 1, fig. 6). In
section (Pl. 3, fig. 32) each is seen to consist of a simple
tube of ectoderm divided into two by a median partition of
thin and apparently structureless membrane. No internal
layer of connective tissue, as described by Herdman in Colella,
could be distinguished, the processes appearing to be exclusively
ectodermal, like the ‘‘stolons” of Botryllus (Hjort, ‘ Mitt,
Zool. Stat. Neap.,’ x, p. 589), a point which is of interest in
connection with the probable origin of buds from them.
Numerous blood-corpuscles occur in the cavity of these tubes.

10 Ww. T. CALMAN.

Sexual Organs.—The sexual organs lie on the dorsal side
of the intestinal loop, the ovary being external or dorsal to the
testis. In all the individuals examined several ova were found
in various stages of development, the largest being about
‘6 mm. in diameter. They are contained in follicles opening
by short canals into the oviduct (Pl. 3, fig. 27). No very
young ova were seen, the smallest being about ‘1 mm. in
diameter, and nowhere did the ovary present in section the
bilateral T-shape described by Maurice in Fragaroides and
by van Beneden and Julin in Clavellina. Inside the wall of
the follicle in all save the youngest there is a loose layer of
cells outside the egg-membrane (PI. 3, fig. 28). This seems
to correspond to the inner of the two layers into which the
follicular wall splits in Clavellina according to van Beneden
and Julin (‘ Arch. de Biol.,’ vi, p. 358), though in the latter
case the splitting does not take place until the ovum is nearly
ripe. Inside the egg-membrane are the characteristic and
problematical “ test-cells,” with large nuclei each containing
several nucleoli, and scanty protoplasm. The vitellus becomes
more and more coarsely granular as the egg matures, and in
the oldest eggs it is broken up into numerous irregular yolk-
masses. The oviduct is a thin-walled tube running up alongside
the vas deferens (PI. 2, figs. 15 and 16).

On the dorsal surface of the animal’s thorax, near its junc-
tion with the neck and rather to the left of the mid-dorsal
line, there is attached by a narrow neck a spherical thin-walled
vesicle of about ‘5 mm. in diameter (PI. 1, fig. 3). The wall
of this vesicle is composed of two layers, the outer continuous
with the ectoderm of the thorax, while the inner seems to be
continuous with the wall of the oviduct. It is difficult to make
out the exact relations of the parts, but I believe that the lumen
of the oviduct opens into the vesicle. This organ occupies the
position of the incubatory pouch possessed by so many Com-
pound Ascidians, and its apparent connection with the oviduct
naturally suggests that it represents that structure, though I
have never found eggs or larvee in it in a single instance.

In nearly all the specimens examined hardly a trace of

ON JULINIA. al

testis could be found, although the vas deferens was usually
full of spermatozoa. In one or two cases, however, the testis
was developed. It consisted (Pl. 3, fig. 26) of a number of
follicles lying in the connective tissue in the intestinal loop,
connected by branching tubes with the vas deferens and
showing groups of developing spermatozoa in each. In one
specimen (badly preserved, unfortunately) the follicles are
considerably larger than those figured, and the lower part of
the vas deferens is very much distended (to °3 mm. diameter)
with spermatozoa. No definite relation between the states of
maturity of ovary and testis such as would suggest the
occurrence of protandry or protogyny could be demonstrated.

The vas deferens is usually about ‘06 mm. in diameter, and
is composed of cubical epithelium, probably ciliated. Its
hinder part is much convoluted, though not spirally coiled as
in the Didemnide; in front it pursues an undulating course
to open close beside the anus.

The spermatozoa are linear, slightly curved, about ‘001 mm.
in length, and bear long flagella.

Buds.—Numerous buds in various stages of development
are scattered in the substance of the test just below the layer
of ascidiozooids. All of them, even the youngest, lie quite
free and unattached, usually alongside a vascular process (Pl.
1, figs. 5 and 6). In no case did I succeed in tracing a con-
nection with the parent animal, and the origin of the buds
must therefore be left undecided—more particularly since
Hjort (‘ Mitt. Zool. Stat. Neap.,’ x, pp. 588 and 589) has
contested Herdman’s account of the process of ‘stolonial ”
budding in the Botryllidz. In view of the obscurity still
involving many points in the process of bud development in
the Tunicata, it did not seem advisible to attempt a detailed
examination of our very minute and imperfectly preserved
specimens. One point of interest, however, is the presence of
several large cells, apparently ova, in all the buds examined
(Pl. 3, figs. 29—31). The appearance of these buds with their
contained ova closely resembles those figured by Kowalevsky in
Distaplia (‘Arch. f. mikr. Anat.,’ x, 1874). Herdman also

12 W. T. CALMAN.

notices the early appearance of ova in the buds in several cases
(e.g. Colella), and Hjort (1. c., pp. 604 and 605) has recently
called attention to the migration of the egg-cells from the
parent animal into developing buds in the case of Botryllus.
It is possible that we have here to do with an instance of such
migration, but on the other hand, if the buds arise, as seems
probable, from the “ vascular processes,” it is very difficult to
see how the ova could have reached their destination.

The youngest buds (PI. 1, fig. 6, and Pl. 3, fig. 29) consist
of a two-layered vesicle, the space between the layers contain-
ing loosely scattered cells and one or two ova. Older buds
(Pl. 1, fig. 5, and Pl. 3, fig. 31) show the two peribranchial
cavities developed at the sides of the branchial sac. In the
latter the endostyle can be seen in sections as a pair of ridges
near its anterior end. The neural tube lies dorsally to the
branchial sac, but its communication with the latter could not
be traced. In sections of the abdominal region (Pl. 3, fig.
30) the rudiment of the heart is seen in its pericardium, and
on either side of the latter are the two “epicardial tubes,”
which, as has been already stated, are in this genus absent or
rudimentary in the adult, but which in Fragaroides persist as
a great space running the whole length of the body. A long
“tail” projecting from the hinder end of the older embryos
(Pl. 1, fig. 5) is probably the outgrowing vascular process.

Systematic Position.—The unnamed genus referred to
by Professor Herdman as (?) ignotus, and already men-
tioned above, is so similar in general appearance as well as
in habitat to the form here described that the two must be
nearly related, if not identical. The poor state of Professor
Herdman’s specimens, which accounts for the brevity of his
description, may also account for certain discrepancies, as for
instance when Professor Herdman describes the tentacles as
numerous and equal, and the nerve ganglion as spherical in
form. Our species, however, certainly does not belong to the
Polyclinidz, to which Herdman refers his (?) ignotus
“from the general appearance of the colony and of the
ascidiozooids.” The genus being certainly nondescript and

ON JULINIA. 13

requiring a name, I give the animal at the same time a new
specific one in the face of what uncertainty remains as to its
identity with Professor Herdman’s type.

As regards the affinities of the new genus, we are confronted
by the fact that the diagnoses of recognised families appear to
be somewhat artificial, and certainly do not lend themselves
very readily to the reception of new forms. Taking, however,
the received families as they at present go, Julinia is obviously
excluded from the Polyclinide by the absence of a post-
abdomen. This negative character, together with the distinct
separation of thorax from abdomen, are characters common to
the three families of Distomide, Didemnidz, and Diplo-
somide. From the last two families our species is excluded
by the absence of retractile muscles in the vascular processes,
by the large anal languets (absent in Giard’s definition from
the Diplosomidz), by the want of the calcareous spicules so
characteristic of the Didemnidz, and by the characters of
the testis and vas deferens, which latter in the Didemnide
is spirally coiled round the single large testis. Narrowing
down our genus accordingly to the Distomide, we find that
it agrees with that family in the numerous spermatic vesicles
of the testis, as well as with individual genera of that family in
its incubatory pouch and large atrial languet. The incubatory
pouch is common to Colella and Distaplia, while the atrial
languet is absent in Colella. The general sum of the
characters, then, seems to bring our genus nearest to Dis-
taplia, with which latter genus it further agrees in the
characters of the branchial sac with its four rows of long
stigmata crossed by intermediate transverse vessels. It differs
chiefly from the definition of Distaplia in the form of the
colony, as well as in the ascidiozooids being embedded in the
test instead of forming prominent knobs or lobes.

The following is the diagnosis of the new genus:

JULINIA, gen. nov.

Colony cylindrical, excessively elongated; ascidiozooids com-
pletely embedded in the fleshy or gelatinous test; systems

14 W. T. CALMAN,

distinct, with well-developed common cloacal cavities ; buccal
orifices small, six-lobed ; common cloacal orifice large, irregu-
larly lobed; zooids with thorax and abdomen united by a
narrow neck; atrial languet very large ; vascular process well
developed; branchial sac rather large, with four rows of long
stigmata crossed by narrow intermediate bars ; dorsal languets
large, three in number, placed to the left of the middle line;
stomach with internal longitudinal folds.

Species.—Julinia australis, n. sp., with the characters
of the genus.

Habitat.—Antarctic Ocean.

EXPLANATION OF PLATES 1—3,

Illustrating Mr. W. T. Calman’s paper “On Julinia, a New
Genus of Compound Ascidians from the Antarctic Ocean.”

Reference Letters.

atr. 1. Atrial languet. 6. Bud. 6.7. Buccal lobes. 4. s. Buccal siphon.
b. sph. Buccal sphincter muscle. 6/. c. Bladder cells of test. 47. g. Blood-
corpuscles. 47. s. Branchial sac. c. c/.0. Common cloacal orifice. cl. 0.
Cloacal orifice. c/. . Cloacal nerve. d./. Dorsal languet. d.. Dorsal
ganglionic nerve-cord. d@.s. Dorsal branchial sinus. d. ¢. Dorsal tubercle.
ect. Ectoderm. ed. Endostyle. epe. Rudiment of epicardial system (?) in
adult. epe. ¢. Epicardial tubes in bud. 7. Outer layer of follicular wall of
ovary. jl’. Inner layer of follicular wall of ovary. gi. Subneural gland.
ht. Heart. ine. p. Incubatory pouch. ixé¢. Intestine. ind. 7. Interserial
lamina of branchial sac. in¢. g/. Intestinal gland. zt. d. Duct of intestinal
gland. 7. m. Longitudinal muscle-bands. x. g. Nerve gland. as. Gsophagus,
es. f. @sophagus funnel. ov. Ova. ovy. Ovary. ovd. Oviduct. pdr. Peri-
branchial cavity. p.c. Pigment corpuscles. per. Pericardium. p. ph. d.
Peripbaryngeal band. *.p. ‘‘ Raphé postérieur.” s¢. Stomach. ¢.c. “Test-
cells” of ovum. ¢.c’. Test-cells. ¢. Trabecule from branchial sac to
mantle. ¢r. 6. Intermediate transverse bars of branchial sac. ¢s. ¢. Tubules
of testis. v.d. Vas deferens. v. p. Vascular process.

ON JULINIA. 15

PLATE 1,

Fie. 1.—General view of colony slightly reduced.

Fic. 2.—A single system of ten individuals x 7. In the centre of the
system is the irregularly-lobed cloacal orifice; on the outer side of each
ascidiozooid the endostyle is seen shining through.

Fic. 3.—A_ single ascidiozooid seen from the left side x 20. This figure
illustrates the form of the branchial sac, the convoluted endostyle, and the
relations of the incubatory pouch and atrial languet. In the abdomen the
ovary and part of the vas deferens are seen on the dorsal side of the intestinal
loop.

Fig. 4.—Ventral view of abdominal region, showing the oblique folds in the
stomach wall, and the position of the heart and part of the vas deferens in the
loop of the gut. The origin of the vascular process is seen on the lower side
of the intestine.

Fic. 5.—Bud lying beside vascular process.

Fie. 6.—Younger bud near club-shaped end of vascular process.

Fic. 7.—View of part of common cloacal chamber seen from within, showing
four atrial languets pointing inwards and meeting together. .In one indi-
vidual of the system the oral region is also shown. The roof formed to the
common cloacal chamber by these converging atrial languets is not to be seen
from the outside, it being covered in its turn by a second roof of test-
substance.

Fie. 8.—Oral region of an ascidiozooid seen from within, showing the six-
lobed buccal orifice, the hexagonal ring of twelve tentacles, the peripharyngeal
band connected dorsally with the dorsal tubercle and ventrally with the endo-
style; the nerve ganglion, the first row of stigmata, and the trabecule con-
necting the stigmatic wall with the atrial wall are also seen.

Fie. 9.—Transverse section through one half of the colony x 4, showing
the position of the ascidiozooids, the distribution of the vascular processes
through the matrix of the colony, and the zone of problematic vesicles referred
to on p. 3 as possibly representing degenerated zooids.

Fic. 10.—Cell elements from the test supposed (p. 4) to be phagocytes.
Zeiss, ~zth hom. imm., 2.

Fie. 11.—Spicules from test. Zeiss, D. D., 2.

16 W. T. CALMAN.

PLATE 2.

Fic. 12.—Internal view of dorsal portion of branchial sac, showing the
four rows of stigmata, the transverse vascular bars which cross them, the
unsymmetrical position of the three dorsal languets, and the interserial
lamelle with which these latter are connected.

Fic. 13.—The ceesophageal funnel and its connection with the endostyle.

Fig. 14.—Semi-diagrammatic transverse section across the thorax at the
level of the cloacal opening, showing the proportions of the endostyle and the
relations of the dorsal sinus and the degenerate rudiment of the dorsal nerve-
cord.

Fic. 15.—Transverse section through the constricted neck between thorax
and abdomen, showing the relations of csophagus and intestine to vas
deferens and oviduct.

Fie. 16.—Transverse section through upper portion of abdomen, showing
the stomach and intestine in relation to vas deferens and oviduct, and to the
heart, pericardium, and (?) epicardial tubes.

Fic. 17.—Transverse section of an interstigmatic bar, showing the con-
tained blood-sinus, the characters of the ciliated epithelium on its lateral
edges, and of the flattened epithelium on its inner and outer sides. Zeiss, 73th
hom. imm., 2.

Fic. 18.—Surface view of an interstigmatic bar.

Fic. 19.—Transverse section of esophagus, showing its folded walls and
columnar ciliated epithelium.

Fic. 20.—Transverse section of endostyle, showing the epithelial charac-
ters of its lips, side walls, and central groove. Zeiss, KH, 2.

Fie. 21.—Ventral section of nerve ganglion and associated organs, showing
the buccal sphincter, the ganglion with origin of the cloacal nerve, the gland
of Julin, and the course of its duct. The long and well-preserved cilia in the
duct are all pointing inwards. Zeiss, H, 2.

Fic. 22.—Transverse section of the dorsal branchial blood-sinus, showing
dorsally the rudimentary dorsal nerve-cord, and laterally, near the attachment
of tle sinus to the thoracic walls, its longitudinal muscular bands. LH, 2.

Fic. 23.—Longitudinal section of the abdomen, showing the valvular lip
of the cesophagus where it enters the stomach, and the entrance into the latter
organ of the duct of the intestinal gland. Some tubules of the testis are
again seen within the intestinal loop.

Fic. 24.—Longitudinal section of part of the intestinal gland. The gland
is seen lying against the wall of the intestine, while its duct passes off in the
opposite direction towards the stomach. Zeiss, ;',th hom. imm., 2.

Fie. 25.—Gregarines from stomach. Zeiss, ;';th hom. imm., 2.

ON JULINIA. 7

PLATE 3.

Fic. 26.—A portion of the testis, showing its racemose follicles containing
spermatozoa. Zeiss, th hom. imm., 2.

Fie. 27.—Part of the ovary, showing one ovarian follicle and a portion of
another, both in connection with the oviduct. D D, 2.

Fic. 28.—Peripheral layer of ovum with its test-cells, surrounded by the
two-layered follicular wall. Zeiss, 4,th hom. imm., 2.

Fic. 29.—Transverse section of a young bud, showing included ova.
Zeiss, D D, 2.

Fic. 30.—Transverse section of abdominal region of older bud. In this
section we see, besides the cesophagus, the intestine, and a large ovum, the

developing heart within its pericardium and the two associated epicardial
tubes. D D, 2.

Fic. 31.—Optical longitudinal section of a bud of about the same age as
that represented in Fig. 30. The branchial sac and the paired rudiments
of the atrium are here shown. D D, 2.

Fic. 32.—Transverse section of a vascular process, with its surrounding
space and adjacent portions of the test-substance. Large bladder-cells, small
test-cells, and pigment corpuscles are seen within the matrix of the latter.

VOL. 37, PART 1.—NEW SER. B

¥E AVANT? :

Ff oy Ay}

iraiaii Yaa etre {pl Trip BAO iE Aber i il iL ho mits | }

; 2G Pitt et) ‘ pl] ~ woygit fs
4

hy F wil Se ivi 5 fj ey JCS i> Dl wow? Yi tO) | =e ik

é
1) [ Phare ai i j toac thi SAP AT ah ol ae

gibb 7st a ea ein off 2h ieee hl ‘ 7 f Uae
Deitel ee ie chthobiok fi ep
7 wauily fohulsde ware inl 2s ae
. aU) ae
Praia ll habs a) rt | BaD
ne iy | pili Did Mic pend ew ae
hi ives: B pit) iih4 : j (ele } ‘i Pew EOE
P eb
) ¥
Bia Pe Ati ’ aS rT Ha i he
Pabsialh 04, fon ’ ee Tea ih ae,
P = ee | | s rt as ME
H/ f 4.84, u Rory 74 ee =e
fiaie ray isu nau! } i ;
i eee) Al Wael | TT ita
E i.
i, 3 -
, -
= ‘2
. -
i 2 ie
z e
q eee oe
of : os
e ¥ vv = :
' re) si" ' 4
= } @ g ° , v ye
° q 7 oe nih j
: Be :
oe ae
Fi be :

HERMAPHRODITISM IN MOLLUSOA. 19

Hermaphroditism in Mollusca.
By

Dr. Paul Pelseneer (Ghent).
With Plates 4—6.

ConTENTS.
I. HermMapHropitism In Moutusca.
1. Amphineura. 2. Gastropoda. 3. Lamellibranchia.

II. Puytoceyetic Evotution or THE HERMAPHRODITE GLAND IN
Mo.tusca.
1. Gland with hermaphrodite acini.
2. Gland with separate male and female acini.
3. Gland with separate male and female regions.
4. Separate male and female glands in the same individual.

ILI. Puystonocicat Evotution oF HERMAPHRODITISM IN THE MOoLLUscaN
INDIVIDUAL.

IV. Ortcin or HERMAPHRODITISM IN MoLuusca.
1. Hermaphroditism has arisen from the unisexual state.
2. Hermaphroditism has been grafted on the female organism.

V. Conclusions.

I. HermMaruropitism 1n Mo.uusca.

Hermarxropirism is found to occur in every class of Mol-
lusca except the Cephalopoda and Scaphopoda.

1. Among Amphineura, in all the Neomeniide.

2. Among Gastropoda, in four genera of Streptoneura
(Valvata, Onchidiopsis, Marsenina, and Entoconcha),
and in all the Euthyneura (Opisthobrauchia and Pulmonata).

3. Among Lamellibranchia, in various species of Pecten,

20 PAUL PELSENEER.

Ostrea, and Cardium; in Entovalva, the Cycladide, the
Poromyide, and all the Anatinacea.

1, Amphineura.

(1) Neomeniidx.—In each of the two contiguous herma-
phrodite glands of these animals the ova arise in the axial half
of the organ, and the spermatozoa in the lateral half (1).

(2) Cry ptochiton.—According to Middendorf (2) this genus
is hermaphrodite. I have, however, been able to obtain a
specimen in alcohol of this form (C. Stelleri, from the North
Pacific), and have found it to present a distinctly unisexual
condition, as in Chiton. I have also satisfied myself that
the sexes are separate in Chitonellus larveformis (=
fasciatus) and Cryptoconchus porosus (or monticu-
laris). Thus the diccious condition is universal in the Poly-
placophora.

2. Gastropoda.

(1) Valvata.—The hermaphrodite gland of this genus is
formed of acini, ali producing ova and spermatozoa (8).

(2) Marsenina and Onchidiopsis (of the family Lamel-
lariidze).—While the hermaphroditism of Valvata has been
recognised by a fair number of zoologists, that of the two
genera just named has only been affirmed by Bergh (4).

As the number of monecious Streptoneura is extremely
limited, an examination of these two genera was desirable in
order that Bergh’s discovery might be confirmed and his state-
ment of facts extended. I have only succeeded in obtaining a
single specimen of Onchidiopsis greenlandica, and upon
it I have made the following observations.

The genital gland occupies the posterior part of the visceral
mass. Its duct is provided with an accessory mass formed of
closely packed, narrow, tubular ceca; it then bifurcates, and
the more lateral branch (the right), after first receiving a large
bent pouch (fig. 1, rv), immediately thickens before opening to
the exterior, and again receives on its right side a large pouch
of flattened form. The aperture of this branch is situated in

HERMAPHRODITISM IN MOLLUSCA. 21

the mantle cavity, in front of the anus and more to the right ;
it is the female orifice (v1).

The other branch immediately presents a glandular mam-
millated mass with closely packed lobules, and then extends
forward under the integument over the neck of the animal as
far as a considerable projection of the body-wall on the right
side, the penis, which it traverses from end to end (fig. 1, v).

The structure of the reproductive apparatus thus differs
from that of Valvata in the presence of glands upon the her-
maphrodite portion of the duct, and by the absence of a dif-
ferentiated “ uterus ”’ in the female portion.

The glands of the hermaphrodite duct (fig. 1, x) do not
appear to me to be vitellogenous or albuminiparous glands
(judging, at least, from the imperfectly preserved specimen
which I have examined). I regard them rather as vesicule
seminales.

As for the two appendages of the female duct, the elongated
pouch is a receptaculum seminis (or poche copulatrice),
and the flattened pouch (fig. 1, 111) is the mucous or jelly
gland. Lastly, the glandular mass of the male duct can be
termed the prostate, as in other hermaphrodite Gastropods.

If the presence of two genital apertures could not, in itself,
demonstrate the hemaphroditism of Onchidiopsis, the struc-
ture of the genital gland is sufficient to remove all doubts.
This gland is formed of parallel ceca, bifid at their extremities
(fig. 3): these terminal divisions are ovogenous, while the
proximal portions of the duct are spermatogenous. As a result
of this arrangement the gland appears to be composed of a
superficial or external female portion, and of a deeper or
central male portion. However, these two regions are not
demarcated with any regularity, and in the middle portion
male and female acini can be seen in sections lying side by
side (fig. 2). At all events, the products of the two sexes
either do not arise in the same cecum or they do not arise in
the same region of a cecum.

According to Bergh’s observations, Marsenina appears to

22 PAUL PELSENEER.

be constructed upon the same plan,—that is to say, the genital
czeca are female in their terminal portion.

(3) Entoconcha.—This genus, so degraded by parasitism,
has unfortunately not yet been studied by the method of serial
sections. As is known, there is a female genital gland in the
middle region of the body, and a number of spermatogenic
capsules further back towards the posterior orifice of the body.

(4) Bulloidea.—The genital gland is here formed of acini,
all hermaphrodite (Bulla, Limacina, &c.); but in certain
forms (Acton, Pelta, Lobiger) I have found it composed
of distinct male and female acini.

(5) Aplysioidea.—The same acini produce spermatozoa
and ova (Aplysia). I have found, however, that the male and
female acini are distinct in certain specialised forms ; the former
(male) occupy the central, and the latter the peripheral region
(Pneumonoderma, Clinopsis).

(6) Pleurobranchoidea.—The genital gland is uniformly
composed of hermaphrodite acini in Umbrella. But in
Tylodina and all the Pleurobranchide I find that the acini
are either male or female, and that the latter open into the
former.

(7) Nudibranchia.—Meckel was the first to notice in certain
Nudibranchs that the genital gland is formed of male and of
female acini; but he supposed that these acini of different
sexes were without communication with one another, and that
in the genital ducts there was a male canal embedded in the
female one (5).

Nordmann immediately afterwards showed that in Tergipes
the ovular acini open into capsules full of spermatozoa; but
he took the entire hemaphrodite gland for an ovary merely, and
held the sacs full of spermatozoa to be “ poches de fécon-
dation” (6).

It was R. Leuckart who first determined and correctly inter-
preted the constitution of the hermaphrodite gland of Nudi-
branchs, especially Eolis (7),—i. e. peripheral acini exclusively
ovular, opening into central chambers producing spermatozoa.
His interpretation has been confirmed since for a great

HERMAPHRODITISM IN MOLLUSOA. 23

number of genera by various authors (Hancock, Bergh,
Trinchese, and myself); the same is the case with those
Nudibranchs in which the male and female acini are easily
distinguished under a lens of low power (e.g. Fiona).

It must be noticed, however, that in Eolis (Cory phella)
Landsburgii the acini of the hermaphrodite gland produce
ova in their distal portion and spermatozoa in their proximal
portion (8) ; this arrangement has been recognised as general
in all the Elysioidea (Cyerce, Herma, Elysia, Lima-
pontia).

(8) Pulmonata.—lIn these Gasteropods the genital gland is
formed of hermaphrodite acini both in the Stylommatophora
(e.g. Helix) and in the Basommatophora (e.g. Auricula).

In Siphonaria (Basommatophora), according to Haller
(9), each acinus of the hermaphrodite gland is exclusively of
one sex, either male or female. In S. Algesirz, which I
have studied, I have observed that the conformation of the
gland is precisely analogous to that found in Onchidiopsis,
the Pleurobranchide, and the Nudibranchia,—that is to say,
the peripheral female acini open into the more centrally
situated male acini.

However, this conformation is not so entirely different from
that presented by the other Pulmonata. I know cases, in fact,
in which the wall of the genital gland already shows a distinct
sexual differentiation upon the two sides of the follicles, and
in which the female side exhibits projections which are the
rudiments of acini of this sex (Amphibola).

It follows, then, from what has been said above, that
examples of the various possible modes in which the genital
gland is constituted are to be found side by side in all the sub-
groups of hermaphrodite Gastropods.

8. Lamellibranchia.

(1) Ostrea.—The question of sex in oysters has long been
a subject of controversy, and its solution, which presents
decided difficulties, is not yet universally recognised.

24 PAUL PELSENEER.

The first point capable of immediate demonstration is that
there are both hermaphrodites and dicecious oysters.

a. Hermaphrodite Oysters.

a. Ostrea edulis, Linné, the hermaphroditism of which
was first demonstrated by Davaine (10), and subsequently
confirmed by Lacaze-Duthiers, Hock, &c.

b. Ostrea stentina, Payr. (= plicata, Chemnitz) ; also
recognised as hermaphrodite by Lacaze (11).

c. Lastly, it has been recently stated that the oyster of the
N.W. coast of America is also hermaphrodite ; this, I suppose,
is Ostrea lurida. .

In the genital glands of hermaphrodite oysters the sperma-
tozoa and the ova arise in the same acini, but at different times,
so that the products of the two sexes are not often to be seen
in the same individual. This perfect alternation is the most
striking and distinctive characteristic of the heramaphroditism
of oysters ; it explains how it has been believed, and how at
first sight it would still be possible to believe, that the oyster
is unisexual.

The male products are the first to appear. This protandry
was discovered by Davaine (13), and confirmed by P. J. van
Beneden (14).

B. Diecious Oysters.

a. Ostrea virginica, Lister, from the E. coast of the
United States (15).

b. Ostrea angulata, Lam. (= lamellosa, Broc.), the
‘< Portuguese” oyster, from the Atlantic (16).

c. Lastly, I have made out the separation of the sexes in a
third species, O. cochlear, Poli, from the Mediterranean,
which belongs to the same subdivision (Gryphza) as the
preceding form.

I have not had the opportunity of studying this species alive,
but I have had numerous specimens of it of very different sizes,
collected at different times of the year.

Of all these there was not a single individual which pre-

HERMAPHRODITISM IN MOLLUSCA. 25

sented both ova and spermatozoa in the genital gland at the
same time. Each specimen had the glands either full of the
products of one sex exclusively (fig. 8), or else almost empty
(fig. 7).

Nevertheless it is impossible to suppose that they repre-
sented the successive stages, male and female, of an hermaphro-
dite condition: in the first place, because the individuals with
male glands were no smaller than those with female glands,—
there were female specimens smaller than the males, and males
and females of every size; in the second place, because the
appearance and conformation of the genital glands differs con-
siderably according to the nature of the contents, as in Lamel-
libranchs of different sexes. The glands with spermatozoa are
formed of ramifications having an appreciably constant dia-
meter (fig. 4, 11) ; the glands with ova are more lobulated, and
so present a distinct appearance, which shows that O. cochlear
is dicecious, like O. virginica and O. angulata.

I will add a word here upon the genital apertures of
O. cochlear. They are asymmetrical, that of the left
being the more anterior (fig. 6, rx), considerably in front of
the adductor muscle. The “fente uro-génitale,” discovered
by Hoek in O. edulis (17) (where the genital gland opens
into the urinary slit), has not in this species the simple
appearance which it presents in O. edulis. The genital and
renal orifices, although adjacent, are quite distinct ; the genital
is more anterior than the renal, less distant from the median
plane (fig. 7, v), and directed outwards, whilst the renal
aperture is directed towards the axis (fig. 7, x).

(2) Cardium oblongum.

Lacaze-Duthiers discovered the hermaphroditism of a species
of this genus, C. norvegicum, from the Atlantic. I have
further observed this condition in a Mediterranean species,
C. oblongum, which is grouped with C. norvegicum in a
special division or sub-genus (Levicardium).

In Cardum oblongum the different acini are each of one
sex, either male or female; but the acini of the same sex are

26 PAUL PELSENEER.

not confined toa particular region (fig.9) asin Onchidiopsis,
for example, although the female acini open into the male
acini as in the last-named genus. I have never, however,
noticed ova and spermatozoa in the same cul-de-sac, as Lacaze-
Duthiers found to be constantly the case in C. norvegicum
(18).

(3) Entovalva.

The structural details of the hermaphrodite genital gland
are not known. There exists a single gland, not differentiated
into sexual regions, on each side (19). Probably, therefore, it
is constituted like that of Ostrea and Cardium.

(4) Pecten.

The greater number of species of this genus hitherto exa-
mined have been found to be hermaphrodite, viz. P. glaber
(20), P. Jacobeus (20a), P. maximus (21), P. opercularis
(22), P. irradians and magellanicus (28).

I can myself vouch for the hermaphroditism of P. glaber
and maximus. Moreover I have observed that P. flexuosus,
Poli, from the Mediterranean, also has the two sexes united in
the same individual (fig. 10).

On the other hand, among all the species examined, I have
only met with two diccious forms, viz. P.inflexus, Poli, from
the Mediterranean, and P. varius, Linné, from the Atlantic,
in which latter form this fact was already known (Humbert,
23a).

The hermaphroditism of Pecten is recognised with unusual
facility. The anterior part of the visceral mass may be readily
seen from the outside to be whiter than the posterior part ;
the whiter part is the male region. Examination of fresh
material, or a section, shows at once that the genital products
of different sexes are formed in different regions of the herma-
phrodite gland (fig. 10). At the same time, as is already
known in the case of certain species, there is only one common
genital orifice on each side (opening into the nephridium), and
only a single genital duct, which ramifies in the various parts

HERMAPHRODITISM IN MOLLUSCA. 27

of the ovotestis (P. glaber, maximus, flexuosus). The
male region is thus nearer the genital aperture than the
female region.

(5) Cycladide.

a. Cyclas.—The hermaphroditism of this genus has been
known from the time of v. Siebold (24). But the conformation
of the genital organs has not been completely described even
by subsequent authors (25) ; Stepanoff has merely made known
that one portion of the gland is male and the other female.

In C. cornea (fig. 12) the genital gland has a superficial
position upon each side of the posterior part of the body. It
is of scarcely any extent except in length (from liver to
nephridium), its height and depth being inconsiderable. It is
transversely divided into two portions of different appearance
and size, separated by a constriction in which there is merely
a ciliated duct uniting the two portions (fig. 11, 11). The
most anterior division is more elongated and more appreciably
lobulated than the other; it extends as far as the liver. This
is the male region of the gland ; the other is exclusively female
(he. LL, 11).

The hermaphrodite gland is thus seen to be here divided
into two portions of different sexes ; and these are no longer
in immediate contact (as in Pecten), but already separated
from one another, though connected by the intermediation of
a duct.

Lastly, the posterior (female) portion of the gland is con-
tinued backwards by an hermaphrodite duct which terminates
at the genital orifice. The latter, which has escaped notice
hitherto) is situated outside the visceral commissure, near the
most ventral point of the nephridium (where the aperture of
this organ is placed), and in front of the posterior retractor
muscle of the foot (fig. 12, v1).

B. Pisidium (26) and Corbicula (fide v. Jhering in
litera) are also hermaphrodite. The conformation of the
genital gland appears to be the same in them as in Cy clas (27).

28 PAUL PELSENEER.

(6) Anatinacea and Poromyide.

Several years ago (28) I made known the hermaphroditism
of Tbracia, Lyonsia, Lyonsiella (Anatinacea), and of the
allied family Poromyidez (Septibranchia) ; and at the same
time I confirmed its existence in Pandora, in which it had
already been once affirmed (29). Having thus demonstrated
the monecious nature of all the genera of Anatinacea which I
had examined, I was led to believe in the hermaphroditism of
the entire group, a conclusion which accorded well with
Lacaze’s observations upon Aspergillum (80).

Since that time I have been enabled to study one other
genus of this group, Clavagella, although, in spite of nume-
rous efforts, I have not succeeded in obtaining either Anatina
or Pholadomya, which appear to be rare. I have observed
the same monecious arrangement in Clavagella as in the
other Anatinacea previously studied. There can accordingly
be no doubt as to the hermaphroditism of this group, since no
genus belonging to it has yet been found to be dicecious.

Like all the Anatinacea, Clavagella possesses two testes,
two ovaries, and four distinct genital apertures. The two
testes are symmetrically placed in the pedal projection (fig. 15,
vii), i.e. in the ventral part of the body; the two ovaries,
which are much more voluminous, compose almost the entire
posterior visceral mass as far as the nephridia (figs. 15 and 16).

The genital pores are situated behind; those of the testes
on the sides of the base of the foot, ventrally or internally to
the visceral commissure (fig. 14, vir); those of the ovaries a
little further back, dorsally or externally to the same commis-
sure (fig. 14, Iv).

II. Poytocenretic EvoLution oF THE HERMAPHRODITE
Gianp In MOLLUSCA.

If any other group of hermaphrodite animals is taken into
consideration, one is struck by the general uniformity in the
structure of the genital glands. Inthe Mollusca, on the other

HERMAPHRODITISM IN MOLLUSOA. 29

hand, as has been shown above, there is a very great diversity
in the structure of the hermaphrodite gland.

Indeed, it was Leuckart who was the first to recognise that
among the Euthyneurous (or hermaphrodite) Gastropods the
genital gland is not always formed in the same manner, and
that different types of structure are recognisable in it. The
facts known to us to-day show that such is the case not only
for the Gastropoda Euthyneura, but also for all the remaining
groups of hermaphrodite molluscs—the Amphineura, Gas-
tropoda Streptoneura, and Lamellibranchia.

Four different principal types can be distinguished in the
conformation of the hermaphrodite gland of molluscs, but
certain forms are transitional between one type and another.

1. The undifferentiated gland, i.e. with acini com-
pletely hermaphrodite.—As development shows (vide
infra), the most simple condition of an hermaphrodite genital
gland is that in which the gonadial wall is still undifferentiated,
and produces ova and spermatozoa side by side. This condi-
tion is exhibited in—

(1) Valvata.

(2) A great number of Tectibranchs (e.g. Bulla,
Aplysia, Umbrella).

(3) Almost the entire group of Pulmonata.

(4) Ostrea edulis and stentina.

A commencement of specialisation is observable in Neo-
meniidz, where each genital gland generally gives rise to male
products towards its lateral face, and to ova towards its axial
face.

Amphibola among Pulmonates (vide supra) also fur-
nishes a transition to the following condition.

2. Gland with separate male and female acini (but
as yet without separation into different male and female
regions).—After the undifferentiated condition comes that in
which the gonadial surface shows a clearly marked specialisa-
tion into distinct male and female acini,—the acini, however,
not forming regions of different sex.

30 PAUL PELSENEER.

This arrangement exists in—

(1) Various Tectibranchs (e. g. Lobiger, Pelta).

(2) Pleurobranchide, Tylodina, and Nudibranchia
(except Elysioidea).

(3) Siphonaria.

(4) Cardium oblongum.

It must be noticed here that the female acini open into the
male acini, and that the acini of the same sex tend to group
themselves together, as may already be observed in various
Nudibranchs (where the female acini are the most “‘ eccentric,”
i.e. superficial). This arrangement is clearly marked, and
furnishes in Onchidiopsis and Pneumonoderma (vide
supra) a transition to the next condition.

3. Gland with separate male and female regions
(with a common duct).—The type which now presents itself is
that in which the acini of the same sex are all united together
in such a way as to constitute in the hermaphrodite gland a
male and a female part distinct from one another, these two
parts having, nevertheless, a single genital aperture and a
common duct.

This conformation is characteristic of—

(1) The hermaphrodite species of the genus Pecten,
where the male and female regions are contiguous
and form an undivided hermaphrodite gland
(fig. 10).

(2) The Cycladidz, where it is already observable that
the two regions are fairly separated, and only connected by
their duct (fig. 12), which forms a transition to the next type.

4. Male and female glands in the same individual
entirely distinct from one another, and with special
ducts.—The acini of each sex can form a region absolutely
separate from that constituted by the acini of the other sex,
these two regions each having their special duct, and so form-
ing veritable testis and ovary.

At the same time the vas deferens and the oviduct may
open into a common orifice (Poromyidz), or there may be no
common orifice for the two glands. This last condition, which

HERMAPHRODITISM IN MOLLUSCA. 31

represents the highest specialisation of the hermaphrodite
genital gland, is only found in the Anatinacea among Lamelli-
branchs, and in Entoconcha among Gastropods.

III. PuoystotoaicaL EvoLuTion oF HERMAPHRODITISM IN THE
Mo.uuuscan INDIVIDUAL.

The hermaphroditism of molluscs is not, so to say, self-
sufficient ; in other words, the eggs of one individual have to
be fertilised by the spermatozoa of another.

Speaking generally, the two kinds of products are not ripe
at the same moment in the same individual: an interval of
greater or less duration separates the two periods of maturity
of the different sexual elements. :

Leuckart was the first to make known that in certain Opis-
thobranchiate Gastropods the period of male maturity precedes
that of female maturity (31), i.e. the hermaphroditism of these
forms is protandric.

This protandry ought to be regarded as a general phenome-
non in Euthyneurous Gastropods. This is notoriously the
case in the Pulmonates (32), and it has been recognised in the
various Opisthobranchs which have been studied from this
point of view, viz. Lobiger (33), the Thecosomatous ‘ Ptero-
pods,” e.g. Clio striata (84), Cymbulia (35), Desmopterus
(86) ; Nudibranchs, among which I have observed it in Holis
and Elysia; and lastly I may add Clione limacina (Gym-
nosomata), in which I have noticed that individuals of a length
of 15 millimetres (or less) do not as yet show any ova in their
genital gland, but stages in the development of spermatozoa
only.

Similarly in Entoconcha testicular capsules are only to
be observed in individuals in which the eggs are not yet well
developed (37).

For Neomenia, among Amphineura, protandry is equally
probable (38).

In Lamellibranchs no general statement can be made with
certainty. I have been unable to examine young individuals

32 PAUL PELSENEER.

of Anatinacea, or hatched specimens of Cyclascornea smaller
than 4 millimetres: at this size both spermatozoa and ova are
already to be found, and not infrequently even developing
eggs among the gills. But according to the old observations
of Davaine (89) and P. J. van Beneden (40), Ostrea edulis
is protandric.

The alternating activity of the two sexes in the same herma-
phrodite Molluscan individual is a perfectly certain fact, and
explains various mistakes like that of Saint-Loup, who believed
he had found a unisexual organisation in Aplysia associated
with external sexual characteristics (41).

It must, however, be remarked that this succession of the
two sexual conditions may be more or less rapid according to
the particular genera under consideration, and that, conse-
quently, the alternation is better marked in certain organisms
than in others. It seems to me to be more appreciable where
the hermaphroditism of the genital gland is most complete
(i.e. where the ova and spermatozoa are produced at the same
spot, as in Ostrea, Aplysia, &c.) than where there are acini
or regions of different sex (Nudibranchs, &c.).

As for the general fact that the male precedes the female
activity, even in the absence of physiological observations it
might have been deduced from the morphological constitution
of the genital glands in the adult of a great number of herma-
phrodite molluscs. In those forms, in fact, which possess acini
or regions of each sex, the male part is always nearest the
efferent ducts or the genital aperture: the male products will
accordingly be the first to reach them. This arrangement
exists in Pecten, Nudibranchs, Pleurobranchide, Onchidiop-
sis, Entoconcha, Pneumonoderma, Clionopsis, &c.
Cyclas alone constitutes an exception,—a fact for which I
can offer no explanation.

For the rest, our knowledge of the physiological evolution of
individual hermaphroditism in other groups of the animal
kingdoms shows that protandry is general in all forms which
have been examined from this point of view, e.g. Sponges,
Plathelminthes (Trematodes and Cestodes), hermaphrodite

HERMAPHRODITISM IN MOLLUSCA. 33

Nematodes (Pelodytes, Rhabdonema nigrovenosa),
Myzostomide, parasitic Isopods, Ascidians and Salps, Myxine,
Chrysophris, &c.

IV. Ontocenetic Evo.ution oF tHE HERMAPHRODITE GLAND
IN Mo.uuvusca.

According to the theory of the sexuality of the embryonic
layers, the female reproductive elements should be of endo-
dermic, and the male elements of ectodermic origin. As
regards the Mollusca this view has been maintained by Fol
(“ Pteropoda”’ Thecosomata, 42).

But this theory has not been confirmed for any other
mollusc, even for any of the nearest allies of the “ Pteropods,”
viz. the Tectibranchs (e. g. Aplysia).

I have not had an opportunity of studying the fresh larve
of Thecosomatous ‘‘ Pteropods ;” but in sections of various pre-
served larve I have been able to make out that the “corps
pyriforme” (which, according to Fol, is the testicular part
of the future hermaphrodite gland) neither has the structure
of a testicle nor contributes at all to the formation of the
genital gland. The latter is unique from its origin, as it is in
its final condition. The same is the case in the other molluscs
hitherto studied.

However, according to Trinchese (48), in Bosellia (a near
ally of Elysia) the ova are not produced in the same part of
the body as the spermatozoa, although this author is unable to
explain how the male and female parts of the hermaphrodite
acini eventually combine with one another. Now in Elysia,
specimens of which I have examined at every age from this
point of view, I find that the first ova arise in the same acini
in which spermatozoa alone were present in the stage imme-
diately preceding. Sothat it is impossible in this case to hold
that the hermaphrodite gland results from the fusion of two
male and female parts of independent origin. The same fact
may be observed equally well in the ‘“‘ Pteropod” Clione
(vide supra).

VOL. 37, PART 1.—NEW SER. c

34. PAUL PELSENEER.

On the other hand, in a form where the male and female
acini are sharply separated in the adult state, viz. Pelta
(= Runcina), I find that in the young individual ova and
spermatozoa arise side by side on the walls of a common
pouch.

I conclude from this that the undifferentiated condition of
the hermaphrodite gland is the most primitive, and the morpho-
logical observations recorded above (II) lead also to the same
conclusion. The Nudibranchs and Pleurobranchide (with
acini either male or female) are more specialised than
U mbrella (with hermaphrodite acini) ; Onchidiopsis (with
acini either male or female) is more specialised than Valvata
(with hermaphrodite acini); Pneumonoderma (with acini
either male or female) is more specialised than Aplysia (with
hermaphrodite acini), &c.; lastly, those Lamellibranchs with
male and female glands altogether separated (Anatinacea) are
the most specialised (cf. the closure of the mantle, the reduc-
tion of the foot, the complexity of the gill, involving the loss
of the external layer of the external branchial lamella).

In direct opposition to the theory of the sexuality of the
embryonic layers, we find that the hermaphrodite gland has a
single origin in the mesoderm (which is itself endodermic),
not only in the Opisthobranchs and Pulmonates hitherto
studied—where the gland is not yet divided into regions of
different sex,—but also in Cyclas (44), where two distinct
male and female regions exist (fig. 12, v and xiv). The same
mesodermic origin is, moreover, to be observed in the case of
both the male and female glands of those forms with separate
sexes, such as Chiton, Paludina, and the Cephalopods.
These results are also in complete accord with what one
finds in other groups of hermaphrodite Invertebrates, e. g.
Oligocheta (45), Sagitta, Turbellaria, Plathelminthes,
Hirudinea (46).

Lastly, the fact that the wall of the genital gland is in con-
tinuity with the coelomic epithelium (mesoderm) in Nuculide
(Lamellibranchs), Neomeniide (Amphineura), and Cephalo-
poda, shows that the genital gland, whether male, female, or

HERMAPHRODITISM IN MOLLUSCA. 35

hermaphrodite, is an organ possessing a single mesodermic mode
of origin, as in all the other Triploblastica.

V. Oricin oF HERMAPHRODITISM IN MoLLUSCA.

It has just been shown that, both from the phylogenetic and
ontogenetic points of view, the hermaphrodite condition with
separate male and female glands is the most specialised, and
the undifferentiated hermaphrodite condition (with gland
producing spermatozoa and ova at the same spot) the most
archaic. Now this last state is that which most nearly
approximates to the unisexual condition, since it only differs
in the supplementary production of the elements of the other
sex,—a phenomenon which is sometimes to be observed as an
abnormality in dicecious molluscs (e.g. Anodonta and
Ampullaria).

The question may, then, be asked in the case of the Mollusca,

1. Whether the hermaphrodite state is not derived from
the unisexual; and, in the event of an affirmative reply,

2. Upon which of the sexes the hermaphrodite condition
has become established.

In the following pages I shall try to show—

1. That hermaphroditism is not a primitive arrangement
in the Molluscan phylum (47), and that it has been derived
from the unisexual state.

2. That it has become superimposed upon the female
condition. :

1. Hermaphroditism has been derived from the
unisexual condition.—Let us consider separately the classes
in which hermaphroditism is found and those in which the
dicecious state alone exists.

i. In the classes in which hermaphroditism exists it is
abundantly clear that the forms with separate sexes are the
most archaic, especially in the conformation of their repro-
ductive apparatus (absence of special duct, accessory gland,
and penis) :

a. Gastropoda Docoglossa and various Rhipidoglossa: the
genital gland opens into the kidney.

36 PAUL PELSENEER,

B. Lamellibranchia Protobranchia (Nucula, &c.): the
genital gland opens at the junction of the ccelom (pericardium)
and the kidney.

ii. If we consider now the two classes where only the
dicecious condition exists, we find that they have preserved
many primitive traits, especially in the reproductive apparatus :
the genital gland opens into the celom (Cephalopoda) or into
the kidney (Scaphopoda).

These facts show that in the Mollusca the dicecious con-
dition has preceded the moneccious, and by no means (as
Rouzand has tried to make out, 48) that “les Gastropodes
unisexués se présentent réellement comme les descendants des
Gastropodes hermaphrodites, plus ou moins analogues 4 ceux
qui vivent encore de nos jours.”

The molluscs which are certainly the most archaic are
dicecious. Their genital organs present a conformation analo-
gous to the primitive disposition observed in the ontogeny
of other forms, which can on no account be regarded as the
last term of a retrogressive evolution—evolution not being
reversible. .

But hermaphroditism has been able to establish itself, re-
placing the dicecious state in Mollusca of different grades of
specialisation—

a. In forms where the genital glands open into the ccelom
(Neomeniidz) ;

B. In forms where these glands open into the kidneys
(Pecten) ;

c. In forms where these glands have special ducts, ac-
cessory glands, copulatory organs, &c. (Onchidiopsis,
Euthyneura, &c.).

2. Hermaphroditism has been established upon the
female organism.—The hermaphrodite condition, then, is
secondary in Mollusca. As for the organisation which has
given rise to it, I believe it to be that of the female.

Comparative study of the genital ducts and of their develop-
ment lends support to this opinion, as we shall at once see.

i. In Lamellibranchs, when there is a separate male and

HERMAPHRODITISM IN MOLLUSCA. 37

female orifice (Anatinacea), and when the visceral commissure
(“connectife cérébro-viscéral”) is sufficiently superficial, the
female orifice is found lying outside this visceral commissure
like the genital orifice—be it male or female—of all
the Lamellibranchs (fig. 13). This orifice is accordingly
not a new formation.

But the male orifice is within this commissure (fig. 14),
and so presents relations to which no counterpart can be
found elsewhere ; it is consequently a new formation, and if
it was necessary for a male orifice to be produced as a new
formation it is evident that the starting-point of the herma-
phrodite condition must have been supplied by the female
organism,

ii. Similarly in hermaphrodite Gastropods the male products
are always expelled by a cenogenetic penial orifice, which
is wanting in the archaic diccious forms (Rhipidoglossa,
—as in the Amphineura and Scaphopods), while the female
orifice in the same forms lies in the place occupied by the
genital orifice (whether male or female) of archaic Gastropods
which possess no penis.

And when an approximation of the two apertures, male and
female, takes place—the former approximates to the latter in
Arion; vice versa in other cases—the disposition is secon-
dary, and is accompanied by an increased complexity of the
ducts and their appendages. In the most primitive cases the
two orifices are always widely separated (Bullide among
Opisthobranchs; Auriculide among Pulmonates).

In the development of the Pulmonata the penis and vas
deferens are new formations (48a) which arise secondarily, the
vas deferens only communicating at a late stage with the
genital duct properly so called (physiologically hermaphrodite),
which leads to the female orifice; that is to say, the genital
organs of Pulmonata develop as female organs, which are even-
tually modified as as to become hermaphrodite (48b).

The opinion formulated above (that in Mollusca herma-
phroditism is grafted upon the female organism) is confirmed
by the following fact. In hermaphrodite molluscs, whenever

388 PAUL PELSENEER.

individuals accidentally present a unisexual genital apparatus,
they are always female; it is always the female sex which
reappears.

a. Clio (Hyalocylix) striata, without penis, cited above
(III, note, according to Schiemenz).

6. Cymbuliopsis calceola, without penis (48c).

c. Agriolimax levis, without penis, having never been
male (adult female; 49).

d. Helix aspersa, without vas deferens, penis, and flagel-
lum (49a).

e. Arion intermedius, without male reproductive organs
(49b).

VI. ConcuusIons.

Granted, then, that in Mollusca hermaphroditism has suc-
ceeded a unisexual state, and that it has been superimposed
upon the female condition, is this process paralleled in other
groups ?

1, The origin of hermaphroditism in a unisexual
condition.—I have not been able to make any special investi-
gations upon hermaphroditism in the various groups of the
animal kingdom. It seems to me, however, that if we examine,
by the comparative method, the data which we already possess
upon the subject, we must come to the conclusion—opposed to
the ordinarily received opinion, which possesses the authority
of the names of Huxley, Gegenbaur, Haeckel, Giard, Claus,
&c.—that hermaphroditism is secondary, and succeeds a pri-
mitively dicecious state.

In fact, like myself, who have reached this result for the
Mollusca, so Beard and Delage have already formulated this
Opinion as concerning two other groups, the Myzostomide
(49e) and Cirripedes (50). And Fritz Miller, after a more
general consideration of the subject, also argues against the
primitive nature of the hermaphrodite state, and shows that
in the majority of groups the most archaic forms are unisexual
(51).

A survey of the various subdivisions of hermaphrodite

HERMAPHRODITISM IN MOLLUSCA. 39

animals shows that hermaphroditism is characteristic almost
always of specialised forms (52). I will more particularly
mention fixation, parasitism, fluviatile or terrestrial life, as
specialisations accompanied by hermaphroditism, as the fol-
lowing few examples show :

Pecten, Ostrea, Aspergillum,

mensals

Fixed (58) or very Clavagella; various Serpulide
sedentary (54), Cirripedes, Myzostomide,
Ascidie.
Hermaphrodite Beene Entoconcha, Entovalva, Ces-
animals. EME 0 oa todes, Trematodes, Hirudinea, cer-

tain Isopods, Myxine,
Fluviatile or ter-

: Valvata, Pulmonata, Oligocheta.
restrial

In support of the opinion that hermaphroditism is a spe-
cialisation of the separation of the sexes, it may be remembered
that in a great number of unisexual glands individual variations
occur in which the elements of both sexes are produced, one of
the two kinds of element being an abnormal product. I will
cite merely a few instances:

i, Batrachiaus: female frog (55).

il, Fishes: female herring (56).

iii. Molluses: Ampullaria, Anodonta.

iv. Cheetopods: ova in the testicle of Lumbricus (57), -

v. Crustaceans: female Apus (58), &c.
_ This phenomenon makes it possible to understand that the
hermaphrodite state can easily establish itself in certain
circumstances where it is useful for the same individual to
give rise to the products of both sexes (59).

2. The establishment of hermaphroditism on the
female condition.—I cannot at present affirm that her-
maphroditisim has everywhere grafted itself upon the female
organisation; nevertheless the phenomenon ought to be
exhibited elsewhere than among the molluscs. .

As a matter of fact, there are several other groups in which
the male sex is preserved in the form of individuals, either
degraded or not, when there are no longer any females,

40 PAUL PELSENEER.

but only normal hermaphrodite individuals. This
is the case with—

i, Various Myzostomide.

ii. Certain parasitic Isopods, e. g. the Cryptoniscide.

ili, Various Cirripedes, e.g. Scalpellum, &c.

And the study of one of these Cirripedes has led Delage to
the same conclusion as that which I have deduced from the
study of hermaphrodite molluscs: “Au début, les Sacculines
étaient des animaux a sexes distincts, dont les femelles sont
devenues hermaphrodites beaucoup plus tard” (60).

Lastly, the same seems to be true among osseous fishes in
Serranus, according to Ginther (61), viz. that the herma-
phrodite individuals are transformed females. This has been
confirmed by Brock, who has specially studied the genital
organs of fishes. He concludes that “die hermaphrodi-
tischen Knochenfische sind weibliche Individuen,
in deren Ovarium sich an Stelle einiger Ovarial-
lamelle ein Hoden sich entwickelt hat (62). And,
on the other hand, the vas deferens of Serranus is not homo-
logous with that of other osseous fishes (68) ; it is, accordingly,
like that of the Mollusca Anatinacea and Pulmonata (vide
supra, v, 2, ii), a new formation grafted on the female
condition.

In these different cases the establishment of the hermaphro-
dite condition has been probably characterised by the follow-
ing successive stages :—Production of spermatozoa in a part of
the ovaries; reduction of the size of the males and of their
number (hyperpolygyny) ; then complete replacement of the
female by the hermaphrodite form; and, lastly, the total dis-
appearance of the degraded males.

Summing up.

a. The study of Mollusca, Myzostomide, Crustacea, and
Pisces shows that in these groups the separation of the sexes
has preceded hermaphroditism ; various cases in other groups
tend to show that this is true universally; and the same con-
clusion applies to plants.

HERMAPHRODITISM IN MOLLUSCA. 41

B. In Mollusca, Crustacea, and Pisces, at least, hermaphro-

ditism is grafted upon the female sex.

18.
19.

BIBLIOGRAPHY.

. Husrecut.— Dondersia festiva, gen. et sp. nov.” (‘ Donders-feest-

bundel,’ 1888), pl. ix, fig. 6. Pruvor.— Sur Vorganisation de
quelques Néoméniens des cdtes de France,” ‘ Arch. d. Zool. Expér.,’
sér. 2, t. ix, pl. xxx, fig. 52.

. Mippenporr.— Beitrage zu einer Malacozoologia Rossica,’’ ‘Mém.

Acad. d. Sci. St. Pétersbourg,’ sér. 6, Sci. natur., t. vi.

. Bernarp.— Recherches sur Valvata piscinalis,” ‘Bull. Scientif.

France et Belgique,’ t. xxii, pl. xx, fig. 2.

. Bercu.—‘ Malacologische Untersuchungen,’ Suppl. Heft iv, pp. 261,

273, pl. x, fig. 8; y, fig. 8.

. Mecxeit.—‘“ Ueber den Geschlechtsapparat einiger hermaphroditischen

Thiere,” ‘ Archiv f. Anat. u. Phys.,’ 1844.

. NorpmMann.—“ Monographie des Tergipes Edwardsi,” ‘ Mém. Acad.

St. Pétersbourg,’ 1845.

. Lrvcxart.—‘ Zoologische Untersuchungen,’ drittes Heft, p. 78, &e.
. TrincnEse.— Alolidide e famiglie affini,” ‘ Atti Accad. Lincei,’ ser. 3,

vol. xi, pl. Ixx.

. Hatter.—‘ Die Anatomie von Siphonaria gigas,” ‘ Arb, Zool. Inst.

Wien,’ vol. x, p. 83, pl. ii, fig. 22.

. Davatne.— Recherches sur la génération des Huitres,” ‘Mém. Soc.

Biol. Paris,’ t. iv, 1853.

. Lacaze-Dutuiers.—“ Recherches sur les organes genitaux des Acéphales

Lamellibranches,” ‘Ann, d. Sci. nat.,’ sér. 4, t. ii, p. 217.

. Ann. Mag. Nat. Hist.,’ ser. 6, vol. xi, p. 262.
. DavatnE.—Loc. cit., p. 316,
. vAN BeneDEN.—‘ Sur les organes sexuels des Huitres (O. edulis),”

‘Comptes rendus Acad. Sci. Paris,’ t. xl.

. Brooxs.—* Development of the American Oyster,” ‘Stud. Biol. Lab.

Johns Hopkins Univ.,’ vol. i, part iv, pls. viii, ix.

. Boucnon BranpELy.—‘ Comptes rendus Acad. Sci. Paris,’ t. xev.
. Horx.— Les organes de la génération de l’huitre,” ‘Tydschr. Ned.

Dierk. Ver.,’ Suppl. Deel i, pl. ia.
Lacazz-Duruiers.—Loc. cit., p. 222.

Vorttzkow.—“ Entovalva mirabilis,” ‘ Zool. Jahrb.’ (‘ Abth, f. Syst.
Geogr. u. Biol.’), Bd. v, p. 622.

42 PAUL PELSENEER.

20. Mitne-Epwarps.—“ Observations sur divers Mollusques,” ‘ Ann. d. Sci.
nat.,’ ser. 2, t. xviii, pl. x, fig. 1.

20a. Humpert.— Note sur la structure des organes génitaux de quelques
espéces du genre Pecten,” ‘ Aun. d. Sci. nat.,’ sér. 3, t. xx, p. 338.

21. Lacaze-Dutst1ers.—Loc. cit., p. 208.

22. Futtarton.—“ On the Development of the Common Scallop,” ‘ Highth
Ann. Rep. Fish. Board for Scotland,’ p. 290, pl. viii.

23. Jacxson.— Phylogeny of the Pelecypoda,’’ ‘Mem. Boston Soc. Nat.
Hist.,’ vol. iv, pp. 341, 348.

23a. Humpbert.—Loc. cit., p. 338.

24. von Sresotp,—‘ Arch. f. Anat. u. Phys.,’ 1837, p. 383.

25. Lreypic.—“ Ueber Cyclas cornea, Lam.,” ‘ Arch. f. Anat. u. Phys.,’
1855, p. 59. Sreranorr.— Ueber die Geschlechtsorgane und die

Entwickelung von Cyclas,” ‘Arch. f, naturg.,’ 1865, p. 3.

26, Steenstrup.—‘ Underségelser over hermaphroditismens tilvaerelse i
naturen,’ 1845, p. 69.

27. Ray Lanxester.— Contributions to the Developmental History of the
“Mollusca,” £ Phil. Trans.,’ 1875, p. 1.

28, PrtsenreR.—‘‘ Deux nouveaux Pélécypodes hermaphrodites,” ‘ Comptes
rendus Acad. Sci. Paris,’ t. cx. PELSENEER.—“ Sur existence d’un
groupe entier de Lamellibranches hermaphrodites,” ‘ Zool. Anzeiger,’
1891.

29. F. J. H. Lacaze-Dutuiers,—Loc. cit., p. 211.

80. H. pr Lacaze-Duruiers.— Morphologie des Acéphales,” ‘Arch. d.

Zool. Expér.,’ sér. 2, t. i, p. 216.

81, Levckart.—‘ Zoologische Untersuchungen,’ Heft iii, p. 76.

82. Since Eisie (“ Beitrage zur Anatomie und Entwickelungsgeschichte der
Geschlechtsorgane von Lymneus,” ‘ Zeitschr. f. wiss. Zool.,’ Bd. xix,
p. 310), &c. According to Bazsor (‘Ueber den Cyclus der Ge-
schlechtsentwickelung der Stylommatophoren,” ‘ Verbandl. Deutsch.
Zool. Gesellsch.,’ 1894, p. 57), Agriolimax levis and Limax
maximus furnish exceptions and are protogynous,

88. PELSENEER.— Recherches sur divers Opisthobranches,” ‘ Mém. Cour.
Acad. Belg.,’ t. lili, p. 21.

84, PeLsenEER.—Ibid., p, 24.

35. Lruckart.—Loc. cit.

86. Cuun.—“ Bericht iiber eine nach den Canarischen Inseln im Winter,
1887-8, ausgefiihrte Reise,” ‘Sitzungsber, k, Akad, d. Wiss. Berlin,’
1889, p, 543,

HERMAPHRODITISM IN MOLLUSCA. A3

37. J. Mitter.—‘ Ueber Synapta digitata und die Erzengung von
Schnecken in Holothurien,’ p. 13.

38. Wiren.—‘‘ Studien iiber die Solenogastren,” II, ‘ K. Svensk. Vetensk.
Akad. Handl.,’ Bd. xxv, No. 6, p. 47.

89. Davaine.—Loc. cit., p. 316.

40. van BENEDEN.—‘ Comptes Rendus Acad. Sci. Paris,’ t. x].

41. Sarnt-Loup.—“ Observations anatomiques sur les Aplysiens,” ‘ Comptes
rendus Acad. Sci. Paris,’ t. evii.

42. Fou.—‘Sur le Développement de Ptéropodes,” ‘Arch. Zool. expér.,’
sér. 1, t. iv, p. 205.

48. TrincuEsE.—‘ Descrizione del nuovo genere Bosellia,” ‘Mem. Accad. d.
Sci. Bologna,’ ser. 5, t. i, p. 776.

44, ZircLer.—“ Die Entwickelung von Cyclas cornea, Lam.,” ‘ Zeitschr.
f. wiss. Zool.,’ Bd. xli, p. 562, pl. xxviii, figs. 30, 33.

45. Woopwarp.— Proc. Zool. Soc. London,’ 1893, p. 321.

46. Korscuet and He1pER.—‘ Lehrbuch der vergleichenden Entwickelungs-
geschichte der wirbellosen Thiere,’ pp. 118, 119, 129, 223.

47. This opinion, which I expressed several years ago for the Mollusca in
general (‘Comptes rendus,’ t. cx, p. 1083), is also shared by Brock
for the Pulmonata (‘ Zeitschr. f. wiss. Zool.,’ Bd. xliv, p. 374), and
has been since adopted by Lane (‘Lehrbuch der vergleichenden
Anatomie,’ p. 816).

48. Rovuzaup.—‘ Recherches sur le développement des organes génitaux
de quelques Gastéropodes hermaphrodites,’ Montpellier, 1885, p. 79.

48a. Brocx.— Die Entwickelung des Geschlechtsapparat der Stylommato-
phoren Pulmonaten,” ‘ Zeitschr. f. wiss. Zool.,’ Bd. xliv, p. 368.

48b. Brocx.—Loc. cit., pp. 372, 374.

48c. Pecx.—On the Anatomy and Histology of Cymbuliopsis ealeeelg®
‘Stud. Biol. Lab. Johns Hopkins University,’ vol. iv, No. 6, p. 15.

49, von JuERinc.—‘ Jahrb. d. Malakozool. Gesellsch.,’ Bd. xii, p. 207 (1885).
Srmrotau.— Zeitschr, f. wiss. Zool.,’ Bd. xlv, pp. 655, 656, 661 (1887).

49a, CoLtince.— Absence of Male Reproductive Organs in Two Herma-
phrodite Mollusca,” ‘Journ. of Anat, and Phys.,’ vol. xxvii, p. 237.

49b. Cottince.—Loce. cit., p. 238.

49c, Brarp.—“ On the Life History and Development of the Genus Myzo-
stoma,” ‘Mittheil. Zool. Stat. Neapel,’ Bd. v, p. 578 :—< Herma-
phroditism, probably all hermaphroditism, had its origin in a unisexual
condition.”

50. Detace.—* Evolution de la Sacculine,” ‘ Arch. de Zool. expér.,’ sér. 2,
t. ii, p. 704:—* La séparation des sexes chez Sacculina, et probable-

44.

51.

52.

53.

54.

55.

56.

57.
58.
59.

60.

61.

62.

63.

PAUL PELSENEER.

ment chez tous les Crustacés hermaphrodites, est l’état primitif dans le
développement ontogénétique et phylogénétique.”

F. Miter. ‘ Die Zwitterbildung im Tierreich,” ‘Kosmos,’ Bd. xvii,
pp. 821—334.

In addition to the various examples given further on, it may be remarked
that the only hermaphrodite forms among Echinoderms are the
Synaptids, which are the most highly specialised by the reduction of
the calcareous plates, and the disappearance of the madreporite and
the ambulacral tubes. Similarly in Pisces the hermaphrodite forms
(Serranide) belong to the most specialised types of Teleosteans; see
Cope, ‘ The Origin of the Fittest,’ 1887, p. 331.

Though descended from free ancestors, as their larvee show.

Some Spirorbis, Protula, Pileolaria, Laonome.

Bourne.—“‘ On Certain Abnormalities in the Common Frog (Rana
temporaria),” ‘Quart. Journ. Micr. Sci.,’ vol. xxiv, p. 83, pl. iv.
Voer.— Notice sur un Hareng hermaphrodite,” ‘ Arch. de Biol.,’ t. iii,

p, 255, pl. x

Woonwarp.—‘ Proc. Zool. Soc. London,’ 1893, p. 322.

BeErnarp.— Nature,’ vol. xlili, p. 343.

Although it is yet a very rare practice to compare the phenomena pre-
sented by the two organic “kingdoms,” it may be added here that
hermaphroditism, properly so called, is also a characteristic of specialised
plants, and that it is wanting, speaking generally, in aquatic plants
and the most ancient of terrestrial plants; see especially Errera et
Gevaert, “Sur la structure et les modes de fécondation des fleurs,”
‘Bull. Soc. Botan. Belg.,’ vol. xvii, p. 164.

Dine, —Loc, cit., p. 704.

Ginruer.— An Introduction to the Study of Fishes,” 1880, p. 157 :—
“In the European species of Serranus a testicle-like body is attached
to the lower part of the ovary, but many specimens are undoubtedly
males having normally developed testicles only.”

Brocx.—‘ Zeitschr. f. wiss. Zool.,’ Bd. xliv, p. 374.

Brocx.— Untersuchungen iiber die Geschlechtsorgane einiger Mure-
noiden,” ‘ Mitth. Zool. Stat. Neapel,’ Bd. ii, p. 488.

HERMAPHRODITISM IN MOLLUSOA. 45

EXPLANATION OF PLATES 4, 5, & 6,

Illustrating Dr. Pelseneer’s paper on ‘‘ Hermaphroditism in
Mollusca.”

Fic. 1—Hermaphrodite reproductive apparatus of Onchidiopsis green-
landica, dorsal view. 1, ovotestis; 11, oviduct; III, mucous gland; Iv,
receptaculum seminis ; V, penis; VI, spermiduct ; vil, female aperture ; VIII,
prostate; Ix, sperm-oviduct; x, vesicula seminalis.

Fic. 2.—Transverse section of the ovotestis of Onchidiopsis.

Fic. 3.—Part of the ovotestis of Onchidiopsis, slightly magnified, show-
ing male (white) and female (grey) portions.

Fie. 4.—Ostrea cochlear, male, left view. 1, rectum; 0, testis; 111,
internal plate of left gill; 1v, internal plate of right gill; v, pallial suture ;
vi, adductor muscle; vit, auricle ; vit1, ventricle.

Fic. 5.—Ostrea cochlear, female, left view. 1, liver, seen by trans-
parency through the ovary; 11, auricle; 11, pallial suture; iv, abductor
muscle ; v, intestine; vi, ventricle.

Fic. 6.—Ostrea cochlear, ventral aspect, mantle and gills removed.
I, right genital aperture; 1, right renal aperture; 11, visceral ganglia;
IV, outline of the mantle; v, adductor muscle ; v1, visceral commissure ; VII,
branchial nerve ; vuiI, left renal aperture; 1x, left genital aperture ; x, palps
(the lines of adhesion of the gills to the visceral mass are indicated by dotted
lines).

Fic. 7.—Transverse section of a female Ostrea cochlear. 1, intestine;
1, right lobe of the mantle; 111, nephridium; 1v, branchial axis; v, genital
aperture ; VI, visceral commissure ; VII, external plate of the right gill; v1,
internal (adaxial) plate of right gill; rx, visceral commissure (left half);
x, renal aperture ; XI, reno-pericardial duct; x11, ovary ; XIII, pericardium;
xIVv, ventricle ; xv, left lobe of the mantle ; xvi, rectum; XVII, auricles joined
together.

Fie. 8.—Transverse section of a male Ostrea cochlear. 1, right lobe of
the mantle; 11, branchial axis; 111, testis; Iv, external plate of right gill;
v, visceral commissure; vI, internal plate of right gill; v1, visceral com-
missure ; VIII, genital aperture ; 1x, nephridium; x, testis ; XI, pericardium ;
xu, ventricle (on the pericardial wall of which are modified epithelial cells
= pericardial glands); x11, left lobe of the mantle; xiv, rectum; xv,
auricle ; XVI, intestine.

Fic. 9.—Transverse section of the ovotestis of Cardium oblongum,

46 PAUL PELSENEER.

Fie. 10.—Transverse section of the abdomen of Pecten flexuosus,
showing male and female halves of the ovotestis. 1, intestine.

Fic. 11.—Longitudinal section of the right genital gland of Cyclas
cornea (the anterior part is below). 1, epithelium of the visceral mass; 11,
testis; 111, part of the genital duct between testis and ovary; IV, ovary.

Fic. 12.—Cyclas cornea, left view, without left gill and pallial lobe.
I, anterior retractor of the foot; 11, anterior adductor; 111, palp; Iv, foot;
Vv, ovary; VI, genital aperture; vil, visceral ganglion; viiI, posterior ad-
ductor ; 1x, posterior retractor of the foot; x, kidney; x1, rectum; xu,
auriculo-ventricular slit; x1II, ventricle; xIv, testis ; xv, liver mass.

Fre. 18.—Plan of nervous system, and genital and renal apertures in an
ordinary Lamellibranch 1, cerebral ganglion; 11, visceral commissure; II,
pedal ganglia; Iv, genital aperture; v, renal aperture; vi, visceral ganglia.

Fie. 14.—Plan of nervous” system, and genital and renal apertures in
Anatinacea. I—III, and v, VI, as in Fig. 13; Iv, female aperture; vm, male.
aperture.

Fie. 15.—Transverse section of Clavagella passing through the male
apertures. I, ovary; II, aorta; 111, mantle; Iv, branchial axis; v, visceral
commissure ; VI, spermiduct; vi1, foot; vit, testis; Ix, pad of adhesion of
the reflected internal lamella of the gill; x, right male aperture ; x1, branchial
axis ; XII, Ovary; XIII, auricle; xIv, intestine; xv, pericardium.

Fic. 16.—Transverse section of Clavagella (posterior to Fig. 15), passing
through the female apertures. 1, ovary; 1, aorta; 111, intestine; Iv, ovary;
v, visceral commissure ; VI, foot ; vi1, pad of ciliated adhesion of the internal
reflected lamella of the gill; vii1, female aperture; 1x, branchial axis; x,
nephridium ; x1, auricle; x11, mantle; x11, pericardium.

CEREBRAL CONVOLUTIONS—‘* SALLY.”’ A”

A Description of the Cerebral Convolutions of
the Chimpanzee known as “Sally;” with
Notes on the Convolutions of other Chim-
panzees and of Two Orangs.

By

W. Blaxland Benham, D.Sc.Lond., Hon. M.A.Oxon.,
Aldrichian Demonstrator of Comparative Anatomy in the University
of Oxford ; Lecturer on Biology at Bedford College, London.

With Plates 7—11.

ConTENTS.
PAGE PAGE
1. Historical : : . 48 9. The Frontal Lobe . ai)
2. Introduction . : Pe!) 10. Summary of Results of
3. Description of ‘“ Sally’s” the Comparison of a series
Brain. . : yO of Chimpanzee Brains . 74
4. The Parieto-occipital Fis- 11. List of Chief Papers re-
SICH ae ei. as Do lating to the Chimpanzee
5. The Parietal Convolutions. 57 Brain. : ‘ a6
6. The‘‘Affenspalte” in Chim- 12. The Brain of the Orang-
panzee and Man . =, bg outang . 3 ° 16
7. The Occipital Lobe . 167 Explanation of Plates . 81

8. The Sylvian Fissure, &c. . 68

PREFACE.

Tue brain of the interesting and well-known chimpanzee
“ Sally,” which lived for eight years in captivity at the Zoolo-
gical Gardens in London, has been recently figured and
described by Mr. Beddard (11) in his important memoir on
the anatomy of this ape, which he considers to be T. calvus;
there are, however, several features about this brain which
deserve further notice,

48 W. BLAXLAND BENHAM.

After the brain had been described it was purchased by
Professor Lankester for the Oxford University Museum; and
my attention was particularly directed to it when comparing
the convolutions of the anthropoid apes with those of man, for
the purpose of exhibiting in the museum a series of anatomical
preparations illustrating the relations of man to the apes.

One of the most striking features about “ Sally’s ” brain is
the absence of a well-defined “ occipital operculum,” the
presence of which was supposed by Gratiolet to mark off the
chimpanzee brain very strongly from that of the orang, in
which it is generally absent. With this disappearance of the
operculum is connected certain other modifications in this
region of the brain, the most evident of which is the diminu-
tion in the extent of the so-called “ Affenspalte” or Simian
fissure, and the consequent resemblance of this region of the
brain to that of the orang and of man. Gratiolet placed the
chimpanzee further from man than the orang, but “ Sally’s ”
brain is considerably more human than that of any orang
hitherto figured, and it becomes a matter of great interest to
compare the gyri and sulci in the brain of this specimen with
those of the more ordinary chimpanzee on the one hand and
with those of man on the other. In making these comparisons
I have the advantage over Beddard in that I am able to make
use of the valuable and extensive researches of Dr. D. J.
Cunningham, published in 1892, just after Beddard’s com-
munication to the Zoological Society ; and in the present paper
I shall adopt Dr. Cunningham’s nomenclature, which is in
agreement generally with that employed on the Continent.

My great indebtedness to Dr. Cunningham’s work is
sufficiently evidenced by my constant reference to it.

I. Hisroricat.

Descriptions, more or less detailed, of the brain of the
chimpanzee are fairly numerous. I append a bibliography of
the more important papers ; but amongst those described I
have only been able to light upon two brains bearing any close
resemblance to that of “Sally.” One of these is described and

CEREBRAL CONVOLUTIONS—** SALLY.” 49

figured by P. Broca (6).1 In the figure here reproduced
(fig. 19) it will be seen that the “operculum” is absent on
both sides. The “ Affenspalte,” or Simian fissure (0. 5.), or
“external perpendicular fissure ”’ as he called it, is not inter-
rupted on the left side, as the second “ pli de passage ” does not
rise to the surface. But on the right side the annectant gyrus
(2) becomes superficial, and separates the Simian fissure (a. a.)
into two parts. The first “ pli de passage” (J.) is present on
both sides. It will be seen below that another explanation of
these arrangements is possible. It may be well to note that
the brain belonged to a young male.

The second brain to which reference may be made was
described in considerable detail by Dr. Joh. Miiller (9) in
1888, one of whose figures is here reproduced (fig. 25). In it
the “ operculum ” is absent on both sides, the occipital lobe
and the region of the Simian fissure being more complex than
in the ordinary chimpanzee, but not so folded as in Broca’s
chimpanzee, or as in “ Sally,” for the second annectant gyrus
does not come to the surface on either side. This brain also
was taken from a young male, which still retained its milk
teeth.

Neither of these authors gives any information as to other
anatomical characters of the animal to which the brain belonged
to enable us to state whether it belonged to the so-called
species T. calvus, or T. niger.

As long ago as 1866 Sir Wm. Turner (4) described a
couple of chimpanzee brains presenting characters which at
the time were of peculiar interest, on account of certain views
held by Gratiolet above referred to. One of the brains, figured
on p. 680 of the paper, had but a feebly developed operculum on
the right side, and therewith presented certain other pecu-
liarities, but there is no special resemblance to that of “ Sally.”
The object of the paper was to show that Gratiolet’s opinions
as to the annectant gyri in the chimpanzee were founded on
insufficient material.

1 T have to thank Sir Wm. Turner for most kindly calling my attention to
this paper.

VOL. 37, PART 1.—NEW SER. D

50 W. BLAXLAND BENHAM.

2. INTRODUCTION.

In looking over the stock of chimpanzees’ brains in the
museum of the Royal College of Surgeons,—for which privi-
lege I wish to express my thanks to Prof. Stewart, who most
kindly gave me every assistance in the matter,—I came across
a specimen in the museum stores which bears considerable
resemblance to that of “ Sally,” as well as to that figured by
Turner, in that it has lost its occipital operculum, but only on
the right side, as it is well developed on the left. This brain
belonged to the. late Prof. John Marshall, but has not been
described by him so far as I have been able to ascertain. It
came into the possession of the Royal College of Surgeons
in 1891, but I have been unable to trace its origin. I have
thought it worth while to give a figure of the right hemisphere
here (fig. 20).

On the supposition that the peculiarities of “ Sally’s ” brain
are specific, and characteristic of T. calvus, I wrote to Prof.
Herdman, of Liverpool, for I heard that he had purchased one
of Garner’s chimpanzees which died on its arrival in England ;
it was possible that this was T. calvus, and Prof. Herdman
most generously acceded to my request to be allowed to
examine the brain, and forwarded it to me. I was, however,
disappointed, for it does not present characters marking it off
for that of ordinary chimpanzees (see fig. 36), and it is as yet
uncertain whether the specimen from which it was taken is
T. niger, T. calvus, or a third species. In his letter to me
Prof. Herdman writes: ‘“‘ Garner declared that this animal was
different from calvus, and also from niger. . . . I saw the
dead head (after being skinned), and thought it might be
calvus.”

Thus we have no positive evidence to show whether the
peculiarities of ‘ Sally’s”’ brain are characteristic of the
species T. calvus; that they are not due to the age or sex of
the animal seems certain from the specimens of Broca and
Miiller, and it can scarcely be maintained that they are con-
nected with ‘ Sally’s” greater intelligence, for both Broca’s

CEREBRAL CONVOLUTIONS—‘* SALLY.” 51

and Miiller’s specimens were quite young, and no particular
intelligence has been attributed to them. The only sugges-
tion remaining is that in the species T. calvus the brain does
present peculiarities, and that Broca’s and Miiller’s specimens
belonged also to this species.

On the other hand, we are still uncertain that there is more
than one “species” of chimpanzee. Mr. Beddard, from a
careful study of the muscular anatomy of “ Sally’s ”’ limbs, from
a comparison of the skull, brain, and other organs, considers
that the animal belonged to the species T. calvus, and shows
reasons for distinguishing it from the animal described, ana-
tomically, by Gratiolet and Alix, and named T. aubryi.
Nevertheless, the differences in muscular anatomy, &c., re-
corded in the ordinary chimpanzee and in these two other
forms, can scarcely be held to establish with absolute certainty
the specific distinctness of these three forms: the variations in
muscular anatomy in man appear to be every bit as great as in
these, and till sufficient specimens of the black-faced bald
chimpanzee, known as T. calvus, on the one hand, and of the
flesh-coloured, black-red haired form (T. niger), have been as
fully and as carefully described as Beddard has done for
“Sally,” we shall not be entitled—at least we ought to
hesitate—to be positive about the “ specific ”’ differences of
these three forms.

As we shall see, the convolutions of the chimpanzee’s brain
are subject to variations as wide as occur in man, and some of
them are of importance. Such features, too, as the shape of
the ear are extremely variable, every stage in the folding of
the margin being exhibited, from the unfolded condition to a
state closely resembling man’s ear.

For the present, at least in this paper, the chimpanzee
“Sally ” is regarded as being merely a “ variety ” of the single
species I’. niger.

3. Description oF “Sauty’s” Brain. Ya

The general features, dimensions, and weight of “ Sally’s ”
brain have been described by Beddard; 1 have but little to

52 W. BLAXLAND BENHAM.

add. He found that the brain “ weighed, after removal of the
pia mater and after an immersion of four months in spirit,
83 oz.; it had been allowed to dry for an hour and a half
before weighing, it was then damp, but not wet.”

He gives the weights of brain of two other chimpanzees
(T. niger) as 63 oz. and 63 oz.

The brain described by Miiller weighed 64 oz. (213 grammes),
whilst that described by Symington (10), though not bearing
any particular resemblance to ‘‘ Sally’s” brain, weighed as
much as 84 oz. after preservation in spirit ; both these authors
tabulate the weights of the brains previously recorded. Mar-
shall’s specimen (1861) (2) weighed under similar circumstances
as much as9 oz. But the weight of the brain fresh is con-
siderably more; thus Symington’s was 13 oz., Marshall’s 15
oz., and Chapman’s (7) 10 oz. 10 grains. In all these cases
the animal was young, and Symington suggests from analogy
with man, and from comparison of cranial cavities of young
and adult chimpanzees, that the weight of the brain in the
latter is 15 oz.

I have not deemed it necessary to weigh the various brains
to which I shall have occasion to refer.

Beddard gives the length of ‘ Sally’s” brain as 103 mm.,
and that of the hemisphere alone as 100 mm. Miller’s
measured 92 mm. and Symington’s 95 mm., so that “ Sally’s ”
appears to be the largest hitherto carefully measured, as one
would expect, since the animal is the oldest chimpanzee whose
brain has been described.

The general appearance of the brain is more like that of man
on asmall scale than is that of an ordinary chimpanzee (see
figs. 1, 2,3). The frontal lobes are wider and more rounded
anteriorly. But we must not lay too much stress on this point,
for there is considerable variation in the shape of the frontal
lobes when many chimpanzee brains are examined. There is
one polut which Beddard remarks upon, and which differen-
tiates “‘Sally’s” brain from any other chimpanzee’s brain
described and figured, viz. the very slight “ keel” or ridge on
the inferior or orbital surface of the frontal lobes where they

CEREBRAL CONVOLUTIONS—‘* SALLY.” 5g

touch in the median line. In none of the brains which I have
examined is this keel absent ; it is present in the “ Sally ”-like
brain in the Museum of the Royal College of Surgeons referred
to above; it was present in the brain described by Miiller.

A front view of “ Sally’s ” brain thus presents a great contrast
with that of any other chimpanzee hitherto examined, and
recalls that of an orang or man (see figs. 5—7).

Since it is the posterior part of the brain which is so
characteristically developed, I will commence with this region.

4. Tue Parieto-occipitaL Fissure.

This fissure in “ Sally’s ” brain presents certain features of
interest, chiefly in that it is divided into two portions, one of
which, the mesial or “internal” parieto-occipital, is confined
to the mesial surface; the other extends on to the upper sur-
face of the hemisphere, and may be termed the “lateral” parieto-
occipital fissure, as the term “external”? has been used in
rather a different sense from that which I wish to indicate here.
Viewed from the upper surface (figs. 2, 10), the parieto-
occipital on this left side appears as a deep, well-marked fissure
about 15 mm. long, extending rather obliquely outwards after
coming on to the surface, which it cuts at adistance of 75 mm.
from the anterior end. This superior or lateral portion of the
parieto-occipital is bounded by an “ arcus parieto-occipitalis ”’
(Gratiolet’s ‘seconde plide passage”’), which is, as usual, marked
out laterally by the ramus occipitalis of the intra-parietal (p*.)
and anteriorly by a small mediad of this branch (n.), the pre-
arcal branch. But there is another portion of the parieto-
occipital fissure which does not reach the upper surface, and
which Beddard overlooked. It is seen in fig. 2 lithographed
from a photograph, and also in fig. 10, just in front of the
lateral portion. When the mesial surface of the hemisphere is
examined (fig. 14), it is seen that the first-named portion of the
furrow (lat. p. 0.) is nearly entirely cut off from the remaining
vertically disposed part of the furrow (mes. p. o.), for the
anterior limiting gyrus passes very deeply down at a., and
disappears from view, so that the superficial and the vertical

54 W. BLAXLAND BENHAM.

portions of the parieto-occipital are connected by a short but
deep sulcus.

The vertical portion of the fissure, after thus joining the
superior portion, forks dorsally as shown (at 6. and c.) ; below
the fork it passes straight downwards on the mesial surface,
and is continued on to the tentorial surface of the hemisphere,
where it joins the calcarine fissure—the gyrus cunei being quite
deep. We thus have the <-shaped arrangement of this
system of fissures which Cunningham regards as typical for
man, but which he denies to the apes.

On the right hemisphere (figs. 2 and 11) the lateral part of
the parieto-occipital is entirely separated from the mesial
vertical portion, and is completely surrounded by gyri (see fig.
15) ; its mediad end dips for only a very little distance down-
wards in the mesial face, and the vertical portion—the main
part of the fissure (mes. p. 0.)—comes up to the upper surface of
the hemisphere in front of the fissure in question. So entirely
cut off is the upper part of the parieto-occipital that I at first
took it for a part of the Affenspalte, but its depth and other
relations, especially when compared with the arrangement on
the left side, show it to be a portion of the parieto-occipital, as
Beddard describes it. He, however, overlooked the gyrus
which cuts it off from the vertical part of the fissure, which also
he did not recognise.

The separation of the parieto-occipital fissure into two dis-
tinct portions appears to bea rare occurrence. I do not find it
mentioned in descriptions of chimpanzees’ brains, and only rarely
is it recorded for man. Cunningham does not say very much
about it. He figures the “normal” arrangement on p. 43, where
the fissure, at a point some distance below its upper end, gives
off two branches, one anterior and one posterior, to the main
track, which is separated from the calcarine fissure by a super-
ficial gyrus cunei; but he represents no gyrus intercuneatus.

What has happened in “Sally”? The hinder of the two
branches of the parieto-occipital appears to have been cut off—
partially on the left, entirely on the right—from the rest of the
fissure. Cunningham represents several varieties in the arrange-

CEREBRAL CONVOLUTIONS—“ SALLY.”’ 55

ment of the parieto-occipital in man, but in none of them do we
find quite these relations. In three instances in which the fissure
is duplicated (see his figures on p. 40), the portion which cuts
into the upper surface of the hemisphere lies in front of the
vertical portion, and not behind it asin “ Sally.” In one case
(his fig. 18) this “doubling” is brought about by the gyrus
cunei rising to the surface and the “stem” of his <-shaped
fissure being prolonged into the cuneus. This was only met
with in 4 out of 127 hemispheres examined. In the other case
(his fig. 19), which he found only once in the course of his re-
search, the gyrus intercuneatus is oblique and becomes super-
ficial, the upper part of the parieto-occipital being prolonged
downwards in front of the main part of the fissure.

Now, in “Sally” it appears to me that the cause of the
* doubling” is the same, viz. the rising up of the gyrus
intercuneatus (a. in my figures), but the upper part of the
original fissure cuts downwards into the cuneus, behind the
main part of the fissure, which extends into the precuneus and
only just reaches the surface.

This part of the parieto-occipital fissure lying on the upper
surface and separated from that on the mesial surface may be
termed the ‘‘ superior or lateral parieto-occipital.”

I wish to avoid the term “ external,” which would perhaps
be more appropriate, since the term “ external parieto-occipital”’
has been used in a variety of senses. The only true sense in
which the expression should be used, as Cunningham and
others have pointed out, is to refer to that part of the vertical
incision which appears as the upper surface of the hemisphere,
and varies in extent, naturally, with the depth of the incision.
But the term has had two other significances attached to it,
especially by English writers, namely, to imply (1) true
Affenspalte (Bischoff’s “ fissura perpendicularis externa,” in the
foetal human brain)—in this sense it is used by Beddard; and
(2) the groove between the anterior free edge of the operculum
and the parietal lobe.

We will now examine the conditions of the parieto-occipital
fissure in the normal chimpanzee brain.

56 W. BLAXLAND BENHAM.

In a specimen (947 f) in our Museum (see fig. 22) the
operculum is normally developed. When this is turned back
the Affenspalte is exposed, and in front of it the deep parieto-
occipital fissure is seen, extending about three quarters of an
inch across the brain (fig. 23). On the mesial face this descends
for nearly half an inch (fig. 24), but is cut off from the rest
of the parieto-occipital by a conspicuous gyrus (a.). The main
portion of the fissure lies in front of it, and only just cuts the
edge of the hemisphere (mes. p. 0.). The anterior portion of
the fissure (mes. p. 0.) does not branch. Here, when the
operculum is removed, we have the same arrangement as in
“ Sally.”

In another specimen (fig. 26), in which the edge of the
operculum is directed obliquely backwards at its mesiad end,
we have a further development of the gyrus intercuneatus,
so that the very deep “lateral” parieto-occipital (lat. p. 0.)
scarcely bends round on to the mesial face at all, and lies

almost entirely below the operculum, cutting off the intra-
parietal from reaching the Affenspalte. In fact, it looks at
first sight as if it were the “ Affenspalte,” and as if the gyrus
marked a. were the “ first annectant gyrus.” Such, indeed, it
appears to be, for in many cases the “external perpendicular
fissure ” of authors is not the same fissure as is here called
« Affenspalte,’ although it has been homologised with it.
The “internal” or mesial parieto-occipital (mes. p. 0.) enters
the calcarine fissure, the gyrus cunei being, as far as I can see,
absent (fig. 27).

In a brain referred to by Rolleston,' in which on the right side
the operculum is less developed than usual (see fig. 30), the
parieto- occipital fissure is very deep, and extends for nearly an
inch across the surface of the hemisphere (fig.32). Here it is
limited by a well-defined ‘‘ annectant ” (arcus). Seen from the
mesial surface (fig. 21) there is no trace of a division of the
fissure into two portions. On both right and left sides there is
but a single, nearly vertical cleft—cut off by the gyrus cunei

1 Rolleston, “On the Affinities of the Brain of the Orang-outang,’
‘Nat. Hist. Rev.,’ 1861.

CEREBRAL CONVOLUTIONS—‘* SALLY.”’ 57

from the calcarine fissure—but presenting no lateral branches,
nor gyri coming to the surface.

This seems to be the simplest condition, and closely resembles
the arrangement in the lower monkeys; some other of our
specimens of chimpanzees exhibit this condition ; in others we
have seen the gyrus intercuneatus interfering with the continuity,
and giving rise to two more or less independent fissures.

In a brain at the College of Surgeons (No. 1338, I a*), the
left hemisphere exhibits this simple arrangement, whilst the
right (1338, I a) possesses the more complicated condition.

In the human brain, the separation of the parieto-occipital
fissure into two appears from Dr. Cunningham’s researches
to be rare, but in the brain (950) exhibited as a typical
human brain in the Oxford Museum, and represented here by
fig. 16, we have on the left side an arrangement repeating,
I believe, that in “ Sally,” namely, the uprising of the gyrus
intercuneatus (a.) so as to separate a lateral parieto-occipital
(/at. p. o.) from the remainder (m. p.o.). On the right hemi-
sphere the parieto-occipital cuts into the upper surface for
about an inch, forming the true “ external parieto-occipital,” and
this part is continuous with the vertical portion. But in
the left hemisphere there lies in front of the “ transverse
occipital fissure” or Affenspalte, between it and the vertical
part of the parieto-occipital, a transversely placed fissure
about an inch in length (dat. p. 0.), bounded anteriorly by a
gyrus which is partially overlapped by the “ arcus parieto-
occipitalis.”

5. Tue PartetaL ConvoLurions.

The fissure of Rolando (“ fissura centralis ”) has the usual
wavy character ; it does not enter the Sylvian fissure on either
side, nor does it pass on to the mesial face of the hemisphere.
The distance of the upper extremity of the fissure from the an-
terior end of the hemisphere (between vertical plates) is 45 mm.
There appears to be some confusion in Beddard’s paper with
regard to the position of this fissure of Rolando. In the plate
xxiii, fig. 3, the index line from F, R. is carried to the upper

58 W. BLAXLAND BENHAM.

end of the calloso-marginal fissure. This appears to be a slip,
yet his measurement of the fronto-Rolandic length given on
p. 200 is 54 mm., which is really the distance of the calloso-
marginal fissure from the anterior end of the cerebrum. The
fissure is thus in front of the middle of the cerebrum and not
behind it, as Beddard states.

In the normal chimpanzee here figured (fig. 1) the length
of the cerebral hemisphere is 95 mm., the fronto-Rolandic
length is 55 mm., so that the Rolandic fissure is behind the
middle of the hemisphere.

The intra-parietal fissure of Turner may be divided, ac-
cording to Cunningham, into three constituents: (1) An an-
terior post-Rolandic vertical fissure, or “ sulcus postcentralis.”
(2) A more or less horizontal furrow, or “ ramus horizontalis ”
(“‘interparietal” of Ecker), passing backwards and (3) as a
prolongation which bounds the arcus parieto-occipitalis and
passes on to the occipital lobe—the ‘‘ ramus occipitalis.”

These three constituents may or may not be continuous in
man, and to them is added, usually as a separate furrow, a
“superior postcentralis,” lying more or less parallel to the
Rolandic furrow, and dorsal of the inferior postcentralis.

In “ Sally,” as will be seen (figs. 10, 11), this is also the
arrangement; the inferior postcentralis (p’.) is continued a
short distance upward beyond its junction with the ramus
horizontalis (p°.), and above this fissure is a small but well-
marked furrow, representing the superior postcentralis (p*.) ;
this is continuous on the left side with an oblique furrow (p. s.)
lying parallel to and above the ramus horizontalis, whilst on
the right side the corresponding furrow is parallel with the
long axis of the brain.

This secondary sulcus (p. s.) in the superior parietal lobe
is represented in Cunningham’s drawing (by the letter c.), and
is a very constant furrow in the chimpanzee, though variable
in form.

The superior postcentralis (p.) is in most of the published
figures of chimpanzee brains, as well as in several of those exa-
mined by me, a much more definite feature (sce figs. 30, 36,

CEREBRAL CONVOLUTIONS—*“ SALLY.”’ 59

20), having a course more nearly parallel with the Rolando
than is the case in“ Sally’s” brain. In both of Cunningham’s
figures (pp. 203, 204) this furrow continues the direction of
the inferior postcentralis (p1.).

Cunningham states (p. 230) that the union of superior post-
centralis (y?.) with inferior is the usual condition in chim-
panzee ; and finds them separated in only two of the brains
out of eight examined, and that only on one side. In “Sally,”
in Miiller’s chimpanzee, and two others from the Royal College
of Surgeons, as well asin Herdman’s, I find the fissure marked
by Cunningham as upper part of superior postcentralis,
separate from the rest of the system, and in all these instances
it loses its parallelism to Rolando, becomes more or less oblique,
and enters into connection with the furrow ps. (secondary
sulcus of sup. par. lobule).

6. Tur Recion or THE AFFENSPALTE.

The most interesting feature in “ Sally’s”’ brain, however, is
the condition of the hinder end of the intra-parietal fissure.

In the normal chimpanzee brain, and in the lower apes,
the hinder part of the parietal lobe is overlapped by a com-
paratively thin fold of the occipital lobe, known as the
“ occipital operculum.” This lies flat against the parietal
lobe, and thus conceals a greater or less amount of the latter.
The deep cleft, extending nearly directly backwards and slightly
downwards, existing between the operculum and the parietal
lobe, is known as the ‘‘ Affenspalte,” or Simian fissure (figs. 27,
29, show its direction well). But some confusion has resulted
from the application of this term to the superficial groove
between the edge of the operculum and the parietal lobe.
This superficial groove has also wrongly been termed the
“external parieto-occipital” fissure, since it is apparently
continuous with the “internal” fissure of that name. The
true ‘‘ Affenspalte’”’ lies behind this latter fissure, and the two
are independent.

The parieto-occipital fissure is a vertically placed fissure in
the median surface of the hemisphere, into which it extends

60 W. BLAXLAND BENHAM.

for a considerable distance, and as such is visible also on the
upper surface for a greater or less extent, according as the
incision is deeper or shallower ; that part of the parieto-occipital
(= Huxley’s occipito-parietal) which is seen in the upper
surface is the only fissure to which the term “ external
parieto-occipital ” is properly applied.

I have shown above that this part of the parieto-occipital
may become separated from the so-called “internal ”’ parieto-
occipital ; but it can always be distinguished from the true
Affenspalte, or Simian fissure, or subopercular furrow.

In “ Sally’s ” brain (figs. 10, 11) the ramus occipitalis (p*.)
sends off a small mediad branch (m.) in front of the parieto-
occipital fissure, as is very generally the case in chimpanzees,
the existence of which was first mentioned by Rolleston and later
described by Sir Wm. Turner (4). This small fissure, there-
fore, is of historical interest ; further, it helps to mark out
definitely the so-called “pli de passage supérieur externe”
which Gratiolet claimed for chimpanzee’s brain.

The ramus occipitalis in ‘ Sally”? now curves outwards,
forming the outer limit of the “ pli de passage ” or annectant
gyrus or “arcus parieto-occipitalis,” and enters the ‘ Affen-
spalte.”

The arrangement on the two sides of the brain is not iden-
tical. The above description refers to the right side.

On the left side (fig. 10) the ramus occipitalis appears to
bifurcate, a condition represented in some of Cunningham’s
figures. The shorter, outer or laterad fork enters the “A ffen-
spalte;” the more conspicuous mediad fork (z.) passes back-
wards on to the occipital lobe in nearly a straight line.

On the right side the ramus occipitalis also appears to bifur-
cate, but this is not the case. Owing to the fact that the
parieto-occipital fissure has become divided into two portions—
the “lateral”? portion lying on the upper surface (lat. p. o.),
whilst the “internal” portion is almost confined to the mesial
surface and only just reaches the upper surface (mes. p. 0.)—
the fissure (v.), which at first sight might be taken for the
mediad fork of the ramus occipitalis, is, as we have seen, in

CEREBRAL CONVOLUTIONS—“ SALLY.” 61

reality the przparieto-occipital branch of the intra-parietal
fissure, which forms the anterior limit of the “ arcus parieto-
occipitalis.” This branch (.), which Beddard’s figure repre-
sents as joining the parieto-occipital, is separated by a gyrus
from this, but as it is overlapped by the superior parietal lobule,
its true relations are not seen in a view from above (as, for
example, in a photograph).

After giving off this small branch the intra-parietal fissure
(p*.) curves slightly outwards and enters the Affenspalte, the
well-marked obliquely-transverse groove so evidently the
remains of the more extensive fissure of the normal chimpanzee.
The hinder boundary of the Affenspalte on each side is a low
ridge only slightly above the level of the opposite side of the
furrow ; this is the remains of the operculum.

In describing the terminations of the ramus occipitalis,
Cunningham states that the two forks into which it divides are
frequently almost at right angles to the main stem and nearly
in line with one another, giving rise to a transversely directed
fissure, which he identifies with the “ occipitalis transversus ”
of Keker.

The condition of affairs on “Sally’s” left hemisphere (fig. 10)
is, indeed, not very unlike the left side of the figure of the human
brain given by Ecker, in which the fissure marked there o., into
which the intra-parietal falls, consists of two parts, one laterad
of the intra-parietal, which soon forks; the other, mediad, lies
behind the parieto-occipital. The question arises, do these
two parts correspond to the two fissures in “ Sally’s’’ brain, the
laterad (Affenspalte) and the mediad (z.), which I have described
as a continuation of the ramus occipitalis? In other words,
may the “ occipitalis transversus ’? be composed of two parts
originally independent? This is a view held by Eberstaller.

There is no doubt but that in the apes the bifurcation of the
intra-parietal is an independent thing which may or may not
become connected with the Affenspalte. In some cases the
outer, in other cases the inner fork enters the Affenspalte, or
both forks. Thus in Cunningham’s figure, p. 203, right side, the
laterad fork is independent, the mediad joins the Simian fissure

62 W. BLAXLAND BENHAM.

but again leaves it, whilst, as we see in “ Sally ”’ on the left side,
it is the laterad fork which enters the Affenspalte. Cunning-
ham shows that this is the case in the ape, and one of our
chimpanzees (fig. 31) somewhat resembles the left side of the
brain of Cebus albifrons figured by him in illustration of
the point (on p. 223 of his memoir).

Amongst our specimens of chimpanzees it is very frequently
the case that the intra-parietal does not bifurcate, or if it does
the resulting figures are coextensive with Affenspalte. The
ramus occipitalis in our specimens usually passes under the
operculum and enters the Affenspalte. In one case, however
(see figs. 22, 23), the parieto-occipital fissure makes so deep an
incision in the hemisphere that it cuts off the intra-parietal from
the Affenspalte. We have here a condition similar, as I believe,
to that described by Dr. Cunningham for Cebus capucinus,
which he uses to support his view as to the homology of the
“occipitalis transversus” with the bifurcation of the ramus
occipitalis. According to him the fissure regarded as ramus
occipitalis lies nearly at right angles to the rest of the intra-
parietal fissure, and the parieto-occipital fissure falls into the
system at this angle (loc. cit., fig. 45, p. 222).

But it appears to me, from observations on C. robustus,
that another explanation may be given of the fissure he regards
as “ ramus occipitalis,” for in our specimen of this monkey,
if the occipital lobe be pressed backwards it is seen that the
parieto-occipital fissure is so deep that it extends more than
halfway across the hemisphere, and receives, nearly in its middle,
the intra-parietal. So that what in C. capucinus is regarded
by Cunningham as “ramus occipitalis” is in C. robustus
the outer part of the parieto-occipital. The Affenspalte, as
Cunningham shows, is quite independent of this transversely
placed fissure, as it is too, in our chimpanzee (947 f) just
referred to.

Now in the human brain do we find any close similarity
between the arrangement of the convolutions in this region
aud those of “ Sally” ?

I have already referred to Kcker’s diagram of a generalised

CEREBRAL CONVOLUTIONS—‘* SALLY.” 63

condition of the human fissures, which is closely similar to
“ Sally,” the left hemisphere in each case being compared.
Again, the left side of the brain figured by Bischoff (and repro-
duced in fig. 17) is like the right hemisphere of “Sally” (fig. 11)
so far as the “ Affenspalte”’ is concerned. The same oblique
direction is to be noted. On the right side of the same brain,
too (fig. 18), we have a very similar arrangement, but the
intraparietal fissure (p‘.), before entering the “ Affenspalte ”
(transverse occipital) sends inwards a short branch (z.) behind
the “ arcus parieto-occipitalis.”’

In a human brain in the Oxford Museum, to which I
have already referred, we have on the left side (fig. 16) a con-
dition of the transverse occipital which seems to support
Cunningham’s view. The intra-parietal appears to curve round
the ‘“arcus,” and nearly reaches the mesial fissure, where a
rather irregular furrow passes backwards from it; this small
portion (z.) resembles a part of the “ mediad fork” of the ramus
occipitalis of ‘‘Sally’s” left side (on fig. 10). The ‘‘transversus
occipitalis ”’ in this human brain would then be, according to
Cunningham, the upper (mediad) fork of the intra-parietal.
But the general direction of the “ occipitalis transversus ” is so
very similar to that on the right side of “Sally’s” brain that it
is possible to explain this irregular furrow (z.) by supposing it
to be an independent adventitious furrow on the occipital lobe,
such as occurs in “ Sally’s ” left hemisphere, which there has
gained a connection with one of the forks of the ramus
occipitalis.

It is still, in fact, a debateable point as to whether there is
in the adult human brain a representative or homologue of the
“* Affenspalte ” of the ape. There are, indeed, three chief views
on the subject: (1) Ecker regards his “sulcus occipitalis
transversus ” as such; (2) Eberstaller (to quote from Cunning-
ham) draws a distinction between the upper and lower limbs of
the sulcus occipitalis transversus of Ecker. In it we can dis--
tinguish a medial and a lateral segment by the point of union
with the sagittal portion of the intraparietal furrow. The
former (z. in my figure) bounds the arcus parieto-occipitalis

64 W.-BLAXLAND BENHAM.

behind, without, in the majority of cases, reaching the mesial
border; the latter is the sharply descending end of the arch-
like “ fissura interparietalis.” Ecker’s view, originally shared
by Cunningham (‘ Journ. Anat. Physiol.,’ 24), isnow combated
by him, chiefly, it appears, on embryological grounds, partly
on those of comparative anatomy. (8) He would regard the
transverse occipital as merely a portion of the “ intra-parietal
system ” which has nothing to do with the Affenspalte. It is
due to the bifurcation, in fact, of the ramus occipitalis ; the
Affenspalte, according to him, is only very rarely represented
in man. The “ fissura perpendicularis externa” of Bischoff
makes its appearance in the human fetus about the fifth month
(according to Ecker and Cunningham) ; it is a “complete”
fissure impressing itself upon the ventricle, so that the outer
wall is bulged inwards by it. Some time later, apparently
during the sixth month, it disappears. Now all these authorities
are agreed that this external perpendicular fissure of the human
foetus is the homologue of the “ Affenspalte” of the ape. But
it is only a transient furrow, though “ complete,” and its place
is occupied later on (seventh or eighth month) by the terminal
bifurcation of the intra-parietal fissure, which is not a “ com-
plete ” fissure, that is, it does not leave its impress upon the
wall of the ventricle ; this second or replacing fissure becomes
the “ transverse occipital.”’

For these and other reasons Cunningham denies the possi-
bility of the suggestion that the latter fissure can be the homo-
logue of the Affenspalte.

Nevertheless it must be borne in mind that with regard to the
“‘completeness”’ of the Affenspalte, Cunningham (p. 69) “cannot
tell whether or not it (the elevation of the wall of the ven-
tricle) exists in the anthropoid brain. Specimens of these are
so valuable that I am unwilling to destroy those I have got, even
in a determination of a point of this importance.” He has
only observed the bulging inwards of the outer wall of the
ventricle at the bottom of the Affenspalte in the Sooty Mangaby
and in Cebus.

As the Oxford Museum possesses several chimpanzees’ brains,

CEREBRAL CONVOLUTIONS—“ SALLY.” 65

Professor Lankester very generously gave me permission to
dissect one of them to ascertain what was the fact with regard
to this point. I sliced the hemisphere in planes parallel to the
median plane till I cut through the posterior cornu of the lateral
ventricle. I find that the “ Affenspalte”’ does not causea
bulging in the roof or side of the ventricle. In the figure (fig. 9)
there appears to be such a bulging, but this is a portion of the
‘ calcar avis’ passing on the roof of the ventricle, and the slight
ridge (#)just in front of it lies distinctly in front of the Affenspalte.
Miller, who represents dissections of the brain of a chimpanzee,
does not figure or describe any such bulging. But even if there
were such an elevation of the outer wall of the ventricle in the
lower apes, it does not seem to me a consequence that the higher
apes would present one. Even if in the ordinary chimpanzee
with a well-marked operculum and deep Affenspalte the impress
of the latter is exhibited by the outer wall of the ventricle, it does
uot necessarily follow that in the condition in which the Affen-
spalte is present in “ Sally,” the same impress would be left.

As the operculum disappears, and the Affenspalte becomes
proportionately shallower, it seems quite possible that the
bulging of the ventricular wall should get less and less
marked.

In the human brain the “ occipitalis transversus,” as we
have seen, replaces the “ fissura perpendicularis externa ”’ after
the disappearance of all trace of the latter. It does not how-
ever follow, because there is no actual continuity of the two
grooves, that there is no genetic relation between them. The
obliteration of the earlier furrow is probably due to the growth
of the nerve tissue bringing about a thickening in the wall of
the ventricle, and “ filling up” (if we may use the expression)
the furrow; at the same time the bulging on the inner face
becomes smoothed out. But the furrow again makes an effort
—inheritance is too strong for the vegetative growth,—and
succeeds in obtaining its permanent position as the “ occipitalis
trausversus,” which is relatively shallower and does not cause
any bulging inwards of the wall of the ventricle.

We know that ontogeny does not by any means follow

VOL. 37, PART 1.—NEW SER. E

66 W. BLAXLAND BENHAM.

slavishly the exact steps of phylogeny: the mode of formation
of the heart in mammals, for example, differs from that of birds,
nevertheless the two hearts in their entirety are surely homo-
logous and homogenetic; yet the ontogenetic condition cannot
represent a functional phylogenetic stage.

In the ontogeny of some Vertebrates the cesophagus
becomes closed at a certain period of development, and it
again opens out to form the permanent esophagus. We do
not deny the homology (homogeny) of these two structures.

It appears to me that the facts which have been gathered
together with regard to the Affenspalte in the chimpanzee, and
the “ occipitalis transversus ” in man,—the fact also that there
is considerable variation in position and extent of the fissure in
both animals, and that parallel variations occur in both,—are
so strongly in support of Ecker’s view, that the mere fact of a
slight difference in ontogeny is not sufficient to overturn the
homology.

Further, what is the course of development of the “ Affen-
spalte” in the ape and monkey? Weare, I believe, absolutely
ignorant on the matter.

In the human brain (fig. 17) it seems at first sight pretty
evident that the sulcus occipitalis transversus is formed by the
bifurcation of the intra-parietal (p*.) ; the angle formed by the
two branches of the fissure lends its weight to this view. But
compare with it fig. 18. Here p*. gives off a branch (z.)
behind the arcus parieto-occipitalis, which, there is little
doubt, is the mesiad fork of the intraparietal, whilst the laterad
fork enters the transverse occipitalis. This fissure itself closely
resembles the “ Affenspalte” in ‘ Sally” (fig. 11). Thus
Cunningham’s interpretation would apply to fig. 17, and
Ecker’s to fig. 18.

Now, in some chimpanzees, such as that represented by figs.
30 and 32, we have an “‘ Affenspalte””? below the operculum
(fig. 32) into which there falls the ramus occipitalis (p‘.),
and at this point an angle occurs in the “ Affenspalte.’’ There
is, in fact, a very close resemblance to the condition in the
human brain (fig. 17), which Cunningham would explain as

CEREBRAL CONVOLUTIONS—* SALLY.” 67

bifurcation of the intraparictal. How are we to distinguish
between the two? We can scarcely regard the subopercular
furrow of the chimpanzees (fig. 32) as anything but the Affen-
spalte. Therefore, why should we not apply the same name to
an apparently similar furrow in fig. 17?

In fig. 31, again, we have an arrangement somewhat like
that in the human brain (fig. 18), viz. the ramus occipitalis
(p*.) gives off a post-arcal branch, which is distinct from the
(greater part of the) Affenspalte. The figs. 17 and 18 repre-
sent two sides of a human brain, and figs. 31 and 32 the two
sides of a chimpanzee’s.

I do not pretend to decide the matter. It would be great
presumption on my part, with my slight experience of human
cerebral topography, to express any decided opinion on a
matter on which such authorities as Eberstaller, Ecker, and
Cunningham are at variance. I merely wish to bring forward
other facts and arrangements of these fissures which appear to
me to have some bearing on the question.

7. Tue Occiritat Lose.

The occipital lobe itself is more furrowed in “ Sally’s”
brain than in the ordinary chimpanzee, and bears some resem-
blance to that of Miuller’s brain.

The old terms given by Gratiolet to the gyri in the neigh-
bourhood of the parieto-occipital fissure, and to which so much
importance was at first attached—the annectant gyri, or “ plis
de passage”—are now gradually being discarded. Since
Rolleston and Turner showed their presence in chimpanzee,
where they are usually concealed by the operculum, the dif-
ferential importance of them has gone; and the names, though
still sometimes employed, appear to be giving way to more
descriptive and systematic terms.

The pli de passage supérieur externe and pli occipital supé-
rieur are now called by one name, “ gyrus occipitalis primus ”
(Ecker), or “ arcus parieto-occipitalis”’ (Eberstaller).

This arcus parieto-occipitalis is well seen on the right hemi-
sphere, where it curves firstly round the lateral extremity of

68 W. BLAXLAND BENHAM.

the “lateral parieto-occipital,’ and is continuous with the
gyrus occipitalis primus, which bends round the mediad ex-
tremity of the Affenspalte.

The gyrus occipitalis secundus (O 2) (deuxiéme pli de pas-
sage externe) lies behind the Affenspalte (in fact, is the opercu-
lum) and curves round its laterad extremity, becoming con-
tinuous with the “ angularis.”’

The third occipital gyrus lies below the oblique fissure (the
sulcus occipitalis inferior), and also passes into the third tem-
poral gyrus.

The calcarine fissure just cuts the edge of the occipital
lobe, but is more extensive on the left than on the right hemi-
sphere.

8. Tue Sytvian Fissure.

The Sylvian fissure in “Sally’s” brain is less vertical than
in the brain figured by Miller, though, as Beddard has re-
marked, it is more vertical than in common chimpanzees.
The “Sylvian angle” in Dr. Cunningham’s sense is 55° in
“ Sally ;” the brain received from Professor Herdman has
rather a smaller angle, viz. 50°, whilst normally it is, according
to Cunningham, 54°5°.

With regard to the “anterior limb” of the Sylvian fissure,
Beddard has quite rightly figured it as a short, nearly hori-
zontal fissure (Pl. 23, fig. 2, F.s.a.). A comparison of a series
of chimpanzee’s brains led Cunningham to believe that it is
the anterior boundary of the fronto-parietal operculum, and
therefore is the homologue of the ‘ramus ascendens” in man.
He gives reasons for his conclusion that in Troglodytes the
only opercula covering the island of Reil are the fronto-parietal
and the temporal. The two anterior ones, frontal, and orbital
of man, are absent; and he suggests that part of the orbital
lobule in the ape may represent the island of Reil in man.

In “Sally” the insula is distinctly limited anteriorly by a
portion of the frontal lobe, which is at a much higher plane
than the insula, and separated from it by a deep fissure; but
it does not sensibly overlap the insula. However, in other

CEREBRAL CONVOLUTIONS—* SALLY.”” 69

chimpanzee brains this portion of the frontal lobe below the
“anterior limb ” is much more developed, and forms practically
an operculum—overlapping, though only to a slight extent, the
insula. .

Cunningham states that the submerged portion of the insula
presents very little trace of an anterior boundary, and that it
ascends gradually along an inclined plane until it finally reaches
the free surface of the frontal lobe. I do not find this to be
the case at all generally. In two brains in the Oxford Museum
this overhanging anterior lobe exhibits more or less distinct
traces of a subdivision into two lobules.

The most distinct case was photographed, and is represented
in fig. 8. Here the “anterior limb” of the Sylvian fissure is
forked ; one branch is nearly vertical (Sy’.), and forms as usual
the boundary of the fronto-parietal operculum; it is the
“‘ramus ascendens ” of the anterior limb. Just below it there
is another slight but distinct fissure (Sy’’.) separating a small
triangular lobe from the rest of the anterior boundary of the
insula (fr. op.). This little lobe, I suggest, is the “ frontal
operculum” or pars triangularis, and the second fissure is the
homologue of the “ anterior horizontal limb” of the Sylvian
fissure. Below it the remainder of the lobe will then corre-
spond with the orbital operculum, which here distinctly over-
laps the insula.

A second brain (Oxford Museum, 947 f.) shows somewhat
similar conditions, but in a less degree. These may be com-
pared with Cunningham’s fig. 1, pl. iv, of the brain of a human
foetus of the eighth month.

The brain represented in fig. 8 resembles still more closely
the figures given by Schafer in ‘ Quain’s Anatomy’ (tenth
edition, vol. iii, part 1). He shows that on the right side of
this human brain the two branches of the anterior limb of the
Sylvian fissure practically join at their commencement, i. e.
the ‘pars triangularis”” or frontal operculum is here less
developed than on the left side, where it entirely separates the
two rami. The greater development on the left side of the pars
triangularis may be in relation with the “centre of speech.”

70 W. BLAXLAND BENHAM.

It appears to me particularly interesting to find in the chim-
panzee a condition so closely resembling that on man’s right
side.

Cunningham regards the inferior frontal convolution of
the ape as representing non-covered insula, and the “ sulcus
fronto-orbitalis” as the homologue of the anterior limiting
sulcus of the insula, and compares it with foetal human brains
at a certain stage. But the position and relations of the
fronto-orbital fissure (e. 0.) with regard to the “ anterior
limb ” of Sylvius (limit of fronto-parietal operculum) seem to
me to be very different. In most of the chimpanzees it passes
upwards from the orbital surface of the frontal lobe some
distance in front of the anterior end of the Sylvian fissure,
and curves upwards and forwards, extending considerably above
the so-called “‘ anterior limbs.” It is easy to understand how
the folding backwards—formation of orbital operculum—in
the fig. 9, pl. iv of Cunningham’s memoirs would give rise to
the state of things represented in his fig. 1 of the same plate
representing human foetal brain.

Here the fissure supposed to correspond with the fronto-
orbitalis joins the “‘ anterior ascending limb” of the Sylvian
fissure, and then curves backwards. From the facts referred
to above, and represented in the photograph (fig. 8), it seems to
me that the sulcus fronto-orbitalis may have some other
meaning. May it not be part of the system of orbital furrows
which has extended upwards ?

9. Frontat Lose.

A system of furrows, more or less parallel to, and in front
of, the fissure of Rolando, constitutes the ‘ preecentralis.”

This precentral fissure arises, according to Cunningham, in
three pieces—vertically, a “ preecentralis inferior,” and a “ prze-
centralis superior,” with a horizontal ramus, which is usually
connected with the upper part of the former.

On the right hemisphere of “Sally’s”’ brain (figs. 11, 12) the
sulcus precentralis inferior (p. c. 7.) is divided into an upper
and a lower portion by a submerged gyrus situated just above

CEREBRAL CONVOLUTIONS—“ SALLY.” 71

the junction of the fissure with the s. frontalis secundus (f?.) ;
the upper portion of this vertical fissure appears to represent
the “ramus horizontalis” (f4.). The pracentralis superior
(p. ¢. 8.) is a very well marked, transversely-placed furrow,
seen best in the view of upper surface. The lower-end lies
behind the upper end of the ramus horizontalis. It is in free
communication with the distinct ‘sulcus frontalis primus ”
(or superior) (f1.), which runs straight forwards and bifurcates
anteriorly.

Lying in front of this is an oblique furrow, with its posterior
end directed towards the mesial fissure, which it almost
reaches: the anterior end bifurcates, and lies in front of, but
mediad of, the end of the sulcus frontalis primus. What is
this fissure ?

It is represented in the left hemisphere by a smaller and
unbranched furrow.

Is it a portion of the frontalis primus ? or is it a representa-
tive of the “sulc. frontalis mesialis ” ?

According to Cunningham, the disjointed portions of the
sulcus frontalis primus lie in exactly the opposite direction,
i.e. the posterior ends of each are laterad (outside) of the ante-
rior end of the preceding. From the fact that these disjointed
pieces are slightly variable, it will perhaps be safer to leave
the matter open, for hitherto the “s. frontalis mesialis”’ is
characteristic of the human brain, and has not been met with
in the ape.

In front of this, again, is a curved furrow which is probably
another portion of the s. front. primus.

The s. frontalis secundus (or inferior) has been carefully
identified by Cunningham, and he has conclusively shown that
the sulcus rectus of the monkey’s brain is the homologue of
the s. frontalis inferior, and not of the s. frontalis medius
as Eberstaller believed. Cunningham therefore agrees with
Gratiolet’s views, which are at variance with some more
modern views, that the inferior frontal lobe is present in the
apes.

In “Sally” the sulc. frontalis inferior (secundus) (fig. 12, f?.)

72 W. BLAXLAND BENHAM.

is well developed. Posteriorly it is continuous with the pre-
centralis inferior, whence it runs nearly directly forwards for
some distance and then bends downwards in front of the orbito-
frontalis (e. 0.), when it divides into a fore-and-aft nearly hori-
zontal fissure (w.), and can be traced forwards to the frontal
pole. The anterior branch appears to be the fronto-marginalis
of Wernicke.

In his description of this fissure in man, Cunningham states
that the hinder limb of the fork lies between the two anterior
limbs of the Sylvian fissure. Granting that the s. frontalis
inferior is identical in man and apes,—and in “ Sally ” it closely
resembles the arrangement figured by Cunningham on p. 248
for man,—the hinder branch of the fork ought to have the
game relation to the fronto-orbital furrow (e. 0.) as it has in
man to the anterior horizontal limb of the Sylvian fissure, if
Cunningham’s identification of the fronto-orbitalis is the true
one. But here in “Sally” it has not this position; a fact
which is adverse to Cunningham’s interpretation of these other
furrows. But in other chimpanzees this fissure marked “ w.” is
not related to the “s. front. inferior,” but to the ‘‘ medius,”
as in fig. 35, where the s. frontalis inferior is short and its
anterior end is surrounded by a curved fissure—part of the
“ medius ” (f. m.)—which passes downwards and divides near
the orbital surface of the brain into a fore-and-aft horizontal
furrow (w.), which can be traced round to the frontal pole.
This appears to be the s. fronto-marginalis of Wernicke,
and this brain resembles in this respect the one figured by
Cunningham (fig. 68, p. 290). In other cases the “ medius ”
runs in the same direction, but does not divide (fig. 38).

The s. frontalis medius in “ Sally” is unconnected with
the s. precentralis inferior, aud is quite a distinct fissure on
both hemispheres.

Turning now to the left side of “Sally’s” brain (figs. 10, 18),
we find the sulcus precentralis inferior (p. ¢. 7.) is a well-marked
continuous furrow extending from nearly the Sylvian fissure
upwards for about half the surface of the hemisphere. There
appears to be no ramus horizontalis, unless it is represented by

CEREBRAL CONVOLUTIONS —** SALLY.” 73

a small oblique furrow (f.), recalling the “ precentralis
medius”’ of some authors.

Cunningham in some cases finds in man the ramus hori-
zontalis nearly vertical, or composed of.a forked furrow. This
he considers is merely an extreme condition to which the
term ‘“‘ preecentralis medius’”’? has sometimes been applied.
The other frontal furrows are normally developed.

It may be worth while to refer to some of the variations in
the arrangement of these furrows in our other chimpanzee
brains. In most of them the ramus horizontalis is continuous
with the preecentralis inferior; it is most characteristically
developed in the brain represented in figs. 338, 34. In fig. 38,
again, this latter furrow is very well developed, but has no
auteriorly directed horizontal branch. Probably the upper part
of it, together with the short posterior horizontal (/.) branch,
represents the ramus horizontalis. In fig. 28 the ramus
horizontalis is continuous with what appears to be a portion of
the frontalis primus. It is in both these cases more or less
vertical.

The frontalis secundus (f%.) is continuous with the precen-
tralis inferior, except in fig. 28 and on the brain represented
at fig. 30. In the former the frontalis secundus is connected
with a downwardly directed vertical furrow, which is very much
more distinct in fig. 33. This condition may be compared with
that of an eighth month human fcetus figured by Cunningham
on p. 250.

In the brain represented in figs. 30, 34, on the right side
the precentralis inferior (p. ¢. 7.) is provided with a large ramus
horizontalis (2.) composed of an anterior sagittal and a posterior
constituent directed obliquely upwards. The s. frontalis
medius (f. m.) and the “inferior” (f*.) are separate from the
precentralis and are nearly simple furrows ; but the “inferior”
bifurcates at its extremity. This bifurcation probably represents
the s. fronto-marginalis. 8. frontalis primus (/1.) is in three
separate pieces, the hindermost being continuous, as usual,
with the s. preecentralis superior (p. ¢. s.).

On the left side of the same brain (fig. 33) a very peculiar

74, W. BLAXLAND BENHAM.

association of furrows occurs, somewhat like that in fig. 35.
The ramus horizontalis is connected, as usual, with the s.
precentralis inferior. The s. frontalis medius is a curved
fissure, concave downwards, lying some little distance in front
of the precentralis inferior; from this is given off a branch
which runs forward—a modification of the ordinary forking
of this fissure. But where is the ‘‘ inferior ’’?

Just below the medius and lying parallel to the ramus hori-
zontalis is a very short fissure (separated from the preecentralis
inferior by a gyrus, which just reaches the surface), crossing
the top of a vertical fissure of considerable extent. This latter
passes downwards nearly to the anterior limb of Sylvius.
This fissure appears to represent the sulcus diagonalis, and
the horizontal one at its upper end the s. frontalis inferior.
We have in this chimpanzee brain a condition resembling the
foetal human brain of eight months figured on p. 250 by
Cunningham—the “ inferior,” however, being shorter in the
chimpanzee.

The brain represented by fig. 38 is of interest in that the
s. frontalis secundus is divided into two pieces by a broad
gyrus, one portion remaining continuous with the precentralis
inferior—a condition resembling that of a seventh month
human foetus figured by Cunningham on p. 276, fig. 63; and
it is probable that the part of the curved longitudinal fissure
in fig. 33, labelled? f2., is a similarly disjointed portion of
the “ frontalis secundus.”

It is unnecessary to enter into further detail, as an examina-
tion of the figures will show the interpretations which I put
upon the various fissures.

10. Resunts oF THE CoMPARISON OF A SERIES OF CHIMPANZEE
BraIns.

1. The parieto-occipital fissure, originally a simple and
single incision on the mesiad side of hemisphere, frequently in
the chimpanzee and more rarely in man becomes divided into
two fissures by the gyrus intercuneatus becoming superficial.
Of these two portions, the superior, or lateral parieto-occipital,

CEREBRAL CONVOLUTIONS—“ SALLY.” 75

may come to be more or less entirely on the surface of the
hemisphere (figs. 11, 16, and 31), whilst the lower or mesial
portion lies on the mesial surface. This “lateral parieto-occi-
pital ” is independent of the A ffenspalte, and is not synonymous
with what is frequently termed in text-books the “ external
parieto-occipital.”

2. The occipital operculum in the chimpanzee presents very
great variations in its size; the two hemispheres of a given
brain frequently present differences in respect of this feature
(cf. figs. 10, 11, 19, 20, 25, 26, 30). In “Sally” this operculum
is practically absent, and the subopercular groove or ‘ Affen-
spalte ” is fully exposed.

3. This “ Affenspalte,” or Simian fissure of the ape, seems to
be homologous with the “sulcus transversus occipitalis” of
Ecker, and to be independent of the bifurcation of the intra-
parietal fissure.

4, In some chimpanzees there exists a ‘ramus ascendens ”
and a “ ramus horizontalis” of the anterior limb of the Sylvian
fissure ; these enclose a “ pars triangularis” or frontal opercu-
lum overlapping the insula (see fig. 8). Further, I find that
the anterior boundary of the insula is more or less distinctly
marked, and that there is a distinct and sudden elevation of
the frontal lobe at the anterior margin of the insula. In
other words, an “orbital operculum” is present in some
chimpanzees, though it may be feebly marked in some speci-
mens.

5. There is a fissure in “Sally’s”’ brain which may possibly
be the homologue of the sulcus frontalis mesialis (figs. 10 and
11, “? f!.”), though I have provisionally taken it to be a dis-
jointed portion of the sulcus frontalis primus, since the former
sulcus has not hitherto been recognised in any ape brain.

6. Various arrangements of the frontal fissures in chim-
panzees resemble those recorded in man, either adult or
foetal.

76 W. BLAXLAND BENHAM.

11. List oF THE MORE ImpoRTANT MeEMOIRS DEALING WITH
THE BRAIN OF THE CHIMPANZEE.
1. Gratrotet.—‘ Mémoire sur les Plis cerebraux de l’homme et des Pri-
mates.’

2. Marsuatt, J—‘ On the Brain of a Young Chimpanzee,”’ ‘ Nat. Hist.
Review’ (new ser.), i, 1861, p. 276.

8. Biscuorr, T. W.— Die Grosshirnwindungen des Menschen,” ‘ Abh. d.
k. Bayer. Akad. d. Wiss. Math. Phys. Classe,’ x, 1866.

4, Turner, W.—‘‘ Notes more especially on the Bridging Convolutions in
the Brain of the Chimpanzee,” ‘ Proc. Roy. Soc. Edin.,’ v, 1866.

5. Biscuorr.—‘ Ueber das Gehirn eines Chimpanse,” ‘Stzber. d. Wiss.
Math. Phys. Classe d. k. Akad. Bayerish.,’ 1871, Miinchen.

6. Broca.— L’Etude sur le cerveau du gorille,” ‘ Rev. d’Anthrop.,’ 1878.

7. Cuapman.— On the Structure of the Chimpanzee,” ‘ Proc. Acad. Nat.
Sci. Philadelphia,’ 1879, p. 52. (Gives literature.)
8. Parker.— On the Brain of a Chimpanzee,” ‘ Medical Record,’ 1880.
9. Mijttmr.— Zur Anatomie des Chimpanse Gehirns,” ‘Arch. f. Anthro-
pologie,’ xvii, 1888, p. 173.
10. Symineton.—‘‘ On the Viscera of a Female Chimpanzee,” ‘ Proc. Roy.
Phys. Soc. Edin.,’ x, 1890, p. 297.

11. Bepparp.— Contributions to the Anatomy of the Anthropoid Apes,”
‘Trans. Zool. Soc.,’ xiii, 1898.

12. Cunnincuam.—“ Contribution to the Surface Anatomy of the Cerebral
Hemispheres,” ‘Cunningham Memoirs,’ vii, Roy. Irish Acad., 1892.

12. Tue BRAIN OF THE ORANG-OUTANG.

The arrangement of the frontal convolutions in this ape is
so varied, and the fissures have in many specimens such pecu-
liar relations, that Dr. Cunningham, in the few remarks he
makes on the subject (loc. cit., p. 295), confesses his un-
certainty as to the accuracy of the identification which he
places upon these fissures.

As the Oxford Museum possesses two brains of the orang,
and as these differ from one another, as well as from
that described by Cunningham, it has seemed to me worth
while to place on record the plan of these convolutions, and to
endeavour in the light of these new variations to arrive at a
more accurate determination of these fissures. Even if I am

CEREBRAL CONVOLUTIONS—‘* SALLY.” Pari

not successful in this—and as so great an authority as Dr.
Cunningham has confessed his inability to thoroughly establish
his identifications I feel some hesitation in believing that I
should be more successful—I shall at any rate have added
new facts to the comparatively small amount of knowledge that
we possess concerning the orang’s brain.

Of the two orangs’ brains one is historical, being that de- |
scribed and figured by Rolleston in 1861 (2). It is very well
preserved.

The other brain, however, is rather soft and slightly injured
on the right side. It is, further, rather distorted, the hemi-
spheres being pressed over to the left side, so that nearly all
the temporo-sphenoidal lobe is thrust under the lower surface,
against which it is flattened. It appears to have belonged to
a young individual, as it is not nearly so large as those hitherto
figured. The cerebral hemispheres measure 84 mm. in length,
but it is useless to give other measurements owing to the
distortion of the brain.

This brain will be referred to as Brain No. 1, that described
by Rolleston as No. 2.

Orang’s Brain No. 1 (figs. 42—45).

The Frontal Lobes.—On the left side (figs. 42 and 44) the
s. precentralis inferior (p.¢.7.) is continuous above with an
obliquely placed fissure, directed upwards and forwards (A.),
which I would identify as the ramus horizontalis, since the poste-
rior end of the s. frontalis medius lies below the anterior end of
it. About midway along its length, the s. prec. inferior gives
off an anteriorly directed furrow, which I would identify as a
part of the s. frontalis secundus (f?.). Such an arrangement I
have described above in a chimpanzee (fig. 38); and which,
moreover, is very similar to that figured by Cunningham for an
adult man on p. 262 of his memoir.

The s. preceutralis superior (p.c.s.) seems to be in two
portions, the lower of which, lying parallel to the ramus hori-
zontalis, is continuous with the s. frontalis primus (/'.), which
in this hemisphere is represented by only this one fissure.

78 W. BLAXLAND BENHAM.

The s. frontalis medius (f.m.) is a simple fissure passing
forwards nearly horizontally, its hinder end having precisely
the typical position below the ramus horizontalis, from which
it is separated by a deep-lying gyrus.

The s. frontalis secundus (f?.) is in two parts, the chief of
which lies just below the s. frontalis medius and passes forwards
to the frontal lobe, giving rise to the “s. fronto-marginalis ”
of Wernicke. The other part of the fissure to which I have
referred above is in continuity with the s. preecentralis inferior.

On the right side of this brain (figs. 43 and 45) the furrows
on the frontal region have rather a different arrangement. The
sulcus precentralis superior (p.c. s.) is single and of consider-
able length, occupying the position of the two furrows of this
name on the left side. It lies parallel to the fissure of Rolando,
and at about half its length is joined by a longitudinal fissure,
which appears to be the main part of the sulcus frontalis
primus (f1.).

The sulcus precentralis inferior (p.c.i.) is also fairly well
developed; upwards it lies in front of the sulcus precentralis
superior, and below it nearly reaches the Sylvian fissure, from
which it is separated by a deep gyrus. Passing from it ante-
riorly are two horizontally directed furrows, the upper of which
appears to be the ramus horizontalis (/.), the lower and shorter
is prohably a part of the sulcus frontalis secundus (f*.).

Where, now, is the sulcus frontalis medius? It ought to lie
below the ramus horizontalis. There is a furrow in this
position which curves upwards round the anterior end of the
ramus horizontalis, part of which possibly represents this
sulcus frontalis medius. Unfortunately the brain is so ill pre-
served in this region that the exact arrangement of fissures
on the lower part of the frontal lobe cannot accurately be
made out.

Orang No. 2 (figs. 39—41).
In Rolleston’s specimen the upper part of the right hemi-
sphere has been removed, but is preserved. The following is
the arrangement :—On the left side (figs. 39 and 41), lying in

CEREBRAL CONVOLUTIONS—“* SALLY.” 79

front of and nearly parallel to the upper part of the fissure of
Rolando is a conspicuous fissure which passes for some dis-
tance forwards, nearly parallel with the median fissure, for
about half the length of the lobe. This appears to repre-
sent a portion of the sulcus frontalis primus (f1.), together
with the uppermost part of the sulcus precentralis superior
(p.¢.8.).

At a point a short distance behind its anterior termination
a fissure passes downwards and slightly backwards, nearly
parallel with the fissure of Rolando, and before it ceases gives
rise to another fissure directed nearly horizontally forwards
(h.). <A triradiate fissure results; the lower, posteriorly di-
rected limb ends freely, the upper limb joins the sulcus fron-
talis primus. This triradiate fissure appears to represent the
“ramus horizontalis ” which is connected with the vertical
limb of the sulcus precentralis inferior (p. c. 7.).

The sulcus precentralis inferior (p.c.7.) passes downwards
and backwards in the usual way, and near its lower end is
apparently connected with a branch of the sulcus fronto-
orbitalis (e. 0.), from which, however, it is separated by a deep-
lying concealed gyrus.

Behind the precentralis inferior lies a particularly long
sulcus diagonalis (d.) which reaches very nearly up to the
ramus horizontalis, and occupies a similar position and has
about the same extent as the same fissure in the orang’s
brain figured by Cunningham, and reproduced here on fig. 46.

The arrangement of these fissures is very similar to that
figured for a human foetus of the seventh month on p. 268,
fig. 60, of Dr. Cunningham’ memoir.

This system of fissure is no doubt very peculiar—unusual,
though not unknown; connections take place in such a way
that we have a practically continuous series of furrows repre-
senting sulcus precentralis superior, sulcus frontalis primus,
ramus horizontalis, and sulcus diagonalis.

Between the upper end of this composite furrow and the
Rolandic fissure is a short, obliquely placed furrow (a.), which
appears to be the ‘sulcus precentralis marginalis.” The

80 W. BLAXLAND BENHAM.

direction of this furrow does not seem to agree with the view
that it is an isolated portion of the s. frontalis primus.

Beyond the anterior end of the “ sulcus frontalis primus ” is a
rather short furrow, lying nearer to the median fissure, to which
it is parallel. It does not appear to be a separate portion of the
frontalis primus, for such portions are usually oblique, and the
upper lateral end is placed below, laterad of, the end of the
upper part of the frontalis primus. It appears to be possible
that it represents the “sulcus frontalis mesialis,” which
has hitherto been regarded as peculiar to man, and has not
been recorded for any anthropoid brain, and I do not wish to
insist upon this interpretation. The fissures in the orang are
so variable that it is quite possible that this one is merely a
portion of the s. frontalis primus.

The s. frontalis medius and s. frontalis secundus are com-
paratively short but distinct furrows having a longitudinal
direction. Both are independent of the vertical fissures.

The “‘s. fronto-orbitalis,” which in Cunningham’s figure is
so conspicuous, is in both our orangs almost limited to the
orbital surface, only a small part coming on to the outer
surface. This portion presents the peculiarity that it appears
to pass into the sulcus precentralis inferior, from which,
however, it is in reality separated by a deep gyrus.

Parietal and Occipital Lobes.—With regard to the
“ Affenspalte”” our two orangs illustrate two different condi-
tions. In the small brain (No. 1) there is on each side a well-
developed operculum, and the intra-parietal passes into the
Affenspalte below it. But in Rolleston’s specimen, as he
described, the operculum is less developed; the “arcus occi-
pitalis,” or annectant gyrus, comes to the surface, so that the
parieto- occipital fissure is cut off from the Simian fissure.

The arrangement of the “ intra-parietal system ” is exhibited
in the diagrams; the right hemisphere of the one brain
resembles the left hemisphere of the other, and vice versa.
There is nothing special to be noticed about them. Rolleston
fully described the specimen thirty-three years ago.

CEREBRAL CONVOLUTIONS—* SALLY.” 81

List OF THE MORE IMPORTANT WORKS DEALING WITH ORANG’S
BRAIN.

1. GratioteT.—Loc. cit. supra.

2. RottEston.—‘‘On the Affinities of the Brain of the Orang-outang,”’
‘Natural Hist. Review,’ 1861.

3. Biscuorr.—‘ Ueber das Gehirns eines Orang-outang,” ‘Stzb. Math.
Phys. Class., 17 June, 1876, d.k. Bayer. Akad.,’ i.

4, CunnincHamM.—Loc. cit. supra.

EXPLANATION OF PLATES 7—11,

Illustrating Dr. W. Blaxland Benham’s paper, ‘‘ A Description
of the Cerebral Convolutions of the Chimpanzee known
as ‘Sally ;’? with Notes on the Convolutions of other
Chimpanzees and of Two Orangs.”

The following abbreviations have the same significance throughout the figures.

; a@. Gyrus intercuneatus (except fig. 19). aff. “Affenspalte” or Simian fis-
sure. arcus. Arcus parieto-occipitalis, or “ seconde pli de passage” of Gratiolet,
or first annectant gyrus. calc. Calcarine fissure. c.c. Corpus callosum. . m.
Calloso-marginal fissure. d. Sulcus diagonalis. e. 0. Sulcus fronto-orbitalis.
J'. Sulcus frontalis primus. /?. Sulcus frontalis secundus. /. m. Sulcus fron-
talis medius. fr. op. Frontal operculum, or pars basilaris. 4. Ramus horizon-
talis of the precentral system. J. P. Intra-parietal fissure. J. R. Insula of
Reil. «. ¢. Inferior transverse fissure of Rolando. &. Keel of the orbital
lobule. /at. p. o. Lateral portion of the parieto-occipital fissure. m. p. 0. or
mes. p. 0. Mesial portion of the parieto-occipital fissure. x. Prearcal branch
of the intra-parietal fissure. 0.1. Superior occipital convolution. 0. 1.
Second occipital convolution. O. 111. Third occipital convolution. op., opere.
Occipital operculum. ord. op. Orbital operculum, or pars orbitalis. p!. Sulcus
postcentralis inferior. y?. Sulcus postcentralis superior. 3. Ramus horizon-
talis of the post-central system. p*. Ramus occipitalis of the post-central
system. p.c.7z. Sulcus precentralis inferior. p.c.s. Sulcus precentralis
superior. P.O. Parieto-occipital fissure. y.s. Secondary fissure in the
superior parietal lobule. #. or 7. Fissure of Rolando, or sulcus centralis.

VoL. 37, PART 1,.—NEW SER. F

82 W. BLAXLAND BENHAM.

S. or Sy. Sylvian fissure. S’. Its anterior limb. sy’. Ramus ascendens of the
anterior limb of the Sylvian fissure. sy. Ramus horizontalis of the anterior
limb of the Sylvian fissure. ¢emp. Temporo-sphenoidal lobe. 7’. Parallel
fissure. w. Sulcus fronto-marginalis of Wernicke.

Fig. 1.—Dorsal view of a brain of an ordinary chimpanzee. (From a
photograph, reduced.) The operculum is well developed, and conceals the
*« Affenspalte.”

Fic. 2.—Dorsal view of the brain of “ Sally,” for comparison with Figs. 1
and 3. ‘The operculum is very feebly developed, and the “ Affenspalte ”
is fully exposed. The parieto-occipital fissure is divided into two, more dis-
tinctly on the right than on the left side. The position of the fissure of
Rolando is to be noted, compared with Fig. 1. The frontal lobes are very much
fuller than in the ordinary chimpanzee, so that the shape of the hemispheres
more nearly resemble those of man. (From a photograph, reduced.)

Fic. 3.—Dorsal view of the “ Hottentot Venus ;” from a photograph of
Gratiolet’s figure, reduced to the same size as those of the chimpanzee.

Fic. 4.—Side view of “Sally’s ” brain, from a photograph.

Fig. 5.—Front view of an ordinary chimpanzee’s brain, rather larger than
natural size. The orbital region of the frontal lobe on each side is greatly
compressed, forming a “ keel” (4.).

Fic. 6.—Front view of “ Sally’s” brain, in order to exhibit the almost
entire absence of the “keel.” Much less of the temporo-sphenoidal lobe
is seen.

Fic. 7.—Front view of a European brain, for comparison with Fig. 6.
Reduced. The keel is very similar to that in “ Sally.”

Fie. 8.— View of the “island of Reil” of the right side in a chimpanzee
(Oxford Museum, No. 947 /.). The temporo-sphenoidal lobe has been pulled
downwards. Sy’. The “ramus ascendens,” and Sy". Ramus horizontalis of the
anterior limb of the Sylvian fissure. 7. op. Frontal operculum. ord. op.
Orbital operculum. (From a photograph.)

Fic. 9.—View of a longitudinal section of a chimpanzee’s brain (Oxford
Museum, No. 947 7). (From a photograph.) The section passes along the
posterior cornu (p. corn) of the lateral ventricle, and is intended to show
the relation of the Affenspalte to the wall of the ventricle. cal. av. Portion
of the calcar avis.

Fics. 10—15 illustrate the surface anatomy of “ Sally’s” brain. They are
drawn life size, and the position and shape of the sulci are represented with
the greatest care.

Fig. 10. View of the left hemisphere, as seen when looked at from above
and slightly laterally. The arrangement of the sulci on the frontal,
parietal, and occipital lobes are shown.

Fig. 11. View of the right hemisphere, as seen when looked at from
above and slightly laterally.

CEREBRAL CONVOLUTIONS—“ SALLY.” 83

Fig. 12. View of the right hemisphere, seen laterally but also with the
upper surface shown, so that the result is a projection of the surface.
The shaded portion indicates the orbital surface of the frontal lobe.

Fig. 13. The left frontal lobe, viewed laterally but with the dorsal surface
shown. Rather less of the orbital region is represented than in Fig. 12.

Fig. 14. Inner (mesial) surface of the hinder part of the left hemisphere,
in order to exhibit the condition of the parieto-occipital fissure (P. O.),
which is nearly divided into two parts (/at. p. 0.) passing on to the
upper surface, and (mes. p. 0.) at a.; the gyrus intercuneatus, however,
does not come quite up to the surface. 4. andc. Branches of the
mes. p. 0. in the preecuneus. In this and the following figure it will be
noted that the calcarine and parieto-occipital fissures meet.

Fig. 15. Inner (mesial) surface of the hinder part of the right hemi-
sphere. The gyrus intercuneatus (a.) is here superficial, and divides
completely the parieto-occipital fissure into the mesial (mes. p. 0.) and
lateral (at. p. 0.) portions.

Fic. 16.—Dorsal view of the left hemisphere of a human brain (Oxford
Museum, No. 950), reduced from a photograph, in order to illustrate the con-
dition of the parieto-occipital fissure. It is to be compared with the right side
of “Sally’s” brain. m.p.o. and J. p. o. are the two parts of the parieto-
occipital fissure. a. is the gyrus intercuneatus. The “ transverse occipital ”
is labelled “‘ Affenspalte ” for comparison.

Fie, 17.—The hinder part of the left hemisphere of the human brain figured
by Bischoff (tracing), for comparison with that of “Sally,” more especially
with regard to the condition of the ‘‘Affenspalte ” or transverse occipital.

Fic. 18.—The hinder part of the right hemisphere of the human brain
figured by Bischoff (tracing), for comparison with that of “ Sally.”

Fic. 19.—A tracing of the chimpanzee brain figured by Broca. a. a. a., 0.6. b.
The Affenspalte on right and left side respectively, divided into two. 2. 2.
The “arceus.” J. J. J. The occipitalis primus.

Fie. 20.—The right hemisphere of the chimpanzee’s brain presented by
Professor John Marshall to the Royal College of Surgeons in 1891. On the
left side an operculum is well and normally developed, and hides the Affen-
spalte ; but on the right side this fissure is exposed by the diminution of the
operculum, and in this resembles “Sally.”

Fic. 21.—The inner (mesial) surface of the hinder part of the right hemi-
sphere of the brain, No. 947 a., represented at fig. 30. The parieto-occipital
fissure is quite simple.

Fic. 22.—The left hemisphere of a chimpanzee’s brain (Oxford Museum,
No. 947 f.), to illustrate the condition of the parieto-occipital fissure and
neighbouring parts. The operculum conceals the Affenspalte and some portion
of the lateral parieto-occipital,

84 W. BLAXLAND BENHAM.

Fic. 23.—The hinder portion of the same brain (947 /.), with the operculum
turned back and the “ Affenspalte” exposed. The intraparietal fissure
(I. P.) enters the parieto-occipital (/a¢. p. o.), which is separated from the
Affenspalte by a large gyrus—a part of the arcus. The shaded parts indicate
those which are concealed by the operculum in Fig. 24.

Fie. 24.—The mesial surface of the hinder part of the same hemisphere
(947 7), to further illustrate the relations of the parieto-occipital fissure. In
this instance this fissure does not meet the calcarine.

Fic. 25.—A tracing of the chimpanzee’s brain figured and described by
Miller.

Fic. 26.—The hinder part of the right hemisphere of a chimpanzee (Oxford
Museum, 947 2.). The operculum is intermediate, as regards the size and
development, between the usual condition and that of “ Sally ;”? the “ Affen-
spalte ” is, however, concealed. The “lateral parieto-occipital” (/ad. p. o.)
is separated from the mesial parieto-occipital (mes. p. 0.) by the gyrus inter-
cuneatus (a.).

Fic. 27.—The mesial surface of the same brain (947 7.), to further illustrate
the condition of the parieto-occipital fissure. The shaded part is the under
surface of the occipital lobe. 4. c. y. are three branches of the mesial portion
of the parieto-occipital. a. The gyrus intercuneatus. This figure also shows
the distinction of the affenspalte from the parieto-occipital fissure. Note
that the calcarine and parieto-occipital fissures meet.

Fic. 28.—Projection view of the dorsal and lateral surfaces of the lett
frontal lobe of the chimpanzee exhibited in the Court (Oxford Museum,
No. 910).

Fie. 29.—The mesial surface of the hinder part of the right hemisphere of
a chimpanzee’s brain (exhibited in the Oxford Museum, No. 910), to illustrate
the arrangement of the parieto-occipital fissure, which does not join the
calcarine. The shaded portion indicates the lower surface of the occipital lobe.

Fries. 30—34 illustrate the brain of a chimpanzee (Oxford Museum, No.
947 a.). Natural size.

Fig. 30. Dorsal view. It illustrates two stages in the diminution of the
operculum. On the right side a condition very similar to that of
“Sally” is seen on turning back the small operculum, as in Fig. 32.
On the left the operculum is larger, and the condition of the fissures
somewhat peculiar.

Fig. 31. The hinder part of the left hemisphere, with operculum turned
back. The shaded parts indicate the portions ordinarily concealed by
that structure. The parieto-occipital, which is undivided, passes for a
considerable distance over the surface of the hemisphere. The Affen-
spalte is divided into two pieces, the outer of which receives the intra-

CEREBRAL CONVOLUTIONS—* SALLY.’’ 85

parietal fissure (p*.). It is possible that another interpretation might
be placed on these fissures.

Fig, 32. The hinder part of the right hemisphere, with the operculum
turned back. The shaded portions are those now exposed. Cf. Sally’s
brain.

Fig. 33. Projection view of lateral and dorsal surfaces of the frontal
lobe of the left side. The ‘frontalis secundus” is apparently
separated into two parts, the hinder (/?.) being joined by the
* diagonalis ” (d.).

Fig. 34. Projection view of the lateral and dorsal surfaces of the
right hemisphere.

Fig. 35.—Projection view of the dorsal and lateral surfaces of the right
frontal lobe of a chimpanzee (Oxford Museum, 947 7.). The dotted ring
indicates a slightly injured area. The precentralis inferior (p. c.7.), the
ramus horizontalis (4.), and the frontalis secundus (/”.), are continuous.

Fic. 36.—Dorsal view of the’ right hemisphere of Professor Herdman’s
chimpanzee. Natural size. It is quite normal. As the membranes were not
completely removed it was not possible to completely trace out all the fissures
in detail.

Fie. 37.—Right frontal lobe of the same brain (Professor Herdman’s).

Fie. 88.—Projection view of the dorsal and lateral surfaces of the left
frontal lobe of a chimpanzee (Oxford Museum, No. 947 /).

Fics. 39—46 illustrate the surface anatomy of the orang’s brain.

Figs. 89—41 refer to Rolleston’s orang, referred to in the text as II.

Figs. 39 and 40 are dorsal views of the right and left hemispheres.

Fig. 41. Projection view of the left frontal lobe. The right side of the
brain had been dissected, but the upper surface had been sliced off and
is preserved. The character of the Affenspalte and neighbourhood
may be contrasted with that in Figs. 42 and 43. The peculiar con-
dition of the “precentral fissure” is noteworthy. a@. Sulcus pre-
centralis marginalis.

Figs. 42—45 refer to the small orang in the Oxford Museum, referred
to in the text as I.

Fig. 42. Dorsal view of left hemisphere. Natural size.

Fig. 43. Dorsal view of right hemisphere. Natural size. The shape of
the frontal lobe is noteworthy.

Fig. 44. Projection view of the left frontal lobe.

Fig. 45. Projection view of the right frontal lobe. The dotted portion
is injured.

Fig. 46. A copy of Cunningham’s fig. 10, p. 294, for comparison of the
frontal convolutions with those in other orangs. I have retained
Cunningham’s lettering.

i, ge IA. Abe hela

‘ 7) a

hh ae: fcicey ite Oe snr bite fi! mew ih] th Cea, ae

i a ot ay WORT ah eee

au! prey! 1 i, 1p} La ridden tary i} ' ¢ 4) ] , dt BS wit “~i =
; o

een eee eG un se ee aw raya. po :

- ~ = 44
inline? wih lo poearore en fora Gandia : ‘ > zak 4
Grains zie! bi aS Mae Ste 4 } reli wv
ACP ft pon que ( \ T ; :
_ - 4
oer By Biowiens .th* a te Tle ae r

ni! 7 ’ ( : ( ; : ‘ ‘ .
vii} i a. ane } ' Tar ° ee 7,

j Livezie cn . hap ee te esr es |

ti eit | 4 vy oer Gee) Le ae x

hol wine ania) le Aiud Cus H et (item
? gah iicea Ail? imdpmowen neers 0 SLE ft ib aPTP 2s ¥ rites
‘ ‘ 7) or
. 7 : phe Ma ba TTS) aL) gent Met OF a Ph ea wi pais ae
j Hix 4) i apn Lhe F: nua mH 3 } $ niry bs { Ai a 5
5 . t A Go nA i edd PP cay A uM ’ Pot ay PE th
: ; MA Brojgctel bt See aT pila rigelapll ip
Dy (ea Oe Pe SW LL & itlase fs) LS er
aa sat pry ee Reh ear forse Det Ga
ae att ty alt riers uitas Width eA. A i Wierd Wot bs ay |
, bAe tue ia Oks Jie See dil ot jie?
>, ee re ar me ae) ea nie dey, ys! ry y
Vora ck ew OR See it: beste Te
MRT ROI > SV 4) ee GE a UeulAi oy Se y p
: Ol eae hina < -
poly able haan? MP ae A $2 ee 7, wt alent thts 7 a

| EN, i het ee
> o risa Loxaten ot eee 7 wept '
Poiiela at) “pdt Wrala dy chris AiG LP ane
. F ae » Aronian: xii

. ie ; x Ayn erin fit. waite ‘ se gat 8 mee

Wi fot ihe. ioe hetieme Aig ail Nites 7 pee le ‘ me
na) fo4 ; Pa

ye . ' may indo Ob! BRR ie AE ead byt Nate! : } ’ ie mat a

equa sail ue nt dag Ss ped {seid}
ee ie 2 ws “Willie, ; Weer

ON THE CELLULAR THEORY OF DEVELOPMENT. 87

On the Inadequacy of the Cellular Theory of
Development, and on the Early Development
of Nerves, particularly of the Third Nerve
and of the Sympathetic in Elasmobranchii.

By
Adam Sedgwick, F.R.S.

Ir is now more than ten years ago since I first pointed out
the inadequacy of the cellular theory of development. That I
did so in a very guarded manner need hardly be said; but
now, after ten years of mature work, I feel justified in giving
a stronger expression to the views which I then formed, and
which all my subsequent work has amply confirmed. My
words then (in 1883) were as follows :—‘‘In short, if these
facts are generally applicable, embryonic development can no
longer be looked upon as being essentially the formation by
fission of a number of units from a single primitive unit, and
the co-ordination and modification of these units into a
harmonious whole. But it must rather be regarded as a
multiplication of nuclei and a specialisation of tracts and
vacuoles in a continuous mass of vacuolated protoplasm.”
Again, in 1888, in the preface to my “ Monograph on the
Development of the Cape Species of Peripatus,”? I wrote:
“It would appear, indeed, that in Peripatus the cells of the
adult, in so far as they are distinct and sharply marked off
structures, are not, as appears to be generally the case, present
in the earliest embryonic stages, but are gradually evolved as
development proceeds. In other words, the cell-theory, if it

1 *Studies from the Morphological Laboratory of the University of
Cambridge,’ vol. iv, part 1.

88 ADAM SEDGWIOK.

implies that the adult cells are derived from embryonic cells
which have been directly produced by the division of the
ovicell, does not apply to the embryos of Peripatus.”

In the days when these words were written it was a general
belief among leading histologists and physiologists that the
connections which were known to exist in some cases between
adult cells had arisen secondarily, and that the primary con-
dition brought about by the cleavage of the ovum was a com-
plete separation from one another of these units, of which
the body was supposed to be composed. There has been, no
doubt, a change of opinion since those days, and although
many biologists would still maintain that cleavage is complete
and results in the formation of separate units which later
become connected, there is a constantly increasing number
who would consider themselves misrepresented if one imputed
to them this belief not long ago universal, and the belief which
was supposed to follow from it, that the first stage in the
evolution of the Metazoa was a colonial Protozoon. But, as
I have said, opinions have changed since those days, and I
quote my words, written then, to show that I have long held
the view which I am now expressing, and that I was among
the first to attack a theory which had even then passed its
stage of usefulness, and is now holding men’s minds in an iron
bondage. For although opinions have changed on this im-
portant subject, and although there are some who think that
they have escaped from the domination of this fetish of their
predecessors, yet as a matter of fact the cellular theory of
development is still rampant, still blinds men’s eyes to the
most patent facts, and still obstructs the way of real progress
in the knowledge of structure.

In order that I may not be met with the statement that
such a state of things exists only in my own imagination, that
I am putting up a dummy merely to knock it down again, it
is necessary that I should give some proof that this hypothesis
has still the power which I ascribe to it. What is the cellular
theory of development? I am not concerned with what its
authors held; what we want to know is, what is the present

ON THE CELLULAR THEORY OF DEVELOPMENT. 89

form and extent of it? What is the point of view which it
compels its votaries to take ?

It is not easy to answer this question ; it is, in fact, as diffi-
cult to answer as that other question so often asked of the
teacher by his pupil—what is a cell? The source of the dif_i-
culty is that we are dealing with a kind of phantom which
takes different forms in different men’s eyes. There is a want
of precision about the cell-phantom, as there is also about the
layer-phantom, which makes it very difficult to lay either of
them. Neither of these theories can be stated in so many
words in a manner satisfactory to every one. The result is
that it is not easy to bring either of them to book.

To answer the question—what is the cellular theory of deve-
lopment ?—the best plan will be to consider for a moment the
ideas which are taught to the student of biology, and which
influence him in his future work. We tell him that the cell is
the unit of structure, that an organism may consist of a single
cell, or of several cells in association with one another: we
draw the most fundamental distinction between the two kinds
of organism, and we divide the animal kingdom into two great
groups to receive them. Asa proof of the importance which
we attach to this feature of organisation we assert that a man
is nearer, morphologically, to a tapeworm, than a tapeworm is
to a parameecium. We tell him that the various structures
present in a protozoon are all parts of one cell, whereas in a
metazoon the various parts are composed of groups of cells
which differ from one another in structure. Finally, when we
ask him in the examination to tell us the principal differences
between hydra and vorticella, we consider that he is very
inadequately prepared if he does not sum them up by saying
that hydra has tissues composed of definite cells and is multi-
cellular, while vorticella is without definite cellular tissues
and is unicellular. Carrying on the idea thus implanted in
his mind as to the fundamental importance of the cell, we tell
him about the neuro-epithelial cell and the myo-epithelial
cell, and we point out their primitive distinctness,—an idea
which is still further impressed upon him when he studies the

90 ADAM SEDGWIOK.

connection between nerve and striated muscular fibre. Finally,
when he comes to study embryology, the importance and dis-
tinctness of the cell meets him at every step, from the complete
cleavage which he is led to believe is primitive, to the develop-
ment of nerves according to the views of His.

So much for the student in the schools: now for the investi-
gator in the laboratory. He studies the ovum and maintains
its absolute isolation in the organism ; or he examines epithelial
cells and draws them as isolated structures separated by sharp
boundary lines ; or he lahours to prove the continuity between
the nerve and muscle, or between the nerve and secreting cell:
so much is he dominated by the idea of separate cells that he
considers that the burden of proof rests rather with the man
who asserts such continuity than with him who denies it.
Or, if he be an embryologist, he will talk of, and figure, the
proliferation of cells at the primitive streak ; he will describe
the nascent ganglion cell sending a process from the developing
spinal cord into the anterior root, and he will figure it; he
will talk of mesenchyme cells, and figure them for the most
part separate from one another.

I take it that this is a not unfair account of the training a
zoologist receives at the present day, so far as the cell is con-
cerned, and of the ideas which dominate him in his later work.
He believes that the cell is the unit of structure, and that it
forms the basis of organisation in the Metazoa; it is the func-
tions of the cell and the relations which it enters into with
other cells which forms an important subject of current biolo-
gical investigation. Who, then, can deny that the cellular
theory of development is still a living power in the school of
biology ? That it blinds men’s eyes to the most patent facts,
and obstructs the way of real progress in the knowledge of
structure, it will now be my endeavour toshow. For this pur-
pose I shall deal on this occasion with the origin and structure
of two tissues of the Vertebrate embryo—the so-called mesen-
chyme and the system of peripheral nerve-trunks. My results are
the product of many years’ work, and will, I hope, be published
in greater detail and with figures on a future occasion.

ON THE CELLULAR THEORY OF DEVELOPMENT. 91

The So-called Mesenchyme Tissue of Elasmobranch
Embryos.

This tissue is always described as consisting of branched
cells lying between the ectoderm and the endoderm. The
cells are spoken of as being separate from one another, and
from the adjacent ectoderm and endoderm, excepting at points
where they are supposed to arise from one of the primary
layers. And not only are they described as being separate
cells, but they are actually drawn in the author’s figures as
separate from each other. This is, perhaps, the best instance
that can be given of the bondage in which the cellular theory
holds its votaries. For what are the facts? The separate
cells have no existence at all! In their place we find, on
looking into the matter, a reticulum of a pale non-staining
substance holding nuclei at its nodes. It is these nodes, with
their nuclei, which are drawn by authors as the separate
branched cells of the mesenchyme, and they are constrained
by this theory, with which their minds are saturated, not
only to see things which do not exist, but actually to figure
them. Another erroneous view due to the same cause is the
view that this mesenchyme tissue is not continuous with the
ectoderm or with the endoderm ; whereas, as a matter of fact,
the opposite is the case, for the primary layers are simply parts
of this reticulum in which the meshes are closer and the nuclei
more numerous and arranged in layers. These are facts of
which anyone with an unbiassed mind can convince himself by
the simple inspection of a Selachian or an Avian embryo, and
they would have been recognised long ago had it not been for
the dominating influence of the cellular theory of development.

The current views as to the origin of this tissue show just
as conspicuously the influence of the same theory. It is said
to arise by the budding-off and migration of cells from the
walls of the embryonic cclom, from the primitive streak, and
from the neural crest ; and the space between the ectoderm and
endoderm into which these cells migrate is described as being
empty of structural elements. What are the facts? The

92 ADAM SEDGWICK.

space between the layers is never empty; it is always
traversed by strands of a pale tissue connecting the various
layers, and the growth which does take place at the places
mentioned is not a formation of cells, but of nuclei
which move away from their place of origin and
take up their position in this pale and at first
sparse reticulum which exists between the layer. As this
reticulum, which has always existed, becomes infested with
nuclei it increases in bulk, and forms the conspicuous reticulate
tissue which is by some authors called mesenchyme. The
primitive streak, the walls of the ccelom, and the neural crest,
and, as Goronowitsch! has shown, parts of the ectoderm, are
growing points where nuclei, not cells, are produced. These
facts I described long ago in the development of Peripatus,
and it is the recognition of the same processes taking place in
the Vertebrata in an even more conspicuous manner that has
induced me to again call attention to their importance.”

The Origin of Nerve-trunks and the Fate of the
Neural Crest.

If there is one point more than another on which the cellular
theory of development has led anatomists completely astray, it
is uponthis one. We may take it that the new views upon the
origin of the peripheral nerves began with Balfour’s discovery
of the structure which is generally called the nerve crest.
Before that discovery nerves were supposed to develop in situ
in the mesoderm; after it, there were two principal views as
to the origin and growth of nerves: one of these was that cells
of the central organ grew outwards as strings to the periphery ;
while, according to the other, nerve-fibres are the elongated

1 ¢Morpholog. Jahrb.,’ Bd. xx, 1893.

2 At the same time Peripatus shows certain features more clearly than
the Vertebrate; I would refer especially to figs. 24d and 26d on pl. v of
my Monograph, in which, while the so-called ectoderm and endoderm are
obviously parts of the same layer, or tissue : they are separated by a region in
which the vacuoles are larger, the protoplasmic strands less numerous, and
nuclei are conspicuous by their scarcity.

ON THE CELLULAR THEORY OF DEVELOPMENT. 93

processes of cells either of the central organ or of the ganglia.
Both these views are erroneous; and if both were not inspired
by the cellular theory of development, they were both promul-
gated at a time when that theory was at its zenith. The earlier
view, that nerves were developed in situ from the mesoderm,
was much nearer the truth.

The nerve crest does not, as was first stated by Balfour and
afterwards by all authors on the development of nerves, give
rise exclusively, or even principally, to nerves and ganglia. It
gives rise to nuclei which spread out in, and add to the meso-
blastic reticulum, which at all times, i.e. from the very begin-
ning, exists between the layers, and to nuclei which become
the nuclei of the rudiments of nerve ganglia. The nerves are
developments of the reticulum; they are elongated strands of
the pale substance composing the reticulum, with some of its
nuclei; and their free ends branch out into the fibres of the
reticulum, and are added to by the latter falling into the line
of the growing nerve. Neither they nor the ganglia appear
until the nerve crest is breaking up. The reticulum further
gives rise certainly to smooth muscular fibres, connective
tissues, and blood-vessels, and probably also to striated
muscle. It is also continuous with all the so-called epithelial
tissues of the embryo; indeed this latter substance is to
be regarded as consisting only of one or more layers
of nuclei embedded in the outer part of the reticulum,
which is rather denser than elsewhere in correspond-
ence with the greater density of the nuclei. Nerves
are a gathering up, so to speak, of the strands of the reticulum
into bundles, and are formed in that way; or, to put the matter
in another way, nerves are a special development of the reti-
culum along certain lines. These special developments are
generally marked by an increase in the number of nuclei, such
increase being particularly great in the neighbourhood of the
ganglia.

To sum up the matter, the nervous and muscular tissues are,
as they were in Peripatus (see my Monograph, p. 131),
special developments of the same primitive reticulum, a com-

94, ADAM SEDGWICK.

munity of origin which renders their adult relations perfectly
intelligible. Further, I have no hesitation in saying that His’
descriptions of the development of nerve-fibres as processes of
central or ganglionic nerve-cells, does not apply to Selachians;
inasmuch as nerves are laid down long before any
trace of nerve-cells can be made out. The neuroblasts
of His and of other authors are nuclei lying in a substance
which, after death caused by the ordinary reagents, has usually
a fibrous structure. This substance is continuous with, and
therefore a part of, the reticulum outside. The cell-processes
which have been described as growing out from the neuroblasts
are merely parts of this reticular substance, the fibres of which
become arranged more or less.in the direction of the long axis
of the nuclei, and the meshes correspondingly drawn out and
narrowed. Many of His’ drawings even show that this is so,
and an inspection of the specimens leaves no doubt at all about
the matter. Inshort, the development of nerves is not
an outgrowth of cell-processes from certain central
cells, but is a differentiation of a substance which
was already in position; and this differentiation
seems to take place from the medullary walls out-
wards to the periphery, both in the anterior and
posterior roots, and to precede, or to proceed pari
passu with, the development of other tissues. The
nerve crest is, then, to be regarded as a centre for the growth
of nuclei, which spread into the body of the embryo and be-
come concerned in the formation of many tissues, nervous
tissues amongst the rest. There are many other such centres
for the production of nuclei; for instance, I may mention the
walls of the ccelom, the caudal swellings, and in the Amniota
the primitive streak. All these centres of growth are in
so-called epithelial tissues. This is, of course, necessi-
tated by the fact that Selachian embryos are at one stage
composed entirely—or almost entirely—of these so-
called epithelial tissues; as are many embryos, e. g. those
of Peripatus and of Amphioxus.! These facts will be dis-

1 The significance of this epithelial structure of the young embryo—this

ON THE CELLULAR THEORY OF DEVELOPMENT. 95

puted by many morphologists, but they are easy of proof by
the simple inspection of good preparations to minds not warped
by the cellular theory as ordinarily taught. In fact, had it not
been for the undue persistence of this hypothesis beyond the
time of its fruitful life, they would have been recognised long
ago, and much needless waste of labour in trying to make the
facts of nerve-development conform to the theory would have
been saved.

The nerve-crest in Selachians (Scyllium, Acanthias,
Raia, and Pristiurus) is, as I pointed out some time ago
(* Notes on Elasmobranch Development,” ‘ Quart. Journ. of
Micr. Sci., vol. xxxiii), from its first appearance, in three
pieces.! The first of these pieces reaches from the region of
the fore-brain to the hind brain. The posterior limit of it is
marked in older embryos by the root of the trigeminal nerve.
It gives rise to the reticulum of the front part of the head,
and contributes to that of the mandibular arch. The
following nerves are formed within its limits :—The trigeminal
and its branches, which include the so-called ramus ophthal-
micus profundus with the ciliary ganglion and the third
nerve (see below). Very possibly other nerves, viz. the
fourth, the sixth, and the olfactory, may be also developed
from this part of the reticulum, but I have no observations on
this point.

The manner in which these nerves are laid down may be
described as follows:—When the nerve-crest, which in this
region of the head very early spreads ventralwards on each
side of the brain, is breaking up into the reticulum, certain
tracts of it remain unaltered and characterised by a greater
density of nuclei. These tracts mark the course of the future
nerves and the sites of the future ganglia. They them-

collection of the nuclei at the surfaces, as it may be described—I hope to con-
sider in another place. Now, I may merely hint that it is probably due to the
impress of some well-marked larval phase in earlier stages of evolution (see my
article on “ von Baer’s Law, &c.,” in ‘ Quart. Journ. Mier. Sci.,’ vol. xxxvi).

1 Goronowitsch (‘ Morph. Jahrb.,’ Bd. xx) has recently found the same fact
for the bird, but he makes no reference to my results on this point.

96 ADAM SEDGWIOK.

selves continue to break up, but a kind of core remains
which constitutes the foundation of the future nerve and
ganglion. The Gasserian ganglion, the ophthalmicus pro-
fundus, the mandibular branch of the fifth, and the ciliary
ganglion thus gradually emerge from the remains of the nerve-
crest—are, so to speak, crystallised out of it. At first they
have the form of dense cords of nuclei; but they soon acquire
some of the non-staining fibrous substance, which makes its
appearance as arule in their central portions, so that fora
time sections of these nerves exactly resemble in appearance
sections of the nerves of Invertebrata, e. g. Peripatus,
Chiton, &c. This description holds for an embryo of 35 mm.,
beyond which stage I have no observations. The nuclei which
have peeled off, leaving the nerve-trunk below, give rise to the
muscular and connective tissues of the parts concerned, the
reticulum of which is freely continuous with that of the nascent
nerve, especially at the free end of the latter. It thus be-
comes apparent that these tissues—nervous, muscular,
connective, and vascular—are all developed in continuity.
While the Gasserian ganglion, the mandibular branch of the
fifth, the ophthalmicus profundus, and the ciliary ganglion all
crystallise out of the nerve crest; the third nerve does not do
so. It arises as a differentiation of the reticulum formed by
the breaking up of the nerve crest, and it first makes its
appearance as a forward projection of nuclei from the ciliary
ganglion. This, by a gradual differentiation of the reticulum,
extends itself until it reaches the base of the mid-brain, with
which it becomes continuous by means of an increase in the
pale fibrous strands which pass between the medullary wall
and the reticulum. The third nerve is at first a cord of nuclei
and rather dense pale substance. The third nerve,’ there-

1 The continuity of the embryonic tissue which will give rise to the
nervous and muscular tissues is well seen in the embryo of Peripatus
capensis, and I have already hinted at this fact in my Monograph on the
development of that species at pp. 131 and 138, and figured the tissue as
nerve musc., pl. x, fig. 5.

2 Tt will be evident, if my observations are correct, that I have found an
earlier stage of the third nerve than Dohrn describes in his sixteenth study. In

ON THE CELLULAR THEORY OF DEVELOPMENT, 97

fore, presents this interesting and remarkable pecu-
liarity in Scyllium and Acanthias; it grows or is
differentiated from the ciliary ganglion to the floor
of the mid-brain, and not in the opposite direction, as has
hitherto been supposed. The proof of this is to be found in the
fact that in a Scyllium and Acanthias embryo of 10 to
11 mm. the third nerve can be seen projecting forwards from
the ciliary ganglion, and ending in front in the reticulum,
short of the floor of the mid-brain. The ciliary or profundus
ganglion is at one time—when it is first laid down—in contact
with the ectoderm. Later it is shifted inwards, but remains
connected for a time with the ectoderm by a cord of cells,
which eventually disappears. This point has been seen by
van Wijhe.

The embryonic medullary wall! is connected with the
reticulum by pale fibres similar to those which compose
the reticulum, and the nerve-roots, both anterior,
posterior, and cranial, are special enlargements of
such connecting strands. They are formed at a time
when no structures which could be called cells by any but a
fanatical devotee of the cellular theory are present, either in the
medullary wall or in the ganglionic rudiments, and in a manner
which, if closely followed, renders it quite impossible to speak
of growths one way or the other, excepting that one can make
one assertion—the pale fibrous substance which marks the
nerve appears both in the anterior and posterior roots and in
the cranial nerve-roots next the central organ, at a time when
the white matter (which is composed of this pale fibrous sub-
stance) first appears as a thin layer, and in continuity with
such white matter. The differentiation outwards pro-
ceeds from this point, and the free end of the nerve-
rudiment always ends by branching out into the fibres

my fuller paper dealing with this subject I hope to examine Dohrn’s results in
detail.

1 Tnasmuch as the nerve-crest is derived from the medullary wall and gives
rise to mesodermal structures, the medullary wall itself gives rise, in part, to
mesoderm.

VOL. 37, PART 1.—NEW SER. a

98 ADAM SEDGWICK.

of the reticulum. The only exception to this rule is the
third nerve of Scyllium and Acanthias (and probably
others), which is undoubtedly differentiated from the ciliary
ganglion to the floor of the mid-brain; but this is, perhaps,
more an apparent exception than a real one, because the
ciliary ganglion belongs to the fifth nerve and the order of
fibrous differentiation is normal, viz. from the root of the fifth
nerve, through the ciliary ganglion, to the floor of the mid-
brain. I commend this observation on the development of the
third nerve to the physiologist, with a view to a renewed
investigation of its functions. It is rendered the more inter-
esting by the fact that in Lepidosiren it is commonly stated
that the area of the third nerve is supplied by the ophthalmic
branch of the fifth, the third nerve being absent.1

I have already, in my ‘ Notes on Elasmobranch Develop-
ment,’ stated my reasons for believing that the views put
forward by Hensen as to the origin of nerves were nearer the
truth than those of any other zoologist. I have, in this paper,
shown not only that the network required does exist, but also
how it arises, and how it gives rise to the rudiments of the
peripheral nerve-fibres. Minot, in his ‘ Human Embryology,’
p. 624, says that Hensen’s theory of the origin of nerves
“cannot be adopted because the outgrowths of the nerve-fibres
have been observed; moreover, Altmann has pointed out that
the fibres seen in the embryonic mesoderm are really pro-
cesses of the mesoderm cells, and, as shown in the excellent
fig. 2 of his plate, are quite distinct both from the ecto-
derm and endoderm.” (Theitalics are mine.) This passage
is, according to my work, full of errors; for I maintain, as
the result of long and careful observation, extending over
many years, that the outgrowth of nerve-fibres from cells in
the ganglia and medullary wall not only has not, but cannot

1 This statement rests on Hyrtl’s work. It must, however, be remembered
that his specimen was confessedly rotten in its nervous tissues, and by the
fact that v. Wijhe (‘ Nied. Arch. f. Zoologie,’ Bd. v, 1882) has found the
third nerve in Ceratodus. Parker does not deal with the brain and nerves
in his memoir on Protopterus.

ON THE CELLULAR THEORY OF DEVELOPMENT, 99

be observed ; that the fibres in the embryonic mesoderm are
not processes of mesoderm cells (as they are always figured),
which have no existence, but are parts of the reticulum which
has always existed from before cleavage onwards, connecting
together the various parts of the developing ovum, and that
this reticulum is not separate from ectoderm and endoderm,
but freely continuous with both, they being but parts of it.
The almost universal practice of drawing this reticulum as
composed of separate branched cells is a most remarkable
instance of the manner in which a theory can blind men’s eyes
to the most obvious facts.

Before concluding this general account of my work, I may
mention one or two other points of general interest which I
have noticed. Firstly, I may mention that in Scyllium there
are a number of anterior roots next the head, varying in
number from three to five, according to the age of the embryo,
without posterior roots. They no doubt give rise, as has beeu
suggested by others, to the so-called anterior roots of the
vagus. Secondly, Balfour was quite correct in the account he
gave of the origin of the sympathetic ganglia in Elasmobranchs.!
The ganglia arise as swellings on the posterior roots of the
spinal nerves, and soon become removed from the latter, so as
to form isolated masses connected with the spinal nerves by a
cord. These masses eventually become united longitudinally
into a chain. I may add to Balfour’s account this fact, viz.
that no sympathetic ganglia are found within the area of
extension of the vagus ganglion. Or, if I am not correct in
applying the term “ vagus ganglion” to the posterior part of
the vagus—the part which lies dorsal to the gill-slits and gives
off the branchial nerves, it would be better to say that sym-
pathetic ganglia are not found in the region of the branchial
slits, but begin immediately behind these structures. Thus,
in an embryo of 22 mm. the vagus ganglion and branchial

1 T have not examined mammals on this point, but I think Paterson’s
memoir (‘ Phil. Trans.,’? 181) does not carry conviction. On the contrary,
there is, 1 think, in it internal evidence which inclines me to the view that he
has not got to the bottom of the matter.

100 ADAM SEDGWICK.

region of the fore-gut ends at the level of the fifth anterior
root, and the first posterior root and the first sympathetic
ganglion occur at the level of the sixth anterior root. In
older embryos, in which the branchial region extends much
further back and overlaps a number of fully-formed spinal
nerves, the original sympathetic ganglia which were formed in
connection with the ganglia of these spinal nerves thus over-
lapped are found to have disappeared. The first sympa-
thetic ganglion appears always to be just behind
the branchial region, as in the adult, and sympa-
thetic ganglia are formed in Scyllium in connec-
tion with nerves which are without a posterior root.

Gaskell reproaches v. Wijhe with not knowing the true
meaning of a sympathetic ganglion, and one is tempted to ask,
does Gaskell himself know much more about it, or throw any
light upon the question? He says (‘Journal of Physiology,
vol. x, p. 162) that a sympathetic ganglion is the ganglion of
the anterior root of a spinal nerve which has travelled to a
variable distance from the central nervous system. As Dohrn
(seventeenth study, ‘Naples Mit.,’ Bd. x) very properly in-
sists, this view is at variance with the known developmental
history of the ganglion—which I am able to confirm so far as
its nuclei are concerned, and with the reservations necessitated
by the views set forth in this paper—and I am now able to
state that it is at variance with the fact that sympathetic
ganglia are entirely absent from those spinal nerves in which
the posterior root fails to reach its full development. In fact,
one may say of these ganglia that they are always absent when
the posterior roots are not developed.

With regard to the fate of the neural crest described in this
paper, I should mention that I strongly hinted that it gave rise
to nuclei which entered the reticulum in my ‘ Notes on Elas-
mobranch Development,’ p. 581, published in 1892; and that
Goronowitsch arrived independently at the conclusion that it
broke up into mesenchyme in the bird, and published his
results at some length in 1893 (‘ Morph. Jahrbuch,’ Bd. xx) ;
but Goronowitsch failed to recognise the reticulum, and he was

ON THE CELLULAR THEORY OF DEVELOPMENT. 101

unable to appreciate the full significance of the facts he de-
scribed in their bearing on the question of the origin of nerves.
Platt approximated to the truth with regard to the third nerve
in her account of it as growing from the ciliary ganglion to the
brain, but retained the error of her predecessors in regarding
it as a cellular object, and not as a differentiation of the reti-
culum.

Minot has a characteristic comment on Platt’s statement.
He says (‘Human Embryology,’ p. 639), “This view rests
probably on erroneous interpretation of observation, for it
cannot be admitted that a motor nerve is formed by
ganglionic fibres”! (The italics are mine, as is also the
note of admiration.)

=

) Sera) Wa) Cb till ae

el ae eo ee a
rise 2. Ge SRG ankle As
or 6 Vor Wee BRR
Bynes (ie Re |) *,
jixhaw his eerie PP. el
bea BR Te HAE od

sa rip Jy a ey

-

€ ' a a
; ‘ 4 j et
i
J i t
ives a >
’
0
‘
a
i
€ v i
+
=
*
_
Fy A
-
> > a
a ®

ON BENHAMIA C@OCIFERA, N. SP. 108

On Benhamia cecifera, n. sp., from the
Gold Coast.

By

W. Blaxland Benham, D.S¢c.Lond., Hon. M.A.Oxon.,
Aldrichian Demonstrator in Comparative Anatomy in the University
of Oxford ; Lecturer on Biology in Bedford College, London.

With Plate 12.

Some time this year I received a large specimen of Ben-
hamia from Professor Jeffrey Bell for examination and identi-
fication. My best thanks are due to the authorities of the
British Museum for their permission to make the examination.

The worm, which turns out to be a new species, to which I
give the name Benhamia cecifera, was collected by Captain
Torry at Axim, in the Fantee Country, on the West Coast of
Africa.

We already know a number of species from this side of
Africa, as well as from the East Coast and inland, and we know
pretty certainly that this continent 1s the home of the genus.

This new species is of considerable size, measuring 510 mm.
(20 inches), with a diameter of 17 mm. at the clitellum, and
gradually diminishing posteriorly. Its average diameter is
about 12 mm., and it tapers only very slightly to within a few
segments of the hinder end, where it rounds off. There are
310 segments.

The colour of the worm is dirty brown (in spirit), and the
hinder end is not very sensibly lighter than the rest of the
body.

In each segment there is a circle of small dark brown pig-

104 W. BLAXLAND BENHAM.

ment spots at the level of the chete, giving the impression
at a casual glance that the chetz are in a circle round the
body. A similar style of pigmentation has been noted by
Michaelsen in the case of B. affinis and others.

The dorsal pores are large and conspicuous, commencing
after the 4th segment ; as usual these are absent (or obliterated)
in the clitellar region. When the animal was handled a con-
siderable quantity of a brown fluid issued from the pores.

The chxtz, which appear as black dots, have the usual
arrangement, and present no pecularity of shape or ornamen-
tation. They are invisible, if indeed they are present, in the
2nd segment. As I had only one specimen, which had to be
returned to the British Museum, I could not, of course, examine
the body-wall microscopically. The four couples are equi-
distant from one another, being about 2 mm. apart. Some
of the anterior segments, except the first three, are biannu-
lated, the first annulus being nearly twice as large as the
second, and bearing the chete.

The prostomium is not dovetailed into the peristomium ;
the latter is not more than half the size of the 2nd segment.

The clitellum covers eleven segments, viz. x111 to XXIII in-
clusive ; the intersegmental groove is still distinct detween the
13th and 14th segments, but elsewhere, at any rate dorsally
and laterally, the limits of the segments composing the cli-
tellum are not recognisable, except by bands of darker pig-
ment. As in all species of this genus, there is a ventral area
in the middle of the clitellar region which in this instance
presents specific characters. In Segments x11, xiv, and xv the
clitellum is ‘‘ complete,” but in the remaining segments it ceases
ventro-laterally, 1. e. just below the outer series of chetz, and
forms a distinct margin (fig. 1, m.) limiting a more or less rec-
tangular depression, the “ ventral field.”’ This is wider ante-
riorly than posteriorly, being about 12 mm. across on Segment
xvit and about 8 mm.on Segment xx11. Anteriorly the ventral
field is bounded by the hinder margin of Segment xv.

Within this “ventral field” lie the four prostate pores, a
pair on each of the Segments xvii and xix, The two pores of

ON BENHAMIA CM@CIFERA, N. SP. 105

each side are connected by a longitudinal groove, as in other
species of the genus. But what marks this species as distinct
from all others hitherto examined is the presence and arrange-
ment of numerous small pits, no doubt having some function
in relation to copulation. The figures give a better idea of the
arrangement of the pits than any detailed description. It will
be seen that for the most part they form transverse rows on
Segments xv, XVI, XVII, XIX, Xx, xx1,xx11. There is a single
median pit on Segment xxii, and on the 18th segment
three short longitudinal rows occur, one median and a pair
of lateral rows.

“ Copulatory organs ” (“ pubertats Tuberkeln”’) have been
noted by Michaelsen in various species, especially B. affinis
and B.inermis. In the latter they are paired, and occur on
segments in front of the clitellum, but they appear from his
figures and descriptions rather as tubercles than as pits, and
have altogether a different arrangement.

These pits in the present worm mostly have a well-defined
margin, not raised above the general level of the surface. In
others there appears to be a papilla on one side of the pit, pro-
jecting to a greater or less extent into the cavity of the pit, as
shown in fig. 2, B. C. But in no case do the papille project
to the exterior, nor are they visible except on careful inspection.

Two pairs of similar pits occur in relation to the sperma-
thecal pores, viz. on Segments vir and vi (fig. 1, p’. p’.). The
anterior pair lies on the hinder margin of the segment, the pos-
terior pair on the anterior margin of the second annulus—this
one is more laterally placed than the former. In the opened
worm there were no sacs or other structures projecting into the
cavity of the worm’s body ; the pits appear to be limited to the
body wall.

The spermathecal pores have the usual position on the
boundaries of Segments vii-vi11 and vii1-1x, in a line with the
inner series of chetz.

The oviducal pore is small, but it is visible just in front of
the ventral (inner) chetz of Segment xiv. On Segment x11
a pair of circular whitish areas, each with a central depression,

106 W. BLAXLAND BENHAM.

is situated just behind and slightly laterad of the inner chete.
These probably have some copulatory significance. They are,
however, different from the “‘ pits” above mentioned. Internally
in this segment is a low but rather extensive muscular promi-
nence occupying the whole length of the segment, which, I
presume, is the wall of some structure opening by the above-
mentioned pore.

Turning now to the internal anatomy, there are the
following points to be noted. As in other species of the genus,
certain of the anterior septa are very much thickened, viz. the
septa 9/10, 10/11, 11/12, 12/13, and 13/14; the next two septa
are also thick, but less than the foregoing five. In front of
the first of these thickened septa, the viscera are wrapped
together by a delicate membrane, due no doubt to the thin
septa of these segments being pushed backwards by the large
gizzards ; owing to the tenuity of the septa here, and to the fact
that they are ruptured in merely turning the organs aside, it is
difficult in a dissection to trace them out. The first septum
in the body is distinct, and separates Segments 1v and v; con-
sequently septa 5/6, 6/7, 7/8, and 8/9 are represented e the
enveloping membrane just referred to.

The Alimentary Tract.—Amongst the internal organs
the most noticeable feature of novelty is presented by the
intestine. The two gizzards, characteristic of the genus, are
of large size, and appear to belong to the Segments v and v1.
The cesophagus is narrow, and provided with the usual three
pairs of “ calciferous glands” lying in Segments xv, xvi, and
xvi1. Whether carbonate of lime is present, as has been deter-
mined for other species, I am not in a position to affirm, as I
did not think it worth while to explore the glands. As is
frequently the case, the first gland on either side is the
smallest. In the genus Benhamia the wall of these glands
presents a variety of patterns, owing to its folding ; these have
not been figured carefully, but have been described by
Michaelsen and by Horst as horizontal folds in some cases,
vertical folds in others. But in the present species the foldings
are very claborate, so that the surface of the gland, as well as

ON BENHAMIA C@CIFERA, N. SP. 107

its shape, suggest a much convoluted cerebral hemisphere (see
fig. 4).

Each gland is, on the whole, reniform, and connected to the
narrow cesophagus by a short, narrow, but distinct duct. The
glands are supplied by a large vessel from the dorsal trunk,
and are further closely connected to the septa.

In the 20th segment the gut enlarges suddenly and
forms a wide, thin-walled, dark brown coloured “ sacculated
intestine,” constricted of course by the septa. In Segment
xxix, however, the intestine takes on the peculiar character
(unknown in its details in any other earthworm) which is
referred to in the specific name “ cecifera.” In each segment,
commencing at the 29th, the intestine gives rise to a finger-
shaped diverticulum on each side, which is directed upwards
and arches over the dorsal blood trunk (fig. 3, c.).

These ceeca extend as far as the 52nd segment, the last half-
dozen gradually diminishing in size, till the intestine resumes
the ordinary sacculated condition. Each such cecum arises
by a comparatively broad base from the upper surface of the
side of the main gut, and gradually narrows as it curves
upwards, to end in a blunt rounded apex. The apices of the
pair of czeca of any segment overlap each other above the dorsal
vessel. These ceca are thin-walled, and quite unlike the well-
known ceca which lie in the 26th segment of the ordinary
Pericheta.

The generative organs conform to the usual Benhamia
type.

There are two pairs of sperm-sacs, in segments x1 and x11,
each being much divided and having a lobulated appearance.

The two pairs of prostates are very long, narrow, yellowish
tubes of the usual acanthodriloid pattern; the muscular duct is
very delicate at its origin from the gland, enlarging in its
slightly winding course, to pierce the body wall: they lie in
segments xvii and x1x (fig. 5).

I could detect no penial chet; these special chete have
been noted as lacking in other species of Benhamia. It is a
point on which I laid some stress in my note on “ The Genera

108 W. BLAXLAND BENHAM.

Trigaster and Benhamia,”’! where I pointed out the various
differences between the two genera. Some of these differences
have been bridged over by various species of Benhamia,
though in my opinion the genus Trigaster still stands as a
valid genus.

Though I could find no penial chetz on careful examination
of the partially dissected individual, I should not like to affirm
that they do not exist. There are several strong bands of
muscle in the Segments xvii to xxi (fig. 5), arising close to
the nerve-cord and passing outwards to be attached to the
body-wall near the dorsal midline; they have a similar ar-
rangement to those figured by me for Moniligaster (indicus)
robusta, A. G. B.,? and it is possible that special cheetz may lie
below these muscles—but I did not wish to do too much injury
to the specimen.

The spermathece, which lie in Segments vir and viii,
differ in size, as has already been noted in several species.
The anterior pair are smaller than the posterior pair, and
whereas the latter is provided with a small ovoid diverticulum,
those of the 7th segment are without one (fig. 6). Each
spermatheca consists of a thin-walled, somewhat ovoid sac,
with a longish and comparatively stout, muscular, glistening
duct, which widens as it approaches the body-wall.

I found no structures in these segments corresponding with
the pits noticed in the account of the external anatomy.

The vascular system calls for a few remarks, but I have
not attempted to make a complete study of this system ; never-
theless, from what I have observed and from what previous
observers have noted, I believe the vascular system of Ben-
hamia and Acanthodrilus would repay a careful study on
the lines of and in relation to Dr. Bourne’s thorough account
of the vascular system in Megascolex® and Moniligaster.*

1 «Ann. Mag. Nat. Hist.,’ 1890, p. 414.

? Quart. Journ. Mier. Sci.,’ vol. xxxiv, pl. 32, fig. 3.

$ «On Megascolex ceruleus, &.,” ‘Quart. Journ. Mier. Sci.,’ vol.
Xxxil, p. 49.

4 “Moniligaster grandis,” ‘Quart. Journ. Mier. Sci.,’ vol. xxxvi.

ON BENHAMIA C@OCIFERA, N. SP. 109

In Benhamia cecifera the dorsal vessel gives origin to
four pairs of very large hearts, lying in Segments 1x, x, XI, XII.
These may be, and I suspect are, connected also with a supra-
intestinal trunk, as Horst has described in certain species of
this genus; but I could not, without injury to the worm,
determine the point.

In Segments vir and vir two other pairs of hearts arise ;
these are smaller, and differ in their appearance from the
posterior pair, being more irregular in their swellings. The
large posterior hearts remain greatly distended till close to the
ventral trunk, to which they are joined by a short narrow
vessel (fig. 8). The heart in Segment vii has a less diameter
and a longer, narrow vessel connecting it with the ventral
trunk. The heart in Segment vir is much shorter. About
midway between the dorsal and ventral trunks it suddenly
narrows ; this narrow part opens into a small spherical dilata-
tion, whence arise two delicate vessels ; one goes forwards to
the side of the second gizzard, the other continues downwards
to enter the ventral vessel. This anterior heart recalls most
curiously the condition of the heart in Segment rx of Megas-
colex figured by A. G. Bourne (loc. cit. pl. 8, fig. 4).

The dorsal vessel terminates here in Segment vir
after giving rise to the hearts (figs. 7, 8), but its place appears
to be taken by a pair of “latero-longitudinal ” vessels (I use
this term topographically, and with no implication as to their
homology with those so named by Bourne in Moniligaster).
These (J. 7. v.) arise, as far as I could determine, as branches
of the dorsal trunk just in front of the third pair of hearts,
i.e. in Segment 1x, and each runs forwards to the pharynx,
giving off vessels to the gizzard and other structures.

The blood-vessels on the body-wall present a feature which I
have not seen noticed in any previous account. In each seg-
ment there is a small vessel having a circular direction along
the inner face of the body-wall. I do not know its relation to
the main trunks, but, as it is readily recognisable with the
naked eye: on each side of its course, it gives rise to a fairly
regular series of small globular dilatations ; these were more

110 W. BLAXLAND BENHAM.

readily noticeable along the dorsal wall, but could he traced
all round (fig. 9). Viewed under the microscope it is seen
that each dilatation is connected by a short narrow vessel to
the circular vessel; and from the dilatation a vessel passes
away, which soon dips into the muscular coat of the body wall
and then subdivides. I have not traced them further ; but the
regular arrangement and large size of these dilatations, which
recall those previously described and figured in the case of
Lumbricus, Trigaster, &c., is deserving of mention.

The nephridial system consists of a series of “ micro-
nephridia,”! arranged in a row around the inner surface of the
body-wall, near the hinder septum of most (probably all) of the
segments. This nephridial system is specially abundant in
Segments x11, xiv, and xv, and most markedly towards the
ventral surface of these segments.

SUMMARY.

Benhamia cecifera is, then, characterised (1) by the
number and arrangement of the peculiar copulatory pits on the
ventral field; (2) by the possession of a number of peculiar
finger-shaped ceca arising from the intestine; (3) by the
form of the foldings of the wall of the calciferous glands, and
possibly (4) by the termination of the dorsal vessel in Segment
vil; (5) by theextent of the clitellum ; (6) by the position of
the first dorsal pore between Segments rv and v.

AFFINITIES.

In its size (length 510 mm.) and number (310) of segments
it might be grouped with certain other large species of the
genus found on the West Coast of Africa, viz. B. Schlegelii,
B. Buttikoferi, B. rosea, and B.inermis. But it differs
from each of these in the above characters. Firstly, it is
marked off by the forward position of the first dorsal pore ;
this structure has been found of specific value in our European

1 This term, employed by Vejdovsky, is preferable to my term of “ plecto-
nephridia,” for we have no evidence that in every case there is a true network
any more than we have a real network in all “ capillary networks.”

ON BENHAMIA C@OCIFERA, N. SP. Tit

earthworms, Allolobophora and Lumbricus, and is
probably of similar value here.

In extent of the clitellum (Segments x11t—xxt11 inclusive)
it recalls B. rosea, Mich., in which it ceases on the 22nd
segment. But in this species the paired spermathecal and
oviducal pores are connected by transverse grooves; further,
penial chete are present, and the shape of the ‘ ventral field ”
is quite different. Again, no mention is made of copulatory
pits or tubercles.

To B. inermis, Mich., our present species presents a greater
likeness, in that there are no penial chzetz but there are
copulatory tubercles (‘‘ pubertats-tuberkeln ”’) ; these are, how-
ever, arranged on quite a different plan. There appear to be
about twelve or more pairs of them; the first pair on Segment
vu, then follow seven pairs situated at the intersegmental
grooves of as many segments, all in front of the clitellum ;
there are four more pairs within the clitellar region. Each pit
leads into a sac projecting into the body cavity. This associa-
tion of the presence of copulatory papillz and pits with the
absence of penial cheete is worthy of note.

Another large species, B. itoliensis, is known only from
its anterior end. This species was found at Itoli, on the
Victoria Nyanza, and though in general it appears to resemble
B. cecifera, yet the possession of penial chetz and absence
of copulatory pits (as well as other features) serve to distinguish
the two from one another.

112 W. BLAXLAND BENHAM.

EXPLANATION OF PLATE 12,

Illustrating W. Blaxland Benham’s paper, “On Benhamia
ceecifera, n. sp., from the Gold Coast.

Fic. 1.—Ventral view of genital segments of Benhamiacecifera. (Nat.
size.) The segments are numbered. All the genital pores are shown. spth.
Spermathecal pores. pt. p?. Copulatory pits in their neighbourhood. cop.
Copulatory (?) organs of a different nature on x1v. m. Latero-ventral margin
of clitellum bounding the “‘ ventral field.” The arrangement of the copulatory
pits, represented as small dots, is shown.

Fie. 2.—Three copulatory pits from the “ ventral field.’ In B. and Cia
papilla (p.) is more or less exposed in the pit; in 4. no papilla is to be seen.

Fic. 38.—A portion of the intestine exhibiting the characteristic finger-
shaped ceca (c.), arching upwards over the dorsal blood-vessel (D. v.) ; in
two cases the blood-vessels (v.) to this ceca are shown.

Fic. 4.—The three pairs of calciferous glands (g/.), with their vessels from
the dorsal trunk (d. v.).

Fig. 5.—The prostate duct (d.), with copulatory muscles (m.), which are seen
passing into the next segment. pr. A portion of prostate. .c. Ventral
nerve-cord. spt. Septum.

Fic. 6.—The two spermathece of the right side. d. Muscular duct. div.
Diverticulum.

Fie. 7.—The three most anterior pairs of hearts, lying in Segments vir,
vill, and 1x. The dorsal vessel (D. v.) is seen ‘o terminate in Segment vit.
spt. Septum 9/10.

Fic. 8.—The same three hearts seen partially from the side, showing the
relations ventrally. 7. 7. . Latero-longitudinal vessel arising in Segment 1x.
g. Its branch to gizzard. g’/. Branch from first heart to the gizzard. s.
Swelling from which this.arises. V.v. Ventral vessel.

Fic. 9.—A small portion of the body-wall of a segment, to show the dilata-
tions on the circular vessel (c.), as well as the position of the micronephridia
(z.). S'., S*. The anterior and posterior septa.

EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 1138

A Re-investigation into the Early Stages of the
Development of the Rabbit.

By

Richard Assheton, M.A.

With Plates 13—17.

INTRODUCTION.

Ir is now nearly fifteen years since Ed. van Beneden'
published his account of the development of the early stages of
the rabbit ; and although I do not see that one can as yet
describe every detail of even the earliest embryology of the
rabbit strictly epigenetically, still it seems probable that van
Beneden took far too little heed of the extrinsic causes which
may direct the course of development.

The effect upon the development of the presence or absence
of such structures as the albuminous layer or zona radiata has
been almost entirely ignored, apparently because they are
matter outside the ovum. So, again, the size and structure of
the uterus have hardly received their proper share of attention.

The present paper and my other “ On the Causes which lead
to the Attachment of the Mammalian Embryo to the Walls of
the Uterus,” tend, I hope, to show how many of the details of
the earliest stages may be ascribed to the direct maternal
influences. That is to say, the inherited force is an energy

1¢Ta maturation de l’ceuf, &c., Bruxelles, 1875; ‘La formation des
Feuillets chez le lapin,’” ‘ Arch. de Biologie,’ vol. i, 1880.

VOL. 87, PART 2,—NEW SER. H

114 RICHARD ASSHETON.

which would of itself produce not a specific embryo, but an
amorphous monster unless directed by the influence of (in the
rabbit) inanimate coats and the walls of the uterus.

I have come to the following conclusions upon certain dis-
puted points.

(i) Van Beneden’s description of the segmentation I consider
to be inaccurate.

(ii) I find no trace of van Beneden’s blastopore.

(iii) I find no trace of a “ gastrulation.”

(iv) There is no evidence in support of Robinson’s! specula-
tions concerning the existence of a hypoblastic wall to the
blastocyst surrounded subsequently by the epiblast.

(v) Rauber’s layer fuses with the inner layer of epiblast as
described by Balfour? and Heape.®

(vi) This fusion has but slight morphological significance,
since its existence and disappearance are caused mechanically
by ontogenetic conditions.

(vii) The growth round of the hypoblast is apparent only,
being due to the presence of a zone of specially active epiblast
surrounding the embryonic disc, which zone is to be considered
to be the equivalent of the trager in other rodents.

I have endeavoured to show how important is the albuminous
layer, and how I believe that it is possible to account for
many details of change up to the end of the sixth day by
strictly epigenetic processes; and since these processes during
this time are almost all directed by environment as between
the embryo and the maternal influences rather than between
cell and cell of the embryo itself, it follows that the palin-
genetic features of the development must be reduced to a
minimum.

This paper is based upon the examination of upwards of 300
embryos between the 24th and 168th hours. The embryos have
been examined fresh, and after treatment with various reagents
—Perenyi, osmicacid, picric acid, Flemming, Hermann, chromic

1 Quart. Journ. Mier. Sci.,’ vol. xxxiii.
2 «Comparative Embryology,’ vol. ii.
3 «Quart. Journ. Mier. Sci.,’ vols. xxiii and xxvi.

EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 115

acid, &c., and sections have been examined of all stages from
the 30th hour onwards.

A large portion of the work for the present paper, and the
three or four other papers I hope to publish shortly, was carried
on while I held the post of Demonstrator of Zoology in the
Owens College.

To the kind consideration of the late Professor Arthur Milnes
Marshall Iam indebted for many opportunities for work which
I should otherwise have found difficult to obtain.

The remainder of the work has been done in the Morpho-
logical Laboratories of Cambridge through the permission of
Mr. Sedgwick, to whom I wish to express my gratitude for the
said permission and for his kindness in reading my papers and
offering many valuable criticisms.

CHAPTER I.

SEGMENTATION OF THE Ovum.

Since the accuracy of van Beneden’s account of the early
stages of the development of the rabbit depends to a great extent
upon the way in which the earliest cleavage planes succeed one
another, I think it advisable to give in some detail the results
of my own researches on this point. In fact the very first
segmentation division, according to van Beneden, gives rise to
an important question, which is as follows: is there any
essential difference between the two first segments, and do the
descendants of one segment give rise to the epiblast cells, and
the descendants of the other to the hypoblast cells ?

To this van Beneden replied that there is always a difference
in size between the two first segments, and that the smaller
of the two is the more granular, and that from that one are
derived all the cells of the inner mass, and that the larger
clearer segment gives rise to cells which gradually surround
the descendants of the small segment and form the wall of the
future “ blastodermic vesicle.”

From the descendants of the small segment van Beneden

116 RICHARD ASSHETON.

derived the hypoblast and subsequently the mesoblast ; from
the descendants of the larger, the epiblast.

As far as the conversion of the whole of the inner mass into
hypoblast and mesoblast is concerned, van Beneden has changed
- his opinion, admitting that Kolliker’s (and others—Balfour,
Heape) careful research in this matter is a truer account than
his own.

Again, Heape, on the mole, has shown that a different inter-
pretation may be placed upon the fact as seen in the rabbit,
although he finds that there is a difference in the size of the
two first segments.

The results of my researches show that there is, if not
always at any rate very frequently, a difference in size between -
the first two segments ; sometimes the difference is very marked,
but more usually it is to be found only by careful measurement.
In one rabbit I examined the following results were obtained.

Rabbit 48, was killed 244 hours after coitus. In the thin-
walled part of the Fallopian tube of the right side I found four
embryos at about 30 mm., 40 mm., 42 mm., 50 mm., from the
funnel mouth of the Fallopian tube respectively.

No. 1, the highest up, had not as yet divided. The ovum
was spherical with a distinct nucleus, the ovum not occupying
the whole space within the zona radiata.

No. 2, the second on the right side, was divided into two
segments, but this I did not measure.

No. 3, the third and lowest on the right side, was also in two
segments. I could detect no differences at all in colour or
texture, but on measuring the longest and shortest diameters
of each the following results were obtained.

One of the segments measured as follows: longest diameter
‘1 mm., shortest diameter (076 mm. The other segment:
longest diameter ‘096 mm., and shortest ‘068 mm.

The specimens were examined and measured in the fresh
condition in a drop of aqueous humour of the rabbit. The
measurements were taken with a micrometer eye-piece, No. 8,
and Zeiss D objective, which gave a magnification of 54, mm.
for each division in the eye-piece.

EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 117

In giving the measurements I have reduced these to actual
measurements, giving them in decimal fractions of one milli-
métre.

No. 4 was also in two segments, but was not measured ; each
segment was composed of an inner denser and outer clearer
part. In the Fallopian tube of the left side I found six ova,
which were scattered along an area between two points
35 mm. and 65 mm. from the upper end of the Fallopian
tube. The whole length of the Fallopian tube was about
100 mm.

Three of the ova were in the wide thin-walled upper part,
and the other three were just within the lower thicker-walled
portion of the Fallopian tube. No. 1, that is to say the one
nearest to the funnel end, was unsegmented. It appeared, how-
ever, to be on the point of dividing (fig. 1). It did not occupy
the whole of the cavity within the zona radiata. It was
spherical, and two polar bodies lay between the ovum and the
zona, the one larger than the other. The nucleus, however,
though not very distinct, undoubtedly seemed to be double or
dumb-bell shaped.

The diameter of the ovum was ‘11 mm.

No. 2 (v. fig. 3) was in two segments. The difference in
size between these two was very marked. It was quite obvious
without measurement. This specimen showed the greatest
difference I have hitherto met with. There was no perceptible
difference in density or in colour, either while fresh or after
treatment with nitrate of silver followed by picro-carmine and
aniline blue-black. The measurements of the two segments of
this specimen were as follows :

Of the larger the longest diameter was ‘101 mm., the
shortest diameter ‘(088 mm. Of the smaller the longest dia-
meter was ‘(08 mm., and shortest was ‘064 mm. The nuclei
were clear and round.

No. 3 was in two segments, but far more equal in size.
Nuclei round and clear and distinct. Measurements were—

The larger segment . : aod) ome. < 078: mm;
The smaller segment. . : SUG sos 6 “OSA 5,

118 RICHARD ASSHETON.

I could detect no difference except as regards size.

No. 4. I could make nothing of this one; it may have been
an unfertilised ovum or pathological.

No. 5 was in two segments. Here again the only difference
I could detect was in size, as follows:

Larger segment . : : . ‘086 mm. Xx ‘072 mm.
Smaller segment : : =; “0840-— 5. << 2066

2

No. 6, the lowest down, that is to say the nearest the uterus
on the left side, was still unsegmented, being spherical. Both
polar bodies were visible between the ovum and the zona pellu-
cida.

The ova from another rabbit, No. 51, were examined fresh in
aqueous humour of the rabbit.

In the left Fallopian tube I found only one ovum, situated
almost midway between the two ends. This specimen was not
yet segmented, but the ovum had retracted from the zona
pellucida, which seems to be a sign that it was a fertilised ovum.

In the right oviduct I found four ova.

No. 1, the one nearest the funnel opening of the Fallopian
tube, was in two segments. I could determine no difference in
character, but on measurement a slight difference in size was
made evident.

Larger segment : ; . 0924 mm. x ‘070 mm.
Smaller segment, ; : s», “0980:. 55) xX Db4nts

After being carefully examined and measured while fresh,
this specimen was transferred to a 2 per cent. solution of
osmic acid for thirty seconds, and subsequently stained with a
mixture of aniline blue-black and picro-carmine, and mounted
in glycerine jelly—a very pretty preparation. At no period could
I detect any difference except in size between the two segments.

In this specimen the two polar bodies were placed far apart
from each other, a peculiarity I have not noticed in any other
(v. fig. 4).

No. 2 was also in two segments, and was examined, measured,

EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 119

and drawn in aqueous humour. The result of measurement
was—

Larger segment . : : . ‘099 mm. x ‘068 mm.
Smaller segment . F : 5 UO SK BOLO

No.3 was also in two segments, and was examined, measured,
and drawn in aqueous humour. Result of measurement :

Larger segment . : : . ‘LOO mm. x ‘079 mm.
Smaller segment. F ; i, OS Geb ne CORE OS

No. 4, the nearest specimen to the uterus, was as yet unseg-
mented, possibly an immature or unfertilised one, as the ovum
completely filled the space within the zona pellucida. I could
not find any trace of polar bodies.

All three specimens from this rabbit, which were in two seg-
ments, I examined most carefully, to find, if possible, evidence
of the marked difference in appearance described by van
Beneden. I treated all with 2 per cent. osmic acid and used
weak staining solutions of picro-carmine and aniline blue-
black, of picro-carmine, of Beale’s carmine and _ aniline
blue-black respectively, but to no purpose. They were then
mounted in glycerine jelly, and now, after the lapse of
a year, I still fail to find any difference between the two
segments.

From another rabbit I obtained two specimens in the two-
segment stage, but was unfortunately only able to make a very
cursory examination. One of these I thought I distinguished
as having a larger darker segment and a smaller clearer one,
but I could only examine it while lying on the folds of the
Fallopian tube, the unevenness of which might easily have
caused one segment to appear darker. So I cannot advance
even this instance in support of there being a marked difference
in appearance between the two segments.

Another rabbit, with ova aged 254 hours, gave me four, of
which two were in the two-segment stage, and two showed four
segments. Here again there was no perceptible difference. I
did not measure these while fresh. One I have since measured,

120 RICHARD ASSHETON.

and find that the longitudinal and transverse axes of the one
measure exactly the same as those of the other.

A tabular view of the measurements of these two first cleavage
segments may be of interest, showing how the size of one
specimen may vary with that of another, and the variation in
size of the two segments of one and the same specimen.

Larger Segment. Smaller Segment.
Longest Shortest Longest Shortest
Diameter. Diameter. Diameter. Diameter.
Rabbit 48—
Specimen No. 1 100 mm. X ‘076 mm.| *096 mm. x ‘068 mm.
‘ se TOM a x O88 y; “080° ,, «x02 |,
Me. ey Ati 100 ;, x 078 ,, | 098 ,, x ‘074 ,,
ah heel ‘086 5. x O72 . | "Usa. © sone
Rabbit 51—
Specimen No. 1 0924.4 XScO7Ou. 092025. Sex 0G
aR ak 099 ,, x 068 ,, | 090 ,, x -060 ,,
ee 100 5, © "079 4, | “098 4, x 07s

Each segment of each specimen when examined immediately
after the death of the animal showed a denser, more granular
inner portion, and a clearer, almost hyaline outer layer, the
nucleus being situated in the denser inner portion (v. fig. 2).
This difference is more evident in the early stages of segmen-
tation, up to the time that there are twelve to sixteen seg-
ments, than later.

Stage with Four Segments.

The second plane of cleavage seems to be at right angles to
that of the first. It appears that the two segments divide
about, if not exactly, at the same time. ‘This occurs about
twenty-five or twenty-six hours after coition. The time between
the segmentation of one into two and between two into four
appears to be nearly two hours. Since there may be some
difference in size between the two primary segments, it follows
that there is also very frequently a difference in size between

-

EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 121

the cells of the four-segment stage, and, as one would expect,
it may appear that two segments are small and two larger, as
in fig. 6.

Development between the 26th and 72nd Hours.
Stages of Segmentation between Four-Segment
Stage and Commencement of Cavity of the Blas-
todermic Vesicle.

The further cleavage of the four segments does not occur in
each segment at the same moment. A stage of five segments
or even seven may often be found.

In a rabbit killed 27+ hours after coitus (Rabbit 36) I found
in the right Fallopian tube two ova in the usual locality for
this age, which is about 40 mm. above the uterus; one of
them was in five segments, the other in seven. The five-
segment one is shown in fig. 7.

In this specimen it will be seen that two (Z2.) spheres are
almost exactly the same size, and that these are considerably
larger than two (S*.) of the remaining three, and slightly larger
than the third (S*.). This specimen I surmise may have been
one in which the two primary segments were unequal, and that
each of these divided into the approximately equal spheres,
and that at the moment of examination one of the daughter
cells of the primary smaller (?) one had divided again into two
very nearly equal spheres (S*.).

It should be noted that in van Beneden’s account of the
process the larger primary segment gives rise to the more
rapidly dividing daughter cells forming his so-called epiblast ;
while in this case it seems to be the descendant spheres of the
smaller primary segment which appear to be the more ready
to undergo division.

Fig. 14 is a camera lucida drawing of the other embryo of
the same Fallopian tube, seen as a transparent object. In
this specimen one segment (L?.) is larger than the others. The
four marked S*. were approximately equal, and slightly
larger than the two (L*.). May this be interpreted as follows ?
The ovum which gave rise to this embryo divided into two

122 RICHARD ASSHETON.

slightly unequal spheres. The smaller of these two divided
into two, each of which has divided into two, the result being
four approximately equal spheres (S°., S®°., S°., S*.). The larger
primary spheres divided into two, one of which is still undivided
(L?.), the other having divided into two (marked Z?.). Here
again there seems to be a tendency for the descendant spheres
of the smaller primary to undergo division first.

In the Fallopian tube of the left side of the same rabbit I
found six embryos, of which some were in seven segments,
others ineight. I had not time, unfortunately, to measure all
these or draw them carefully. The embryos of this rabbit
seem to be unusually far advanced for their age.

The stage with eight segments is a very common one to find
between the 29th and 44th hours. This may be accounted for
by there being a rather long resting stage after the production
of the eight segments.

One rabbit (Rabbit No. 19) presented a very curious condition
of its Fallopian tube, a condition, I believe, that has been
noticed before, but I cannot remember by whom. The rabbit
was a very large, healthy English doe. Both Fallopian tubes
were almost filled with ova. Onexamining the Fallopian tubes
with a lens before opening I imagined I had come across a
marvellous find of embryos—in all about fourteen “ ova” were
to be seen shining through the wall of the left Fallopian tube.
Of these only three turned out to be embryos, and were
mingled with at least twelve apparently disintegrating un-
fertilised ova. The right Fallopian tube likewise contained a
number of disintegrating ova, as well as four perfectly normal
embryos.

Some of these unfertile ova were more thickly coated with
albumen than is usual for normal eggs, possibly due to their
having been a long time in the Fallopian tube. It is a matter
of curiosity why these unfertilised or pathological ova had not
passed down the Fallopian tube, but had allowed the fertile
ova to pass them, as one at least had succeeded in passing the
whole twelve bad ova. One can hardly believe that all these
ova had left the ovary at the same time and together with

EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 123

the ova which became fertilised and developed into normal
embryos.

The corpora lutea are often very difficult to make out at
this early stage, but I am pretty certain that there were cer-
tainly not more than five corpora lutea in one ovary and six in
the other.

Fig. 9 is drawn from one of the fertile ova of this rabbit. I
have reproduced it here to show the appearance of the denser
inner and clearer outer layer of protoplasm of each sphere, de-
scribed above as occurring in the earlier stages.

Possibly this appearance is constant in later stages, but as
the cells get smaller the difference becomes less easy to distin-
guish.

I have avoided the use of the terms macrosphere and micro-
sphere, because a morphological significance is sometimes given
to these terms which I think is not advisable, at any rate in the
case of the rabbit.

At the same time it is necessary, in discussing van Beneden’s
description and deduction, to pay especial attention to the
question of the fate of the segments derived from the two first
spheres.

It is hardly necessary to point out the very great difficulty
in determining the fate of the descendant segments of the two
primary segments when size is the only character on which we
can rely, and when even in size there may be no difference.
Again, if the ovum may sometimes divide into two segments
of different size, may not the two primary segments also
segment unevenly?

I can see no way of determining the question except by
watching the division of one specimen, and this is, as far as I
have tried, impossible. Even if it were accomplished, it must
be under conditions which can hardly be called normal, and
therefore not a very safe ground on which to base either facts
or theory.

Certainly, in cases where the primary division is so unequal
as in fig. 8, the subsequently formed segments could, on the
whole, probably be classed as larger ones derived from the

124 RICHARD ASSHETON.

large primary segment (fig. 3, Z.), and smaller ones derived
from the small primary segment (v. fig. 3, S.). If, then, we rely
upon the only character at present possible, that is to say,
upon the size of the segment, we are bound to conclude that
not only does the process of segmentation proceed with great
irregularity, but that also there is no evidence of the descend-
ants of the larger primary segment ever forming a cap and
subsequently enveloping the descendants of the smaller primary
segment.

I believe, rather, that descendant segments of the larger
primary segment become intermingled quite irregularly with
the descendants of the smaller primary segment, for this un-
doubtedly affords a better explanation of the appearance of
embryos like that shown in fig. 12 or fig. 15.

Fig. 11 is an outline drawing of a specimen taken from a
rabbit which was killed at the completion of the 39th
hour. In this animal seven of the eight embryos found were
in the eight-segment stage. It is extremely difficult to mea-
sure the segments with sufficient accuracy to be of any service,
and although I measured them I shall not give the results. In
this case, as shown by the fig. 11, there was very little dif-
ference in size, and none, so far as I could judge, in texture.
There was one curious feature which is worth mentioning, but
I do not attach any importance to it. The larger polar body
was visible between the embryo and the zona radiata. The
smaller of the two was inside, or rather mingled with the seg-
ments of the embryo. It is quite possible that this frequently
occurs, for the polar bodies seem very often to disappear en-
tirely. It would be of interest to determine whether, when
this is the case, the polar bodies, ever under the altered con-
ditions, acquire a renewed activity and give rise to segments
whose descendants become part of the embryo.

I have as yet no evidence as to which sphere commences the
next series of cell division.

At about the 47th hour the embryo has the typical morula
form, and is made up of a number of segments (sixteen to
twenty), which very frequently present great diversity in size.

EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 125

Nor is there any regularity, as far as I have observed, as to
the location of the large and small spheres. Figs. 12, 15, and
17 show this very well.

Fig. 12 was drawn with camera while the specimen was still
fresh in a drop of aqueous humour of the rabbit.

After the drawing had been made the specimen was placed
in Perenyi’s fluid and subsequently freed from the zona pellu-
cida with fine needles. The segments were then separated one
from another. In all there were seventeen segments, and of
very different sizes.

Fig. 18 shows three of the segments thus separated. Except
in size I could detect no difference.

Another specimen from the same Fallopian tube was placed
in 4 per cent. solution of silver nitrate for two minutes, and
after having been washed in water and exposed to the sun for
a few hours was embedded in paraffin and cut.

Fig. 20 is from a section through about the centre of this
specimen. The nuclei of the individual segments are not at
all distinct, excepting where they have been cut almost through
their centres, as no other stain except silver nitrate has been
used.

(Of all methods of fixing the early segmenting stages, I
believe none answer so well, as far as concerns the preservation
of the correct shape of the spheres, as a weak solution of silver
nitrate allowed to act for not more than two minutes. Next
to silver nitrate I believe osmic acid, 2 per cent., is best.)

Fig. 15, which is a specimen from a rabbit killed sixty-six
hours and a half after coition, exhibits in a very remarkable
way the great difference in size which may sometimes occur
between the several segments.

From the forty-fifth to the seventieth hour the segmentation
proceeds slowly, and, I am inclined to think, sometimes very
irregularly, as shown by the last-mentioned specimen (fig. 15).

In sections of these stages I do not notice anything particu-
larly remarkable, except that I have completely failed to find
any constant character whereby the inner cells can be distin-
guished from the outer. Fig. 21 is of a section through the

126 RICHARD ASSHETON.

centre of a rabbit embryo aged seventy-six hours and a half.
This was preserved in Perenyi’s fluid, stained with borax
carmine, cut and mounted in series. The one I have drawn is
the fourth of eight which pass through the actual embryo itself.

The embryo at this stage is composed of a number of cells
or segments, as far as I can see all similar in character, though
varying a good deal in size.

The cells in the centre are no doubt pressed closely
together in the living state, the several clefts, with the possible
exception of one, being artificial. This just-mentioned excep-
tion, the more regular and continuous slit marked C. BL., isin
all probability the first commencement of the slit which ulti-
mately enlarges into the cavity of the blastodermic vesicle.

In only one specimen have I found certain cells to take the
stain better than others. I have drawn two figures (18 and 19)
from the series of sections in which this occurs, in order to
show how such an appearance as that described and figured by
van Beneden (pl. iv of his paper) may arise, and at the same
time to show how the interpretation put on it by him cannot
be held to be sound.

In this specimen the embryo has contracted very much, and
is lying quite free from the zona pellucida and albuminous
layer.

In fig. 18 the majority of the cells at the surface are seen to
be slightly darker than the mass inside, with the exception of
two at the point 2. This is undoubtedly hke van Beneden’s
figure of an optical section (pl. iv, fig. 1). But if we look at
another section, fig. 19, we find here that again the cells of the
surface layer have mostly stained darker, with the exception of
one at 2’.

Hence we are obliged to believe that in this specimen there
existed at least two of van Beneden’s blastopores, for the
light-coloured cells which show at the surface in fig. 18 are at
almost the opposite pole to that at which the lighter-coloured
cells of fig. 19 show at the surface.

In sections of other specimens of about this age (seventy-
two hours) fixed with osmic acid 2 per cent., I have failed also

EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 127

to find differences of any value between the inner and outer
cells. The same may be said of sections of specimens treated
with silver nitrate.

Fig. 16 is a drawing through the centre of a specimen from
a rabbit seventy-two hours after coition, fixed with nitrate of
silver + per cent., exposed to the light and stained with picro-
carmine.

Summary up to the Seventy-second Hour.

The original description of the segmentation stages by
Bischoff I believe to be in the main correct. I cannot find
any evidence to support van Beneden’s view of the origin of
the inner mass of cells from a smaller, more darkly staining
primary segment only ; or for the origin of the outer layer of
cells from a larger, more lightly staining primary segment
only ; or, again, for the growth of the descendants of one of
the two primary segments round the descendants of the other
primary segment.

It must be borne in mind that the very fact of van Beneden’s
description of the later stages of development (the origin of
the mesoblast and hypoblast from the inner mass, and that of
the epiblast entirely from the outer layer) having been shown
to be wrong by several authors (Kolliker,' Heape, Balfour, &c.),
made it almost certain that van Beneden’s description of the
segmentation stages was incorrect, or, at any rate, that the
interpretation he put upon the supposed facts could no longer
be held to be sound.

My account is briefly as follows :

1. The ovum about the twenty-fourth hour after coition
divides into two segments, one of which is usually larger than
the other, there being much variation in this respect.

2. Each of these segments again divides about the twenty-
sixth hour after coition, each dividing very nearly at the same
time. These four segments now resulting may vary in size.

3. The third series of divisions takes place about the twenty-

1 A, Kolliker, ‘ Festschrift zur Feier des 300 Jahrigen bestehens der Julius-
Maximilians- Universitat zu Wiirzburg,’ 1882.

128 RICHARD ASSHETON.

eighth hour after coition. There is now less unanimity of action
in point of time of the division of the several cells, so that it
is fairly common to find embryos with five or seven segments.
Here, again, four cells may be distinctly smaller than the
remaining four, or all may be almost exactly equal in size.
There is no difference to be detected in the character of the
contents of the spheres.

4. The segments continue to divide with less and less regu-
larity, so that the descendants of one primary segment become
mingled with those of the other, there being much difference
in size between the several segments.

5. The result of this continued activity is the formation of
a solid “morula” of cells of varying size but similar
character, and it is impossible to apply the term epiblast or
the term hypoblast to any part of the embryo as yet.

The ovum when it leaves the ovary is enclosed within a
sheath or protective investment, the zona radiata. As the
ovum proceeds down the oviduct this protective investment is
further strengthened by the deposition on its outer surface of
a thick coat of some albuminous substance which is secreted
by certain cells of the epithelial lining of the oviduct, and
which forms a tough strong membrane which has an important
influence on the future mode of development.

CHAPTER II.
Tur Formation or THE BLAsToDERMIC VESICLE.
The Fourth Day (73rd to 96th Hours).

The most noticeable feature of this day’s events is the
commencement of a cavity within the morula, which cavity
enlarges enormously.

What the object of this cavity is we can pretty safely infer,
as also we can pretty safely conclude that the causes which
bring it about originated in the organism in connection with
the diminishing size of the ovum of the distant ancestral
animal; but how this cavity is actually produced in the

EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 129

embryo at the present day is a question of very great dif-
ficulty.

It can hardly be looked upon as a cavity comparable to the
segmentation cavity of Amphioxus or Amphibia, or the Lam-
prey or Ganoids, for it has a fate different from that in any of
those animals. The fate of the segmentation cavity of the
above-mentioned animals is to disappear and take no part in
the formation of the cavities of the adult.

The cavity of the blastodermic vesicle whose formation we
are about to discuss never disappears, never, at any rate, in
the rabbit, as part of it remains as the cavity of the alimen-
tary canal of the adult. Part of it undoubtedly is comparable
to the archenteron; possibly all of it is, for it becomes the
gut-cavity of the adult.

The first cause which produces the cleft that subsequently
enlarges into the cavity of the blastodermic vesicle may be a
more active growth of the outer layer of cells. Undoubtedly
there is after this time a more active growth of the cells of
the outer layer. They increase much more rapidly than the
cells of the inner mass. .

It is not easy to explain why the energy which up to a
certain time causes a solid morula should do so no longer.
Why should not the morula steadily increase in size, but be
stilla morula? Although I believe that the subsequent vast
increase in size of the blastodermic vesicle cavity is due to the
diffusion inwards of fluid derived from the uterus, still this can
hardly be the cause of the first starting of a cleft as in fig. 21.
This seems to be best explained by the assumption of an in-
crease in rate of growth of the outer cells over the inner cells
of the morula. That such an increase does exist the following
table provides evidence.

In a median section through an embryo:

Hour. Number of cells cut through in outer layer. Inner layer.
76th . : : : Loe, : . ; 27
82nd . : . : 24. : : : 24
Ser OMe sweets: be i Bi; : - teh

100th . : ° . 4A 4 : : 28

VOL. 37, PART 2,——NEW SER. I

130 RICHARD ASSHETON.

This, though rather a rough method of investigation, shows
sufficiently well that there is a proportionally greater rapidity
of increase of the cells on the outside over that of the inner
cells.

Growth of the embryo must surely be dependent upon
nourishment from without, when the bulk of the mass increases
as it does from the 70th hour. It is the cells upon the outside
of the morula which are in the most favourable position for the
acquisition of such nourishment from the fluids of the uterus
or Fallopian tube. No doubt during the early stages of seg-
mentation the energy of division is derived from the nutriment
—yolk, &e.—contained within the ovum itself. As this be-
comes used up the embryo will become more dependent upon
external sources. This will mean that the externally placed
segments will become more favourably placed for growth than
the internally placed segments. May not this gradual ex-
haustion of intrinsic nutriment be a determining factor in the
cessation of the increase of the embryo as a morula and causa-
tion of the first commencement of a cavity ?

This may give rise to the first cleft, but in itself it can
hardly account for the large increase in size of the blasto-
dermic vesicle. The cells are so very delicate it is hardly con-
ceivable that they could cause the great expansion of the very
tough albuminous wall. It seems far more likely that the
force which causes the expansion is due to an osmotic current
being more rapid inwards than outwards, either simple or
more probably assisted by the vital activity of certain cells of
the embryo, as is supposed in the case of diffusion in intestine,
and suggested by Heape in connection with the mole. Or the
diffusion process may be a simple physical process, but the
nature of the liquid after entering the cavity may be so changed
by the activity of the cells as to render its diffusibility less
when once inside than before its entrance from the uterus.

However this may be, the most noticeable fact of the
development of the rabbit embryo during the fourth day is the
commencement and enlargement of the cavity of the blasto-
dermic vesicle.

EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 1831

Up to the moment of the beginning of the cavity there does
not appear to be any pressure upon the walls of the embryo by
zona radiata and albumen layer ; in fact, the embryo may often
be found to be slightly retracted from the zona radiata in the
fresh state.

The diameter of the cavity within the zona radiata is, in the
unsegmented ovum, about 0°11 mm., and up to the time that
the cleft appears it has not greatly enlarged, measuring only
about 0°12 mm.

On the appearance of the cleft the cells of the outer layer
become pressed hard against the zona radiata and flattened,
but no great increase in diameter of the outer border of the
zona radiata is as yet recognisable, although the thickness of
the zona radiata is diminished, apparently being of a com-
pressible nature, and therefore becomes compressed between
the pressure from within the blastodermic vesicle and the
resistance afforded by the tough albumen layer from without.
At this time, being in the very firm lower part of the Fallopian
tube, the resistance afforded by the albumen layer is very
likely aided by the walls of the Fallopian tube itself. In this
condition the embryo usually passes into the uterus, although
sometimes specimens may be found in the uterus (but very
rarely) in which no cleft has as yet appeared. The embryos, I
think, pass rather suddenly through the last 4 to 6 mm. of the
Fallopian tube at some time between the 75th and 80th hours
after coition.

No increase takes place in the thickness of the albuminous
layer after entering the uterus, but by the stretching of it
caused by the expansion of the blastodermic vesicle it rapidly
thins. The zona radiata thins so much as to be hardly per-
ceptible by the end of the fourth day.

Van Beneden’s figures, 5, 8, 9, 10, 11, on pl. iv of his paper,
taken from optical sections, represent very well the appearances
presented by the embryos during these changes.

As van Beneden’s figures represent optical sections, it is, I
think, advisable to give a series of figures drawn from real sec-
tions, as none have hitherto been published.

182 RICHARD ASSHETON.

Figs. 20 to 34 are all camera drawings, and each is magnified
465 times.

Figs. 20 to 25 show sections through the whole embryo, the
subsequent figures through the embryonic disc, or a part of it
only.

Fig. 20, and also fig. 16, are sections of silver nitrate
preparations.

In both cases there is no difference between the cells at
the surface and those towards the centre as regards their
colouring or nuclei. Those inside are certainly more com-
pressed than those on the surface. So also it frequently
happens that a large segment may be so placed that although
the greater part of it may be said to belong to the “inner
mass,’’ yet a small part of it may appear on the surface, as 2.
in fig. 20.

If such a specimen is examined in optical section, this
individual cell might very likely give rise to such an appear-
ance as van Beneden describes, and lead to the idea of an inner
mass partially surrounded by an outer layer. The difference
in colour which van Beneden describes is totally absent in real
sections, and I can find no greater opacity of the inner mass in
optical section than what can be equally well explained by the
greater thickness through which the light must of necessity
pass in viewing a sphere in optical sections. I am not able to
offer any explanation of the condition figured by van Beneden
in his second figure ; I can only say that I have have not been
able to find it.

Fig. 21 is a section through an unusually large specimen.
This specimen was preserved in Perenyi. This specimen I
believe shows the earliest stage in the formation of the cavity
of the blastodermic vesicle (C. BL.).

There is as yet no evidence of an internal hydrostatic
pressure, but the outer cells form a compact layer, and seem
at one point to be as it were lifted away from the inner cells,
leaving a slight cleft (C. BL.).

In fig. 22 this cleft has increased considerably. It must be
noticed that it does not extend through more than about 240°.

EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 13838

Through the remaining 120° the inner mass is still as much part
of the wall of the vesicle as the outer layer of cells.

In fig. 23 the cavity shows a further increase in size. The
outer layer of cells (O. ZL.) now show signs of having become
stretched, due, I believe, to the rapidly increasing hydrostatic
pressure within the blastodermic vesicle. The walls of the
vesicle are now, while living, firmly applied to the zona radiata
and albumen layer. The space shown between the vesicle and
the zone in the figure is due to reagents. The same remark
applies to all the figures from the 22nd onwards. I have not
been able to distinguish sharp lines of division between the
outer layer cells in section after the commencement of the
blastodermic vesicle cavity. In surface view there are certain
lines of divisions, which, as van Beneden has shown, are very
clearly brought out by silver nitrate.

In this fig. 23, which I believe to be a median section, there
is a fairly clearly defined line marking the inner mass from the
outer layer, which is less perceptible in fig. 22, and which
becomes very distinct in later stages, such as in fig. 28.

It is now possible to distinguish clearly an inner mass as
separate from the outer layer. This separation seems to be
caused simply by the tension being more acute in the outermost
of the segments of the mass J. M. in fig. 22, causing these seg-
ments to be more stretched than the remainder, because they
are more directly united with the already separatedand stretched
cells O. L.

Lines of division between the segments of the inner mass
can usually be found, but varying greatly in definition. I
cannot say to what extent these several segments may be really
divided. The inner mass is certainly connected in some way
with the outer layer, but whether by direct protoplasmic
union I am quite unable to say.

Fig. 24 presents no new characters excepting a tendency to
flattening of the inner mass. Im figs. 25, 26, 27, the inner
mass is flattened out still more, so that it presents a lenticular
form in section instead of a circular outline as it did in figs.
22 and 23. How does this change in form come about ?

134 RICHARD ASSHETON.

Van Beneden (p. 29) describes the process thus :—“‘ La
masse cellulaire endodermique s’aplatit 4 la face profonde de
Vectoderme et la surface de contact entre les deux feuillets
s’étend peu a peu.”

This is a quite correct account of the appearances but in that
the description tends to suggest an inherent inclination on
the part of the inner mass cells to flatten themselves; it
is, I think, right to look for some other explanation.

The inner mass is attached in some way to the outer layer,
and accordingly it seems only probable that this mass should
be drawn out with the stretching of that part of the outer
layer to which they are attached, during the expansion of the
whole blastodermic vesicle by the increasing hydrostatic pres-
sure within.

Figs. 22—27 are all sections of preserved specimens, and so
mostly crumpled to a certain extent.

If, however, the angle subtended by the arc of the wall of
the blastodermic vesicle to which the inner mass is attached
is measured in optical sections of fresh specimens of ages
corresponding to my figs. 22 to 27,it will be found that the
angle is nearly constant. This angle is about 80° (see my
diagram, fig. 41). Up to this time there is no indication of a
separation into what one may call epiblast and hypoblast.
There is a separation into outer layer and inner mass, but this
I believe to be of no palingenetic importance whatever.

The separation I have tried to show is due to the individual
circumstances of ontogeny entirely. So also, and as I shall
endeavour to show in the next chapter, I believe the ultimate
separation of epiblast and hypoblast is also devoid of all palin-
genetic influences such as any form of invagination, or epibole,
which is usually spoken of as “ gastrulation.”

The several segments of the inner mass change their shape
in response to the change of their individual environment.
In figs. 20, 21 they are compressed and hence polygonal. As
the pressure is removed during the growth of the cavity C. BL.,
those towards the cavity become rounded (figs. 22—27), which
character tends to become universal in figs. 28, 29, with excep-

HARLY STAGES OF DEVELOPMENT OF THE RABBIT. 185

tion of some which subsequently come under the influence of
tension, to which I shall refer again.

CHAPTER: if.
Toe APPARENT EXTENSION OF THE HyPpoBLast.
The Fifth Day (97th to 120th hours).

In the description of the shape and growth of the blasto-
dermic vesicle, and the discussion which follows, I call the
part of the blastodermic vesicle to which the inner mass is
attached the upper pole, the part immediately facing this and
furthest removed from it the lower pole. The zone midway
between the poles I call the equator.

The part whereat the central portion of the inner mass is
attached to the outer layer, that is to say that part where
there are three layers of cells, I call the embryonic disc, but I
do not thereby mean to say that only that area and none other
takes part in the formation of the adult.

The events of the fourth day are concerned almost entirely
with the enlargement of the embryo.

During the fifth day this enlargement continues, but other
important developments now take place.

It is during this day that we can first detect a separation into
the two well-known embryonic layers, epiblast and hypoblast,
and as to the mode of origin of these layers I fear I cannot
quite agree with the description and explanations of any former
writers on the subject. As I shall point out in a future paper,
I can see no necessity for the occurrence of any folding in or
growing in of cells from without, usually known by the term
‘“‘ vastrulation,” and most certainly I can find no evidence of
it whatever. As far as I know, only one investigator has stated
that he has found actual evidence for this in the case of the
rabbit. This is Keibel, in his work “ Zur Entwickelungs-
geschichte der Chorda bei Saiigern,” ‘ Arch. fiir Anat. und
Physiol.’ 1889. In this Keibel only found his “ blastopore ”

136 RICHARD ASSHETON.

in one specimen, in which the albumen layer and zona radiata
had been removed (thereby rendering it very likely that a tear
may have occurred), and he was quite unable to find it in other
specimens. If such an aperture really existed, it would cer-
tainly be perceptible in surface views of specimens stained with
silver nitrate. Van Beneden makes no mention of such an
aperture. I have hunted carefully but can find none,
either in surface views of specimens treated with silver
nitrate and other reagents, or in series of sections; in
fact, I should be greatly surprised to find such an aperture
at the time given by Keibel for its appearance. It
seems to me quite useless to look for it, since it is sup-
posed to represent a turning inwards, such as can be seen in
Amphibian ova, for the Mammalian embryo at this age is
utterly unlike the Amphibian embryo at the corresponding
age, and therefore the conditions which lead to the apparent
turning in as seen in Amphibia are hardly likely to be existent
in the case of the Mammalian embryo.

It may be well to point out that this is not the view of
so great an authority as Dr. Oscar Hertwig, who says,
_ p. 90 of his text-book, when speaking of this spot, “ Von
dieser Stelle aus, nehme ich an, hat sich schon auf einen
noch friiheren Stadium das untere Keimblatt durch Um-
schlag eines kleinen Bezirks der einblatterigen Keimblase
entwickelt.”

The changes that occur during the fifth day affect both inner
mass and outer layer.

The outer layer consists throughout the fifth day of flattened
lenticular cells, whose boundaries are easily recognisable in
surface views, and I have nothing to add to van Beneden’s
careful and accurate description of them. They are of very
various shapes and sizes, the lines which mark their boundaries
may be at any angle to each other. The boundaries may
number six, five, or four, the most usual form being the bexa-
gonal.

As the vesicle enlarges the cells become rather smaller, and
become much less regular in outline, so that by the seventh

EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 137

day the cells of the outer layer may be almost circular, or may
have any number of sides curved or straight.

Changes in the inner mass.—These are very important
during the fifth day.

Up to now the cells of the inner mass have been very uni-
form in character. During the process of flattening of the
mass (a stretching out as above suggested) the individual cells
have likewise become flattened, and each is somewhat lenticular
in shape (figs. 27 and 28).

The inner mass presents an approximately circular outline
when viewed from above or below, as in fig. 38, but here and
there single cells (HY. J.) seem to jut out somewhat beyond the
others. So we may say at this moment that the embryonic
disc is several layers of cell thick at the centre, but thins out
toward the periphery (figs. 26, 27).

The whole embryo is now spherical. Early on the fifth
day, somewhere, as a rule, between the 96th and 100th hours,
the cells which were noticed before as jutting slightly
from the sides of the “ inner mass ” may now be seen to be
quite separated from it, as in the diagram fig. 39, HY. L.,
standing out clear and round. These are not easily seen in
the perfectly fresh specimen, but are brought out beautifully
by certain reagents, as, for instance, Flemming’s strong solu-
tion, or a filtered mixture of one part Perenyi with one part
picro-carmine ; indeed, almost any reagent that stains slightly
the cells and not the albuminous layer.

At the same time the cells of the centre of the inner mass
have become more flattened, and can now be seen to form a
patch of cells, nowhere more than two cells thick. That is to
say, the embryonic area is composed of, in all, three layers of
cells, the outer a definite membrane continuous with the
general wall of the vesicle, and two other looser layers of cells
slightly flattened but much thicker and rounder than the cells
of the outer layer.

Beyond the periphery of this mass, a number of cells very
much rounder may be seen scattered irregularly over the inner
surface of the thin outer layer, extending over an arc of

138 RICHARD ASSHETON.

about 60° from the upper pole in all directions (fig. 28,
HY. @,):

From this moment the apparent changes for some hours
seem to concern the innermost layer of cells of the embryonic
disc, and the straggling cells.

The hypoblast, as a perfectly definite layer, is formed by the
time that the blastodermic vesicle measures ‘5 mm. in diameter,
that is, by about the 102nd hour after coition. It is not, how-
ever, as yet by any means a continuous membrane, it is a net-
work or a fenestrated membrane. For this reason in section
it appears to be represented by isolated cells lying beneath the
embryonic disc (v. fig. 29, HY.).

Certain cells can be detected in earlier stages which from
their being more lenticular, and more inwardly placed, can
probably be described as hypoblast cells (v. fig.28, HY.). This
stage is about the ninety-eighth to one hundredth hour, when
the diameter of the blastocyst measures about ‘36 mm. Also
all the apparently—and in many cases I believe actually—
isolated cells (HY. I. in figs. 28, 29, 87—40) can from the
moment of their apparent wandering be termed hypoblast.

By the separation of the cells round the periphery of the inner
mass, and certain others of the more inwardly placed of the
inner mass into what we may call hypoblast, there are left
cells in between the hypoblast and outer layer, which do
not show the same tendency to be flattened or drawn out as the
others. As is well known, this intermediate cell layer forms
part of the epiblast, and so may be termed the inner layer of
epiblast. To this I shall refer later.

I wish now to refer to the straggling cells (HY. J.) of the
inner mass, and consider why they apparently wander round
the inside of the blastodermic vesicle. That they arise from the
inner mass I have no doubt.

I have explained above the flattening of the inner mass as
due to being drawn out by the general expansion of the walls
of the vesicle, and to this may also be due the original isola-
tion of certain cells round the periphery of the inner mass as
seen in figs. 28, 37, and 38.

EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 139

If the cells of the inner mass do not multiply quickly enough,
and if they are not connected together firmly, and if they
adhere by some means or other to the outer layer, then, as
the outer layer expands, there must be a tendency for the
general separation of the cells of this inner mass, and this
will be most apparent at the edges. That is to say,
the centre, originally thicker than the edges, will become
thinner, and at the periphery the edges will be drawn out
to an irregular or regular fringe according to less or more
regular local growth and expansion of the outer epiblastic
layer.

If the spread of these is due to this cause, and if the growth
is equal over the whole surface of the sphere, then, in a section
taken through the centre of the sphere and the centre of the
embryonic disc, the are along which these isolated cells are
found should subtend an angle equal to that subtended by
the compact embryonic disc up till now; that is to say, an
angle of about 80°. But on measurement at a rather later
stage, 102 hours after coition, represented in fig. 89, the
angle is found to be very considerably wider than 80°, in fact
somewhere between 110° and 130° (the limit of these cells
in question being very irregular), and by the eighth day an
angle of 200°.

Hence it would seem that this does not account for the
apparent growth round. If, however, we have any reason to
suppose that the cells of the outer layer multiply more rapidly
and thus allow the expansion of that part of the sphere to take
place quicker around a zone bounded roughly by the edges of
the compact embryonic area on the one hand and some other
parallel line about the original equator on the other hand, we
could still consider that the suggested cause is the correct one.
I think we have, for these reasons :

The blastodermic vesicle of 96 hours is a sphere. The
blastodermic vesicle of 120 hours is no longer a sphere, but
is a body of such a nature that a horizontal plane taken
through its longest diameter will not pass through the equator,
but will be nearer the upper pole than the lower.

140 RICHARD ASSHETON.

The circumference of this plane and of all planes taken
parallel to it are, I think, as yet circles. There is nothing so far
to indicate which will be anterior or posteriorend of the embryo.

The blastodermic vesicle of about the 140th hour is in shape
of the same character, but more markedly so. .

In the blastodermic vesicle of the 175th hour sections taken
parallel to the equator are now no longer circles, but are
ovoidal. The fact that there are now a long and a short axis
may be entirely due to the pressure upon the vesicle exerted
by the walls of the uterus; but I want to point out more
especially that the horizontal sections are not true ellipses,
but have one end larger than the other. The large end corre-
sponds approximately, but by no means always exactly, with
the future posterior end of the embryo.

A median longitudinal vertical section of this stage is very
instructive, for the asymmetry is very marked. One end, the
future posterior end, is much more bulky than the anterior
end. The embryonic area is always placed nearer to the
anterior than to the posterior end.

Fig. 42 shows four stages, namely, at the 100th hour, the
125th, 140th, and 175th. A. is the more anterior, P. the more
posterior end. The thick black line in each case is the outer
layer of cells or epiblast; the thin black line represents those
cells of the inner mass which become separated to form the
hypoblast. It is quite clear that a very great change in shape
as well as size has occurred between the 96th and 168th hours.
How has this been produced ?

I have argued that there is a hydrostatic pressure within the
blastodermic vesicle causing it to expand. The pressure is
sufficient to cause the stretching and flattening out of the cells
of the wall of the vesicle, and also to cause the stretching and
expansion of the very tough albumen layer. On the other
hand, we know that the pressure is not so great as to rupture
either the vesicle wall or the albumen layer; in other words,
the vesicle wall and albumen layer are sufficiently strong to
resist the hydrostatic pressure for, at any rate, some con-
siderable time. The albumen layer becomes continually more

EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 141]

and more stretched, and quite late it does, as a matter of fact,
rupture, but at a time (the ninth day) that does not con-
cern the present question. No addition is ever made to the
thickness of the albumen layer after the embryo leaves the
Fallopian tube; it is an inanimate structure. The cellular
wall of the blastodermic vesicle is, on the contrary, living, and
capable of adding to itself at any part of its area. Like the
albumen layer, it becomes greatly stretched and becomes very
thin (v. figs. 21—29), and unless it received additional matter
(i.e. multiplication of the cellular units) it would rapidly thin
out altogether. After about the 100th hour the cellular wall
ceases to get any thinner. Up to this moment we must suppose
that the rate of increase of hydrostatic pressure has been in
excess of the rate of addition of material to the cellular wall of
the vesicle and so has stretched it, but from now there is no
appreciable thinning out of this cellular wall (v. figs. 27, 28,
29, and 34).

This means, I believe, that the increase of cellular tissue
just balances the increase of hydrostatic pressure, and so the
vesicle grows in size, the thickness of the cellular wall remaining
unaltered. Now there is no reason, as far as I can see, to
doubt the albumen layer being equally tough on all sides of
the embryo. The albumen layer is not secreted by the embryo,
but is applied by the Fallopian tube.

Although preserved specimens and sections are not well
adapted for this purpose, still my figures show that there is no
regularity at all in these preserved specimens, such as a thicker
part of the albumen layer being present over any one part of
the embryo in stages up to the 96th hour (v. figs. 16, 20, 21,
22, 23, and 24), and in the fresh specimens the outer and
inner limits of the albumen layer present in optical section
true concentric circles. Therefore I do not think we can
attribute any subsequent difference in thickness of the albumen
layer to a difference in texture acquired by that albumen layer
during its deposition. At a later stage we do find a difference
in thickness occurring as a constant character.

The difference I find is as follows:

142 RICHARD ASSHETON.

The most marked difference concerns the part of the albu-
men layer adjoining the embryonic disc. This is in the
later stages very much thicker than elsewhere. Figs. 30 and
31 are portions of the upper and lower poles of an embryo
taken from a rabbit of the one hundred and forty-fourth
hour.

The albumen layer is at the upper pole about twice as thick
as it is at the lower pole. This is a perfectly constant feature,
and becomes more and more marked the later the stage.

Hydrostatic pressure exercises its influence equally in all
directions. If it encounters less resistance in one direction
than in another it will cause greater effects in that one direc-
tion. I have argued above that at any one moment (between
one hundredth and one hundred and sixty-eighth hours) the
hydrostatic pressure within the vesicle on the one hand, and
the living cellular wall of the vesicle together with the non-
living albumen layer on the other hand, are in a state of
equilibrium. The blastodermic vesicle is always taut, but does
not rupture. But the next moment the hydrostatic pressure
has increased on the one hand, and the albumen layer has
become thinner and the cellular wall has increased its material
—and still the equilibrium is maintained. It seems to me to
be clear that the degree of rapidity of increase of the cellular
wall must be a factor in the resistance afforded to the hydro-
static pressure by the two walls of the vesicle.

If this is so, then it follows that at any area where the
increase of cellular tissue is greatest, there the hydrostatic
pressure will exert its greatest influence, and there the albumen
layer will, as a consequence, be thinnest. This is, of course,
dependent on the close attachment of the cellular layer to the
albumen layer. I am bound to confess I do not find it easy
to prove that there is no sliding of the albumen layer over the
cellular layer; but, on the other hand, I see no evidence to
suggest that there is such a sliding.

Accordingly, I take it that the great thickness of the albu-
men layer adjoining the embryonic disc over that adjoining
the lower pole shows that there has been less stretching and

EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 148

therefore less growth at the embryonic pole than at the lower
pole.

It may be said that this is due to the fact of the cellular
wall at this point being practically several layers thick. But
it must be remembered that the cells of the inner cell-mass
are very loosely arranged (figs. 28, 30, 35) at this time, and
certainly do not give me the impression of being under great
tension, as are the outer cells.

As far as the tension is concerned, I believe the outer layer
has to bear it very nearly all, even in the embryonic disc.
Sections show how very thin it is here.

Granted that they are only very equally stretched, it follows,
I think, since the albumen shows not the same amount of
stretching as elsewhere, that the multiplication of cells must
have been going on more slowly in the embryonic region.
Now from measurements of the albumen layer it will be seen
that the thinnest place is not usually at the lower pole, but
somewhere between the equator and the upper pole. This,
then, marks out as an area a zone of more rapid multiplication
of cells, that area which lies just outside of the embryonic
disc.

Now we know that upon the eighth and many subsequent
days there is a very great activity evinced by the cells of a
zone immediately surrounding the embryonic disc; it is the
zone called by Duval the ectoplacental area. If it is an area
of very great activity upon the tenth day, of great activity upon
the ninth and eighth days, when does it begin to be an area of
activity ? Why not upon the seventh or sixth, or even on the
fifth day ?

I believe that it does begin as early as the fifth day, and that
it is to the presence of this area of more active cell division
round the embryonic disc that the shape of the blastodermic
vesicle is as I have described it to be (v. fig. 42), and that it is
to this that the apparent growth round of the hypoblast cells
is due.

The ectoplacental area is, as is well known, much more
strongly developed all round the posterior end of the embryo

144, RICHARD ASSHETON.

than anteriorly. Although present to a slight extent at first
anteriorly, it never develops greatly. This means that the
zone of special activity is at a later stage more intense poste-
riorly than anteriorly. So apparently it is at an earlier stage,
and causes the greater bulkiness of the posterior end of the
blastodermic vesicle of the seventh day, and causes the appa-
rent throwing forward of the embryonic disc (v. “175 hr.,”
fig. 42).

Table of Measurements of Thickness of Albumen
Layer at Different Parts of the Vesicle in Mil-

limetres.
Five Days. Six Days Four Hours.
Embryonic area , ; ‘0060 : : 0058
Placental zone, 4. . : 0024 ; : "0022
Placental zone, P. . ; 0030 ; : "0024
Lower pole ; : : ‘0036 : 5 "0034

To sum up the above section:

(1) There is no growth round of the inner mass cells over
the surface of the outer layer cells in the sense of a migration.
It is only an apparent growth round produced by the more
rapid growth of a zone of the wall of the vesicle immediately
surrounding the embryonic disc, in which zone the marginal
cells of the inner mass lie.

(2) The presence of such a zone accounts in a great measure
for the shape assumed by the blastodermic vesicle during the
fifth, sixth, and seventh days.

(3) A zone of such a character undoubtedly exists during
the eighth, ninth, and following days, giving rise to the ecto-
placenta.

(4) Such a zone of activity accounts for the varying thick-
ness of certain parts of the albumen layer.

To return for a time to the question of the growth round of
the inner mass cells. It must be remembered that the cells
never completely surround the cavity of the blastodermic
vesicle. The lower pole is always, in the rabbit, one-layered
only as long as it exists.

EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 145

There is only one other alternative that will account for the
apparent growth round of the inner mass on the inner wall
of the blastodermic vesicle, as far as I can see, and that is
actual active migration of the cells in question. Of course it
is difficult to bring evidence to show that they have not
migrated, since it is not possible, I fear, to follow the process
in any one and the same specimen. At the same time I cannot
find any evidence to prove that they have actually migrated.

If a cell does migrate, like an Ameceba for instance, one
would expect to find evidence of protoplasmic protuberances or
pseudopodia. Of this I can find no trace. The majority
of the cells seem at first (figs. 38, 39) to be quite isolated
from each other, and to be approximately spherical, whether
examined in the perfectly fresh condition or after treat-
ment with various reagents. They are, it is true, slightly
flattened on the side by which they adhere to the vesicle wall
(fig. 28, HY. J.).

Certain of the cells here and there are connected by threads
of protoplasm, but this, I think, is not a sign of pseudopodic
activity, but merely indicates the final stage in division between
the two cells. I have no doubt that these cells divide rapidly
after a time, though I do not think much activity of division
takes place during the first few hours after the apparent
migration begins.

If one of these inner rounded cells is undergoing the process
of division, then, as the wall on which it rests expands, the two
dividing halves of the inner cell will be pulled apart, and a
strand of protoplasm connecting the two cells may remain for
some time.

Of course it is just possible, I suppose, that these rounded
inner cells might migrate by means of a rolling motion con-
sequent on ‘*‘ streams” of protoplasm within them, as do some
protozoa. But is such a phenomenon known anywhere in the
metazoan body? I cannot think we are justified in assuming
this without evidence, for when examined in the fresh con-
dition no such protoplasmic activity can I notice.

When the cells are so isolated as I believe them to be during

VOL. 37, PART 2,—NEW SER. K

146 RICHARD ASSHETON.

the 98th to 108th hours or thereabouts in the rabbit, I cannot
conceive that the migration can be accomplished otherwise
than by an actual migration, for which there is no evidence, or
else by a process such as I have attempted to describe above.
This I believe to exist. If the inner layer was not composed of
isolated cells, but was a compact membrane, then it might creep
round by means of its own interstitial growth, although I do
not think it would in that case be a thin smooth membrane.

Such is the account of the growth round of the hypoblast in
the mole. Heape makes no mention of any isolated cells at
the edges; he says of the hypoblast, ‘‘it extends laterally by
virtue of the multiplication of its cells, which at the same time
become much flattened.” It seems to me to be much more
likely that the layer would be flattened if they were drawn out
aloug with the epiblast cells ; but I have no evidence at present
whether the same cause can be attributed to the spread of the
hypoblast in the mole as I suggest for the rabbit.

In fig. 42 the dotted lines radiating from the centre of the
smallest blastocyst indicate the amount of growth of the
several segments up to the ove hundred and seventy-fifth hour.

Until the one hundredth hour I imagine the growth of all
parts of the wall of the blastocyst to be equal.

From that moment there is a greater activity in an area
around the embryonic disc, which causes the inner layer to be
apparently carried further round the anterior of the blastocyst.

This zone of activity by the one hundred and fortieth hour
has become more marked still, and now shows itself to be more
intense posteriorly than anteriorly, which latter character is
plainer still at the one hundred and seventy-fifth hour.

Eventually the activity culminates, upon the addition of
resistance afforded by the walls of the uterus, in the production
of the ectoplacental area, or part of it.

To this I have referred in another paper, and also to the
hypothesis that this zone of activity, in the absence of a tough
albuminous coat, results in certain other rodents in the pro-
duction at once of a heaping up of cells—the trager.

In the embryonic disc region there is very much less activity

EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 147

exhibited. The outer layer in that region is very much
attenuated, and shows very little sign of activity. The inner
layer consists all this time of rounded or ellipsoidal cells, and,
like the outer layer of epiblast of that region, shows very little
if any sign of activity.

So little activity of division does there seem to be in the
inner layer of epiblast, that there is a distinct tendency for the
several cells to become slightly separated as in fig. 30, which
gives rise to the very irregular and speckled appearance of the
embryonic disc of the one hundred and twentieth hour.

Very probably this is caused by the slight stretching of this
region. It is more noticeable at the edges than towards the
centre.

Whether there is any palingenetic meaning in this double-
layered condition of the epiblast I have discussed in another
paper. For the present, I think the right view to take of the
condition is that derived from the study of the actual way in
which the separation has originated, and to regard it as a con-
sequence of ontogenetic circumstances only.

The Outer Layer of Epiblast.—This is by far the most
active of the embryonic layers of the fifth day. It is in an
active condition of growth during the whole of the day, and
thereby allows of the expansion of the vesicle. The character
of the cells seems to be just as it was during the latter part
of the fifth day.

The Inner Layer of Epiblast.—This layer seems to be
a region of rest throughout the whole of the sixth day. There
is very little sign of cell multiplication. The cells are more
or less circular in outline when viewed from above, and
oval when seen laterally. They are rather scattered, and
thus give rise to the speckled appearance noted above. They
seem to be clearly separated from the outer epiblast above
them and from the hypoblast below them, and since they
are either quite separated from each other, or connected
only by fine strands of protoplasm at certain minute spots,
they are simply pulled apart by the expansion of the blasto-
dermic vesicle, and are not individually stretched and flattened,

148 RICHARD ASSHETON.

as are the cells of the outer epiblast and hypoblast, where they
are more intimately connected one with another. In the latter
case, that of the hypoblast, cells with a tendency to the features
characteristic of both layers of epiblast are to be found.

CHAPTER IV.

CHANGES THAT OCCUR DURING THE SixtH Day (120TH To

1447H Hours).

If we may call any period of the development of an animal
unimportant, it is to the period between the 120th and 144th
hours of the development of the rabbit that this epithet might
be applied.

During this day the blastodermic vesicle increases very
greatly in size, and assumes very markedly the shape described
in the last chapter as being characteristic of the later stages of
the development of the vesicle, prior to its attachment to the
walls of the uterus.

The blastodermic vesicle is no longer a sphere. It will be
found by measurement to have longer and shorter equatorial
axes, and a polar axis which is of less length than the shortest
equatorial axis.

The following measurements are taken from specimens of the
earlier part of the sixth day :

L t Short 3

equatorial equatorial Polar axis. pt re

axis. -
No.1. 5 days 5 hours . .| l'1 mm. | 1:05 mm. | 1:00 mm. | 0°55 mm.
B02. Bois Beek heel BB gy |) 088) Re 0 -vee

(1) Hypoblast of the Embryonic Disc.—This is now a
continuous layer, sufficiently so as to show lines of demarcation
between the cells when treated with silver nitrate. This con-
tinuous membrane extends a short distance beyond the peri-
phery of the inner epiblast layer. The cells composing this
membrane are completely flattened.

(2) Hypoblast beyond the Embryonic Disc.—The

EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 149

scattered hypoblast cells have now become much more numer-
ous, and are scattered more evenly over the portion of the wall
upon which they are to be found. Many of them, possibly
all of them, are now undoubtedly connected by more or less
fine protoplasmic threads.

These scattered cells, although such conspicuous objects
during the fifth day, are now extremely difficult to make out,
and can very easily escape notice.

They are more numerous now nearer to the embryonic disc,
and merge gradually into the continuous layer just described.
They are now very much more flattened.

Towards the line of their outer limit they present more the
characters of the former day, being fewer and rather rounded
and more isolated.

Consideration of the Extent to which the Shape
of the Cells of the several Layers may be attri-
buted to Mechanical Causes.

At this time we find cells of two very different types, with
cells showing all intermediate stages between the two.

The first type is the rounded, almost completely isolated cell,
such as those of the inner layer of epiblast, or near the outer
limit of the hypoblast ; the second type is the flattened or
stretched cell of the outer layer of epiblast or embryonic hypo-
blast, continuous with its neighbours around all its edges; and
thirdly, forms of cells intermediate between these two types.
How far can we hold this difference of form to be due to the
environment of the individual cell apart from its own inherited
tendencies ?

Whether there is any cell in the embryo at this time quite
separated from all others I am not certain. Of course they are
all contiguous to one or more other cells, but possibly there is
an actual protoplasmic union between one cell and another.
This is certainly the case very frequently with the cells of
the hypoblast, which at first sight seem to be quite separate.

In sections, and in surface views, there is no distinct connec-
tion to be made out between the rounded cells of the inner

150 RIOHARD ASSHETON.

epiblast layer. But when the embryonic disc is broken up in
such a way as to scatter and tear apart the various cells, then
there undoubtedly appear to be strands passing from some of
these rounded cells to others of the same layer. At the same
time I cannot say positively whether these strands are really
connections between the cells in question, or whether they are
fragments and shreds derived from the tearing apart of the
hypoblast or outer epiblast layer, between which they are
placed. But however this may be, the cells of the inner epi-
blast layer are, at the time I am speaking of, either isolated or
else connected only at certain spots of small area. These are
of the rounded type.

At the outer limit of the hypoblast there are also cells, some
of which, I believe, may be quite isolated; others are connected
to each other by a few strands of protoplasm. These approach
very closely to the rounded type of cell. This is the type of
cell which I believe to be the most natural, by which I mean
the least influenced by its environment. This is the type of
cell which first comes into existence in the segmentation of the
ovum, when, within the protecting investments, the cells, at first
uninfluenced by pressure or tension from without, or from each
other, assume their natural or spherical contour. As the seg-
mentation proceeds, the inner segments become the more com-
pressed, and assume polygonal forms.

After the establishment of the cavity of the blastodermic
vesicle, the outer cells, by the pressure from within increasing
more rapidly than their rate of multiplication, are drawn out
into thin plate-like cells.

These outer layer cells are from the first connected with each
other by their edges, and form a continuous membrane, a
condition without which in all probability the formation and
enlargement of the blastodermic vesicle could not be produced.

As long as this tension within is maintained at arate greater
than the rate of multiplication of the cells, the cells retain
their flattened condition.

What of the inner mass cells? Upon the removal of the
pressure of the outer layer, the more outwardly placed of the

EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 151

inner mass cells re-assume, at any rate on their free surfaces,
the rounded contour which is natural to them.

As the blastodermic vesicle expands, the inner mass, which
is adherent to the wall of the vesicle either by actual proto-
plasmic connections or otherwise, is drawn out into a lenticular
shape.

I have tried to show that there is a zone of the wall of the
vesicle which, by greater activity of multiplication of cells,
admits of more rapid expansion of that part. Upon this zone
rests the edge of the lenticular inner mass. The expansion of
the zone is in direction radially from the embryonic poles.
Hence the outermost cells of the inner mass, i.e. the cells at
the edge of the lenticular mass, will tend to be separated more
rapidly than the innermost. This will tend to isolate these
cells from others of the inner mass. Let us suppose that all
the cells of the hypoblast layer are dividing at a uniform rate.
I think it is reasonable to suppose that the existence of strands
connecting cells of this kind together have their origin in
past cell divisions. Accordingly on this supposition the con-
necting strands will be more numerous and the nuclei nearer
together, and the meshes of smaller area in the embryonic disc
hypoblast than in the hypoblast outside that area. The hypo-
blast cells of the embryonic area will differ from those of the
extra-embryonic area in this way :

(i) The embryonic cells will have more and shorter, and so
presumably stronger, strands connecting them with their neigh-
bours than will the extra-embryonic hypoblast cells.

(ii) The embryonic hypoblast cells will have connecting
strands upon all sides, whereas the outermost extra-embryonic
cells will have them upon one side only.

Is it possible for the flattening of the embryonic hypoblast
cells to be due to their becoming stretched by the tension
produced by the extra-embryonic hypoblast cells (with which
they are in direct connection by means of the strands just men-
tioned) being removed in all directions by the rapidly expanding
zone of the outer epiblast? If so, it is possible to account for
all the shapes of the cells composing the embryo at this age.

152 RICHARD ASSHETON.

As I have stated on a previous page, the hypoblast of the
embryonic area is a network at first. Also, I believe that
at first many of the outermost cells of the extra-embryonic
area are really isolated. These will be under less tension than
those near the embryonic pole, as they will, if they are connected
at all with other hypoblast cells, have connecting strands upon
one edge only. Hence these preserve for a longer period their
rounded contour.

The ultimate conversion of the isolated cells into a network
(or a series of networks) and of the networks into thin
continuous membranes, and of thin continuous membranes into
columnar membranes, would seem, therefore, to be but the
result of increase of rate of multiplication over rate of
expansion.

The inner layer of epiblast cannot be said to have come into
existence as a layer until after the formation of the hypoblast.
Until that moment it formed, together with the future hypo-
blast, the inner mass. It was impossible, except in as far as
could he premised from their position, to say from their
characters which would be inner epiblast cells and which hypo-
blast cells (v. fig. 28). What I believe takes place is this. Those
cells of the lenticular inner mass which, being at its edges, are
removed by the expansion of the wall of the vesicle, and those
which are in direct connection with the cells so removed,
become by virtue of their position the future hypoblast ; the
remainder become the inner layer of epiblast. That is to say,
those cells of the inner mass which are not influenced by the
expansion of the vesicle, as above described, and are accordingly
upon that part of the wall, though not actually as yet part of it,
which is least affected by the expanding forces, become the
inner epiblastic layer.

Of all the cells, therefore, in the embryo at this time, these
(the inner epiblast layer) are least affected by external causes.
These cells are the more free to assume their natural shape,
which I believe to be spherical, and are only slightly flattened
between the two layers, outer epiblast and hypoblast.

EARLY STAGES OF DEVELOPMENT OF THE RABBIT, 153

CHAPTER V.

CHANGES THAT OCCUR DURING THE SevenTH Day (145TH To
168tH Hovrs).

The Fate of the Outer Layer of Epiblast (or
Rauber’s Layer) in the Embryonic Disc.

The embryos have now grown to such a size as to cause
them to respond more effectually to the impulses set up by
contractile movements of the muscular walls of the uterus,
and therefore we find them much further advanced along
the uterine tube, and more scattered. They have not, how-
ever, taken up a permanent position as yet, for although this
may occur in some cases during the last few hours of the seventh
day, more usually it does not take place until the early part of
the eighth day.

It will be best to describe the course of events in the three
layers separately.

Outer Layer of Epiblast.—Very little need be said of the
greater part of this layer, no change except such as has been de-
scribed as occurring during the fifth and sixth days takes place.
But special attention must be given to that part of the area which
lies over the patch of inner layer of epiblast, i.e. embryonic
disc.

Inner Layer of Epiblast.— During the early part of the
seventh day the cells of this layer are just as described in
the preceding chapter. They extend now over an area of
about °6 mm. The general outline of the mass is still circular.
Each cell is distinct and rounded, with very large nucleus ; and
with nearly all stains that I have used they stain only lightly.

Early on the seventh day these cells show signs of greatly
increased activity. They multiply, become pressed together,
and now form a very compact layer at the same time as certain
changes occur in the outer layer of epiblast.

The course taken by these two layers during the next few
hours, and its significance, have been very differently described

154 RICHARD ASSHETON.

by different authors; indeed, very opposite views have been
held during many years. It is an interesting question from
the extreme difficulty of the investigation and from the morpho-
logical problems connected with its solution. As regards the
actual facts, there have been three quite distinct accounts.
Most observers have noticed three layers at this stage: (1) an
outer thin layer; (2) a middle thick layer; (3) an inner thin
layer. These accounts are very briefly as follows:

(1) Van Beneden maintained that the inuer layer is not
epiblast at all, but is mesoblast, and that the outer layer
becomes thickened over the embryonic disc area, and gives rise
by itself to the epiblast of the embryo.

(2) Rauber, Lieberkuhn, and later Kolliker and others,
hold that the outer layer is quite transitory, and that during
the seventh day it splits up, degenerates, and disappears
entirely, taking no part in the epiblast formation of the embryo.
The epiblast, they hold, is wholly derived from the inner layer
of ovoid cells, which van Beneden took to be mesoblast.

(3) Balfour and Heape contend that both layers persist as
the epiblast of the embryo, but the two layers fuse together
by the growth downwards of the outer layer cells in amongst
the inner layer cells.

There cau be no doubt now that van Beneden was wrong.
The middle layer certainly forms part if not all of the perma-
nent epiblast, and this is acknowledged by van Beneden himself
(‘ Arch. Biologie,’ vol. v).

It requires some careful examination to determine whether
Rauber, Lieberkuhn, and Kolliker, on the one hand, or Bal-
four and Heape, on the other, were right, or whether all were
wrong.

The surface views given by van Beneden (pls. v and vi) and
Kolliker (pl. ii, figs. 13, 15, 16, and 19) are, I think, quite
correct. The interpretation van Beneden put upon his figure
he no doubt now admits to be wrong. The interpretation put
upon them by Kolliker I believe to be correct, namely, that the
large areas are cells belonging to the outer epiblast layer, and
the smaller ones are cells belonging to the inner epiblast layer. I

EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 155

have drawn a figure (fig. 36, Pl. 17) from a specimen aged six
days five hours. This is very similar to Kolliker’s aged six
days. It isa silver nitrate specimen. This was drawn when
the extreme upper surface was in focus. On focussing down it
is possible to find small cell outlines under the large areas, but
finer than those of the cells (HP. F.). On focussing still further
down, the outlines of the hypoblast layer are visible as very
fine, very wavy lines, forming as a rule more regular areas.
The small cells (EP. I.) are present throughout the embryonic
disc, but in some places are much more marked than others;
that is to say, in some places the silver nitrate produces the
characteristic black marks between the cells more strongly
than in others. For this there must be a reason, and this
seems to be that in some parts the small cells are at the
surface (#P. J.); at others (HP. O.) they are covered by some
other body, this other body being certain cells of the outer
layer of epiblast (Rauber cells).

This is an intermediate stage. LHarlier the outer layer is
quite continuous over the whole of the embryonic disc, as
shown in Kolliker’s figs. 11 and 12, pl. 11 (see also my sections
of these stages, figs. 27, 28, and 30). In these the outlines of
the outer layer are distinct as large polygonal areas. In the
earliersstages the outlines of the inner layer of epiblast are
not sharply defined in silver nitrate specimens, apparently
because they are but loosely arranged. But in Kolliker’s figs.
11 and 12, which represent specimens at the latter end of the
sixth or beginning of the seventh day, the small cells, i. e. the
inner layer of epiblast which now forms a more compact layer,
show outlines which are visible below the outer layer cells,
although faint.

I entirely agree with Kolliker that this outer layer (Rauber’s
layer) now becomes broken and, as it were, torn up into
isolated cells or little patches of cells, and I think that the
cause of this may be as follows. I have above given reasons
for supposing that the cells of the embryonic disc region are
comparatively inactive during the fifth day and first part of
the sixth day. The outer layer of cells (Rauber’s layer), I

156 RICHARD ASSHETON.

pointed out, has been stretched to a very high degree of
tenuity.

From the middle of the sixth day the embryonic disc area
becomes an area of increased activity.

We have at this moment, therefore, an outer extremely
attenuated membrane under high tension due to the hydro-
static pressure within. Each “ cell’’ of this membrane repre-
sents an individual minute centre of protoplasmic activity.
These are so drawn out that a few will extend over a con-
siderable area.

Closely pressed against these is a layer of rounded or but
slightly compressed cells, the inner epiblastic layer. Hach cell
of this layer also represents an individual minute centre of
protoplasmic activity, and it will be seen that in any given
area of the embryonic disc for one of these little centres of
activity of the outer layer (Rauber’s) there are three, four, or
five of the little centres of activity in the inner epiblastic
layer.

The hypoblast, although it will tend to press the inner
layer of epiblast cells firmly against the Rauber layer, need
not necessarily be affected by what I am going to describe,
because at no time does the hypoblast, after its initial separa-
tion from the epiblast, appear to be at all firmly attached to
the epiblast. Note the readiness in which it stands away
from the epiblast in sections (v. all my figures, and those of
Kolliker).

To return to the two epiblast layers. It is clear now that
if the two epiblast layers of the embryonic disc acquire an in-
creased activity, then if for any given area the inner epiblast
contains, say, three times more cells than the outer epiblast,
then the inner epiblast will increase its bulk, roughly speaking,
three times more rapidly than the outer epiblast.

This might produce several effects. It might produce a
heaping up of cells; it might produce an arching inwards of the
inner layer of epiblast ; it might produce an extension of the
inner epiblast by sliding over the outer epiblast (or rather be-
tween the outer epiblast and hypoblast) ; or it might cause the

EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 157

further stretching of the outer epiblast, or it might cause its
rupture.

I believe it causes the rupture of the Rauber layer ; whether
the several fractures between the cells are sharp, or whether, as
is possible, the cells are pulled apart so as to produce large
meshes through which the inner layer cells come to the sur-
face, I cannot say. If the latter case, although the Rauber
layer as a membrane would be obliterated, still the continuity
of protoplasm would not be broken ; but I think, from considera-
tion of all the appearances, the fractures are fairly clean.

We now come to the question of the fate of the Rauber layer
cells. Kolliker says they disintegrate and disappear; Balfour
and Heape say they pass into the inner layer of epiblast. I
am fairly confident that the latter view is correct.

If my description and hypothesis are correct, the Rauber
cells, as soon as they become broken up, are no longer under
such a high degree of tension. If they have any vitality in
them they will, as they grow, be able to assume other shapes
than ‘‘squamous.” They will become intermingled with the
cells of the inner epiblast layer. Their nuclei will always form
a knob, and thereby perform the part of the thin edge of a
wedge to the whole cell, and so on recovering what we must
allow was their more normal shape, lost only under pressure of
the fluid within, will tend to become intermingled with the cells
of the lower layer. No doubt some will get settled earlier than
others, and often upon the eighth day it is possible still to
detect a Rauber cell not yet accommodated. Many stages of
this process, I believe, are shown in my figs. 32, 33, and 84.
These are sections from the same specimen, which was taken
from a rabbit at the 153rd hour, preserved in Flemming’s strong
osmic-aceto-chromic fluid, and stained with a weak borax car-
mine. In fig. 34 to the left is the thin layer of outer epiblast
(formerly continuous with Rauber’s layer alone) (EP. O.),
lying between the albumen layer (ALB.) and hypoblast (HY.).
To the right is the fused or fusing inner epiblast and “ Rauber
cells” or outer layer of epiblast. The greater part is stained
only slightly, and contains numerous large rounded nuclei

158 RICHARD ASSHETON.

(EP. I.). The outline of cells is hardly visible. Here and
there, as though filling up interstices, are cells which stain
much more darkly (HP. O.), and in the character of their
nuclei appear more like the cells of the outer epiblast beyond
the border of the embryonic disc. Some of these are only
little wedge-shaped bodies on the surface (fig. 32, EP. O.),
others pass right through; others seem to send flaps over the
surface of the inner layer cells (fig. 833, HP. OR.). The last-
named cell (fig. 33, HP. OR.) is very curious. It will be
noticed there are several rather like it. These almost look as
though they were being enclosed quite accidentally by the
inner layer of cells. It seems to me quite possible for this to
be very often the case. Take such a spot as that to which the
line EP. O. runs in fig. 80; here, of course, the bend in the
outer layer is in all probability artificial. But if a gap existed,
as frequently happens, between the cells of the inner epiblast,
then if at a neighbouring spot the outer layer became ruptured,
the tension would be removed, and the Rauber cell might very
well become enfolded quite passively, as I believe is taking
place in fig. 33, EP. O.R. There is no reason why a cell
thus enfolded should die, in all probability it would grow and
multiply like any other cell of the region.

(It may be observed that of all the cells in the whole blasto-
dermic vesicle at this time, none are so badly placed for nutri-
ment [provided we allow that nutriment is all the time being
received from the fluids of the uterus] as the cells of the outer
epiblast layer over the embryonic disc. For they here lie be-
tween the thickest part of the albumen layer and the thick
inner epiblast layer, which may account for the period of in-
activity of this layer at this moment.)

I cannot leave this question without a short discussion of
the morphological bearings of the events connected with the
fusion of these two layers.

First of all, I wish to call attention to Heape’s description
of the mole’s development. The mole is very like the rabbit
in its developmental history.

Just after the separation of the hypoblast layer, Heape

EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 159

describes the appearance of a cavity between the outer layer of
epiblast and the inner layer of epiblast. It is hardly a cavity,
for it is partially filled by very “stellate” cells. With the
formation of this the inner layer of epiblast becomes arched
inwards, a process termed “ temporary inversion of layer ”’ by
Heape. Subsequently this arch flattens out again, and it and
the outer layer cells and stellate cells between them all fuse
together to form the permanent epiblast.

I have not studied the mole, but from Heape’s description
this seems to be an almost exact parallel to the process which
occurs in the rabbit, with this exception: whereas in the
rabbit the increase in activity of the inner layer of epiblast
gives rise to a rupture of the outer layer of epiblast, in the
mole one of the alternatives suggested above is taken, and the
inner epiblast bulges inwards, leaving a loose space threaded
across with “stellate ” cells between it and the zona radiata.
Subsequently on the expansion of the whole blastodermic
vesicle the plate flattens out again and the “stellate” cells
and other outer layer cells become intermingled with those
of the plate and form the permanent epiblast. Heape
regards this as a kind of inversion, and the stellate cells as
oWirager.’”

I cannot agree with Heape in considering the stellate cells
he mentions as being equivalent to trager cells, and certainly I
do not think that the “ Rauber cells ” of the rabbit are in any
way connected with “ trager””—but of this I shall say more in
a later paper.

As regards the meaning of the fusion of the two layers, I do
not see that it need necessarily have any morphological signi-
ficance at all. It may be merely an accident of development.
At the same time I cannot entirely neglect certain occurrences
in another group of Vertebrates, and have discussed them in
another paper.

By the union of the two layers the embryonic dise acquires
a very much more distinct outline, which is now practically
circular; its outline is considerably more regular than before
the junction just described has taken place.

160 RICHARD ASSHETON.

Hypoblast.—(1) Hypoblast of embryonic area. This
seems to have become in very slight measure changed. It is
now undoubtedly a continuous membrane in the region of the
embryonic area. This condition seems to extend a distance
from the embryonic area equal to about the diameter of the
embryonic area, beyond which it becomes a network and passes
insensibly into—

(2) The straggling cell portion of hypoblast. This part of
the hypoblastic layer retains its irregularly scattered condition
of the sixth day, but certain features may be remarked upon
which were unnoticed before.

The cells are more thickly scattered about; they are more
irregular, having entirely lost their rounded form, and are more
flattened. Many in all parts may be seen to be connected to-
gether by fine filamentous strands, not only in the close
proximity of the embryonic hypoblast, but also near its
periphery.

Again, the outer limit of this zone is much more marked, and
is, in fact, now rendered very plain indeed. It forms a well-
marked edge, very irregular it is true, but an almost if not
quite continuous edge. Along this edge the cells are slightly
crowded, and rather elongated in the equatorial plane of the
vesicle. What I mean may be made out from fig. 40, and an
idea of the general history of events connected with the deve-
lopment of this part of the hypoblastic layer may be derived
from the four figures 37—40.

The extent of area covered by the two parts of the hypoblast
is now rather more than half the whole area of the inside of
the wall of the blastodermic vesicle.

Another point of interest may be noticed. There is a strong
tendency for this line of limit to be thrown into small folds or
waves, as shown in fig. 40, Pl. 17.

These two characters may be observed thoughout the area
under discussion. If an embryo of this age is cut in two,
and the cut edge examined under a high power, these characters
are seen very clearly. That is to say, the cells forming this
limiting line are themselves rather more rounded or ‘ hog-

EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 161

backed,” and the joining strands are curved and even arch
away from the epiblast, and in some cases undoubtedly the
cells themselves seem to stand away from the surface.

It is, however, quite possible that the sinuosity of the line and
arching away may be the result of reagents, as I have not
examined this edge in a fresh specimen with success,

EXPLANATION OF PLATES 138—17,

Illustrating Mr. Richard Assheton’s paper, “ A Re-investigation
into the Early Stages of the Development of the Rabbit.”

List oF REFERENCE LETTERS.

A. Anterior end of blastodermic vesicle. 4ZB. Albuminous layer acquired
in Fallopian tube. C. BL. Cavity of blastodermic vesicle. HM, D. Embryonic
disc. #P.J. Inner layer of epiblast. 2#P.O. Outer layer of epiblast.
EP. OR. Cell of outer layer of epiblast becoming “accidentally ” included
in the inner layer of epiblast. HY. Hypoblast of embryonic disc. AY, J.
Hypoblast of region beyond embryonic disc. J. /. Inner mass of cells of
blastodermic vesicle. JZ. Larger of the first two segments. JZ?. Supposed
second generation of layer of first two segments. JZ*. Supposed third genera-
tion of layer of first two segments. O.Z. Outer layer of cells of blastodermic
vesicle. P. Posterior end of blastodermic vesicle. 2. B. Polar body acci-
dentally enclosed. PZ, 8. Sinuous protoplasmic junction between two hypo-
blast cells. S. Smaller of the first two segments. S?. Supposed second
generation of smaller of first two segments. §°%. Supposed third generation
of smaller of first two segments. 2., 2’. Deceptive appearances suggesting a
van Beneden blastopore. Z. Zona radiata.

All the figures excepting Figs. 37, 38, and 41 have been drawn with the
help of a camera.

All those figures of which the magnification is 465 times were drawn with
Powell and Lealand’s 4 apochromatic oil immersion ; the others with Zeiss or
Reichert’s lenses.

VOL. 37, PART 2,—NEW SER. L

162 RICHARD ASSHETON.

PLATE 13.

Fic. 1.—Fertilised ovum. Rabbit 243 hours. x 165.

Fie. 2.—Ovum in two segments, from same rabbit as above. 243 hours.
x 165.

Fic. 3.—Ovum in two segments, showing great difference in size, from
same rabbit as above. 243 hours. xX 165.

Fic. 4.—Ovum in two segments, from rabbit 24 hours. Polar bodies sepa-
rated. x 165.

Fie. 5.—Ovum in two segments, from rabbit 253 hours; drawn after mount-
ing in Canada balsam. xX 165.

Fic. 6.—Embryo in four segments, from same as preceding. 254 hours.
x 165.

Fic. 7.—Embryo in five segments. 273 hours. xX 165.
Fie. 8.—Embryo in eight segments. x 165.

Fic. 9.—Embryo in eight segments. x 165.

Fic. 10.—Embryo in seven segments. xX 165.

Fig. 11.—Embryo in eight segments, showing internally-placed polar
body. 89 hours. x 165.

Fic. 12.—Embryo of 47th hour, showing contrast in size of the several
segments (seventeen segments). X 165.

Fic. 13.—Isolated segments of same specimen (47th hour). x 165.
Fic. 14.—Embryo (27% hours) in seven segments.

PLATE 14.

Fie, 15.—Embryo (663 hours), showing great difference in size of seg-
ments,

Fic. 16.—Section of a specimen preserved in silver nitrate } per cent.;
stained picro-carmine, and cut in paraffin, 72 hours. x 465.

Fic. 17.—Embryo, showing contrast in size of segments.

Fic. 1§.—Section of embryo (72 hours) preserved in Perenyi; stain, borax
carmine.

Fie. 19.—Another section of the same embryo. Both showing what might
be a van Beneden blastopore, but on opposite sides!

Fie. 20.—Section through rabbit embryo of the 47th hour. Cut in paraffin.

Preserved 4 per cent. silver nitrate 2 minims, water and sunlight + hour; no
other stain. xX 465,

EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 163

PLATE 15.

Fic. 21.—Section through rabbit embryo of the 77th hour. Removed
from Fallopian tube. Preserved in Perenyi; stained borax carmine; cut in
paraffin. This seems to be an unusually large specimen. x 465.

Fie. 22.—Section through rabbit embryo of the 80th hour. Removed
from uterus. Preserved in Perenyi, and stained in borax carmine; cut in
paraffin. x 465.

Fie. 23.—Section through rabbit embryo of the 83rd hour. Removed from
uterus. Preserved in Perenyi, and stained in borax carmine; cut in paraffin.
xX 465.

Fic. 24.—Section through rabbit embryo of the 80th hour. Removed from
the same uterus as Fig.22. Preserved in Perenyi, and stained picro-carmine ;
cut in paraffin. x 465.

Fie. 25.—Section of embryo of rabbit of the 80th hour. Taken from uterus
(same as Figs. 22 and 24). Preserved in Perenyi, and stained in borax
carmine, and cut in paraffin. x 465.

PLATE 16.

Fre. 26.—Section of the embryonic disc of rabbit of 96th hour. Taken from
the uterus and preserved in Perenyi, and stained in aniline blue black ; cut in
‘paraffin. x 465.

Fie. 27.—Section of embryonic disc of rabbit of 96th hour. Taken from
the uterus. Preserved in silver nitrate 4 per cent. 25 minims; water and
sunlight 2 hours ; alcohol; borax carmine; cut in paraffin. x 465.

Fie. 28.—Section of the embryonic dise of a rabbit of the 100th hour.
Taken from the uterus, and preserved in Perenyi. Stained in borax carmine,
and cut in paraffin. x 465. ,

Fie, 29.—Section through the embryonic disc of a rabbit embryo of the
103rd hour. Preserved in Perenyi; stained borax carmine. X 465.

Fic. 30.—Section through a portion of the embryonic dise of a rabbit
embryo of about the 140th hour. Preserved in Perenyi; stained picro-
carmine. The hypoblast is now nearly a perfect membrane in the embryonic
area. xX 465.

Fic. 31.—A portion of the lower pole of the same section, x 465.

Fies. 32, 33, and 34.—Portions of sections of the embryonic disc of a
rabbit of the 153rd hour. Preserved Flemming strong formula; stained in
borax carmine. The fusion of layers is taking place. x 465.

164 RICHARD ASSHETON,

Fic. 35.—Portion of a section of the embryonic dise of a rabbit of the
125th hour. The hypoblast here is seen as consisting of cells with very con-
spicuous central portion, but these are connected by fine and apparently dis-
continuous strands—really a network. x 465.

PLATE 17.

Fic. 36.—Surface view of a portion of the embryonic area of a rabbit of
about the 150th hour. Prepared with nitrate of silver. This corresponds to
the sections Figs. 32—34. The shading is diagrammatic, and represents only
my interpretation of the areas as defined by the silver nitrate lines, which are
drawn by camera.

Fic. 37.—A diagram of the embryonic area of the rabbit of the 96th hour.
Fie. 38.—A diagram of the embryonic area of the rabbit of the 100th hour.

Fic. 39.—A surface view of a portion of the wall of the blastodermic
vesicle of the rabbit of the 103rd hour. Seen from within. Preserved
Flemming; stained borax carmine. Xx 350.

Fic. 40.—A portion of the wall of the vesicle, showing the termination of
the inner layer at the 144th—150th hours. x 350.

Fic. 41.—A diagram to show the constancy of the angle subtended by the
inner mass during the earlier stages of the growth of the blastodermic vesicle.

Fie. 42.—Camera drawings of the blastodermic vesicles of a rabbit of the
100th, 125th, 140th, and 175th hours, as seen from the side, arranged so as
to show the supposed variation in rate of expansion of the different zones
under the influence of—(1) increase of hydrostatic pressure within; (2) resist-
ance (mechanical) of cellular wall of vesicle and albuminous layer, which are
supposed to be nearly constant factors ; and (3) the vital activity of the cellular
wall of the vesicle, which by hypothesis is supposed to vary along certain
zones and areas.

FUSION OF EPIBLASTIC LAYERS IN RABBIT AND FROG. 165

On the Phenomenon of the Fusion of the
Epiblastic Layers in the Rabbit and in
the Frog.

By

Richard Assheton, M.A.

With Plate 18.

In my paper upon the early stages of the development of the
rabbit I have given evidence in support of the views held by
Balfour and Heape concerning the fate of the outer layer of
epiblast over the embryonic disc of the rabbit embryo of the
seventh day. The two layers of epiblast gradually fuse together,
and cells from each take part in the formation of the permanent
epiblast.

In the description I have given of the process, and in the
attempt I have made to explain how the fusion is brought
about, I have regarded the phenomenon as being entirely
accidental and of no morphological importance.

It is, however, only right to point out that there is a fusion
of two epiblastic layers in certain Amphibians which seems to
have a deeper meaning, and to which in some way the condi-
tion in the rabbit may be comparable.

As far as I know, there is no account published of this
phenomenon, and I am not aware that anyone else has
noticed it. Therefore, it seems to me, a short account of the
facts as they appear in young embryos of Rana temporaria
may be of some interest.

For the purpose of following the fusion of the two epi-

166 RICHARD ASSHETON.

blast layers in the frog, it is best to examine the sections un-
stained.

In Rana temporaria the epiblast is from a very early
period divided into two layers—an outer called the epidermic
layer, and an inner called the nervous layer. If a section is
taken, say, transversely though the neural plate of a tadpole
about the time of the folding up of the neural folds, a section
is obtained of which fig. 1 is a drawing.

The two layers are seen to be very sharply and distinctly
divided, the outer or epidermic layer of epiblast (Z. EP.) is a
single cell in thickness. The cells are much more deeply
pigmented than are the cells of the inner or nervous layer of
epiblast (H. NE.) which forms a much thicker layer. The
cells of this layer are very closely packed in the region of the
neural plate through which this section is taken; so much so
as to render it impossible at this stage, at any rate, to distin-
guish the boundary of the cells—if, indeed, there are any
distinct boundaries. The nuclei are large, but are only seen
with difficulty without staining. Some indication of the
boundaries of the cells is to be seen in the slight increase of
pigment along certain lines.

In fig. 2, which is taken at a slightly later stage, but before
the neural folds have completely closed, the cells of the
epidermic layer may be seen to have become much elongated,
their inner borders being no longer truncated, but mostly
pointed, and seem to be growing into the mass of nervous
epiblast.

In fig. 3, which is from a section of a tadpole of about
3 mm. to 34 mm., in which the neural tube is now completely
closed and separated off from the skin, the appearance of the
darkly pigmented cells is very remarkable and instructive.

I can see no reason to doubt that the darkly pigmented cells
of the epidermic epiblast, seen to be elongated in fig. 3, have
by this time elongated and passed right through the mass of
nervous cells (or nuclei) and spread out into fine filaments on
the further side of the nervous layer. I cannot say for certain
whether these fine filaments anastomose or not, but, however

FUSION OF EPIBLASTIC LAYERS IN RABBIT AND FROG. 167

that may be, they seem to form that which may, as a whole,
be regarded as a reticulum, which is comparable to the myelo-
spongium of His.

The outlines between the cells of the nervous layer are quite
imperceptible in unstained specimens of this stage. The
nuclei of the nervous layer are only seen with great difficulty.

The boundaries of the epidermic layer cells can in most
cases be readily perceived, owing to the deeper pigmentation
of the cells of this layer. The processes (PR.) are sharply and
clearly defined, and are heavily loaded with pigment.

The bodies of the epidermic epiblast cells remain as yet
lining the interior of the neural canal, although there is cer-
tainly a tendency for the nuclei in some of them to move more
inwards. Also there seems to be a tendency for the epidermic
cells to become pressed apart by the nervous cells. This is
more marked at this stage in the spinal cord than in the brain,
as may be seen in fig. 4.

I think we may take these appearances as a conclusive proof
of the occurrence of an intimate fusion of the two epiblastic
layers in the frog.

Observe the condition of the auditory vesicle in fig. 3. Here
the walls of the vesicle are composed entirely of the light-
coloured elements. There is no trace of the dark, deeply-
pigmented strands such as are to be seen in the central nervous
system. This is a further piece of evidence that the dark
strands in fig. 3 are derived from the epidermic layer. For, as
is well known, in the frog the nervous layer alone gives rise
to the auditory vesicle—so it is a most significant fact that the
dark strands should be entirely absent in this case.

After this stage (3—5 mm.) it is extremely difficult to trace
the fate of the epidermic and nervous cells respectively. Up
to now, the fact of the greater pigmentation of the former has
rendered the inquiry easy.

The origin and meaning of the pigmentation is obscure, but
its presence in the frog seems to be due to two causes, separate
at any rate chronologically.

Firstly. Pigment is present in the unfertilised ovum as a

168 RICHARD ASSHETON.

superficial layer covering the upper pole. Hence for a long
time we find the superficial layer of cells after segmentation
to be more deeply pigmented than the more internally situated
segments.

Secondly. Pigment seems to be in some way connected with
the actual protoplasmic activity, as it appears internally wher-
ever division of cells takes place.

So also I believe the intensely black appearance of the pro-
cesses of the epidermic epiblast which I have been describing
is in some way connected with their intense activity just evinced
by their growth inwards—which seems to be very rapid, and
rather sudden.

When at a later period certain groups of nerve cells show a
similar intense activity, there is a similar deposition of pigment
in and around their processes, as, for instance, in the develop-
ment of the ganglion habenula as shown in fig. 5. This pig-
ment in each case becomes greatly lessened after the period of
intense activity of growth has passed by.

The Question of Early Separation into Neuroblastic
aud Spongioblastic Elements.

Although the evidence to be drawn from the figures accom-
panying this paper is far from being conclusive, yet I think
it tends very strongly towards the inference that the epidermic
layer of epiblast in the frog gives rise to the spongioblastic
elements, and the nervous layer to the neuroblasts.

His has shown how the spongioblastic network precedes the
development of neuroblasts and nerve fibres and forms an
irregular network with angular processes by no means unlike
the dark strands in my fig. 3. These processes are not at all
like the processes of nerve cells, which are always more
tapering and less knotty. Further, at this stage there cannot be
found any trace of definite nerve-fibres. These do not appear until
later, till the tadpole has attained a length of about 6—64 mm.

I can only interpret these figures as showing the con-
version of the epidermic layer of epiblast into a supporting

FUSION OF EPIBLASTIC LAYERS IN RABBIT AND FROG. 169

framework in between the cells of the nervous layer. It
is quite possible that we ought to regard the nervous layer of
epiblast as comparable to the germinal cells of His rather than
to the fully-developed neuroblasts.

If this is correct, we then must conclude that in the frog
there is a very early separation of the neuroblastic from the
spongioblastic elements.

Comparison between Rabbit and Frog.

Is it then possible that the condition in the rabbit has its
parallel here in the frog? It is true that the epiblast is
double only over a certain area of the embryo in the rabbit,
whereas in the frog it is double throughout.

In the frog the nervous layer soon becomes much thickened
along the future dorsal surface of the embryo, and over the
rest of the embryo the nervous layer becomes reduced to a
layer of one cell only in thickness, like the epidermic layer.

Now, although it is extremely difficult to trace the history
exactly, I am almost sure that the area over which the inner
layer of epiblast cells in the rabbit is found, corresponds to
that area in the frog over which the nervous epiblast remains
thick, or becomes thicker—i. e. to the neural plate.

In the frog the whole of the neural plate does not become
folded up to form the neural tube, but the outer lateral portions
of the anterior part remain outside of the tube, giving rise to
the ganglia of the anterior cranial nerves.

I have endeavoured elsewhere to bring evidence to show that
the epiblastic wall of the anterior part of the body of the
rabbit embryo includes more than the double-layered part of
the embryo, i.e. more than the so-called “ embryonic disc.”

Whether the ‘‘ embryonic disc”’ goes to form the neural tube
and ganglia of anterior cranial nerves as in the frog, or whether
it forms only the neural tube, I have no evidence to offer.

The embryonic disc is precisely in the same position relative
to the primitive streak as is the neural plate to the blastopore
of the frog.

I am well aware that the epiblast is not at first double in all

170 RICHARD ASSHETON.

mammals. For instance in the opossum, according to Selenka,
it isa single layer. But it also is not always double in the
Amphibia. In Triton it is at one time only a single layer of one
cell in thickness. Why there should be this early differentia-
tion into spongioblastic and neuroblastic elements in one and
not in another so comparatively closely allied animals it is not
easy to guess. Possibly it may be that, since the spongio-
blastic elements of the nervous system are the first to show
activity of growth in the nervous system (His, and above), then
in those animals in which, owing to various individual causes,
the epiblast is many-layered, the outermost layer of “cells” or
centres of activity being, by reason of its external position
more favorable to processes of respiration and so in a con-
dition more favorable to active growth, it will be this layer of
epiblast that will take on itself the earliest phase in the further
development of the nervous system.

Although I think the rabbit’s condition can be quite well
explained without reference to the above, on the other hand
there may be some deeper meaning in it, for which reason I
have thought it best to make notice of the condition in the
Anura.

To make this parallelism complete and certain it is necessary
to show that in the rabbit the cells derived from the outer
layer give rise to the spongioblastic tissue, and those derived
from the large inner layer cells to the neuroblastic tissue.
- This I cannot do. After a while the cells which have origi-
nated in each of the layers become equally active, but I have
as yet been unable to trace their fate respectively.

FUSION OF EPIBLASTIO LAYERS IN RABBIT AND FROG. 171

DESCRIPTION OF PLATE 18,

Illustrating Mr. Richard Assheton’s paper, “On the Phe-
nomenon of the Fusion of the Epiblastic Layers in the
Rabbit and in the Frog.”

List oF REFERENCE LETTERS.

AU. Auditory vesicle. 2. NH. Nervous epiblast. #. HP. Epidermic
epiblast. M#. Neuroblast. PJ. Stalk of pineal gland. PR. Processes of
epidermic layer cell. 7. #. Band of nerve-fibres passing from ganglion
habenula to the ventral side of the brain.

Fig. 1.—Transverse section through a portion of the neural plate of a frog
embryo (Rana temporaria) during the process of folding up of the neural
plate. Unstained. x 165.

Fie. 2.—Transverse section through corresponding region, but at a slightly
later period, of a frog’s embryo. Unstained. x 165.

Fig. 3.—Transverse section through the corresponding region after the
separation of the neural tube from the skin. Tadpole, 33 mm. Unstained.
x 165.

Fic. 4.—Transverse section through the spinal cord of the same specimen
as Fig. 3. Unstained. x 165.

Fic. 5.—Transverse section through ganglion habenula of a frog tadpole

of 13 mm., showing deeply pigmented neuroblastic processes. Aniline blue-
black. x 350.

il

2S ‘er had io WYRE 9.0% co. oerae vo wahey
" palin - li eo S - ie
2 » a : an i 7
en Ba, cn mali . 7
— 7 : . Ss are ve 7 ‘1! Bit ik “*
ere Morro and ae
= i. Oe E
4 eg 'F ale i @ el seer re ey h ti Lose yi +1 | i yo ae
; = awa Ati sala eu widen! ee P| Aid 7 ' Hwisad slit 73 Liat} ‘Sait
1 tee hie alt UT Bre idaast
—s i ; Tr) see) ter Gant : *

Or iatin lf a a | snes tiie NAO ive? yonOtn a
> joa bheviayd SL? Josie bettie Tr see iviGina Gwe iy
ed, Moule, ao ee tide hy hate A Ua qa shoe
7 rite) Gi) 10 Ohie [earee Ag ey

7 path So Tal Ler ull ii add ¥ dguvndl idtinde dow Teele
Aislin oy uitinlyl iv Gobel) MOA T4108) 4
ehhh 2 [= dle t
3 Toile a de ind Bian mk) ir eri thy oma glk ot win wn Devedh: ae
08) Ko della wall 2 creyee eli) Sug) OP EG
: Add unith ey Posty witmoueriay wih wis Stet) Cuil, a oie A eh ey;
; AL Wasa’ i agra at Anil es mi) merry! Lifiss faipud- ad! woe
7 : «palate mnimeaes Yi Sor bax ha ie eget: Gulove 4 Konan 5 ;
fis © ; eet se Se ee i
aa : me nd Spar a a Miyareaiiy rtiane® Hane” Pati nailin teat g
:

5 - odiig tt Hiaahia(s eee qe xd bit Bit entice “dl es Boer in
. _ | peu ag %~
' _—
- * 4 - rt neal
>

ATTACHMENT OF MAMMALIAN EMBRYO TO UTERUS. 173

On the Causes which lead to the Attachment
of the Mammalian Embryo to the Walls of
the Uterus.

By

Richard Assheton, M.A.

With Plate 19,

I.

Tue First ATTACHMENT OF THE RaAspsit EMBRYO TO THE
WALLS OF THE UTERUS.

By the end of the eighth day, if not actually attached to the
walls of the uterus, the embryos have become definitely located,
and their presence is made evident from the exterior of the
uterus by a bulging in the wall opposite the mesometrium,
Frequently by this age they are actually attached to the uterus,
and cannot easily be extracted without damage.

It must also be noted that the embryos by this time no longer
lie anyhow in the uterus, but when definitely located the position
of the embryo to the uterus is such that the embryonic area
is always towards the mesometric wall of the uterus.

This is a very important fact. It is probably necessary for
the development of the rabbit that it should be thus situated.
It would be very awkward if the embryo became fixed in any
other position. The shape of the uterus at this stage, and the
shape of the blastodermic vesicle at this stage, are so beauti-
fully adapted one to the other as to render any other position
almost impossible. The blastodermic vesicle by this time is a
slightly elongated body whose lower pole is semicircular in
transverse section, while the upper pole is much flattened. If

174 RICHARD ASSHETON.

a piece of uterus is blown out with water, then hardened, and
a transverse section cut, the cavity will be seen to be bounded
by a semicircular wall on the abmesometrial side, and, by
reason of the two largely developed “ placental” lobes, a very
flattened wall on the mesometrial side.

Fig. 1, Pl. 19, is a figure of the section of the uterus during
the early days of pregnancy, or in such part of the uterus in
which no embryo is present during rather a later stage in the
earlier part of pregnancy. JM. is the mesometrium; C. is the
cavity of the uterus.

The general external outline of the uterus is circular; so is
the inner muscular system, but the latter is eccentrically
placed to the outline of the uterus.

Within these muscular coats, which form the most resisting
part of the uterus, comes the soft connective tissue and epi-
thelial lining of the uterus. The loose connective-tissue layer
seems to be the most yielding, and, as a consequence of the
blastodermic vesicle within the cavity of the uterus, the folds
or lobes, as they appear in transverse section (PL. L., PP. L.,
OP. L.), diminish in size, and as the body within increases in
size, one by one they begin to disappear. The obplacental
lobes have quite disappeared by the middle of the eighth day,
and the periplacental lobes can no longer be described as lobes
or folds after the beginning of the ninth day. The placental
folds being the largest, and also upon the less expansible side
of the uterus, become much flattened but never entirely dis-
appear.

About the time that the blastodermic vesicle becomes de-
finitely located, its shape is in transverse section as fig. 4.

The lining of the uterus is seen in fig. 1 to be thrown into
folds. When the blastodermic vesicle (fig. 4) is inserted into
the cavity, the folds become pressed out to a greater or less
extent : thus the folds called obplacental are entirely obliterated
(OP. L., fig. 2), the periplacental are nearly obliterated, but
the placental lobes being very much larger, and also being
supported by a larger mass of muscular tissue, are hardly com-
pressed at all.

ATTACHMENT OF MAMMALIAN EMBRYO TO UTERUS. 175

The cavity of the uterus is thus bounded on the mesometrial
side by an almost straight line, while the opposite wall is
circular. A body having the shape shown in fig. 4, in passing
down the tube by virtue of the continual slow contractions of
its walls, could hardly fail to become so fitted as to lie with its
flatter surface against the flatter wall of the uterus.

Since the flatness of one surface of the blastodermic vesicle
is caused by the embryonic dise and changes connected there-
with being upon that surface, it follows that in the rabbit the
embryo always comes to lie up against the mesometrial side of
the uterus, which is, under the circumstances, by far the most
favorable position for its future development. For in this
position it will be less exposed to the tension which must
necessarily arise upon the opposite side in the subsequent
expansion of the uterus by the accumulation of fluid within
the blastodermic vesicle.

The blastodermic vesicle does not by any means become
attached with the centre of its flat surface always exactly
adjoining the median cleft between the two placental lobes.
It may sometimes be a little to one side or the other, but it is
never very far out.

As the blastodermic vesicle expands, it eventually fills up the
whole cavity of that section of the uterus in which it happens
to be. With further expansion it necessarily exerts a pressure
upon the walls of the uterus, and this pressure is made apparent
without by a swelling, or protuberance, occurring upon the
side of the uterus away from the mesometrium.

The first visible sign of the commencement of the formation
of the placenta—or, at any rate, of those structures which
eventually cause the placenta to come into existence, appears
very shortly after the appearance of this swelling. I allude to
the papille and protuberances of epiblast.

I said just now the first visible sign because I believe that
the outgrowths are due really to a continuation of those
causes which have given to the blastodermic vesicle its present
form and shape, and described in a former paper.

The immediate cause of the origin of the papille seems to be

176 RICHARD ASSHETON.

the pressure which now is applied to the vesicle from without
by reason of the resistance offered by the elasticity of the walls
of the uterus to the internal hydrostatic pressure.

First of all, let us consider what is the nature of these
papilla-like growths.

Sometimes during the eighth day, it may certainly be as
early as the seventh day four hours, when the uterus is swollen
very considerably by the pressure of the contained blastodermic
vesicle, here and there it may be noticed in transverse sec-
tion that over all the lower surface of the vesicle certain of
the epiblast cells are no longer so much flattened, but the
nuclei appear rounded instead of oval in section, and the proto-
plasmic part of the cell much more distinct and granular—
altogether more comfortable-looking (v. a., fig. 5). A little
further along, at 6., may be seen a couple of such nuclei ina
mass of granular protoplasm. At c. a group of three or four,
or more, of such nuclei in a mass of granular protoplasm.
Outside this may be seen the torn remnants of the much
attenuated albumen layer. The portion of the walls of the
vesicle here figured was part of the lower pole of the blasto-
dermic vesicle of an embryo of seven days four hours. Al-
though I did not succeed in taking it out of its place in the
uterus unhurt, it was, nevertheless, not yet attached to the
walls of the uterus at any point. It may be noted that here
there is no trace of an inner layer or hypoblast.

In fig. 6, another piece of the wall of a vesicle of about the
same age cut in situ in the uterus, this piece shows a por-
tion of the side of the blastodermic vesicle, or, at any rate, a
portion not so far removed from the embryonic pole as that
drawn in fig. 5. Here in fig. 6 the same features are to be
seen in the epiblast as described for fig.5. There is here, how-
ever, a well-developed hypoblastic layer.

Fig. 8 is a section of the upper part of the wall of the vesicle,
of that part which closely adjoins the embryonic disc. This
specimen is from an embryo from the same rabbit as that
from which fig. 5 was drawn, and almost exactly the same
size. Here it will be noticed that not only are the cells

ATTACHMENT OF MAMMALIAN EMBRYO TO UTERUS. 177

here and there large and granular with round nuclei, but the
whole of the epiblastic layer is now composed of thickened cells,
almost columnar with rounded nuclei, and for a given length
there are many more nuclei than there were at that region
during the sixth and seventh days. In fig. 9a piece of the
same region a few hours later is shown, and here the epiblast
is seen to be so thickened as to be actually several cells
thick,—in fact, a proliferation of cells is taking place out-
wards.

Figs. 10 and 11 are still later stages of the same area.

I believe that to understand the first outgrowth of these
papille, both the small ones scattered over the lower pole, and
the area nearer to the embryonic disc, the great change that
the embryo has now undergone in the physical and mechanical
conditions must be taken into careful consideration.

During the fifth, sixth, and seventh days I described in my
previous paper the growth of the walls of the vesicle as being
due to the hydrostatic pressure within it, together with the
multiplication of the cells of the walls of the vesicle, support
being rendered to the delicate wall of the vesicle by the albumen
layer. The cells of the wall (the epiblast) are very much
flattened because of the tension produced by the hydrostatic
pressure.

The hydrostatic pressure is sufficient to keep the cells always
taut, and any increase in size of a cell owing to growth, or any
aggregation of cells by multiplication, is prevented by their
being flattened out by the internal pressure as soon as formed.
Thus all the cells are uniform in thickness, and extra activity
of one part shows itself by extra expansion of that
are of the vesicle.

What happens, however, when the walls of the uterus, by
reason of the great size now attained by the blastodermic
vesicle, affords supports to the walls which was hitherto want-
ing? It must decrease the ratio of hydrostatic pressure to the
rate of growth of the cell wall.

If the amount that the cells are stretched is constant when
the hydrostatic pressure is to the rate of growth of the cell

VoL. 37, PART 2.—NEW SER. M

178 RICHARD ASSHETON.

walls as, say, 10 is to 10, then when the hydrostatic pressure
(P) is to the rate of growth (R) of the cell walls as say 8: 10, it
follows that the cells will not be so much stretched as when
Rad: sas 0.

The actual pressure within the vesicle no doubt does not
diminish, more probably it increases, but the relation of P: R
is altered by the fact that additional resistance is now added
without by the walls of the uterus. At any rate, it upsets the
ratio formerly existing.

The result, I believe, is that now the rate of growth of the
cells of the wall is as compared with the rate of increase
in size of the blastodermic vesicle greater than it was before, so
that when a cell ‘‘ grows” and “ divides ” it no longer becomes
at once stretched out, but forms a rounded granular cell, or
group of cells, as in figs. 5 and 6. The cells which are for
the time inactive will remain flattened, for elasticity is not an
attribute of protoplasm.

This is the case at the lower pole of the blastodermic vesicle
and at the lower sides. In the region near the embryonic
disc (where it has all along been assumed that there is a more
active growth) it is found that almost every cell has evinced
signs of activity, for here the whole area has become thickened,
and the cells far more closely packed than they were, and almost
columnar instead of being flattened (fig. 8).

It is, I believe, usual to describe the first attachment as
occurring between the ectoplacenta of the embryo and the
placental lobe of the uterus. This I am convinced is not an
accurate statement. The first actual attachment is between
the lower parts of the blastodermic vesicle and the periplacental
and obplacental folds.

The exact course of the procedure I am doubtful about, but
I believe it to be a combination of at least two main causes,
but it may involve more; or possibly it is wholly due to only
one of the two I am about to mention.

The first attachment is effected by means of the “ papille,’’
or thickened spots of epiblast of the lower pole of the embryo
already described (figs. 5 and 6, a., b.,¢.). In fig. 7 one of

ATTACHMENT OF MAMMALIAN EMBRYO TO UTERUS. 179

these thickened spots is shown to be on the point of effecting
an attachment. It may be noticed to be wedge-shaped in
section; it is a little blunt cone. Practically each papilla
assumes this shape, and is being pressed against the epithelium
lining the uterus. In this figure (and almost any number
might be drawn showing the same characters) the papilla
certainly looks as though it were piercing the epithelium by
reason of the pressure from within the vesicle.

Of course the actual hydrostatic pressure will be the same at
a, as it is at e., but nevertheless a greater pressure will be
exerted on the uterine wall at the apices of any knobs on the
wall of the vesicle than at the area between them if the uterine
wall is in a state of tension, which undoubtedly it is at this
time.

If we consider that in all probability the uterine epi-
thelium is a softer material than the muscular and con-
nective tissue outside it, it is all the more probable that the
softer (if it is softer) uterine epithelium will give way between
the two.

I think it quite possible that this may be the only neces-
sary cause, and by this means the “ papilla”? reaches the
capillary system of blood-vessels in the uterine connective
tissue, a point of the utmost importance to the welfare of the
embryo.

On the other hand, although I have no doubt that the
additional pressure which exists at the points of these knobs is
of much importance in that it causes very close contact
between the uterine epithelium and parts of the wall of the
blastodermic vesicle, yet it is more than possible that the
breaking down of the uterine epithelium at those points is
not due to mechanical pressure alone, but to a chemical or a
physiological process, such as absorption of the uterine cells
by the vital activity of the cells of the knobs. But the possi-
bility should not, I think, be lost sight of that the first
breaking through of the uterine epithelium at these points
may be entirely due to mechanical pressure alone.

A point to which special attention must here be drawn is

180 RICHARD ASSHETON.

that the moment after the uterine wall comes to take the
place of the albumen layer as a support, then the greater part
of the expansion of the combined vesicles must be that part
furthest removed from the embryo, because it is here that the
uterus wall is infinitely less resistent, and with it will go that
part of the vesicle to which it is attached.

Onthe Importance of the Albumen Layer and Zona
radiata.

Those who have followed my account of the development of
the blastodermic vesicle of rabbit between the eightieth hour
and the time of attachment of the vesicle to the walls of the
uterus at about the 170th hour, must have seen how important
a part the albumen layer plays in producing the form actually
assumed by the vesicle. I pointed out how that the thin cellular
wall of the blastodermic vesiele would be itself quite able to
withstand the hydrostatic pressure within ; and, further, that it
is not until the vesicle has attained such a size as to stretch the
walls of the uterus as to cause them to afford the support
necessary to prevent the bursting of the vesicle, that the albu-
men layer was lost. By this time the albumen layer has
become exceedingly thin, and the disappearance of it is brought
about by its rupture. Traces of it may be found curled up and
crumpled several days later. Once burst, its function, as far as
I can judge, is at an end.

Now it has almost certainly at least one other very important
influence upon the development of the rabbit. By its presence
until the eighth day it absolutely prevents the cellular tissue
of the blastodermic vesicle from coming into close contact with
the cellular tissue of the uterus. The embryo up to the eighth
day is as free probably from the protoplasmic influence of the
mother, as is the egg of a bird after it has been covered with.a
thick calcareous shell. There is also, it will be remembered,
another coat to the ovum, the zona radiata, which is present
from the very first, and has a similar effect to that which the
albumen layer has, but being less thick its effect is more evanes-
cent. The two coats may be considered as one.

ATTACHMENT OF MAMMALIAN EMBRYO TO UTERUS. 181

As regards the rabbit. Ihave shown in considerable detail
how many events are, or at least may be, explained by external
influences (such as albumen layer, hydrostatic pressure, pressure
of uterus, rupture of albumen layer) acting in conjunction with
a simple steady force or energy, the primary centre of cell mul-
tiplication.

In doing so, it will be remembered that in connection with
the actual forms assumed, and phases passed through, in the
rabbit, the presence of the albumen layer, the size of the cavity
of the uterus, and even the shape of the walls of the uterus, had
important consequences ascribed to them.

If my line of argument has been correct, the effects produced
by the albumen layer and the size and shape of the uterus must
be very different or absent altogether in forms in which these
conditions are different or absent. As regards the size of the
ovum itself, there is very little difference amongst mammals.

Let us examine the case of a rat,—a rodent, and so not very
distantly removed genetically from a rabbit. And yet how
different is the form assumed in the earliest stages of develop-
ment! Do the conditions differ from those in the rabbit, and
if so, how?

They differ in these respects :

(i) There is no zona radiata or albumen layer.

(ii) The diameter of the uterus is very much smaller, and the
lumen proportionately smaller still.

(iii) The walls of the uterus are of a more uniform thickness.

In the rabbit the zona radiata is so quickly covered by the
albumen layer that it cannot be said to have in itself much
effect as a support or protective coat.

In the rat there is no zona radiata or albumen layer. The
ovum develops freely in the cavity of the uterus unprotected
by any coat. It would seem to pass very rapidly down the
Fallopian tube. Robinson found one early stage of segmenta-
tion; from this he figures the mesial section, which shows
parts only of twelve segments, so we may conclude that his
specimen was a fairly early stage of segmentation—comparable,
perhaps, to that of the rabbit of the forty-eighth hour,

182 RICHARD ASSHETON.

The segmented ovum of the rabbit lies within a thick, tough
spherical coat, and is spherical. The ovum of the rat is not
thus placed within a spherical mould, and can therefore take
other forms, which from the conditions would seem to be
necessarily disc-like or oval.

The blastodermic cavity (“ Dottersackhéhle” of Salenka,
“ vitelline cavity ” of Robinson) would seem to be produced
by a hydrostatic pressure from within, as in the rabbit, but
owing to there being no resistance of albumen layer the walls
of the cavity of the vesicle offer less resistance. Thus there is
less tension, and this has the effect of producing a shape of no
symmetry except such as is given to it by the walls of the
uterus in which it lies.

In the rabbit an early tendency to rapid growth of the area
immediately surrounding the embryonic disc gives rise to an
expansion of that part of the vesicle; so, also, the same tendency
to arapid growth occurs in the rat round the corresponding
area, but with very different results. A heap of cells accu-
mulates, which gives rise to the irregular mass growing
among the cavities of the uterus, and usually called Trager.
After a while the blastodermic vesicle becomes fixed in the
walls of the uterus, and the condition then is similar to
that in the rabbit, i.e. the necessary resistance is supplied, and
a more or less spherical swelling appears upon the uterus, in
which the embryo develops in safety.

As regards the actual fixing of the rat embryo, this is no doubt
primarily due to the pressure exerted between the irregularities
of the walls of the uterus, the “triger” (and other external
layer cells of theembryo), this pressure being brought about by—

(i) Multiplication of cells of trager and general growth of
embryo.

(ii) Hydrostatic pressure within the blastodermic vesicle.

Thus we see that what corresponds to the trager in the
rabbit becomes obvious in the rat at a very early period;
whereas in the rabbit its presence (essentially as an area of
increased energy) is marked by the continued expansion of
the vesicle, as described in a former paper,

ATTACHMENT OF MAMMALIAN EMBRYO TO UTERUS. 183

Then, again, let us look at the conditions of the mole.
Here is an animal, an insectivore, which is considerably re-
moved genetically from a rabbit. Yet here we find that
the development is extremely like that of the rabbit. It
is true that there is a slight approach to the inversion
condition, but only slight; and I think that what appears
to be an inversion for a short time is very possibly in no
way connected with the inversion such as that of the house
mouse, field mouse, rat, or guinea-pig. The separation figured
by Heape (‘ Quart. Journ. Micr. Sci.,’ 1883, Pl. 29, figs. 24
and 25) seems to be due to quite other causes, namely, temporary
and rather sudden acceleration of growth of the embryonic
disc itself, and not of the area just beyond it. For the cells
which become separated from the epiblast, and which Heape
compares with the trager, afterwards again unite and fuse with
the epiblast and form together the epiblast of the embryonic
disc; whereas in all other forms any cells which have once
become separated as “trager”’ never participate in the formation
of the permanent epiblast. I should say these cells correspond
rather to the outer layer of epiblast in the embryonic disc in
the rabbit, and that in the mole the representatives of the
trager cells would be those which at this moment (Heape,
figs. 24 and 25) form part of the wall of the vesicle beyond the
region of the embryonic disc, as in the corresponding region
of the rabbit.

Until the time of this temporary bending in of the embryonic
area, the development of the mole is exactly comparable to that
of the rabbit. The ovum segments and forms a spherical
morula. A blastodermic cavity and vesicle are formed almost
exactly comparable to the rabbit, and the hypoblast is formed
in the same way; the blastodermic vesicle is a spherical bladder-
like body ; and may not the reason of this be that the ovum
develops up to a certain point under quite similar conditions,
that is to say, surrounded by a thick protective covering of
zona pellucida and ‘‘ mucous layer ” (v. Heape) ? This mucous
layer develops rather differently, being applied according to
Heape in the uterus, and not in the Fallopian tube.

184 RICHARD ASSHETON.

The coat, however, is very much less firm, and offers
apparently less resistance than that of the rabbit, for, later,
after the vesicle has filled the cavity of the uterus the walls
very soon become moulded into the shape of the cavity of the
uterus (v. Heape, Pl. 28, figs. 8 and 9).

Of course it is necessary to take into consideration the fact
that the actual lumen of the uterus in the mole is very much
less than in the rabbit, and that therefore the developing
vesicle will be influenced by the resistance offered by the
muscular coat of the uterus at a time when the vesicle is much
smaller than in the case of the rabbit.

Now to consider another case where we have pretty complete
records, i.e. the guinea-pig.

During the early segmentation stage the ovum of the
guinea-pig is surrounded by a zona radiata. We find that in
the development of the early stages the embryo assumes just
the same form as in the mole or rabbit; but after the embryo
has reached the uterus the zona is ruptured, and from this
moment the embryo is naked and goes through phases which
are certainly more like those of the rat than those of the
rabbit or mole. It can hardly be necessary to point out that
whereas at first the conditions were similar to those of the
rabbit or mole, from this moment they resemble more closely
those of the rat.

In this question I have only gone into any details in the
case of the rabbit. But still, although I can only judge of the
other forms from either rather superficial personal examination
or from the writings of others in which this point has not been
given any special prominence, I think I may be allowed to
make a few general remarks upon such other forms in which
marked divergence of shape and structure occurs in the
method of formation of the blastodermic vesicle and germinal
layers.

Amongst the Carnivora, from observations on the dog by
Bischoff and Coste, and on the cat by Fleischmann, it seems
clear that the condition most nearly resembles the rabbit.

The ova are in all these cases invested with a firm coat,

ATTACHMENT OF MAMMALIAN EMBRYO TO UTERUS. 185

either zona pellucida or albumen-layer, and more probably
both. These are present until as late a period as in the
rabbit. There is no inversion in either case. The shape is
somewhat different in both cat and dog. The vesicles are
spherical until about the time that the resistance of the
uterine walls is encountered, after which time they assume a
much more oval shape. Whether this is due chiefly to the
effect of the walls of the uterus, or whether due in part toa
modification in intensity or of the extent of the trager area, it
is impossible to say without careful investigation.

Amongst the Ungulates we have descriptions of the deer by
Bischoff, the sheep by Bonnet and Coste, the pig by Keibel.

Although the literature on the point I am now discussing is
very scauty for this group of animals, we know two facts.
Firstly, that the vesicle, at first spherical, grows to an
enormous length—as much as 180 mm. in the pig (Keibel),
or the whole length of the uterus in cases where there is
only one embryo. Also we know there is no inversion.

They are surrounded by a zona radiata or some other
investment, but it would seem to be very thin. Bischoff says
there is no albumen-layer in deer.

As regards inversion, the Ungulates, as far as at present
known, behave like the Carnivora and rabbit. But after a
certain point the vesicle lengthens enormously. It seems
probable that this lengthening is due partly to the large
cavity in which the vesicles lie. All the Ungulates of which
we have any record are large. The lumen of the uterus
in the pig is enormous as compared with the lumen in the
uterus of a rabbit.

Also there seems to be evidence to show that the zona
radiata or albumen layer is thinner, and offers therefore less
resistance, and accordingly has less effect in resisting the
pressure of the walls of the uterus, which, it must be remem-
bered, actually lie in contact with each other.

We must also suppose that although the pressure is sufficient
to prevent the vesicle from retaining an approximately spherical
form, it is not as yet sufficient to cause a fixation of the em-

186 RICHARD ASSHETON.

bryo to the walls of the uterus. What actually brings about
the fixation in the embryo in Ungulates I do not know. It
would seem possible, from the frequency of a cotyledonary
placentation, that a number of spots of pressure arise at
considerable intervals amongst the folds of the uterus, and so
bring about local areas where the rate of growth of the cellular
wall is in excess of the rate of expansion of the elongated
vesicle.

In connection with the elongation of the blastodermic
vesicle, I may draw attention to a fact in the development of
the rabbit.

It follows from what I have said, that if the albumen-layer
in the rabbit was less thick, or absent, its development would
be very different. So it happens that after the rupture of the
albumen-layer there is a tendency for the vesicle to elongate
during the eighth and ninth days. The normal shape of the
vesicle upon the ninth day is shown in fig. 3.

It is important for the development of the space (C. BL.)
that after the rupture of the albumen layer the passages (UT.)
should be closed; otherwise, unless the thin wall of the blasto-
dermic vesicle is very much stronger at that point than _-
hitherto, it could not withstand the pressure from within, which
must now be considerable, and would rupture, and the liquid
would escape and the swelling (C. BL.) would collapse. In
the rabbit, no doubt, the mucous membrane of the approxi-
mated walls of the uterus at this point (UT7.) is thrown into
folds and pressed together by becoming a little bent inwards,
and by this means tends to block the passage, but also, no
doubt, the horns of the vesicle (H.) also help to obliterate and
plug up the cavity at these points.

Not unfrequently it occurs that either one of the horns (or
both) extends through this narrow passage (UT.) and is pro-
longed (the albumen layer having ruptured by now) for a short
distance along the cavity of the uterus. Sometimes a horn
seems to be prolonged for a distance of as much as 34 mm., but
the greater part of this is non-cellular (Z.), the origin of which
I have not traced,

ATTACHMENT OF MAMMALIAN EMBRYO TO UTERUS. 187

SUMMARY.
Section l.

(i) The blastodermic vesicle of the rabbit becomes first
attached to the walls of the uterus by its lower pole.

(ii) This attachment of the lower pole is regarded as the re-
sult of a mechanical pressure of certain spots or knobs of
thickened epiblast of the blastodermic vesicle upon the epithe-
lium of the uterus (the pressure being the hydrostatic pressure
within the vesicle), whereby the uterine epithelium is pierced
and the knobs of epiblast become embedded in the connective
tissue below.

(iii) These knobs of epiblast, as also the thickening along
the trager region, are regarded as being the direct result of a
destruction of the equilibrium between the rate of increase of
the hydrostatic pressure within the blastodermic vesicle and
the rate of growth of the cellular wall of the vesicle, under
which conditions the epiblast had hitherto remained practically
at a constant measure of thickness.

The destruction of the equilibrium is brought about by the
additional pressure put upon the expanding blastodermic
vesicle by the resistance of the uterine walls. This is a con-
tinuation of the same series of forces which in my former
paper were supposed to account for the peculiar shape of the
blastodermic vesicle and apparent growth round of the hypo-
blast.

(iv) In the trager region the attachment is effected in a
rather different manner. Here the activity of the epiblast is
greater than at the lower pole, so that here the thickening is
of a more general and more rapid character, which results in a
somewhat extensive and irregular area instead of isolated
knobs of thickened epiblast. Also the soft tissue imme-
‘diately underlying the epithelium of the uterus is here very
extensive.

Thus the conditions are not such as to cause a perforation of

188 RICHARD ASSHETON.

the epithelium. Instead of this, the epiblast of the embryo
becomes very thick and amounts to almost a loose proliferation
of cells, which cells become eventually pushed into the
irregularities and glands of the placental lobes.

Section 2.

(i) The primary cause of “inversion”? is the fact that the
embryonic area is at one time a region of less activity, and is
surrounded by a zone of greater activity in connection with the
future formation of placenta and a space within which the
embryo can develop unaffected by pressure external to itself.

(ii) The occurrence of “inversion” is determined by the
production of a heaping-up of cells at an early stage by this
zone of greater activity and so forcing the embryonic area
inwards.

(iii) Inversion is prevented by causes which impede the
heaping-up of tissues around the embryonic area.

(iv) Foremost among these preventing causes is the presence
of an investing coat, either zona radiata, or albumen layer, or
mucous coat, &c., which—

(1) By hindering a close connection between the cells of
the blastodermic vesicle and the uterine walls ;
(2) By affording a strong support to the walls of the blasto-
dermic vesicle—
allow the blastodermic vesicle to assume, under the expanding
influence of the hydrostatic pressure within, shapes due more to
its own inherent tendencies and less to the effect upon it of the
pressure of the uterine walls.

(v) The thicker and more lasting these coats are, the more
marked are the intrinsic characters of the blastodermic vesicle,
and the longer deferred is the impression of characters due to
the physical effects of the uterine walls.

For instance, both the rabbit and the dog have investing
coats. In the rabbit it is far thicker, so in the rabbit it is
found that the vesicle assumes a shape which can be best
accounted for by intrinsic causes, while in the dog, where
it is much thinner, the uterus seems to be the more potent

ATTACHMENT OF MAMMALIAN EMBRYO TO UTERUS. 189

factor in moulding the shape of the vesicle. In the rabbit the
vesicle attains a far greater size than in the dog before
becoming attached to the uterus, though this may, no doubt,
also be partly owing to the far greater resistance offered by the
uterus of the dog.

(vi) When the lumen of the uterus is large, and where the
investing coat is present, but delicate, the blastodermic vesicle
may grow to a great length, and possibly the development of
villi and consequent placental attachment may be brought
about by local developments of regions of pressure.

(vii) Even in the rabbit, after the rupture of the albumen
layer the blastodermic vesicle may become extended along the
cavity of the uterus to a length of over 20 mm.

EXPLANATION OF PLATE 19,

Illustrating Mr. Richard Assheton’s paper “On the Causes
which lead to the Attachment of the Mammalian Em-
bryo to the Walls of the Uterus.”

List oF REFERENCE LETTERS.

a. Epiblastic papilla. 4ZB. Albumen layer. 4. Epiblastic papilla.
Cavity of uterus. c¢. Epiblastic papilla. C. BZ. Cavity of the blastodermic
vesicle. e. Wall of the blastodermic vesicle. ZC. Ectoplacental cells. ZZ.
External longitudinal muscle layer. HIZB. Embryonicarea. HP. Epiblast.
H. Horn of the blastodermic vesicle. HY. Hypoblast. JC. Internal circular
layer of muscle-fibres. IM. Mesometrium. MHS. Mesoblast. P. LZ. Placental
fold of the uterine mucous membrane. OP. LZ. Obplacental fold of uterus.
PP. L. Periplacental fold of uterus. U7. Cavity of the uterus. U7. EP,
Epithelium of the uterus. Z. Non-cellular continuation of horn of vesicle
attached to the remains of albumen coat.

Fic. 1.—Transverse section of the uterus of a rabbit before any distension
has been caused by the presence of an embryo.

Fic. 2,—Transverse section of the uterus of a rabbit on the seventh day
after fertilisation. The distension is chiefly at the expense of the weaker

190 RICHARD ASSHE'TON,

wall, i.e. the obplacental folds. At this moment the location of the embryo
can only with difficulty be recognised from without.

Fie. 3.—A diagram illustrating a longitudinal section of the blastodermic
vesicle and uterus of a rabbit about nine days after fertilisation. The epi-
blastic papilla and ectoplacental area are omitted. The cavity of the uterus
is seen to be obliterated by means of the horns of the vesicle (H.).

Fic. 4.—A transverse section of the blastodermic vesicle about the time
that it becomes finally located.

Fic. 5.—Section of the lower pole of the blastodermic vesicle of a rabbit.
Seven days four hours. x 465.

Fic. 6.—Section of the side of the blastodermic vesicle of about the same
age. x 465.

Fie. 7.—Section of the lower pole of the blastodermic vesicle of a rabbit
(after attachment) and of the uterine epithelium. One of the wedge-shaped
papille is seen to be piercing the epithelium. x 465.

Fic. 8.—Section of the ectoplacental area of the blastodermic vesicle of
about the same age as that of which Fig. 5 is a section. x 465.

Fic. 9.—Section of ectoplacental area later than Fig. 8. x 465.

Fic. 10.—Section of ectoplacental area still later. x 465.

Fig. 11.—Section of ectoplacental area about the time of its attachment
to the placental lobes. x 465.

THE PRIMITIVE STREAK OF THE RABBIT. 191

The Primitive Streak of the Rabbit; the Causes
which may determine its Shape, and the
Part of the Embryo formed by its Activity.

By
Richard Assheton, M.A.

With Plates 20—29.

Tue following pages are offered as a contribution towards
the elucidation of the mode of formation and the function of
the primitive streak in the rabbit.

Although it is necessary to make incidental remarks upon
the formation of the mesoblast and notochord, I have not given
here a full description of the results of my own observations
upon those points, but I hope to be able to do so, and to dis-
cuss the results of former authors upon this and other animals,
inalatercommunication. In the present paper I have confined
myremarks almost entirely to the rabbit, and to my ownsugges-
tions towards an explanation of certain phenomena connected
with the structure we call the primitive streak in that animal.

Of recent years the theory of concrescence has been again
brought into great prominence in attempts to account for
the growth in length of the Vertebrate embryo. The
present paper will, I hope, tend to show that there is no
trace of such an occurrence in the rabbit, and that the growth
in length of the embryo can quite as well—and to my mind
infinitely more easily—be accounted for by a process of addi-
tion of new cellular units between the pre-existing embryo and
an area of rapid cell-production. I have attempted to show
which part of the embryo is to be regarded as the “ pre-existing
embryo,” and which as being due to the activity of the area of
rapid cell-production—the primitive streak.

192 RICHARD ASSHETON.

I have also tried to explain how the changes in shape of the
primitive streak are brought about.

I begin with a short account of this area of activity (which
I shall speak of as the secondary area or centre of cell-
production—as compared with the process known as the
segmentation of the ovum, which I call the primary centre of
cell-production) which adds certain details of interest to the
accounts already given of the rabbit primitive streak by other
authors, notably by Hensen, Kolliker, and Rabl.

The Earliest Signs of the Formation of the Secon-
dary Area of Proliferation.

Until the middle of the seventh day the embryonic disc is
circular in outline. After the fusion of the inner and outer
layers of epiblast, the outline of the embryonic disc becomes
very much sharper. This state continues until about the
middle of the seventh day, when, although the anterior
periphery of the embryonic disc retains this character, the are
of the hindermost quadrant becomes very much less distinct,
and is no longer a semicircle.

On measuring the embryonic disc very soon after this, it is
found to have long and short axes, the long one usually parallel
to the long axis of the vesicle, the short to the short axis of
the vesicle. The lengthening has been in the antero-posterior
direction. The anterior end presents much the same shape
and features as before, except that its periphery describes an are
of a larger circle than before. It is still sharply defined. The
posterior end is, if anything, more ragged, and is undoubtedly
at this moment thinner than the rest of the embryonic area.

The measurements are as follows:

Blastodermic vesicle, long axis. F : 4°] mm.
My » short axis . : ; Bu 4;
uf » polar axis . : ; 3°86 2s

Embryonic area, long axis . ‘ : M357 'g
bs ‘ transverse axis . . 126 |

This ragged border I believe to be the effect of the setting
up of an area of increased activity of cell-production either

THE PRIMITIVE STREAK OF THE RABBIT. 193

upon, or just within the posterior border of the previously cir-
cular embryonic disc (see PS., Pl. 20, fig. 2).

Whether this activity concerns in the first instance both the
cells derived from the outer layer of epiblast and those
derived from the inner layer of epiblast equally, I cannot form
at present a definite opinion.

Very soon the centre of this ragged area becomes thickened,
and the embryonal area has now the appearance shown in fig. 3.
The original outline of the embryonic disc is fairly distinct, but
of course slightly larger than in the earlier stage (fig. 1). This
thickening, which is due to a greater accumulation of cells at
this point, would seem to be the result of a greater activity of
cell multiplication at this spot.

As to the cause of this sudden increase of activity I have no
suggestion to make. It is, of course, quite possible that the
activity was present from the first, and that it is only owing to
changed conditions that it now becomes evident as a heaping-
up of cells. To give a strictly epigenetic explanation some such
cause ought to be adduced, but for the present I am not able
to suggest any, and so must ascribe it to a palingenetic cause.

It is not possible to point out the exact boundaries of this
area of increased activity, but probably if that portion of the
epiblast which is distinctly thickened is taken to be the area of
increased activity, I think the error will not be great. Up
to now, and for some considerable time to come, there is no
change noticeable in the condition of the hypoblast.

The darkened area PS. in fig. 3 may be taken as the area
of increased activity. The posterior border of this area is
semicircular, the anterior border is conical. The darkest part
of this area is about the position of the posterior border of the
original embryonic disc.

Fig. 4 is a slightly later stage. The anterior portion of the
embryonal area still shows a circular darker part, which I take
to correspond approximately to the original circular embryonic
dise of the sixth day. The secondary area of activity is of the
same nature as in the preceding figure, but its anterior conical
part is slightly more pronounced. On either side of it there are

VOL, 37, PART 2.—NEW SER. N

194 RICHARD ASSHETON.

two parts which are very much less opaque, but still more opaque
than the extra-embryonic part of the blastodermic vesicle.

The specimen of which fig. 4 is a drawing was cut into a series
of transverse sections, that of which fig. 4.4 is a drawing, into
a series of sagittal sections. Fig. 17, Pl. 20, is a very nearly
median sagittal section of the specimen represented by fig. 4 a.
In this drawing A. is the anterior end, and P. the posterior.

The epiblast (ZP.) of the embryonic disc is thickened
throughout. The under or inner surface of the epiblast is even
along the anterior half, while along the posterior it presents a
very irregular, and towards the extreme posterior end, broken
edge.

The hypoblast (HY.) is separate all along, and forms a con-
tinuous layer. Itis specially thickened at the two points where
it underlies the extreme anterior and posterior ends respectively.
There are now between the hypoblast and the epiblast, towards
the posterior end, certain scattered cells, some apparently en-
tirely, others only partially separated from the epiblast. These
are the first mesoblast cells which make their appearance.

The figs. 13—16 on Pl. 21 are from transverse sections of _
the specimen shown in fig. 4. They are drawn on a larger
scale, and, taken together with section 17, demonstrate pretty
accurately the structure of the specimens figs. 4 and 4a.

Fig. 13 is taken through the anterior part of the specimen
along the line 13 in fig. 4. The central portion only is given
in fig. 13 (see fig. 13, A.). This section carries out the evidence
derived from the longitudinal section (fig. 17), and shows that
the under or inner surface of the epiblast is quite smooth ante-
riorly.

There is no median thickness of the hypoblast, and no sign
of mesoblast cells. In fact the anterior region of the embryonal
area has not greatly altered from the condition of the circular
embryonic disc upon the day previous. The posterior end,
however, is very different. Fig. 16 passes through the speci-
men (fig. 4) along the line 16—that is to say, through the very
densest portion of the area.

Here we see the epiblast enormously thickened, and a certain

THE PRIMITIVE STREAK OF THE RABBIT. 195

number of cells seem to be lying separate or forming a loose
network between the epiblast and hypoblast. These are meso-
blast cells.

As the sections pass more posteriorly this primitive streak
knob becomes smaller and ends rather abruptly, as seen in
the sagittal section, fig. 17. But forwards the thickening
of the epiblast is continued much further, getting less very
slowly, as figs. 14 and 15 show, which are taken along the lines
14 and 15 in fig. 4, This narrow band of thickened epi-
blast is seen as the irregular lower surface of the epiblast in
fig. 17. It extends very nearly halfway, along the length of
the embryonal area.

I think it is pretty clear that this is part of the same area
of proliferation as that through which fig. 16 passes, as will
appear more certainly later on. Hence the most anterior
point to which this reaches is a point of very great importance.
This is marked A. PS. in figs. 3 and 4.

The proliferation of cells is so rapid at the posterior end as
to cause an eminence in transverse section (fig. 16; evident
also in fig. 17). The three sections, figs. 14, 15, 16, differ
only in amount of proliferating area and accumulation of
cells proliferated. This I take to mean a difference only in
intensity. Excepting in intensity there is no difference sug-
gested by the structure of the secondary area of activity in the
three sections.

The secondary area of proliferation is now fully established,
and has changed in shape from an ill-defined spot in fig. 2 to
a more and more elongated area in figs. 3 and 4.

In fig. 5 it has still further lengthened. Figs. 21, 22, 23
are sections taken across this specimen at corresponding spots
to figs. 14, 15, 16 respectively, and must be compared one
with the other. It may be noticed that there is here ex-
tremely little difference in the extension transversely of the
area of proliferation. The cells, which appear to be entirely
separated, extend further from the middle line on each side,
but the area of fusion itself (between epiblast and mesoblast)
differs very little. This is especially noticeable in the most

196 RICHARD ASSHETON.

posterior of the sections, namely, figs. 16 and 23. The actual
area of proliferation is about the same.

In the anterior sections, if there is a difference, it is that in
the more advanced specimen (fig. 5) the sections show a smaller
extent of proliferating area than in the younger specimen.

In specimen fig. 5 there is now a very slight groove along
the surface of this area, so slight as to be scarcely recognisable
in sections fig. 22 (P. GR.).

There is another point of interest. Up to this moment the
extreme anterior end of the secondary area of proliferation has
almost shaded away imperceptibly into the epiblast of the future
neural plate, at any rate it has always been the most slender
part of the structure. Now it has become a very well-marked
spot—it forms almost a knob—and is henceforth recognisable
as what other authors have called “ Hensen’s node.” Fig. 21
passes through the centre of this.

Fig. 6 is a later stage still. In this the secondary area of
proliferation has attained about its maximum length, and at
this moment shows the greatest degree of distinction into
different parts: (i) hindermost, a broad area of proliferating
surface, its posterior border forming an arc of a circle,
while anteriorly it tapers off into (ii) the middle region, a
long narrow strip of proliferating area deeply grooved along
its length ; more deeply grooved anteriorly and less posteriorly
(fig. 26). This again passes abruptly into (iii) the most
anterior part of the primitive streak or Hensen’s node, a very
much thickened and compressed region of short extent (vide
section, fig. 25).

This stage is the height of development of the primitive
streak. The secondary area of proliferation is at this stage a
perfectly typical primitive streak, with well-developed primitive
groove.

In the preceding figure (fig. 5) the groove is very shallow
and wide, and easily recognisable in the whole specimen. In
fig. 6 the groove is very deep and narrow, and is only with
difficulty seen in surface views as its walls are nearly approxi-
mated, and it appears when seen by transmitted light as a

THE PRIMITIVE STREAK OF THE RABBIT. 197

narrow dark line instead of a wide lighter line when seen
under similar conditions at the earlier stage.

In all the stages described hitherto there has been no fusion
of the hypoblast with the proliferating epiblast. Im stage fig.
5 at the extreme anterior end of this area, where there is a
special thickening of the primitive streak, the hypoblast is in
very close contact (see figs. 20 and 21).

In this stage (fig. 6) the fusion is complete, and anteriorly
the hypoblast in the middle line and the epiblast and mesoblast
are continuous in the mass spoken of as Hensen’s node.

Figs. 24, 25, 26, and 27 are sections taken along the lines
24, 25, 26, and 27 in fig. 6.

A section taken more anterior still than fig. 24 corresponds
to figs. 13 and 19.

The last three sections (figs. 25, 26, and 27) should be com-
pared with figs. 21, 22, and 23, and figs. 14,15, and 16. Here
again it will be noticed that there is no great increase or
diminution of extent in transverse section of the area of
active proliferation. The most posterior parts differ hardly
at all except in the spreading of the separated cells (meso-
blast).

It must be noticed that there is now a region hitherto not
to be found. The section drawn (fig. 24) passes through this
region, which is just anterior to the primitive streak.

The section through the middle region (fig. 26; compared
with figs. 22 and 15) differs in one particular only (besides ex-
tension of mesoblast). Instead of the area of proliferation
appearing at the surface on a level with the rest of the outer
layer, it lies at the bottom of a deep groove. The sides of the
groove are not “ fused” with mesoblast ; it is only the floor of
the groove that is participating in the addition to the meso-
blast.

Fig. 25, the section through the anterior region of this area
of proliferation, shows an intensified form of the state in the
preceding stage (fig. 21). Here the groove suddenly ceases,
and even in most instances gives rise to an eminence. This
eminence so overhangs the groove posteriorly as to give rise to

198 RICHARD ASSHETON.

a lumen in a section through its posterior half, as shown at
P. GR. in fig. 25.

The origin of this groove—the primitive groove—I shall
discuss in a later paragraph.

The presence of this groove necessitates that the thickened
epiblast of the neural plate anteriorly, and the thickened epi-
blast of the floor of the groove posteriorly, should be at dif-
ferent levels. In this it is possible that any intrinsic growth
of the neural plate may tend to produce an overlapping, and
so give rise to what has been termed a notochordal canal.

The secondary area of proliferation having now attained its
maximum length and most complicated form, it henceforth
tends to return to its original shape, and becomes less and less
like a typical primitive streak.

In fig. 7 no very great change has taken place as regards the
area in question, with the exception that the primitive groove
is not nearly so deep or narrow. It resembles more closely in
transverse section the stage of fig. 5. Also the layer of meso-
blast cells on either side of the primitive groove are thicker
now than in fig. 6.

Fig. 8 shows the area still at its greatest length, and figures
of sections taken through the anterior end and the middle
are given on PL. figs. 80 and 31, and should be compared
with figures of the corresponding sections from the earlier
stages.

The most marked difference is the entire absence
of the groove. Compare fig. 31 with fig. 26.

The apparent canal into Hensen’s node has gone, and the
outline of the embryonal area adjoining this region has
changed shape. In fig. 4 the outline is arched away from the
primitive streak. In fig. 5 it is less arched, in figs. 6 and 7 it
is nearly straight, in fig. 8 it is becoming again curved. From
this moment the length of the secondary area of proliferation
becomes less.

In fig. 10, in which five protovertebre are visible, this area
has diminished to about one third of its former length.

Figs. 32, 33, and 34 are three sections through such a speci-

THE PRIMITIVE STREAK OF THE RABBIT. 199

men. The most posterior, fig. 34, is in no way different from
fig. 27, except that the proliferating area may be slightly
broader. A similar remark may be offered with regard to the
most anterior section, fig. 32, namely, that except the area of pro-
liferation may beslightly broader, it very closely resemblesfig. 25.

In both figs. 32 and 34 the mesoblastic plates on each side
of the fused area are thicker than in figs. 25 and 27.

The section that passes through the middle of the length of
the primitive streak presents the most noteworthy appearance.
In this, fig. 83, the breadth of the area of fusion is very much
greater than in any of the specimens previously described.
There is no groove; in fact, there is, in the place of a groove,
actually aneminence (P. S.). The cells proliferated are much
more numerous and more crowded within a given space as
compared with fig. 31, and still more so as compared with fig. 26.

The total length of this secondary area of proliferation is
now only about one third of what it was when at its maximum
length as in fig. 6. Thus it has apparently become shorter
and thicker.

In an older stage, when there are as many as twelve proto-
vertebrz formed, the foundations of the most bulky portion of
the body have been laid down. The secondary area of pro-
liferation is now less conspicuous. Its length is not much
more than one fifth, if so much, of its greatest attained length.

In stages much later than this it is not possible to observe
it in surface views, as such parts as remain are then placed at
the extreme tip of the tail, after the complete development of
which it disappears.

To sum up—the secondary area of activity arises as a
small spot excentrically placed to the primary centre
of activity. It increases in magnitude, then becomes
elongated, and very much reduced in breadth towards
the centre of its length, and is deeply grooved.
Rather suddenly, after attaining its greatest length,
the groove disappears entirely; the area becomes
much shortened, and thicker at the spot where
it had beenso thin. Now, instead of a groove, there

200 RICHARD ASSHETON.

is a ridge along the median line of the area; it still
furthershortens and diminishes in size, but does not
finally disappear until the last segment has been
formed (or until sufficient material has been added to allow
of the formation of the requisite number of segments).

The above description draws attention to the most note-
worthy features of the primitive streak itself, from the moment
of its first appearance. At no time during its existence, or before
its existence, is there any trace of any form of concrescence
such as Duval has described for the Avian primitive streak. Is
it possible to account for the changes as seen in the rabbit by
a consideration of the conditions under which the blastoderm
of the rabbit is placed ?

The following pages contain what seems to me a likely ex-
planation.

The Process of Elongation of the Primitive Streak.

In the exercise of its functions the secondary area of activity
is essentially centrifugal in its results. While it is, so to speak,
spread out flat upon a plane (as during the stages represented
in figs. 83—10), the most noticeable feature of its action is the
production in every direction of a sheet of mesoblast, of almost
circular outline. When after ten days or so it becomes a freely-
growing knob, it produces a mass of cells, the outline of which
still is circular, the area itself being now also circular in outline.

If the nature of the area, as shown by its action and by the
form assumed when perfectly free, is to be radially symmetrical,
why does it assume a linear form? Why does it become dis-
torted? Is it an ontogenetic distortion, or is it connected
directly with phylogenetic causes ?

I believe the actual elongation, or rather all the changes it
undergoes, including the grooving, is due entirely to the
ontogenetic influences, and has no recapitulatory meaning. In
one sense it may be said to be due to the phylogeny ; but only
in that, unless the mammalia had been descended from animals
which had large yolked eggs, the ontogenetic development of a

THE PRIMITIVE STREAK OF THE RABBIT. 201

rabbit would not, in all probability, have been upon the lines
on which it is—provided a rabbit could ever have been evolved
under other circumstances. What I mean is: a bird ora reptile
has a large egg because the embryo obtains its nourishment
from yolk, the presence of which causes the egg to be so large.

A mammal does not require the yolk as a nourishment, and
therefore the eggis small. A mammal is enabled to develop
within the body of its mother because, we presume, originally
the embryo was protected within the membranes of a large egg
within the oviduct of the mother, and it was only by the
substitution, no doubt very gradual, of (i) placental nourish-
ment instead of the yolk, and (11) fluids exerting considerable
pressure on the uterine walls instead of yolk, albumen, and
shell, that it was possible for a mammal to dispense with
a large-yolked egg and to be developed under the conditions
in which we now find it.

Thus it is to the phylogeny that the conditions under
which development takes place to-day are due. Now these
conditions were imposed upon the development at a compara-
tively very recent period in the evolution of the rabbit.

So when we find that certain forms assumed by certain
centres of activity, which centres of activity themselves
undoubtedly date back to infinitely earlier epochs than the date
of these superimposed conditions, are due entirely to the con-
ditions of this day, we must be very careful in drawing any
morphological conclusions from those forms assumed. If, for
instance, I can show that the changes in form of the primitive
streak from a nearly circular spot to a linear expression, and
its groove which appears during part of its existence, are due to
these present conditions alone, then to draw any conclusions
such as frequently are drawn—as, for example, the theory that
the line-like primitive streak and its groove represent the lips
and aperture of an elongated ancestral gastrular mouth, is
impossible.

If the blastodermic vesicle, at the time that the secondary
area of activity is established, became an object freely suspended
in some fluid, and was under no influence of internal hydro-

202 RICHARD ASSHETON,

static pressure, it seems probable that the immediate effect of
this rapid proliferation of cells at one spot would be to produce
at first a mass of cells or a knob, as in reptiles, and then a
wrinkling of the thin wall of the vesicle, and eventually a pro-
jection or outgrowth. This is, indeed, what takes place a little
later, when, by reason of the attachment to the uterus of the
walls of the blastodermic vesicle immediately surrounding
the embryo, the circumstances are such as to fulfil these condi-
tions of equilibrium, The result at this time is that a definite
projection and rolling over is effected—the tail fold.

But as yet the blastodermic vesicle is unattached and lies
freely within the cavity of the uterus, but is continually ex-
panding in size as described before, owing to the combined
effect of increasing hydrostatic pressure and multiplication of
the cells of its walls.

I believe that the conversion of the almost circular spot in
fig. 2, or pyriform area in fig. 3, to the linear expression of figs.
6 and 7, and back again to the pyriform area of fig. 10, is inti-
mately connected with two facts:

(i) Growth of the embryo taking place from different centres.

(11) Expansion of the whole area, due in this case (rabbit) to
the hydrostatic pressure within the blastodermic vesicle.

Ina former paper, “‘ A Re-investigation into the Early Stages
of the Development of the Rabbit,” I ascribed the apparent
growth round the hypoblast as being due to its being carried
round by the outer wall, due to the combined effect of the
greater area of special activity around the embryonic disc,
aud increasing hydrostatic pressure within. May not a
similar explanation be given to account for the apparent
growth outwards of the primitive streak mesoblast from the
primitive streak? In this case the area of special activity is
much more concentrated and vigorous than in the former, and
is therefore itself evident as an area of rapid growth. All the
cells towards its periphery will be as liable to be removed by
the expansion of the walls as the outer layer cells, as with
them the inner mass cells towards the periphery are, according
to my hypothesis, apparently removed in the former case.

THE PRIMITIVE STREAK OF THE RABBIT. 203

Having drawn attention to this possibility, I must leave the
consideration of the mesoblast for the present.

What are the conditions under which the blastodermic vesicle
exists during the lengthening of the primitive streak and the
formation of the deep primitive groove ?

In the previous paper mentioned above, I have accounted
for various changes in shape of the blastodermic vesicle, as
well as the apparent growth of one layer over another,
by pointing out how these changes could be produced by
increasing hydrostatic pressure within, together with more
rapid cell-division in one part, and less rapid cell-division in
another part.

So, I believe, the same principle holds good with regard to
the lengthening of the secondary area of proliferation into a
streak, and to the production of the median groove.

The blastodermic vesicle is still expanding rapidly by reason
of the increasing hydrostatic pressure. The walls of the vesicle
are thin except at one point, the embryonic disc, where the
wall is thick and compact. Somewhat suddenly, a spot at the
posterior edge of this embryonic disc becomes extremely
active—the secondary area of proliferation. This gives rise,
as we have seen, to a mass of cells produced very rapidly at
one spot (fig. 3), forming a second compact disc.

We have now two very distinct compact masses in the outer
layer or epiblast, one formed by the presence of the inner layer
of epiblast and its fusion with the outer layer, the other by the
very rapid proliferation at one spot caused by the appear-
ance of the secondary growing point or primitive streak.
While this is proceeding, the whole blastodermic vesicle is still
expanding from the hydrostatic pressure within. Now what
effect, if any, will this hydrostatic pressure have in regard to
these two compact masses? In the first place there will be a
tendency for them to part. There will be a tendency to part from
another cause, namely, the counteracting effect of their re-
spective growths ; but this would tend to produce a space (i. e.
mass of cells) between the primary growing point and the ante-
rior border of the secondary growing point (i.e. anterior end

204 RICHARD ASSHETON.

of primitive streak). This takes place, no doubt, and isa very
conspicuous feature at a later stage, as described later, but it
does not account for a lengthening of, and within,
the secondary area of proliferation itself.

The lengthening of the, at first, circular patch into a streak,
i.e. the change from fig. 11 to fig. 12, I believe to be due almost
entirely to the expansion of the blastodermic vesicle by the
hydrostatic pressure within.

The expansion of the vesicle is quite sufficient to allow of
this. It is not easy to make very satisfactory measurements
in support of or against this view. If we take a specimen
about the stage shown in fig. 3, we find the secondary area to
measure not less than 38 mm. antero-posteriorly, and the
diameter of the whole vesicle about 4°5 mm. to 5 mm.

If we take another specimen in which the secondary area has
attained its maximum elongation, we find the area to measure
about 1:08, which is rather less than three times the length it
was in stage fig. 3. The diameter of such a blastodermic vesicle
is about 10 mm. to 12 mm., which is from rather under two
and a half to rather over two and a half times the length it was
in stage fig. 3.

I may, perhaps, make my meaning clearer by reference to
figs. 11, 12, Pl. 20.

In fig. 11 the point M is the centre of the primary area of
cell-production, i.e. the centre of the embryonic disc. The
point N is the centre of the secondary area of cell-production.
The grey represents the circular embryonic disc; the white
represents the proliferating area of the primitive streak.

Really this diagram represents no actual stage. Fig. 3 is
very near it. If my account of the course of events is correct,
there never could be a stage which would be exactly repre-
sented by this diagram, although it theoretically represents the
condition.

Now the tendency of each of these two areas considered
separately would be to expand equally in all directions, from
the centre Min the one case, and Nin the other. But these
two areas are not separate, they even overlap, and therefore

THE PRIMITIVE STREAK OF THE RABBIT. 205

they must have some disturbing effect upon each other; and
the line along which that disturbance will be most marked will
be along a line drawn from the centre of the one to the centre
of the other—from M to N.

I have argued before that while the increasing hydrostatic
pressure within the blastodermic vesicle causes its expansion,
yet a more rapid expansion of one part of the wall than another
may be brought about by the more rapid cellular growth of
that part of the wall. In other words, areas of weakness are
produced upon which the hydrostatic pressure takes more
effect than elsewhere.

So at this stage must there not be a line of weakness between
the two points M and N in diagram 11, an area along, at any
rate, that portion in which the influences of both centres are
at work, in which there will be less resistance offered to the ex-
pansive force of the hydrostatic pressure than elsewhere? It
seems to me that this is extremely likely, and that a figure
such as that represented in diagram fig. 12 must be the result.

The anterior half or portion of the secondary area of pro-
liferation will be drawn out as indicated in the diagram by the
separation of the black dots; the posterior portion, being unin-
fluenced by the primary area, will not lose its radial symmetry.

Similarly the embryonic disc will be influenced, and its
hinder portion drawn out as indicated by the black crosses on
the two diagrams, its anterior border not losing its radial sym-
metry. It must be remembered, however, that the secondary
area of proliferation is in its function of cell-production more
vigorous now than the primary area, and becomes still more
vigorous for the next two days.

Also the visible line of the primitive streak does not even at
this comparatively early stage represent the whole of the effects
of the secondary area of proliferation. No doubt the epiblast
immediately surrounding it owes its origin to the energy of the
area in question, but to what extent it is not easy to determine.
At any rate it is probably not less than that part beneath which
primitive streak mesoblast lies. So that a more accurate
figure of the conditions of the two areas of activity respectively

206 RICHARD ASSHETON.

would probably be such as I have drawn in fig. 43, on Plate
22, to be explained presently. I conclude, therefore, that the
lengthening of the primitive streak is due to the expansion in
an antero-posterior direction of the most anterior part (half?)
of the secondary area of cell-proliferation due to the conditions
under which the embryo is developing.

May not the primitive groove be also due to ontogenetic
conditions ?

It will be noticed that it does not exist during the early
stages of the formation of the primitive streak (vide figs. 1—
4a, and the sections of these).

It will be noticed also that it is at its greatest development
at the time the primitive streak reaches its greatest length, and
from that moment it becomes shallower and very rapidly dis-
appears altogether (vide fig. 8 and sections). It is deepest
and most distinct along that part of the streak which is
thinnest, and gradually shallows and disappears towards the
thicker posterior end of the streak.

The sections through the stages in figs. 4, 5, and 6 show
that the groove commences at the same time as the spreading
of the mesoblast.

The mesoblast is a reticulum connected with the area of pro-
liferation, and is seen to spread out from both sides of the
drawn-out proliferating area.

I have explained above how this spreading out of the meso-
blast may be due to the general expansion of the walls of the
vesicle of this region by the increasing hydrostatic pressure.
If the proliferating area produces cells more quickly than they
can be removed laterally by the expanding walls of the vesicle,
a heap of cells will be the result at this spot, or if the prolife-
rating area is linear a ridge will be formed. If the cells pro-
duced exactly balance those removed, the surface will remain
flat ; there will be no ridge.

But what will happenif the cells produced along the prolife-
rating area are removed more rapidly than they are produced ?
If the deficiency of production is very great, then probably the
cells—few in number—will be torn apart and removed as

THE PRIMITIVE STREAK OF THE RABBIT. 207

isolated cells. But if the deficiency is only slight, surely the
result will be that the meshes of the reticulum of mesoblast
will be drawn out into larger meshes, and the connecting
filaments become finer.

It seems to me that this must mean that a tension between
cell and cell must be produced, and not only between cell and
cell of the reticulum, but between the reticulum on each side
of the primitive streak and the actual proliferating area itself.
Is it not possible for this tension to produce such a groove as
that shown in figs. 22 and 26?

Posteriorly there is no groove. Here the supply of cells is
greater than the removal, and a heap is formed ; but along the
narrowest part of the proliferating area the conditions favour the
formation of a groove as I have suggested, for not only is the
proliferating area very attenuated, but lateral dragging of the
mesoblast is in the same direction along a considerable length.

It follows that upon the above explanation of the lengthening
of the primitive streak and formation of the groove any increase
in activity of the proliferating area might lead to the oblitera-
tion of the groove, and even perhaps a cessation of the increase
in length of the streak.

Also any modification which either prevented the increase of
hydrostatic pressure within the vessels, or which prevented the
portion of the wall containing the embryonal area from being
affected by the increase of hydrostatic pressure, would bring
about the obliteration of the groove and cause the cessation of
lengthening of the streak. In the rabbit it seems probable
that the latter event occurs.

During the stages represented by figs. 1 to 5, that is during
the formation of the greater length of the primitive streak, the
blastodermic vesicle lies freely in the uterus. About the age
represented by fig. 7 the papille are appearing by which the
lower pole becomes attached to the obplacental portion of the
uterus. The ectoplacental area is thickening, but is still quite
smooth and quite free to slide over the surface of the placental
lobes; but between that time and the time represented by
fig. 8, the ectoplacenta has become irregular, the albumen

208 RICHARD ASSHETON.

layer has ruptured, and the ectoplacental region is firmly
attached to the placental lobes.

From this moment the upper pole of the blastodermic
vesicle can only be expanded by the hydrostatic pressure as
much as the placental lobes and the thick mesometrial wall of
the uterus will allow it. As is well known, this portion of the
uterus expands now very little and slowly; by far the greater
part of the swelling of the blastodermic vesicle concerns only
the ab-embryonic pole of the vesicle and the obplacental lobe
of the uterus.

In this way a very large part, if not all, the tension is
rather suddenly removed from the actual embryonal area and
immediately surrounding tissues, and this is exactly coincident
with the attainment of the maximum length of the primitive
streak and with the rather sudden obliteration of the primitive
groove.

No doubt to all appearances the primitive streak of a bird
is at the time of its greatest development extremely like the
primitive streak of a mammal of the type of development such
as we find in the rabbit.

If we accept Duval’s description for the bird, and also the
explanation I have offered above for the rabbit, as being both
correct, we are bound to conclude that it is really only a coin-
cidence that the secondary area of cell-production in each case
should assume a linear form. Such a coincidence is unlikely,
though of course not impossible.

An Attempt to determine which Portion of the
Embryo is derived from the Cells proliferated
from the Primitive Streak.

Those who have followed my description of figures will have
noticed that, according to the account, there is one part.of the
embryo laid down as the immediate result of the segmentation
of the ovum, and that the most conspicuous part of this is a
circular patch—the embryonic disc. Subsequently a renewed
activity of cell-production takes place upon the posterior
border of this embryonic disc, giving rise to new tissue, and

THE PRIMITIVE STREAK OF THE RABBIT. 209

continues active till the formation of the extreme end of the
tail is completed.

Is it possible to determine approximately which parts of the
embryo are formed by the two centres of growth respectively?
Evidence may, I think, be brought to show that the primitive
streak in the rabbit is the growing point of the whole of that
part of the animal which is situate posterior to the head, while
from the primary centre of activity is formed the part of the
embryo anterior to the first protovertebra. In other words, the
secondary centre of growth is responsible for the metamerically
segmented part of the animal.

During the later stages of development the line of demarca-
tion between the two—which perhaps is never absolutely
definable—becomes less and less distinct. For instance, the
heart, which is formed distinctly in the region of the primary
centre of growth, becomes located more posteriorly; while
parts of the nervous system, owing to the more rapid growth
of the neural tube, become moved forwards, so as to bring into
the primary region portions whose origin has been due to the
secondary area of activity.

If we regard the secondary area of proliferation or primitive
streak simply in its functional capacity, neglecting for the
moment all preconceived ideas of its morphological or recapi-
tulatory meaning, and consider it only as we find it in the
rabbit embryo, it seems to me that we are bound to describe
it as the visible expression of an area of intense protoplasmic
activity, which is continuously budding off cells, the most con-
spicuous of which are those from its lower surface, the primi-
tive streak mesoblast cells. These cells, according to my
hypothesis, are rapidly removed from this area, being carried
away by the expanding wall of the blastodermic vesicle, to
which they are closely approximated. This part of the wall is
enabled to respond to the expanding influence of the hydro-
static pressure within, by itself receiving rapid additions of
cells from the proliferating area.

On this hypothesis we may take the outline of the primitive
streak mesoblast as indicating also the outline of epiblast

VOL. 37, PART 2,——NEW SER. )

210 RICHARD ASSHETON.

derived from the primitive streak, at any rate for the stages
up to about the one drawn for fig. 6. At this time mesoblast
cells are separated from the hypoblast, both at points where
there is already existing primitive streak mesoblast, and at
points where up to now there has been no mesoblast, e. g.
under the part of the embryonal area in front of the primitive
streak. After this stage it is extremely difficult to determine
in many places what the origin of the mesoblast cells has been.
As, however, this involves the question of the formation of the
whole mesoblast, I must leave the further discussion for another
paper.

I must now refer again to the figures of the embryonal area
(figs. 3 to6). I have given above my explanation of the change
in shape, and of the process of elongation of the proliferating
area.

Has the whole difference in length between the embryonal
areas (figs. 5 and 6) been due to the elongation of the primi-
tive streak itself?

In the following paragraphs the measurements given are taken
from measuremeuts made upon photographs of the specimens,
drawings of which are given in figs. 3, 4, 5, 6, 8, and 10.
The measurements, reduced to their natural magnitude, are
given in millimetres.

It is evident that growth has taken place somewhere, for the
embryonal area in fig. 6 measured 1°72 mm., while fig. 5
measured but 1°38 mm. It is due hardly at all to lengthening
of the primitive streak area, as the primitive streak area of
fig. 6 measured 1:02 mm., and that of fig. 5 measured 1 mm.
Of course one must not suppose that the primary area of
growth has ceased to produce effects. Its effects hitherto
have been to produce a circular patch. If we suppose it still
to exercise a like influence the diameter of the head plate
will have increased anteriorly the same as it has from side
to side.

Fig. 5 measures transversely ‘94mm. Fig. 6 measures trans-
versely 1:1 mm. Therefore we may say that at least (16 mm,
of the length may have been due to growth of the anterior

THE PRIMITIVE STREAK OF THE RABBIT. 211

portion. So by adding to 1:38 mm. (length of fig. 5) the
sums of ‘02 mm. due to lengthening of the primitive streak
and ‘16 mm. the amount due to growth of anterior end, we
have ‘16 mm. still to account for to produce the length shown
in fig. 6.

Where has this extra length been acquired? In transverse
section it is evident that there is now a region which was not
to be found in the preceding stage.

A section through this region is given in fig. 24. The most
noteworthy feature is the notochordal thickening. A few
sections further forward this thickening is quite absent. It is
surely legitimate to ascribe the appearance of this new region
as being the result of cells budded off from the front end of
the primitive streak. This becomes much more evident in the
later stage (fig. 8). In this the total length of the embryonal
area is about 3°11] mm.

Fig. 6 measured 1°72 mm. There has therefore been an
increase of 1:39 mm. The proliferating area measured the
same as in fig. 6, namely 1:2 mm. There is at the anterior
part of the embryonal area a very well marked circular area.
The neural groove along this area is narrow, while the
neural groove along the embryonal area between the above-
mentioned circular area and the anterior end of the primitive
streak is broad and shallow. The junction between the two
corresponds almost exactly with the circumference of the
circular area at the point Z.

Does this anterior circular patch represent either exactly or
approximately the area of influence of the primary centre of
activity ? I wish to point out the possibility of all that part
between the anterior end of the primitive streak and the
posterior border of the circular anterior area (namely that
part of the embryonal area which is marked by the shallow,
broad, neural groove) having been formed entirely by the
activity emanating from the primitive streak area, or, to be more
exact, the anterior and antero-lateral parts of that area
of proliferation ; the parts of the embryonal area in front of this
region having been formed chiefly as the result of the activity

Ai? RICHARD ASSHETON.

of the primary centre of growth (the most anterior portion
entirely from the primary centre, the most posterior possibly
from both).

The circular area measures transversely 1°2 mm. Therefore
we may suppose that at least ‘1 mm. of the increase in total
length of the whole embryonal area may be due to the activity

_of the primary centre. This leaves ‘97 mm. to be accounted
for.

If we measure 1°2 mm. (the total length assumed to be due
to the primary area of activity) from the anterior end we come
to a spot Z. The space between this spot Z. and the anterior
end of the primitive streak, 4. PS., measures ‘97 mm. This
seems to me to be evidence that this space, Z. to A. PS., is
the area which has been added between the stages fig. 6 and
fig. 8. The question remains, where have the cells composing
this portion of embryonal area originated ? They must have
arisen either (i) from an area of rapid proliferation in front, or
(ii) from a general area of proliferation in situ, or (il) froma
general area of proliferation behind. Absolute proof is, as far
as I can see, impossible, but I will take each alternative
separately, and give my reason for believing that the latter
only can be the true solution.

(i)—(a) From the very first moment of the development
the effect of the primary centre of activity is in the ontogeny
of the rabbit to produce an embryo of a radially symmetrical
figure, whereas the area in question between Z. and A. P. S.
shows no sign of a radial symmetry.

(6) By the time that the area in question is developed, and
while it is being developed, there is no sign of one spot from
which all pre-existing structures could obtain material.

For these reasons I think growth does not take place in front.

(ii)—(a) Growth of the area in question is, comparatively
speaking, rapid, but there is no sign of the rapid multiplication
of undifferentiated cells, every cell within this area is distinctly
either epiblast, hypoblast, or mesoblast.

(6) This area continues to grow rapidly in length, and the
“ mesoblastic somites,’ which are a characteristic feature of

THE PRIMITIVE STREAK OF THE RABBIT. ots

this area, are continually being added on posteriorly, showing
that the differentiation is at any rate taking place from
before backwards.

(c) If this is a centre of growth there should, under the cir-
cumstances, be some sign of a radial disposition of tissue, which
there is not.

For these reasons I think we cannot conclude that there is
any positive proof of a rapid addition of new cell material,
although an arrangement of the mesoblast cells into meso-
blastic somites undoubtedly takes place in situ.

(iii) Since, therefore, we cannot find probable proof of
growth of new segments either anteriorly or in the area in
question itself, and since we do find an area immediately
behind the area in question and continuous with it in which
there is an undoubted rapid proliferation of cells (which are all
similar and connected, and only become epiblast, hypoblast, or
mesoblast according to which region they may adjoin), I think
it may be considered that there is very good evidence indeed
that the growth of the embryo of the area in question has
taken place by the addition of indifferent cells from this area
—the primitive streak.

Although such measurements as the above are useful in
support of my contention, they are not satisfactory, because
every specimen of exactly the same stage is by no means always
exactly the same size.

A very useful landmark is supplied by the band of tissue, or
rather an area which gives rise at first to the mesoblast form-
ing the pericardium, and rather later the endothelial lining of
the heart. This area is clearly marked as a circular band in
fe LOC.

In this stage the neural tube, owing to its more rapid growth,
has become folded forwards so as to hide the anterior part of
the pericardial band PC.

In fig. 28, Pl. 20, which is a median sagittal section of this
specimen, it is seen at the point PC.

In fig. 9, an earlier stage, before the neural tube has become
thrust forward, this band is seen to follow the contour of the

214 RICHARD ASSHETON.

neural tube (NP.). So also, but less distinctly, in the earlier
stage (fig. 8).

In the stages earlier than this it is not perceptible in surface
views, but may be easily recognised at its first appearance in
stages like figs. 4 and 4a.

At this time—its first appearance—it shows itself as a slight
tendency to a more rapid growth of the hypoblast immediately
underlying the anterior and lateral edges of the embryonal
area.

Figs. 17 and 18 show the anterior edge cut at PC.

Sections 13 and 14, if continued to the edges, would show the
same slight thickening of the hypoblast.

Figs. 35—38, Pl. 21, show the subsequent history of this
thickening.

Fig. 35 is a transverse section through the lateral edge of the
anterior part of the embryonal area of a stage between those
represented in figs. 4 and 5. It is in front of the primitive
streak area. There is as yet in thisregion no mesoblast. The
hypoblast (HY.) shows, however, signs of increased activity, and
is thicker than before. This is specially the case at the
edge of the thickened epiblast.

Fig. 36 is a slightly older specimen, still rather younger
than fig. 7. Here the hypoblast (HY.) is thickened over the
whole area, extending under the thickened epiblast, but more
especially so at the edge of the epiblast.

A certain number of cells are to be seen lying between the
epiblast and hypoblast, marked PC. in my drawing. These
I believe to have been budded off in situ from the hypoblast
(HY.). Cells are budded off from the hypoblast all over this
area of the anterior part of the embryonal disc, but more
thickly at this region round the edge of the thickened epiblast
than elsewhere. Fig. 37 is a section from the same region of a
later stage, a stage with two mesoblastic somites (vide fig. 9).

There seems to be here still a slight proliferation of cells
from the hypoblast of this region, but not so great as before.
The cells formerly budded off, which we can call mesoblast,
have become arranged so as to leave a slight cavity between

THE PRIMITIVE STREAK OF THE RABBIT. Al |e)

them (PC.). This cavity gives rise to the pericardial cavity,
the inner thick wall to the muscular wall of the heart, and
peritoneal lining of the cavity around the heart; the outer
thinner wall to the peritoneal lining of the rest of the peri-
cardial cavity. The cells, which seem now to be budding off
from the hypoblast of this spot (marked END. H.), give rise, I
believe, to the first of the cells which, in fig. 38, are seen to be
forming into a tube, which is the endothelial lining of the
heart (END. H.).

Fig. 38 is from an older stage, an embryo with seven meso-
blastic somites, a little older than fig. 10. Although it is not
possible to give a decided opinion, I am inclined to think that
at this stage cells are still being budded off from the hypoblast
to form the endothelial lining of the heart.

At the same time I must mention that at an age intermediate
between figs. 37 and 38 I have found a stage in which a mass
of cells lies between the PC. cells and the hypoblast cells,
showing no trace of a budding off either from the hypoblast or
the mesoblast mass marked PC.

In fig. 88 the endothelial cells may be seen to be attached
to both the pericardial cells and the hypoblast cells, though the
latter give one the impression of being concerned in the pro-
duction of the cells in question rather than the former.

This band, originating as a thickening of the hypoblast at a
time when the embryonal area is hardly at all affected by the
primitive streak activity, seems to be of great use in preserving
the outlines of the embryo due to the primary centre of
growth.

Now no part of the pericardial thickening seems to be
affected in the growth of the embryo caused by the secondary
centre of activity, the primitive streak, excepting that it is
absent posteriorly. Where it is present, that is, anteriorly and
laterally, it forms almost an accurate “‘ segment ” of a circle,
the circumferential boundary of which extends through about
270°, or three quarters of a whole circle.

From this I argue that all the parts of the embryo which are
formed within the outer peripheral boundary of the pericardial

216 RICHARD ASSHETON.

band may be ascribed to the activity of the primary centre of
growth (segmentation of the ovum) ; all posterior to this due
in the main to the secondary centre of growth (primitive
streak). Referring again to fig. 8, it will be seen that this
again indicates the point Z.as the line of demarcation between
the two areas. This spot is marked by a very distinct difference
in the character of the neural groove, which is seen in the later
stages also. It is at this spot that the first protovertebre are
formed (vide fig. 9).

As development proceeds, the fore-brain, whose outline is
in figs. 8 and 9 concentric with that of the pericardial band,
becomes thrust forward (vide fig. 10), so as when viewed from
above to be no longer concentric. So also there is a similar
thrusting forward of the whole of the dorsal portion of the
embryo, including the protovertebre, and the pericardial band
no longer serves as a landmark of the same character.

The Shortening of the Primitive Streak.

Although the primitive streak becomes very much shortened,
Tam not at all sure that its area is diminished during the process.

The hinder end of the streak is about the same width through-
out its existence up to such a stage as fig. 10, and somewhat
later, but the elongated anterior portion is extremely narrow.
After the shortening the anterior part is almost as wide as the
posterior part. So I doubt whether there is any diminution of
area during the contraction.

After the cessation of tension has been effected by the close
attachment of the surrounding walls of the blastodermic vesicle
to the uterus, the result of growth must be soon to cause pres-
sure of the nature of a thrust in the tissues which are most
actively growing. This, we know well, brings about a very
considerable thrust in the direction of the longitudinal axis of
the embryo, causing the head and tail folds.

There is also the tendency that has all along existed for the
secondary area of proliferation to be radially symmetrical. This,
together with the pressure which is known to exist in the direc-
tion of the longitudinal axis, seems to be quite sufficient cause to

THE PRIMITIVE STREAK OF THE RABBIT. 217

bring about the compression of the primitive steak, without
assuming a conversion of it in situ, as it were, into the
embryo.

On Pl. 22 will be found six diagrams which illustrate my con-
ception of the lines of growth between the stages of figs. 1 and
8 of Pl. 20.

Fig. 40 represents the embryonic pole of the blastodermic
vesicle before there is any visible sign of the appearance of the
secondary area of cell-production. The grey central area is the
embryonic disc. The vesicle is expanding approximately equally
in all directions ; this is indicated by the concentric circles
round the embryonic disc. The dotted line is an imaginary
line drawn outside the embryonic area. All parts of the vesicle
outside this line are only slightly affected by the subsequent
origin of the secondary area of cell-production, as seen by the
next diagrams.

In fig. 41 the secondary centre of activity has become esta-
blished, its centre being about the spot marked with a cross in
the posterior region of the embryonic disc (grey). The effect
is as yet slight, leading to a little more rapid expansion of that
part of the embryonic disc indicated by the ellipticity of the
lines of growth. This represents the stage of fig. 2, Pl. 20.

In fig. 42 the activity of the secondary area of cell produc-
tion has become more intense. It is very concentrated, and is
now marked by a heaping-up of cells. The outline of this
actual thickening must not be taken as being the boundary of
the part of the wall of the vesicle due to the secondary centre
of activity, for the outermost cells produced thereby will no
doubt be stretched and flattened.

Fig. 43 represents a stage intermediate between figs. 5 and 6
on Pl. 20. The primitive streak is at its greatest development.
The anterior part of the embryonal area, which owes its existence
to the primary centre of activity alone, is shown to have its
posterior borders distorted as explained in the earlier part of this
paper. It is practically the same condition as that illustrated by
fig. 42, but more pronounced. The outline of the part of the
wall of the vesicle due to the secondary centre of activity is de-

218 RICHARD ASSHETON.

rived from the outline of the primitive streak mesoblast, which
upon my hypothesis would approximately mark this area.

Immediately in front of the anterior end of the secondary
area of activity (i.e. in front of the front end of the primitive
streak) there will be a lesser tendency to expansion, and
accordingly a heaping-up of cells proliferated from the front
end of the primitive streak, forming the thickening known as
*‘Kopffortsatz.” From this moment the increased intensity of
action of the secondary area of cell-production over the primary
is shown by the fact that the outline of the former increases its
diameter four times by the stage illustrated by fig. 8, whereas
the latter’s increase is scarcely perceptible.

Fig. 44 illustrates the conditions of fig. 8. The two centres
of growth have become further removed from each other by the
interposition of new cell material proliferated chiefly by the
secondary area. In other words, growth in length of the
embryo has now very markedly taken place.

This growth in length is represented in the diagram by the
area within the curves of which the one marked 1 is the outer-
most. The region of the embryonal area due to the primary
activity has recovered from its temporary distortion, and has
regained its radial symmetry to a great extent. This can be
detected in the series of drawings figs. 1—8.

The secondary area of activity also, upon the diminution of
the tension to which it had been subjected, tends also to assumea
more radial form (fig. 45, with which fig. 10 may be compared),
Ultimately the counteracting effects of the two centres cause
each to become tilted over, so that instead of lying in the same
plane they lie in different planes parallel to each other, and at
right angles to the original common plane. In fact they
assume their natural positions,

On this explanation it is clear that most, if not all, of the
ectoplacental region is really derived from the primitive streak,
which may perhaps account for its much greater activity than
that part of the blastodermic vesicle which arises directly from
the primary centre of activity.

THE PRIMITIVE STREAK OF THE RABBIT. 219

EXPLANATION OF PLATES 20—22,

Illustrating Mr. Richard Assheton’s paper on “ The Primitive
Streak of the Rabbit; the Causes which may determine
its Shape, and the part of the Embryo formed by its
activity.”

List oF REFERENCE LETTERS.

A, Anterior end. AM. Amniotic cavity. 4. PS. Anterior end of primi-
tive streak. #ND.H. Endothelial lining of heart. 4#P. Hpiblast. AN.
Hensen’s node. HY. Hypoblast. M4. Centre of primary area of cell-pro-
duction. MHS. Primitive streak mesoblast. MHS. HY. Hypoblastic meso-
blast. 2. Centre of secondary area of cell-production. CH. Notochord.
NG. Neural groove. WP. Neural plate. iP. Posterior end. PC. Peri-
cardial band. P. GR. Primitive groove. P. PS. Posterior end of primitive
streak. PS. Primitive streak. Z. Point which marks approximately the
boundary between the results of the primary centre of growth and that of
the secondary centre of growth.

PLATE 20.

Fie. 1.—Surface view of the embryonic dise of a rabbit embryo of the
150th hour. x 18.

Fic. 2.—Surface view of the embryonic disc of a rabbit embryo, at the
earliest moment at which the primitive streak activity is perceptible as a
thickening. Age about 158 hours. x 18.

Fic. 3.—Surface view of the embryonal area of a rabbit embryo, in which
the primitive streak is very evident. Age about 168 hours. x 18.

Fic. 4.—Surface view of the embryonal area of a rabbit embryo, in which
the process of lengthening of the primitive streak is well advanced. Age
about 168 hours. x 18.

Fic. 44,.—Surface view of the embryonal area of a rabbit embryo, slightly
more advanced than the preceding. Age about 168 hours. x 18.

Fic. 5.—Surface view of the embryonal area of a rabbit embryo, in which
the primitive streak has become greatly elongated and is grooved slightly.
Age about 172 hours. x 18.

Fic. 6.—Surface view of the embryonal area of a rabbit embryo, in which
the primitive streak has attained its maximum length, and is most deeply
grooved. This figure has unfortunately been drawn slightly larger than the
photograph from which it was copied, on which the measurements in the text
were made. The error is about 75. Age about 180 hours. x 18.

220 RICHARD ASSHETON.

Fic. 7.—Surface view of the embryonal area of a rabbit embryo, in which
the primitive streak has become much shallower. Growth in length of the
embryo has now taken place. Age about 188 hours. x 18.

Fic. 8.—Surface view of the embryonal area of the embryo of a rabbit.
The primitive groove has entirely disappeared. The ecto-placental region
is now firmly attached to the placental lobes of the uterus. Age about 192
hours. xX 18.

Fig. 9.—Surface view of embryonal area of the embryo of a rabbit. Age
about 196 hours. x 18.

Fie. 10.—Surface view of the embryonal area of the embryo of a rabbit.
The primitive streak has now become compressed to its former dimensions ;
the mid-dorsal portions of the embryo have become thrust forward, and the
head-fold has commenced to be formed. Age about 200 hours. x 18.

Fie. 11.—Diagram to illustrate the relative positions of the two areas of
cell-production.

Fic. 12.—Diagram to illustrate the effect produced upon the two areas of
cell-production by the increasing hydrostatic pressure within the blastodermic
vesicle.

Fie. 17.—A sagittal section through the specimen of which Fig. 44 is a
drawing. It is taken along the line 17. x 100.

Fig. 18.—A median sagittal section through a specimen intermediate be-
tween Figs. 5 and 6. xX 72.

Fie. 28.—A median sagittal section through Fig. 10. x 54.

Ftc. 39.—A horizontal section through a portion of the sheet of mesoblast

surrounding the primitive streak. The portion drawn was a lateral portion.
x 350.

PLATE 21.

Fies. 13, 14, 15, 16.—Transverse sections through specimen Fig. 4, along
lines 183—16. x 175.

Fics. 19, 20, 21, 22, 23.—Transverse sections through specimen Fig. 5,
along the lines 19—23. x 175.

Fies. 24, 25, 26, 27.—Transverse sections through the specimen Fig. 6,
along the lines 24—27. x 175.

Fries. 29, 30, 31.—Transverse sections through a specimen similar to that
drawn in Fig. 8, along lines corresponding to those marked 29, 30, 31,
in Fig. 8. x 175.

Fics. 32, 33, 34.—Transverse sections through the anterior end, the
middle, and posterior end of the primitive streak of a specimen slightly older
than that of which Fig. 10 is a drawing. xX 175.

THE PRIMITIVE STREAK OF THE RABBIT. 221

Fic. 35.—A transverse section through the edge of the embryonal area of
a specimen rather older than Fig. 4, on a level corresponding to the line 13.
x 175.

Fie. 36.—A transverse section through the corresponding region of a
specimen like Fig. 7. x 175.

Fie. 37.—A transverse section through the corresponding region of a
specimen like Fig. 9. x 175.

Fic. 38.—A transverse section through the corresponding region of a
specimen rather older than Fig. 10. x 175.

PLATE 22.

Fie. 40.—Diagram showing lines of growth, i.e. expansion of blastodermic
vesicle of the embryonal area and surrounding parts of a rabbit embryo,
corresponding to Fig. 1, Plate 20.

Fic. 41.—A similar diagram of a slightly later stage, corresponding to
Fig. 2 at the time of the first appearance of the secondary centre of growth,
which is placed eccentrically to the centre of the primary area of growth.

Fies. 42, 43, 44.—Similar diagrams illustrating the increase in importance
of the secondary centre of growth. These correspond to Figs. 3 or 4, between
5 and 6, and to 8.

Fie. 45.—A portion of a similar diagram of the posterior end only of the
embryonal area of a specimen, corresponding to Fig. 10, or a little later.

‘ay

ys

.
WS Rs A ee eh ‘

ty oese ahi Rev rg Bite Dif tis =

Ae CN ee

w > Be - @

a
SE me Shae eat a0 |

~ o os =

“9 rey tohgel Pays si
a 7 : 7

ah ie sd iuher Jah pall wat
art ey

’ a a}
Opt ee PRS “hh pS] Leip Pek yy gh!

5 Pelle wae i a aq heal ith

ti Richly {nas «i was
: Aa vel 5 Welw Lay yon

0 Te ee “dé

<i =i)
Ani tOM OTP et Ti bes A a TD a4 vee a

wane) bcs at, 1) 7 VAL TLD a ; )
nets al hk way -j ia i) | [ iri ‘/) De yid, ith pire

ark NLL Tea Tester Gere Fa eaeRAT) Gin HY elie

eal winle wh Paolo

i

—— ¥
e

> Ging «
ur
>

wae oe

og ee ee ae 7

ro Tee A er? 2S g” i, :
i
dF whh \\ie0 jen

opin nt ee fe ; a
ie ~
l | a 1% saa

SSA TAL T

were 10 cagil ange Ob onion AR ¥

WiiZee wwe © batarael yay, od Ce nis

0? Yt 1 ine

Lal Dias fd ty comstadal te a ‘

wot Lat catty: we ti pel eae

ae asl
‘-

-

a ey

is Podine 7 Ah 7 Ae gah ws

/ “
7 rhe ie
i A bal Ps | cl

pilee gov} 4 ‘> Wy

iuila @ Te pe hiegw

ri '

ae

”

GROWTH IN LENGTH OF THE FROG EMBRYO. aoe

On the Growth in Length of the Frog Embryo,
By

Richard Assheton, M.A.

With Plates 23 and 24.

In some previous papers on the development of the rabbit I
have attempted to show that there are two main centres of
growth, each in itself tending to produce a radially symmetrical
form; but since these two centres of growth are situated
eccentrically to each other the resulting embryo is cylindrical,
and subsequently bilaterally symmetrical.

I endeavoured to show that no concrescence occurred in the
rabbit, and that no theory of concrescence was necessary to
account for the facts.

I wish now to indicate the manner in which two centres of
growth can also bring about the corresponding results in the
frog embryo, without any concrescence of the dorsal lips of the
blastopore.

In a paper in this Journal Dr. Robinson and I discussed the
question of the formation of the archenteron in the frog, and
came to the same conclusion as Moquin-Tandon (in Anura)
and Houssay (in Axolotl), that the archenteric cavity was due
to a splitting amongst the cells in situ, and not to an invagi-
nation or overgrowth of surface cells.

Recently Jordan, and Umé Tsuda and Morgan have made
very interesting communications upon the subject.

The former author, after an able summing up of the evidence
on both sides, concludes (p.331), ‘The evidence thus far adduced

224, RICHARD ASSHETON.

for invagination is, to say the least, inconclusive ;” and on the
other hand (p. 332), “ There is not a shred of evidence to show
that the large cells at first surrounding the mouth of the blas-
topore are not subsequently pushed in by the ingrowth of
ectoblast cells. No positive evidence whatever exists to prove
either the impossibility of invagination or the likelihood of no
invagination. I find it difficult to gather the reasons that
have influenced Houssay and Robinson and Assheton to adopt
the view that invagination does not occur.”

Jordan then describes (pp. 333-4) “ocular evidence that
the small cells around the lips of the blastopore are actually
infolded.”

Morgan, after a description of interesting experiments after
the method of Roux upon the living egg, says, “‘ The statement
of Robinson and Assheton that no portion of the archenteron
in the anura is formed by invagination is certainly incorrect,
as I hope to show in a later paper.”

Of course this latter statement must depend upon the exact
meaning to be attached to the word archenteron.

My own conception of the term archenteron is that cavity
which in the embryo is supposed to represent the digestive
cavity of a hypothetical ancestral “ gastrula,”’ no matter how
this cavity was brought about.

If, however, by archenteron is meant any subsequent pro-
longation of this cavity, such as would represent a post-gastrula
condition ancestrally, then certainly such a statement was
inaccurate.

When Dr. Robinson and I made the statement referred to,
we regarded as archenteron part of the cavity which I now con-
sider to represent a post-gastrula condition. In other words,
I agree with Morgan and others to a certain extent as regards
the growth over of the dorsal lip of the blastopore, and consider
only the most anterior part of the gut cavity of the frog’s
embryo at the time of the closure of the blastopore as being
formed by a splitting, and as representing the true archenteron.

Accordingly, in my opinion, the sentence (referred to above)
by itself accurately describes the facts, but in the context in

GROWTH IN LENGTH OF THE FROG EMBRYO. 925)

which it stood I admit that I now think it inaccurate. My
reasons for so thinking I will now proceed to give.

Again, Morgan and Umé Tsuda say that we “ apparently
at the outset have orientated the embryo wrongly, for they
state the segmentation cavity has a roof which ultimately
becomes the anterior wall of the gastrula; for the anus which
marks the posterior end of the embryo appears at the opposite
side of the ovum,—that is, on the floor of the segmentation
cavity.”

I cannot understand their objection to this paragraph.

Figs. 7 to 11 on Pl. 24 are all placed with what we conceive
to be the dorsal surface (D.) directed towards the top of the
plate.

As regards the ocular evidence of an invagination spoken of
by Jordan, it is a pity that more details are not given of the
observations.

Is it possible to trace a cell, or a spot on the surface some
distance from the lip of the blastopore, to gradually approach
and fold over the edge and so disappear, or do only cells
actually on the edge seem to be affected by the process ?

What is the cause of the invagination? I can quite well
imagine that individual cells at the edge may, by multiplication
of their neighbours or themselves, be pushed over the edge, as
also might cells on the inner edge appear to be pushed out-
wards, if we could see that edge. The splitting theory still
seems to me to be the more probable for the commencement
of the archenteric cavity and its extension forwards.

But, as I shall point out a little further on, there is un-
doubtedly an apparent overgrowth, and I think certainly an
actual overgrowth of the lower pole cells by the dorsal lip of the
blastopore, together with the lateral and ventral lips, as they are
formed at a later period. This process, however, should not, I
think, be compared with the process of gastrulation so called,
or formation of primitive archenteron, but should be considered
to be intimately connected with the growth in length of the
embryo. In other words, to follow the same line of argument
that I have used in the description of the rabbit embryo, the

VOL. 37, PART 2.—NEW SER. Pp

226 RICHARD ASSHETON.

formation of the primitive archenteron is by a process of split-
ting, and is the direct effect of the primary centre of growth;
whilst the continuation of the cavity produced by an overgrowth
is the direct effect of the secondary ceutre of growth, producing
the elongation of the animal.

The splitting process in the frog corresponds in results to the
invagination process of Amphioxus, while the overgrowth of
certain parts of the white pole of the ovum of the frog by the
dorsal, and subsequently lateral and ventral lips of the blasto-
pore, together with the continuation of this process in the for-
mation of the tail, corresponds to the elongation of the gastrula
in Amphioxus, by means of what Hatschek called the polar cells.

I shall now attempt to explain what I believe to be the actual
method in which the splitting is brought about.

The frog’s egg segments, as has been described by many
observers, more rapidly at one pole than the other. ‘This
is, I think, universally supposed to be due to the greater accu-
mulation of yolk granules at the “ lower” pole, which thereby
hinder the segmentation activity at that pole.

If we admit that “ yolk ”’ determines the inequality of the
process known as segmentation, we must admit it also in the
case of each cell. If it is true of the segmented ovum, it is
equally true of the unsegmented ovum. ‘To say that yolk
being more plentiful in one part of a cell than in another
hinders the activity of the protoplasm, is the same as saying
that a cell divides into two parts, which in magnitude are in
inverse ratio to the purity of the protoplasm contained. In
other words, the result of asimple process of cell division, such
as we see in the segmenting ovum, is two cells equally balanced
as regards protoplasmic energy.

Fig. 15 on Pl. 24 is a diagram of a vertical section of the
unsegmented ovum of the frog.

The circles 1 to 7 represent diagrammatically what I imagine
to be the distribution of yolk, as determined from a considera-
tion of the segmented ovum.

The space No. 1 is that region in which segmentation is
most retarded, and so presumably the region in which yolk is

GROWTH IN LENGTH OF THE FROG EMBRYO. 227

most abundant. The space No.2 contains less yolk to a given
area than No. 1, No. 3 less than No. 2, and so on.

For the sake of simplicity we may regard the outer space
only. This may be supposed to contain protoplasm of a
uniform degree of purity. Accordingly division of this space
will be such as to produce two spaces whose areas are equal.
This is about the spot marked by the line (a), and will repre-
sent the third furrow of segmentation, that is the first hori-
zontal furrow.

Similarly, the next horizontal furrows will be about the
spots 6 b, the next at cc cc, the next atdddddddd, and
so on; always resulting in a balance of protoplasmic energy
on each side of the furrow.

In this way the frog’s egg becomes segmented more and
more rapidly in the upper hemisphere than in the lower. For
a considerable time there is an almost complete absence of
horizontal furrows in the lower hemispheres.

This point is very well seen in Umé Tsuda’s figures Iv, v, of
Plate 24, ‘ Quart. Journ. Micr. Sci.,’ vol. xxxv, part 3.

Another effect is that as segmentation proceeds there is a
continual increasing disparity in size between the cells of the
black pole and those of the white. Whereas at first the
superficial area of the cells of the extreme upper pole bears to
the superficial area of the cells of the extreme lower pole the
ratio of 1 to 2, at the time of the commencement of the blasto-
pore it bears the ratio of 1 to 5.

In this way there is a gradual apparent creeping of small
(black) cells over the surface of the egg—though in reality it
is conversion of large cells into smaller in situ, as, I believe,
is now generally accepted.

My diagram fig. 15 gives the idea of no segments in the
white or lower hemisphere of the ovum. This is because it
deals only with horizontal furrows.

The segmentation energy may be said to produce its effects
along the area of least resistance. Is it not possible that the
commencement of the archenteron may be a continuation of
this same process ?

928 RICHARD ASSHETON.

The effect up to now has been to produce a fairly sharp line
of demarcation between small and large cells upon the surface
at the point x in diagram, fig. 15.

On the supposition that this diagram represents fairly accu-
rately the distribution of yolk, it is clear that as this line
advances it encounters greater and greater resistance. May
not a time come when it will find the path of least resistance
to be inwards and backwards, as in diagram 16?

Diagrams 15 and 16 are inaccurate for later stages of seg-
mentation, because they do not show a segmentation cavity.

Fig. 12 is a more accurate representation of a completely
segmented egg.

A. is the black upper pole (the anterior wall of the future
embryo); P. is the white lower pole (posterior end of the
future embryo) ; sg. the segmentation cavity.

The letter A. points to the smallest cells of this stage, y. p.
to the largest.

There is a gradual merging of the one into the other, not
along the surface, for here the line is much sharper, but along
the cells to which y. and a. are directed.

My idea is that the continuation of the segmentation
process is the conversion of, first, the cells y., then the cells z.
into smaller ones, and in this way a layer of small cells will be
produced lying up against the mass of much larger cells y. p.

This layer I have indicated in the fig. 12 by the dotted line.

If a section of an embryo of the stage in which the blasto-
pore is nearly complete is examined, it will be seen that there
is such a layer of cells along the floor of the segmentation
cavity.

Fig. 21 is an outline camera drawing. Such details as are
shown were not drawn by camera. The smallest cells are
those forming the lip of the blastopore.

The point a. represents the at present most anterior limit of
the archenteron. More anteriorly, however, following the
lines a a., a a., there is what I take to be a differentiation of
the yolk-cells, that is a splitting up into smaller cells, which
cells, upon the splitting hypothesis, will eventually form the

GROWTH IN LENGTH OF THE FROG EMBRYO. 229

roof of the archenteron, and come to lie up against the epiblast
now forming the roof of the segmentation cavity, as shown
diagrammatically in figs. 12 and 13.

The differentiation as seen at the point aa. must, on this
hypothesis, be considered to be the effect of the direct continua-
tion of the process of differentiation on the surface of the ovum
(whereby the epiblast is separated), that is a direct continua-
tion of the process of segmentation. This line bounded above
by small cells, below by large cells, constitutes a line of separa-
tion or a split, which is, I believe, the first commencement of
the archenteron, and is a result of the primary centre of
activity, comparable to the events of the first five days in the
development of the rabbit, or to the formation of the gastrula
in Amphioxus. But although corresponding in effects, the
only really homologous feature is the presence of a primary
centre of activity, or process of segmentation of the egg; the
actual directive agencies being in each case ccenogenetic and
entirely different.

The conversion of the narrow slit into a spacious cavity is to
be considered to be due, at any rate in part, to the effect of the
secondary centre of activity, to which I shall refer again.

I must now refer to the experiments made by Roux and
Schultze and Morgan and Umé Tsuda upon the developing egg
by following natural or artificial spots, which experiments I
have myself repeated during this spring.

It is impossible to repeat these experiments without becom-
ing convinced that there is a change of relative position between
certain spots on the ovum,—for instance, the dorsal lip of the
blastopore, and the most inferior spot upon the white pole
of the ovum. These two spots, as seen from without, un-
doubtedly approach one another before the complete formation
of the circular blastopore. But the question to what extent
this approximation is carried, and whether by a concrescence
of the lateral lips of the blastopore, or by a rolling under of the
white pole, or by a growth over the upper lip without concres-
cence, is answered differently by the several observers,

My own suggestions are as follows,

230 RICHARD ASSHETON.

The Secondary Area of Cell Production.

I think that every one will agree that after the closure of the
blastopore, the embryo grows in length by the proliferation of
cells at the spot which formerly formed part of the lips of the
blastopore.

There is very little doubt that rapid growth at this spot takes
place before the final closure of the blastopore ; the question is,
when does this growth begin ?

Again, it must be remembered that growth of the embryo as
a whole, derived from the rapid multiplication of the cells in this
area, is growth in length. It is the secondary area of growth
comparable to the secondary area of growth or primitive streak
of the rabbit.

If there is such a growth backwards of the blastoporic lips
before their closure, there will then be a portion of the future
gut cavity of the embryo that will have been formed, not by a
splitting nor by an invagination, but by a growth backwards of
the blastoporic lips.

Amphioxus, after the completion of the process of invagina-
tion, begins to grow in length. According to Hatschek’s
account this was largely due to the activity of two pole cells.
Recently, however, Wilson has stated very clearly that these
pole cells are “a myth.” They never exist at any time, but
the posterior region of the larva of later stages ‘is rapidly
growing, and numerous mitoses may be observed in all the cells
in the region of the mesenteric canal.”

Now although the process of invagination produces the
double-layered condition of the embryo of Amphioxus, and at
the same time the cavity of the archenteron, yet it is only the
anterior part of the archenteron that is formed in this way.

There is a posterior point of the archenteron which is
formed, not by invagination, but by growth of the blastoporic
lips. This must be so, whether we accept Hatschek’s or
Wilson’s description of the secondary growing point.

The exact line of demarcation between the two parts I have
no means of showing. It is not easy to say at what moment

GROWTH IN LENGTH OF THE FROG EMBRYO. 231

the secondary growing point becomes a functionally active area
in Amphioxus. It is possible for it to become established as
soon as a blastoporic lip is formed, and not before, because a
characteristic feature of this secondary area of proliferation is
that it should produce cellular units to all existing cellular
layers.

I have in a previous paper tried to locate this line of demar-
cation in the rabbit. The moment of origin of the secondary
area of proliferation in the rabbit is fairly well marked. Of
Amphioxus I cannot speak. Can we find it in the frog? The
frog is by no means so simple as the rabbit, but is more
amenable to experiment than is Amphioxus.

The frog differs in one respect from Amphioxus, which is of
importance in reference to the question now under discussion.
In Amphioxus the blastoporic lip is formed at the same
moment apparently at all points of its circumference. In
the frog it is formed at one point first, namely, at the future
dorsal region, and many hours elapse before the lip is formed
ventrally. Hence it is possible for the secondary
area of proliferation to become established much
sooner dorsally than ventrally.

In my account of the frog given above, the production of the
split forming the primitive archenteron, which I believe to
represent the process of gastrulation of Amphioxus, although
no invagination occurs, is to be considered, like the invagina-
tion process in Amphioxus, as the result of the primary area
of proliferation, and of itself would tend to the production of a
radial symmetry. The very moment this split begins, a portion
of the blastoporic lip is thereby formed.

If my supposition is right, that the secondary area of pro-
liferation may be established as soon as there is a mass of
cellular tissue in connection with all the primary layers, it is
clear that possibly the secondary area of proliferation may
start immediately upon the formation of the dorsal lip of the
blastopore, and not delay until the whole blastoporic rim is
completed.

Accordingly, on this view, there will be an extension of gut

232 RICHARD ASSHETON.

cavity anteriorly by means of a splitting, the result of the
primary area of activity, and posteriorly by means of a growth
backwards of the dorsal lip of the blastopore, the result of the
secondary area of activity, comparable to the corresponding
parts in the rabbit, formed previous to the eighth day, and
upon and subsequently to the eighth day respectively. In the
rabbit and in Amphioxus the lining of the archenteron of the
primary area is completed before the secondary area of pro-
liferation has become established, but in the frog afterwards ;
and so the linings of both parts of the gut cavity are formed
together.

Many actual experiments and observations have been made
upon the eggs of the frog with the object of demonstrating the
mode of formation of the blastopore, and the relative position
of the blastopore when it has a completed margin to the
originally black and white poles of the unsegmented ovum.
Such attempts have been made with varying results by Roux,
Schultze, Hertwig, Morgan, and Umé Tsuda, and although the
experiments described are in many cases contradictory, yet
there seems to be no doubt that the dorsal lip of the blastopore
does overgrow a portion of the whiter side of the embryo prior
to the completion of blastoporic lip ventrally.

I have myself made similar experiments in repetition of
Roux, and I am quite convinced that this overgrowth does
occur to a certain extent, but I am equally sure that it is
incorrect to assert that the neural plate is formed entirely
upon the lower (white) pole of the ovum.

The dorsal lip overgrows the white segments, at
any rate apparently, but so do the lateral and
ventral lips as they are formed. It is only because
the dorsal lip is formed first that this part seems
to overgrow the white pole to so large an extent.

The overgrowth is a part of the same process which pro-
duces the lengthening of the embryo. If the whole blastoporic
lip could in the frog be formed at once the embryo would, I
suspect, change rapidly from a sphere to an oval, as does the
embryo of Amphioxus (v. Hatschek, figs. 80 and 34).

GROWTH IN LENGTH OF THE FROG EMBRYO. 233

In the frog the dorsal lip cannot of itself grow outwards and
so produce an oval embryo until the rest of the blastoporic
lips are formed. Unless it remains inactive it must follow the
contour of the ovum. That it does not remain inactive I have
convinced myself, and therefore I agree with the above-
mentioned authors that a portion of the white area passes out
of view of the observer by becoming hidden by the advancing
dorsal lip of the blastopore.

Experiments in marking Parts of the Ovum.

General Remarks.—I find—

(i) That it is impossible to fix the egg in any one position
so as to prevent with certainty the rotation of the ovum
within the vitelline membrane without injuring or distorting
the ovum.

(ii) That, accordingly, any fragment which exudes from the
ovum through the aperture made in the vitelline membrane
when pricking the ovum in order to mark one spot is useless
as a landmark.

(iii) A scar upon the ovum itself, fixed to the ovum and
within the vitelline membrane, is the only mark which can be
relied upon for drawing conclusions as to the relative rate of
growth, and a change of position at different points upon the
surface of the ovum.

(iv) A severe injury by pricking naturally produces much
abnormality of development; whereas a very slight injury,
although admirable for a short observation, is apt to recover
and so get lost and obliterated after many hours.

Some of my own experiments I will now briefly describe.
Outline figures are given upon Pl. 23.

Figs. 1a—1ld show the results of an experiment. Here the
puncture was very small, and made midway between the two
horns of the developing blastoporic lips (fig. la). Three
hours and a half later the blastoporic lips were completely
marked. By this time (fig. 14) the mark was distinctly closer
to the dorsal lips of the blastopore. Three hours later the
mark had approached the dorsal lip still nearer, but the ventral

234, RICHARD ASSHETON.

lip of the blastopore had gained a little upon the mark. After
a lapse of eight more hours the blastopore was very much
smaller, and the mark was found partly covered by the dorsal
lip. In this (fig. 1d) the ventral lip had gained much more
upon the mark than had the dorsal lip.

Figs. 2a—2d are figures of a specimen which had a natural
mark upon the white pole of the ovum. The drawings were
made at 5.30, 8.30, 10.40 p.m., and 8.30 a.m.

I thought the scar was part of the embryo, but upon the
blastoporic rim reaching it the scar became partly scraped off
on to the rim.

Both this specimen and the last show that the apparent over-
growth of the dorsal lip of the blastopore is much more marked
at first than afterwards.

This is well illustrated by figs. 4a,45. In this specimen
a mark was made after the complete formation of the blasto-
pore (fig. 4a) near the centre of the unenclosed yolk. Fig.
4b is the same specimen seventeen hours afterwards. I do not
know to which lip the mark was approximated.

Figs. 83a—3g show a similar apparent overgrowth of the
dorsal lip, and also that this overgrowth is greater during the
earlier period of blastopore formation.

Figs. 54, 5c, and 5d are from embryos which have completed
the closure of the neural plate. All these were, at the moment
of the first signs of the blastoporic lip, pricked near to the
margin between the black and white at the point most distant
from the commencing blastopore, and equidistant with the
latter from the equator of the ovum. Fig. 5d shows the scar
a little to the left of the spot where the blastopore has closed.

Fig. 55 shows the scar upon the side of the embryo about
its middle, both dorso-ventrally and antero-posteriorly.

Fig. 5c shows the scar upon the ventral edge of an unclosed
blastopore. In this specimen the injury was very severe, a
large mass (ewo.) exuded, and, as always follows in such a
case, an abnormal embryo was formed.

Fig. 5a shows the spot near which they were all pricked.

Ten embryos were pricked in the centre of the lower pole of

GROWTH IN LENGTH OF THE FROG EMBRYO. 235

the ovum in the blastula stage before any trace of the dorsal
blastoporic lip could be detected. Of these when preserved, at
which time the normal ones were from 43 mm. to 5 mm. in
length, seven showed no trace of the injury externally, and
seemed to be quite normal. One showed no injury, but was
rather abnormal in shape. Two failed to develop beyond the
blastula stage.

Another specimen was pricked, as shown in fig. 6a, on both
sides of the blastopore, on one side upon the lip, on the other
a slight distance away from the line where the lip was appa-
rently about to develop. Fig. 64 was drawn ten and a half
hours afterwards.

After the blastopore had closed I was unable to detect the
injuries.

There is no doubt that one must be very cautious indeed in
drawing conclusions from injuries made upon eggs. This is
especially so with injuries made upon the more active part of
the embryo, i.e. the more deeply pigmented cells. A very in-
considerable injury is sufficient to produce an abnormality.
Three slightest punctures possible upon the rim of the blasto-
pore equidistant from each other are sufficient to prevent the
closure of the blastopore, while one only, if at all severe, will
have the same effect.

This clearly must be the case, as the closure of the blasto-
pore is an effect of increase of bulk of certain parts of the
walls of the embryo, and if this increase in bulk is, through
pricking these walls, prevented or delayed by the letting out
of matter, and thereby obviating the necessity for the blasto-
poric lip to advance, the blastopore does not close; and so also
injuries to the white pole or yolk plug by allowing the escape
of material from that area, and thereby diminishing the bulk
of the part of the embryo that can be covered by the advancing
lips of the blastopore, hastens the closing of the blastopore.

My own experiments are in part confirmatory, but mostly
contradictory to those of Roux. They are upon the whole
confirmatory of those performed by Morgan and Umé Tsuda.
Roux asserts that the dorsal lip of the blastopore passes over

236 RICHARD ASSHETON.

the lower pole of the ovum through at least 170°. The ventral
lip according to him does not advance at all.

Morgan and Umé Tsuda consider that the ventral lips
and lateral lips advance, but not to so great an extent as the
dorsal lips. They also notice the “first overgrowth of the
dorsal lip of the blastopore is more rapid than the later growth ;
that is, the approach to the points of injury is faster at
first.”

I quite agree with the latter authors that “itseems.....
most probable that the blastopore does not start at the equator
of the egg, but some distance below that circle.”

Now my experiments do not give evidence of an overgrowth
by the dorsal lip of more than 60° or 70° from the moment of
the first commencement of the dorsal lip, and to the closure
of the blastopore. More probably, I think, the apparent
overgrowth is even less.

According to Roux the overgrowth is at least 170° to 180°.

If Roux is right in both his suppositions, namely, that the
dorsal lip moves over the white pole, and to an extent of 180°,
I cannot understand how the last remaining portion of the
blastopore to remain open should show so white a piece
of yolk plug. This piece of yolk plug is as white as any part
of the surface of the ovum of the frog ever is. There is a
considerable amount of variation in the pigmentation of the
unsegmented ovum. It is extremely rare in England to find
eggs in which there is deficiency of pigment over an area sub-
tended by so great an angle as an angle of 120°. An area
where there is almost an absence of pigment is much more
restricted. Very frequently the less pigmented area extends
over a much smaller are. On Roux’s supposition, the part
which remains longest uncovered ought to be grey, if not
quite black. It is, as far as I have observed it, an almost
invariable rule to find the yolk plug at its latest stage intensely
white.

I have only once seen embryos which in the gastrula
stage showed a darkened blastopore, and these were from eggs
which in the unsegmented stage were so intensely pigmented

GROWTH IN LENGTH OF THE FROG EMBRYO. 937

that the lower pole was only slightly lighter in colour than the
upper, and this for an area not greater in extent than that sub-
tended by an angle of 50°. Yet in these the blastopore, though
very dark, was quite as light as the lightest part of the unseg-
mented ovum.

If the centre of the blastopore at the moment it is in the
stage represented in fig. 20 is not either the lower pole, or
some spot extremely close to the lower pole, of the unsegmented
ovum, the intense whiteness of this spot must have been pro-
duced by a disappearance of pigment previously existing. Is
there any evidence of this? I cannot think of any. On the
contrary, there is evidence of increasing pigmentation, as, for
instance, in the epiblast-cells as they form upon the sur-
face, in the cells of the splitting archenteron, and even in the
white cells themselves. If the final position of the blastopore
is, as Roux supposes, at the equator, and at a spot removed 170°
from the spot of the first commencement of the blastopore,
surely the uncovered part of the surface would be dark, if not
black, and certainly not intensely white.

At the moment the definite outline of the ventral lip of the
blastopore is formed the anus of Rusconi thus fashioned is
not of uniform tint. The dorsal part of the area is much
lighter in colour than the ventral part, fig. 19. When, how-
ever, it has diminished to the condition of fig. 20, it is of
uniform tint and intensely white. This is accounted for by the
more rapid closure of the ventral lip from this moment, as my
experiments and those of Morgan and Umé Tsuda demonstrate,
as also may be seen by examination of sections as described in
a former paper (Robinson and Assheton).

Except for a slight advance of the dorsal lip of the blasto-
pore my experiments do not support Roux’s.

According to Roux, injuries made at the point 2 in fig. 5a
ought to have appeared upon the dorsal side in the medullary
folds a little way anterior to the blastopore. Instead of which
one was in the median line ventral to the _ blastopore,
one was laterally placed on a level with the blastopore, and one
laterally placed but far forwards. So, again, if the lips of the

238 RICHARD ASSHETON.

blastopore concresce as Roux assumes, then marks made upon
the lips should show upon the dorsal surface somewhere along
the neural folds. In no case did I ever find this to occur. As
the result of such injuries, I either found the scar upon the
lateral lip of the blastopore when completed, or else as in figs.
6 a, 6 6, further removed from the blastopore but in the same
relative positions.

I never found a defect in the neural plate except when the
dorsal, or near the dorsal lips of the blastopore, was injured.
This spot is unfortunately at the same time the most interesting
to injure, and the most delicate, and most liable to produce
abnormalities which prove very little. There is certainly no
need to assume a concrescence, as the facts can be equally well
accounted for by other means, which to my mind agree far
better with the development of other Vertebrates than does the
concrescence theory.

In other words, I believe that as in the rabbit, so in the frog,
there is evidence to show that the embryo is derived from two
definite centres of growth, the first, and phylogenetically the
oldest, being a protoplasmic activity which gives rise to the
anterior end of the embryo (= gastrula stage) ; the second,
which gives rise to the growth in length of the embryo: which
centres of growth occupy the same relative positions in loca-
tion and in sequence of time, and probably to each are due the
same parts of the embryo.

In the rabbit the area is a spot which assumes for a time a
linear form, but is unaffected by its change of shape in the
functions it has to perform.

So in the frog, although at first crescentic, then circular,
then linear, and ultimately a knob, its function is precisely
the same as in the rabbit, and is unaffected by the change in
form. From the moment of its first appearance it performs
its one function—that of adding on new cellular units to the
previously existing embryo.

One difference of effect is that owing to the manner of its
coming into existence, one portion arising before the other,
that portion—the dorsal—becomes functional before the

GROWTH IN LENGTH OF THE FROG EMBRYO. 239

ventral, and so the dorsal part of the embryo is developed
more quickly than the ventral.

I believe the true way of regarding this area of secondary
proliferation, both in the rabbit and in the frog, is as a single
area, whether circular, annular, or linear, whose sole function
is the addition of cellular units to the posterior end of the
previously existing embryo. Its form is the result of secondary
or ontogenetic causes.

The exact line of demarcation is not easy of determination.
Very careful marking of the dorsal lip might give it as far as
the nervous system is concerned, but organs, no doubt, change
their relative position somewhat as they develop. The brain
is certainly thrust somewhat forwards. I brought forward
evidence to show that in the rabbit this point was about the
level of the first mesoblastic somite. It is, at any rate,
possible that metameric segmentation may be directly due to
this process of elongation. It seems always to be closely
connected with it. If so it may be due to this, that the
anterior mesoblastic somite of the frog is the smallest, and
each for succeeding five or six becomes longer dorso-ventrally
than its preceding neighbour.

For upon my suppositions of the non-concrescence of the
blastoporic lips, and of the unity of the nature of this area
of secondary proliferation, then, since the first part of this
proliferating area to be formed is that part adjoining the
dorsal surface, those parts in the mid-dorsal line, e. g. neural
plate, will be the first, and at first the only part of the
embryo to receive additions from the proliferating area. As
the lateral lips of the blastopore are formed, more and more
of the lateral plates of mesoblast will receive additions, so that
in this way it is possible that the gradual increase in size of
the first six mesoblastic somites in the frog may be connected
with the gradual development of the area of proliferation.

Every one is agreed that there is a certain part of the neural
plate formed on an area anterior to the first commencement of
the blastopore lip. The point in discussion is to what extent
does this pre-blastoporic formation exist ?

240 RICHARD ASSHETON.

Morgan and Umé Tsuda conclude that all except ‘“ the
thickness of the medullary folds”? round the dorsal lip of the
blastopore is formed by the growth of the lip.

Roux shows the same in his figure.

Pfluger, however, thinks it possible that a considerable
length of the anterior part of the nervous system is formed in
this black hemisphere, and with Pfliiger I quite agree on this
point.

I find that the neural plate in normal embryos at the time
it becomes visible on the surface extends through fully 170°, if
not more, while the distance through which the dorsal lip of
the blastopore travels I cannot make out to be more than 70°
at the most; that is, from a spot a little below the equator to
the lower pole, or perhaps a little beyond it.

Figs. 7 to 14 represent diagrammatically the views put
forward in this paper. Fig. 7 is the fully segmented frog’s
egg, the white pole placed to the right of the paper; the black
pole or roof of the segmentation is placed to the left, as repre-
senting the future anterior end of the embryo. All the others,
8—14, are arranged similarly, Fig. 8 represents the stage at
which the dorsal lip of the blastopore has become established.

Up till now there has been but one general centre of growth.
From this moment the secondary centre of growth is in
existence, and we have now the commencement of the con-
version of an embryo radially symmetrical into an embryo bi-
laterally symmetrical. As yet only the dorsal part of this
secondary area of proliferation is in existence, and accordingly
the dorsal part of the embryo is developed more rapidly than the
ventral, as the annexed figure 9 shows. In this figure the
ventral part of the secondary area has just been completed, and
now the whole of the secondary area of proliferation, the homo-
logue of the whole of the primitive streak of the rabbit, is
complete, and new material is added to ventral and lateral
and dorsal parts of the embryo, as diagram fig. 10 illustrates.
The shape now rapidly changes, and the radial symmetry is
lost and the bilateral symmetry acquired, fig. 11.

The neural plate is indicated in fig. 11 by the continuous

GROWTH IN LENGTH OF THE FROG EMBRYO. 241

line ; the dotted line represents only approximately the sup-
posed division between the parts of the embryo derived from
the primary and secondary areas of proliferation respectively.
The subsequent fate of the secondary area of proliferation (or
primitive streak) I have, with Dr. Robinson, described and
discussed in a former paper. The three figures 8, 9, and 10
represent my views of the extent to which the white pole
becomes overgrown by the dorsal lip of the blastopore. The
neural plate is indicated as in fig. 11 by the continuous line.
The extreme anterior end of this part of the epiblast has been
obtained by subtracting the amount due to overgrowth from
the total amount observed when definitely established.

It does not necessarily follow that the distance through
which the edge of the dorsal lip of the blastopore advances
represents the total growth in length due to that part of the
secondary area of proliferation. In order to advance, the lip of
the blastopore has to exert pressure upon the “ yolk plug,”
causing it to be forced inwards. It thus follows that an equal
force must be exerted in the other direction. What effect this
has will depend upon the strength of the resistance offered by
the anterior wall of the embryo.

From the fact that the archenteron is a slit in the stage
represented by figs. 8, 21, and 17, and is a spacious cavity in
the stage represented by figs. 9, 13, and 19, it seems likely
that the growth of the blastoporic lip is rendered evident, not
only by the amount of white yolk plug covered, but also by the
arching up of the dorsal roof of the archenteron. But, on the
other hand, the arching up may be in great part, if not eutirely,
* due to its own interstitial growth ; though I do not think this is
likely, for it seems to me to require the thrusting energy of
the blastoporic lip to account for the obliteration of the seg-
mentation cavity.

If Schultze’s idea of the apparent overgrowth of the white
yolk plug being due to a rolling inwards of the white pole were
correct, ought not the ventral end of the segmentation cavity
to become obliterated before the dorsal? But it is the dorsal
part that first disappears.

VOL. 87, PART 2,.—NEW SER. Q

242 RICHARD ASSHETON.

DESCRIPTION OF PLATES 23 & 24,

sueate Mr. Richard Assheton’s paper ‘On the Growth in
Length of the Frog Embryo.”

CompLete List oF REFERENCE LETTERS.

A. Anterior end. aa. Small cells in floor of segmentation cavity. 47. d.
Dorsal lip of blastopore. 4/. v. Ventral lip of blastopore. D. Dorsal
surface. exo. Exovate. P. Posteriorend. sg. Segmentation cavity. y.
Dorsal lip of blastopore. z. Floor of segmentation cavity. z. Yolk-cells,
which will be overgrown by dorsal lip of blastopore. y. p. Yolk-plug.
x. Spot at which the blastopore commences.

PLATE 23.

Fic. 1.—a—d. Ovum pricked in centre of white between the developing
lips of the blastopore.

Fie. 2.—a—d. Course taken by a natural mark on the white pole.

Fic. 3.—a—g. A mark was made near the centre of the white pole.

Fie. 4.—a, 6. Ovum was marked in centre of blastopore when first out-
lined.

Fic. 5.—a—d. Three embryos marked at the cross in 5a. In 4 the mark
was on the side; in ¢, ventral to blastopore; in d, in which it was very
indistinct, at the side of the blastopore.

Fic. 6.—a, 4. This embryo was marked at the sides of blastopore.

PLATE 24.

Fic. 7.—Frog embryo before appearance of blastopore. The white (or
posterior) pole is placed to the right of the observer.

Fic. 8.—Frog embryo at the time of the commencemeut of the blasto-
pore, which is shown as a dark crescentic groove. The dorsal cap represents
that part of the epiblast which will form the anterior part of the neural
plate.

Fic. 9.—Frog embryo at the moment of the completion of the ventral lip of
the blastopore. The dorsal lip has grown over the white pole through an are
of 50° to 60°. The dotted line indicates that part of the embryo supposed
to be derived from the secondary area of proliferation. The arrows indicate
from which portion of the rim the respective parts have been formed.

Fic. 10.—Frog embryo at the time of first appearance of neural plate,
visible only in sections. The blastopore is much reduced. The ventral lips

GROWTH IN LENGTH OF THE FROG EMBRYO. 243

have closed more rapidly than the dorsal. Dotted line and arrows indicate
same features as in Fig. 9.

Fie. 11.—Frog embryo when the neural plate is a conspicuous object ex-
ternally, and is deeply grooved. The embryo has become very distinctly
elongated, owing to the growth due to the secondary area of proliferation.
This was drawn with a camera. The dotted line and arrows represent my
own interpretation of the facts as in the preceding figures.

Fic. 12.—A section of the same stage as Fig. 8, semi-diagrammatic. The
dotted line indicates the location of the split amongst the cells, whereby the
archenteron is supposed to have originated and become prolonged forwards.

Fic. 13.—A diagram of a stage intermediate between Figs.9 and 10. The
portion of the gut cavity indicated by the dotted line represents that part
which I suppose to be formed by the splitting amidst the yolk-cells. The
future continuation of this slit is represented by a prolongation ventral-
wards of the dotted line. The posterior part of the roof of the gut cavity,
marked with small dots, is that part formed by the active growth of the
dorsal lip of the blastopore, which is shown by the diagonal shading.

Fie. 14.—A diagram of a later stage, such as Fig. 11. In this the seg-
mentation cavity has become entirely obliterated. The secondary area of
proliferation is completed and active all round the blastopore.

Fie. 15.—A diagram to show the sequence of the horizontal furrows during
segmentation of the frog’s egg.

Fig. 16.—A diagram to show in which direction the process of segmentation
will incur least resistance, on the supposition that the yolk is distributed as
indicated by the intensity of the shading.

Fics. 17—20.—Figures of the frog’s egg during the formation of
the blastopore, to show which part of the surface of the ovum forms the
yolk-plug. Lach figure is arranged in the same position.

Fic. 21.—A section of a frog’s egg of a stage intermediate between Figs.
8 and 9. The ovum was drawn with camera.

4 ’ CDW i pila’ | peer © 20-3) =o Libihe © re tu Laat!
‘ i : ‘? >
. wep ji Ley hia) Bis yb Sihl LF ln § MMi Whit. Melo’ fae yt

.

Oks OTHER DOAN Mae Yo HE a BrwoRe

avihei xaeuw baw aae a dwelt od) pale aiden saa lesa
f ped iz aa ajalltast
one divide «egndite ou ae lituadd di oa ty Serene eee: Ott
Eitomityil (ie ens rae wl ergyetare +] Letting these BP time ome
lt a clini: Vee cain unt Woe vk xf vivig | ots is
a pre Waa ba Fr owt faven. Ay’ tee 4. OSes wae
NAT yeas gmt yay, tae | orks et t slp weanaly

=

ior iim DL aae A awe an boot gwar helt Wind
Wiel 7 el pier Sin nwo |e le all ae ek eee ON
aw iit sia ine Uil M) Tis Ti Me MOLW pea «
Wer, Ae late 0 in die ack Gh ae ey Aleta: AAR eae een cee
yi) au, afi } in WA a oe arelarla ? ite: Th we Vie
wt Ai eOL oO Pai? ‘OPE iat! vue cone al ae ijtel
yi ailicaeiaue @ ie Poverty ab Ail etl? Wy Tact tlh
ow yy a, 10" Toe" * YA Jaga Sale wh +,/ale 1» Lt ea
Aitl fo A eoie ovninin vl ey Tpadyy ae at rab ie ' Sede ghee ;
get tal ative &ih ar. i +. or ) VIEL) POM |r ALLE te
wiry ily as “ ah} an Hum rs i Te in irae “et a
Ws ris eraluie O46 orl) chal otha Tit ie ator: (eu: BES ert

*

»
=

-_ ‘Ss ~ *) ares 4 er Olah fie.

HObsiyoite® a1) phos Ce Pn ee aint ue

A Ai Paw te B47 b q' baa) 7 iy i aye * ae, PEN Viog¢ Loe
: ] qa ¥ cul ot yb? a ike

Bie Ter, Ge Via! ‘en a . byt fi r. bl—.0S* VE Spy tt

es knee Mewar! crib Wo ie Stang od feel eno
» ity ‘f Tie L * i= ier hee wy han iiiabh otha

aa. MERA Jewel sale 4 ie ih ney eet ferrrah

“

s Abeta OF w& W had a ated wiht J uk
_ ‘ ;
Ln
|
i my ; : i
; af Bs JAA —E pte, aa)
ss - -
ie 7
.
*
. 7 7
4 (4
at a
a. -

t 169 ORT OG
Ar GO 09

VARIATION OF TENTACULOCYSTS OF AURELIA AURITA. 245

On the Variation of the Tentaculocysts of
Aurelia aurita,

By

Edward T. Browne, B.A.,
University College, London.

With Plate 25.

Ir was a suggestion from Professor Weldon that led me to
examine a large number of specimens of the ephyre and adult
stage of Aurelia aurita for the purpose of finding out the
variation in the number of tentaculocysts, and if a variation
occurred among the ephyrz to see how far it affected the
adults.

All the specimens were collected and preserved at Plymouth
by the officials of the Marine Biological Association, and I
sincerely thank the director, Mr. Edward J. Bles, for the loan
of so many specimens.

The ephyrz are divided into two sets; the first collected
during the spring of 1893, the second specially obtained for
me during the spring of 1894.

The ephyra of Aurelia normally has eight arms, each bear-
ing a tentaculocyst, four perradial bundles of gastric filaments,
and four mouth lappets.

The first table gives the numerical variation of the tentacu-
locysts of 8359 specimens collected in 1893,

VOL. 37, PART 3,—NEW SER, R

246 EDWARD T. BROWNE.

TaBLe I,

The Numerical Variation of the Tentaculocysts of
359 Ephyre collected in 18938.

benteanloestta qpeeenaia: Percentage.
Six 5 : < : 4 : - cy eet
Seven : ; : 8 : . 2) 33
Hight (normal) F . 278 : : « Ts
Nine = . a 29 A : 5 Gel
Ten). ‘ : eos ; : 2
Eleven : ; 19 3 ‘ « 1333
Twelve ; , a ae ‘ : a oe
Thirteen ; ; ; 3 : , = 10:6

It will be seen from this table that no less than 81 speci-
mens (22°6 per cent.) are abnormal in possessing more or less
than eight tentaculocysts, and that the range of variation
extends from six to thirteen tentaculocysts. There are only
12 specimens (3°3 per cent.) with less than eight tentaculo-
cysts, and the remaining 69 specimens (19 per cent.) are
above the normal number.

TasueE II.

The Numerical Variation of the Tentaculocysts of
1156 Ephyre collected in 1894.

Number of Number of

tentaculocysts. specimens. Percentage.
Five . ‘ . ‘ 1 : : _ —
IK vs : : ; 6 , J 5 OS
Seven ; , PINS 5 5 2 ow
Hight (normal) © . . 883 , : Bye (so!
Nine. : F s sb : e os eM
Ten . : : 5 alll . : A
Eleven ‘ 5 100 : : eae (21 b
Twelve : : ara 4 : 5 .) ae
Thirteen 5 : F 3 : : » PAQS
Fourteen : s ; i : é oo

The second table shows in detail the variation of the
tentaculocysts of 1156 specimens collected in 1894. On

VARIATION OF TENTACULOCYSTS OF AURELIA AURITA. 247

comparing it with the first table it will be seen that the
percentage of abnormal specimens is nearly the same. In
the first set 22°6 per cent., and in the second set 20°9 per
cent. of the ephyrz are abnormal.

The decrease is mainly due to a falling off in the number
of specimens with twelve tentaculocysts amounting to 23 per
cent.

By taking a larger number of specimens the range of varia-
tion has extended from five to fourteen tentaculocysts, but
only one specimen of each of the two extremes has been found.

The ephyra with five tentaculocysts (fig. 1) has four perra-
dial arms, equal in size; but the fifth is interradial and about
half the size of the other arms.

The variation in the number of tentaculocysts does not
affect the other organs of the body, which may vary inde-
pendently of one another.

Two specimens have only three bundles of gastric filaments
instead of the normal four; both have four mouth lappets;
but one (fig. 2) of them has six and the other seven tenta-
culocysts. Six specimens have six bundles of gastric filaments.
and six mouth lappets; three possess eleven tentaculocysts
(fig. 8) and the others have twelve.

A few curious abnormal growths of the arms were also
observed. One specimen (fig. 4) has a perfect double arm
with two tentaculocysts, like two arms united together.
Another specimen (fig. 5) shows a bifurcation of an arm, each
branch terminating with a tentaculocyst.

Perhaps the most interesting monstrosity is that which
occurs in a specimen (figs. 6 and 7) with a large outgrowth
on the aboral side of the umbrella. The outgrowth has two
arms, and one of them bears a tentaculocyst. There are also
seven other arms, with tentaculocysts, in the normal position
and a vacant place for two more.

Adult Aurelia.

The adult specimens of Aurelia were also collected at
Plymouth during the summer of 1894, and belong to the

248 EDWARD T. BROWNE.

same generation as the ephyre taken in the spring of that
year.

The umbrella of these specimens varied from 14 to 24
inches in diameter. The tentaculocysts of 383 specimens
were examined, and the number possessed by each specimen
is recorded in Table ITT.

Tasre ITI.

The Numerical Variations of the Tentaculocysts of
383 Adult Aurelia collected in 1894.

Number of Number of

tentaculocysts. specimens. Percentage.
SILK) a. ‘ : 2 : . am “GE
Seven : : 5 ale d ; sa AEE
Hight (normal) : ~ 296 : : dS
Nine. : 3 oy EGS 5 4 to 8656
Ten . 3 16 4°]
Eleven 10 ‘ : 7 oO
Twelve . : ‘ 7 ; ‘ sy 8
Thirteen 0 =
Fourteen 0 oe
Fifteen il 2

There are 87 specimens (22°8 per cent.) with a variation in
the number of tentaculocysts, 20 having less than the normal
number and 67 showing an excess.

On comparing the abnormal number of tentaculocysts of
the adults with those of the ephyra stage, it will be seen from
the percentages that there is only a slight difference. The
ephyre have 22°6 per cent. abnormal in 1893, and 20:9 per
cent. in 1894; the adults show 22°8 per cent. Itis clear from
these figures that the abnormal ephyre do not appear to suffer
from their abvormality, but are able to reach in safety the
adult stage. The figures also show a slight increase of ab-
normal forms in the adult stage. This may be due to an
insufficient number of adult specimens; the small number is
due to their scarcity at Plymouth.

On comparing the 359 ephyre taken in 18938 and the 383
adult specimens taken in 1894, it will be seen that the per-

VARIATION OF TENTACULOCYSTS OF AURELIA AURITA. 249

centages of abnormality are very close for nearly the same
number of specimens. It is probable that ifa thousand adults
could have been obtained at Plymouth the percentage of ab-
normal forms might have been closer than 2 per cent. of the
ephyre taken in 1894, The adult specimens were taken at
random out of large jars; and it is interesting to note how
close the percentage of abnormality of each complete hundred
comes to the mean abnormality. The first hundred showed
23 per cent., the second 22 per cent., and the third hundred
24 per cent. of abnormal forms. An examination of the spe-
cimens does not show that any particular position on the
margin of the umbrella is favoured either by an increase or
decrease of the tentaculocysts.

Eighteen specimens possess seven tentaculocysts, and in
eleven of these the missing tentaculocyst is a perradial one,
and in seven it is adradial. The presence of an extra tenta-
culocyst may either affect the symmetry of a single quadrant
or one half of the umbrella, and in a few cases by being very
close to another not even upset the symmetry.

Ten specimens with nine tentaculocysts show that the
extra tentaculocyst is in one quadrant of the umbrella, and
thirteen specimens have one half of the umbrella containing
five tentaculocysts about equal distances apart, and the other
half possessing the normal four. Five specimens have eight
tentaculocysts occupying their normal positions, and an extra
one only separated from a normal one by a few marginal
tentacles. When the tentaculocysts exceed nine in a specimen
their position is by no means constant, and a different arrange-
ment occurs in almost every specimen. In some the tentacu-
locysts are about equal distances apart, and in others one half
of the umbrella contains the greater number.

A few specimens have three tentaculocysts very close to-
gether, and usually separated by a few marginal tentacles.

One specimen of an adult Aurelia has fifteen tentaculocysts
with the normal number of genital pouches and arms. This
exceeds the maximum number reached among the ephyre.
None of the adults have, however, either thirteen or fourteen

250 EDWARD T. BROWNE.

tentaculocysts. Their absence is probably due to the exami-
nation of an insufficient number of specimens.

A variation in the number of tentaculocysts does not inter-
fere with the other organs of the body. There appears to be
a correlated variation between the number of genital pouches
and buccal arms, and eight specimens show it.

One specimen has three genital pouches, three buccal arms,
and nine tentaculocysts.

Three specimens have three genital pouches, three buccal
arms, and each one has traces of a fourth genital pouch and a
fourth arm. Two have eight tentaculocysts, and the other has
ten tentaculocysts.

One specimen has five genital pouches, five buccal arms, and
eight tentaculocysts.

Three specimens have six genital pouches, six buccal
arms; two have eleven tentaculocysts, and one has twelve
tentaculocysts.

I have not given drawings of these specimens, as somewhat
similar forms have already been figured by Ehrenberg (1)
and Romanes (2). Mr. Bateson, in his recent book, ‘ Mate-
rials for the Study of Variation,’ gives abstracts and figures
from the papers of Ehrenberg and Romanes. He also gives
a table which shows that 26 specimens (1°49 per cent.) out of
1763 adult Aurelia, washed ashore on the Northumberland
coast in 1892, have more or less than four genital pouches.
The Plymouth specimens show 2°08 per cent. with an abnormal
number of genital pouches.

It may be difficult to notice in a few years what effect the
numerical variability of the tentaculocysts has upon the adult
Aurelia. The variation shows a tendency for an increase of
the tentaculocysts, since the specimens displaying an increase
are about three times as numerous as those possessing a
diminished number. Whether this will change the present
characteristic features of the species or not can only be found
out by examining the ephyre and adults at long intervals of
time, and comparing the results with previous records.

VARIATION OF TENTACULOOCYSTS OF AURELIA AURITA. 251

References.

1. Enrensere, C. G., 1834.—‘Abh. K. Ak. Wiss.,’ Berlin, pp. 199—202,
plates.

2. Romanss, G., 1876.—! Journ. Linn. Soc. Zool.,’ vol. xii, p. 528 vol. xiii,
p- 190, plates xv, xvi.

8. Bateson, W., 1894.—‘ Materials for the Study of Variation,’ pp. 426—
429, fig. 128.

DESCRIPTION OF PLATE 25,

Illustrating Mr. Edward T. Browne’s paper ‘‘ On the Varia-
tion of the Tentaculocysts of Aurelia aurita.”

Fie. 1.—Ephyra with five arms. Aboral view. xX 25.

Fig. 2.—Ephyra with six arms, three bundles of gastric filaments, and
four mouth lappets. Aboral view. x 35.

Fic. 3.—Ephyra with eleven arms, six bundles of gastric filaments, and six
mouth lappets. Oral view. x 25.

Fic. 4.—Ephyra with a perfect double arm, seven tentaculocysts, four
bundles of gastric filaments, and four mouth lappets. Aboral view. x 40.

Fic. 5.—A portion of the umbrella of an ephyra, showing a bifurcation of
anarm. X 40.

Fic. 6.—Ephyra with an outgrowth of two arms from the aboral side of
the umbrella. Aboral view. x 35.

Fic. 7.—A lateral view of Fig. 6. xX 25.

. 3
PS oh ie hs eGR REY
- : A =
sel oe Pro

st

Os

: = jy A iasey To “nt ‘D a) Ya) a= a hy —
— ud By oe

4d

onmey S07? pf) >i ay son dort Cee ee
Walivne otha A do mesmo) et ietal

s
.

Py deli lee ge EA io ee lo urieiegh
A 6 Shy “iv ¢ tAjsalb nav Ga ul| gee wey DALe i ae
3 = . , are ie “iG A, ne

ed Baa. Ci 90a oO |:

yess prea ato hs) ora), was fifient

we Wy Woe eye A
We Neen ifad s ge) eu, ie vyign &
- Lh » ay? Mild @dlive GOTO ea
bs
a i’ /
<>
~~ / -
eh ’ ‘
i 7
i
, ee
, _ —_
7 .
as >
-
a 7
; od = = er é
5 =
ie Jal Pr) ¥ 7
2 a
> ‘ _ be oO 7
¢ Ly
: 7
i .
*
° ,
n i .

ene ann: San Ant si

ia Or WaT aioe

1 Pulpoha al

yyy Wut iit

-

COQ ETS, Th anvrataae

7 ——

tr ‘shite i

bd ; aa
; me, 1 ot ah i) ‘<n

a4 7v 2 Z PelidQ Ain ‘fh

onl WP nen
of ok ie

e@

his A

eth aw dile ples
iM eel to eee
YP (ds aot
es atalil Obras ini
pat IO Pe Tie TERE) it feo Ts:

Oe ane
whit jai de ane e ou a
oid
; la Waly to rstsl b-oSy

nV the} i | | a

nar . Mil

| rie

Ps ee! i aril |

ON THE STRUCTURE OF VERMIOULUS PILOSUS. 2538

On the Structure of Vermiculus pilosus.
By

E. 8. Goodrich, F.L.S,,
Assistant to the Linacre Professor, Oxford.

With Plates 26—28.

In 1892 (8a) I published a short description of an interesting
new Oligochete I had found on the sea-shore near Weymouth.
Since then I have obtained more material, which I have worked
at in Professor Lankester’s laboratory at Oxford, and am now
able to give a more complete and accurate account of its
anatomy.

The worm, which I have called Vermiculus pilosus, was
discovered in the rich black mud underlying the sandy surface
of the shore about halfway between Weymouth and Portland.
It lives there in considerable numbers with various species of
Pachydrilus and Tubificids, especially Heterocheta cos-
tata, Clap., from which it is quite indistinguishable to the
naked eye. Unlike the Enchytrzids, it does not seem to
venture to the surface amongst the decaying seaweed and in
the tidal pools, nor is it so often found underneath large half-
buried stones as other Tubificids, Heterocheta for example.
Fig. 1 represents it natural size; it is from about 4 to 6 cm.
long, and of a dull reddish tinge. The body is soft, and its
movements are slow. Under a low power of the microscope
it can be distinguished at once, after a little practice, being
much less transparent than any of its associates. This opaque-
ness, caused chiefly by the structure of the body-wall, the
network of blood-vessels, and the enormous number of coelomic

254 E. 8S. GOODRICH.

corpuscles, renders it very difficult to make out the details or
its anatomy in the living animal.

The prostomium (figs. 2 and 6, pr.) is rather large and
conical. Hach segment, from the second, is provided with two
dorsal and two ventral bundles of sete (fig. 2, J. d. s. and
I. v. s.), placed in a transverse section nearly at the four corners
of a square. Each bundle contains from two to five sete,
generally three. All the sete, both dorsal and ventral, are
alike S-shaped in outline, with a thickening about one third of
the way from the distal end, which is bifurcate (fig. 3) ; they
are of the ordinary Tubificid furcate type.

On close examination in a favourable light it is seen that the
whole worm is clothed from head to tail in a more or less
dense furry covering of hair-like processes, closely set,
extremely fine, and apparently of cuticular origin (fig. 4, pel.).
These hair-like structures are not cilia; they do not move, and
are not protoplasmic. Resembling the so-called “ sense-hairs ”’
or palpocils found on the prostomium and first segments of
most aquatic Oligochetes, they are probably homologous with
these, but developed to an extraordinary degree. The cuticle
itself (fig. 4, c.) presents no peculiarity. The epidermis (figs.
4 and 7, ep.) is formed of regular more or less cubical cells,
with large oval or round nuclei. Below the epidermis is a
narrow layer of circular muscles (fig. 7, c. m.) within which
are the longitudinal muscles (fig. 7, 7. m.). In the transverse
section the contractile fibrils are of different heights, being cut
through at various distances from the middle of the cells which
bear them. Along each side, about halfway between the
dorsal and ventral bundles of sete, the layer of longitudinal
muscles is interrupted by a row of cells forming the lateral
line (fig. 7, 7. l.). Dr. Hesse has recently shown (4) that these
.are the cells to which belong the circular muscular fibres ; my
preparations support this view. Lining the inside of the body-
wall is the cceelomic epithelium, composed for the most part of
large vesicular cells (fig. 7, c. ep.).

The coelomic corpuscles, which as above stated are very
numerous, are spherical and coarsely granular (fig. 28, c. c.).

ON THE STRUCTURE OF VERMICULUS PILOSUS. 255

The vascular system is very peculiar. In fig. 5 the main
blood-vessels have been represented somewhat diagrammatically,
omitting for the sake of clearness the complex fine network
found all over the first ten segments. A large longitudinal dorsal
vessel (fig. 5, d. v.), reaching from the brain to the last segment,
gives off on either side in front of each septum from the 2nd
to the 10th segment a lateral dorsal vessel (fig. 5, J. d. v.),
which soon branches out over the body-wall, reuniting below
into two lateral ventral vessels which enter the longitudinal
ventral vessel (fig. 5, v.v.). One of the lateral ventral vessels,
the smaller, enters the ventral vessel about the middle of the
segment (fig. 5, m. 1. v. v.); the other, and larger, runs into it
towards the hinder end of the segment (fig. 5, p. J. v.v.). In
my preliminary note I stated that there were a pair of “ hearts ”
in the 10th segment; this was a mistake. Indeed, the fact
that there is no direct communication between the
longitudinal dorsal and ventral vessels is the most
striking thing about the blood-system of Vermiculus. The
manner in which the lateral dorsal communicate with the
lateral ventral vessels by means of the cutaneous network is
shown in fig. 9. In the post-genital region only small vessels
are given off at each segment to the body-wall and to the
intestine.

The dorsal longitudinal vessel from the third to the last
segment is dilated in front of the region where each pair of
lateral vessels is given off into a small bulb or swelling, round
the posterior constricted edge of which are ranged large cells
(often as many as eight or nine), forming a system of valves
(figs. 8 and 10, v.). Each cell is attached at its posterior end
only, while its main body can swing backwards and forwards ;
it contains a number of bright fatty granules and a small
spherical nucleus (figs. 8 and 10). The lateral dorsal vessels
increase in size from the 2nd to the 10th segments ; from the
3rd segment to the 10th they have an increasing number of
dilatations and constrictions, each provided with valves similar
to those in the longitudinal vessel (figs. 5 and 10). The
vessel in the 10th segment, which is by far the largest, has as

256 E. S. GOODRICH.

many. as five such sets of valves; the number of cells round
each constriction diminishes the farther it is from the longi-
tudinal vessel. It is this vessel (of the 10th segment) which
supplies the blood to the ovisac, into which it passes directly
from the 11th segment. Both the longitudinal and these
lateral dorsal vessels are contractile.

The non-contractile longitudinal ventral vessel runs straight
up to the 3rd segment, where it divides and joins again, forming
aring. From the anterior border of this ring are given off
three vessels, which join the terminal dorsal branches under
the brain (fig. 5). The ventral system has no valves.

The circulation of the blood (which is red) and the action of
the valves are as follows: The blood is propelled forward in
the dorsal longitudinal vessel by the contraction of its walls;
it is helped in its course by the position of the valves, which
prevent its returning, for when the flow is forwards they
assume the position a’ (fig. 11), leaving a clear passage; when
the blood tries to return they take up the position 0’ (fig. 11),
meeting in the middle and closing the way. Occasionally the
return flow is too strong, and the valves are forced right back
to the position c’ (fig. 11), when they more or less completely
shut off the communication with the side branches. The
course of the blood-stream in these branches is from the longi-
tudinal vessels outwards to the body-wall. As already men-
tioned, coming off from the lateral branches there is over the
whole surface of the body-wall of the first 10 segments under-
lying the coelomic epithelium a fine network of blood-vessels,
which must form a very efficient respiratory system. ‘Thus we
see that, in the main, the course of the blood is up the longi-
tudinal dorsal vessel, down the anterior lateral vessels, spreading
over the body-wall to be aérated, and returning to the ventral
vessel, in which it will flow backwards. Thence it probably
reaches the dorsal vessel again by means of the small branches
given off to the intestine, where nourishment is no doubt
absorbed, and so brought into the general circulation. As will
be seen later, excretion probably takes place from the ventral
vessel by means of the nephridia which cling closely to it.

ON THE STRUCTURE OF VERMICULUS PILOSUS. 257

The alimentary canal is simple, and quite similar to that of
Tubifex. A short way behind the mouth (fig. 2, m.) is a
muscular pharynx in the 2nd segment (fig. 6, ph.), followed by
a slender cesophagus stretching through the 38rd and 4th seg-
ments (fig. 6, @s.), and from the 5th segment backwards a
straight intestine covered with brown chloragogen cells (fig. 6,
int.). The anus is terminal. The pharynx is provided with
irregular masses of glandular cells. From the pharynx back-
wards the alimentary canal is lined by ciliated columnar
epithelium, outside which is a thin layer of circular and longi-
tudinal muscles. In the wall of the intestine is a system of
blood-capillaries forming a plexus similar to that described in
the Enchytrzids and other Tubificids.

The cerebral ganglia form a bilobed brain, deeply cleft in
front and situated in the lst segment (fig. 6, d7.). The struc-
ture of the ventral nerve-cord is somewhat exceptional. On
examining a transverse section taken between the ganglia
(fig. 16), it is seen to consist of an outer covering of ccelomic
epithelium (fig. 16, c. ep.) enclosing the nerve-cord proper and
three strands of muscular fibres, of which the middle one
(fig. 16, m.n.c. m.) is the most important. The nerve-cord
itself is formed of two circular strands of ordinary nerve-fibres,
on the dorsal surface of which run three neurochordal fibres.
Owing to their delicate structure, these enlarged fibres are so
difficult to see in specimens preserved in corrosive and stained
with borax-carmine that at first 1 thought they were altogether
absent. They are, however, plainly visible in sections of
specimens preserved in Lindsay-Johnson fluid and stained with
paracarmine (fig. 15). A transverse section taken through a
ganglion (fig. 15) shows that the ganglionic cells extend on the
under surface of the cord, and rise up at the sides, actually
overlapping the muscular strands (fig. 15, g. c.). In a section
taken at the point where the nerve-cord is passing through
the septum (fig. 14), we see that muscle-fibres pass from the
septum underneath the median muscular strand overlying the
cord (fig. 14, m. s.), while the lateral strands communicate
with those which descend on either side to the body-wall,

258 E. 8. GOODRICH.

The nephridia of Vermiculus are no less peculiar than are
its other organs ; they are situated in Segments 6 to 9 inclu-
sive, and all the segments behind the 11th till near the tail
end. It is interesting to observe a nephridium opening into
the 11th segment in the adult. The general shape of the
nephridium is shown in fig. 25. Stretching back from the
funnel (fig. 25, . f.) we have a narrow duct which passes
through the septum, and gradually swells into a pear-shaped
mass (figs. 25 and 28, p. s. m.), which is slightly pigmented,
contains a large number of granules, and in which the canal is
much convoluted (fig. 28, mph. c.). The main body of the
nephridium (fig. 25, 5. m.) is a more or less irregularly lobed
mass lying by the side of, and spreading over the ventral
blood-vessei, to which it closely adheres (fig. 32). Although
the right and left nephridia in the same segment mingle
over the blood-vessel, their canals do not communicate. From
the main body starts a canal to the exterior (fig. 25, c. ext.),
which opens on the surface in front of the ventral sete by a
minute pore (fig. 25, nph. p.). The internal funnel is small
and somewhat flattened, but has its dorsal lip expanded into a
large, flat, ciliated process of oval shape (figs. 26—29, ce. p.)
Delicate waving cilia are set round the edge of the funnel and
its process (figs. 26 and 27, ew. ci.) ; on the lower or inner
surface of the flat process alone arise numerous long cilia,
forming a flame-like structure with its free end pointing into
the canal (figs. 26, 27, and 29, fl.). These long cilia beat
together with a peculiar quick motion, forming rapidly moving
undulations which, starting from the rim of the process and
travelling towards the tip of the flame, give the whole struc-
ture, with its transverse waves, somewhat the appearance of a
tennis racket with its cross strings. The nephridiostome is
formed of two cells with oval nuclei (fig. 29, n. f.) ; the rest of
the organ consists of a mass of cells with roundish nuclei, and
without distinct cell-outlines, pierced through and through by
the delicate nephridial canal (fig. 30). Except for the few short
blind diverticula given off, especially from that part of the canal
which leads to the exterior (fig. 25, c. cxt.), it does not branch,

ON THE STRUCTURE OF VERMICULUS PILOSUS. 259

and although much convoluted, appears to be continuous
throughout. Here and there this canal enlarges into an
ampulla, in which waves a flame-like bunch of cilia pointing
towards the external aperture (figs. 28, 31, and 382, c. a).
No cilia are found anywhere else but in these
ampullz, which are not very numerous.

The gonads consist of a pair of testes on the anterior wall
of the 10th segment (fig. 12, ¢.), and a pair of ovaries on the
anterior wall of the 11th segment (fig. 12, ov.). The sperm-
ducts are rather short tubes opening by means of widely
opened shallow ciliated funnels into the 10th segment (fig. 12,
sp. f.). A short narrow duct passes from behind the funnel
through the 11th septum, rapidly changes into a thick tube for
about half its course, then narrows again and opens into a
median chamber below the nerve-cord (fig. 12, sp. ch.).
This cavity, which may be called the median spermiducal
chamber, opens to the exterior towards the posterior end of
the 11th segment (fig. 2, m. p. sp.) by an irregular longitu-
dinal aperture. The openings of the sperm-ducts into the
chamber are situated very near together, close to the nerve-
cord (fig. 22, sp. p.)

Fig. 24 represents a median longitudinal section through
the spermiducal chamber ; it shows how the nerve-cord (n. c.)
and the ventral blood-vessel (v. v.) pass over it. From the
structure of its wall with the cuticle, the epidermal cells (ep.),
the longitudinal and circular muscle-fibres (c. m. and J/. m.),
the median chamber is obviously a direct invagination of the
body-wall. Moreover, that this is really the case is proved
by the fact that in the young worm the sperm-ducts
come separately to the surface (fig. 23, sp.), and there
is then no sign of a median pore. The gradual formation of
the median chamber can be traced from its earliest beginning.

The sperm-duct, except the narrow terminal region, is
ciliated. The structure of the wide region is shown in a
longitudinal section in fig. 17, where it is seen that its large
size is chiefly due to the covering of coelomic epithelium cells
(c. ep.), which have become modified into long columnar cells

260 E. S. GOODRICH.

with granular protoplasmic contents. Under this is a thin
layer of circular and a few longitudinal muscles (c. m.),
followed by the ciliated lining epithelium of the duct (w. of
sp.) composed of small granular cells. In the narrow region
of the sperm-duct the ccelomic epithelium is not modified
(fig. 18, c. ep.), while the lumen is bounded by columnar
epithelium, the cells of which are very deep when the duct is
not expanded (fig. 18, w. of sp.). There is no special penial
apparatus, and no prostate.

The oviducts appear to be reduced to a mere depression on
each side of the ventral edge of the 12th septum (fig.
12, ovd.).

The spermathece are two pear-shaped sacs (fig. 12, spth.),
opening in the median ventral line by means of a common
median pore immediately below the nerve-cord at the an-
terior end of the 10th segment (fig. 2, spth. p.). Fig. 20
shows a transverse section of an adult worm in this region
passing through the median pore (spth. p.), at which open
both the right and the left spermathecz (spth.). It is interesting
to note that in the young worm a similar transverse section
(fig. 21) shows that here the two spermathecz (which have all
the appearance of being direct invaginations of the epidermis)
arise distinctly from right and left points some distance from
each other (fig. 21, spth. p.); it is only later that the necks
of the two organs approximate and finally open by a
common aperture. This may easily be confirmed by surface
views of the living worm. ‘Thestructure of the wall of mature
spermathece is shown in fig. 19. Outside is the ordinary
coelomic epithelium (c. ep.), followed by a layer of scattered
longitudinal and circular muscles ; finally we have lining the
interior of the organ a thick epithelium, with large oval nuclei
ranged with their long axis parallel to the surface. The lumen
of the neck of the spermathecz has a pronounced constriction
and enlargement (fig. 12), as in many Tubificide. Spermato-
phores have not been observed.

In the mature worm there are two large sperm-sacs and one
large ovisac. By cutting longitudinal sections of individuals

ON THE STRUCTURE OF VERMICULUS PILOSUS. 261

of varying ages the development of these organs can be traced,
and the result has been diagrammatically represented in fig.
13. The spermatozoa are shed at an early stage of develop-
ment into segment 10, and the anterior septum of this
segment soon bulges out, forming a sac—the anterior sperm-
sac (fig. 13, a. sp.s.). Later on this sperm-sac pushes its
way across segment 9, through its anterior septum into seg-
ment 8. The hinder wall of segment 10 also bulges out,
forming the posterior sperm-sac (fig. 13, p. sp. s.) The ova
are shed into segment 11. The hinder septum of this segment
forms the ovisac (fig. 18, ov. s.), which ultimately pierces as
many as seven or eight septa behind it.1 The posterior
sperm-sac also enlarges, enters the ovisac, and grows back
through five or six segments. It is within the space between
the wall of the ovisac on the outside and the wall of the
posterior sperm-sac within, that the blood-vessels run which
supply nourishment to the ova (fig. 5, v. ov.); it is also, no
doubt, through this space that the ova escape back again to
the 11th segment when ripe.

The clitellum extends over segments 10—12, and part of
segment 13.

SumMMARY AND CONCLUSIONS.

The chief characters of Vermiculus pilosus are therefore
the following:

Four bundles to each segment of furcate sete, generally
three per bundle.

A dense covering of hair-like processes.

A vascular system, containing red blood, and composed of a
dorsal and a ventral longitudinal vessel communicating by
means of lateral vessels which branch on to the body-wall.
The absence of hearts or commissural vessels. An elaborate
system of unicellular valves in the longitudinal and transverse
dorsal vessels.

1 The ovisac pierces the septa at a point nearer to the dorsal blood-vessel
than is represented in the diagram.

VOL. 37, PART 3.—NEW SER. 8

262 E. S. GOODRICH.

A brain deeply cleft in front, and a nerve-cord bearing
muscular strands of considerable size.

A compact nephridium, the funnel of which is peculiarly
modified.

A pair of testes in the 10th, and a pair of ovaries in the
11th segment.

Two short sperm-ducts, without prostate, opening into a
median chamber, which itself opens to the exterior on the 11th
segment by a large median pore. Oviducts rudimentary.

Two pear-shaped spermathece, opening by a common median
ventral pore on the 10th segment.

An anterior and a posterior sperm-sac, and an ovisac.

A clitellum extending over the 10th, 11th, 12th, and part
of the 13th segments.

Many of these characters place this little worm in a very
isolated position. The dense covering of “ sense-hairs,”
although perhaps of no great morphological importance, is
quite unique amongst the Oligocheta. The absence of com-
missural vessels again distinguishes it, I believe, from all its
allies, except the smallest and lowest, such as Molosoma; nor
is it usual among the Tubificide, to which Vermiculus is
doubtless related, to find strong muscular strands on the
nerve-cord.

As for the nephridium, the ciliated process of the funnel is
no doubt of the same nature as the long ciliated processes on
the nephridiostome of Nereis (8) ;! all such structures appear
to be specialised processes of the funnel-cells themselves.
The presence of cilia in special ampulle, and their absence in
the rest of the nephridial canal, is a character which may, I
think, prove to be common to all the so-called Microdrili.
Such ampulle are certainly present in other Tubificids I have
examined, and in the Enchytreids, to the nephridium of which
latter the compact excretory organ of Vermiculus bears no
little resemblance. Professor Vejdovsky (6) has described a
Planarian, Microplana humicola, in which the nephridia

1 T once observed inan Enchytreid long ciliated processes on the nephridio-
stome. exactly like those of Nereis above referred to.

ON THE STRUCTURE OF VERMICULUS PILOSUS. 263

terminate in flame-cells, while the canal itself is provided with
“ flames” at intervals along its course; it seems not impro-
bable that the flame-like cilia of the nephridiostome of Vermi-
culus represent the “‘ flame” of the terminal cell, and the cilia
of the ampulle represent the “ flames ” distributed along the
course of the canal in the Planarian. Possibly the arrange-
ment found in the large nephridia of the Earthworms, described
by Benham (1), in which whole tracts are ciliated, is derived
from some such system as we find in Vermiculus and the lower
Oligochzetes by the extension of the cilia over a great length
of the canal.!

The late development of the median spermiducal chamber
is another of those characters quite peculiar to our worm.
What the function of this chamber may be it is difficult to
conjecture ; perhaps it acts as a sucker during copulation; the
disposition of the muscles would favour this supposition. How-
ever, it must be noticed that besides the formation of the
spermiducal chamber, the apertures of the genital organs them-
selves show a distinct tendency, as it were, to unite in the
middle line. This is clearly seen in the case of the male pores,
which come close to each other; but more especially in the
case of the spermathecal pores, which become actually con-
fluent. As a striking contrast, we may compare such a form
as Heterocheta, in which, as described by Benham (2), the
sperm-ducts and spermathecz open above the ventral setz!

The fact that Bothrioneuron, described by Stole (5), possesses
a median male pore might suggest a close relationship between
the two worms, but they differ in other particulars most
markedly from each other. The sete, nervous system, and
nephridia of Bothrioneuron are all very different from those
of Vermiculus ; it possesses commissural vessels, a prostate, a
complicated system of genital sete, and no spermathecze.

On the whole, it must be concluded that Vermiculus stands

1 Since this was written, Professor A. G. Bourne has in this Journal
(vol. xxxvi, 1894, “On Moniligaster grandis”) described somewhat
similar undulating bundles of cilia, which, however, are attached at both ends,
not hanging freely in the canal as the cilia do in the ampulle described above.

264 E. 8S. GOODRICH.

very much by itself; the shape of its sete, and above all the
situation of its gonads, place it in the family Tubificide, but
its more intimate relationships remain obscure for the present.

March 30th, 1894.

List oF WoRrKS REFERRED TO.

1. Benuam, Dr. W. B. B.—“ The Nephridium of Lumbricus,” ‘ Quart. Journ.
Mier. Sci.,’ xxxii, 1891.

2. Benuam, Dr. W. B. B.—“ Notes on some Aquatic Oligocheta,” ‘ Quart.
Journ. Mier. Sci.,’ xxxiii, 1891.

8. Goopricu, E. §8.—* On a New Organ in the Lycoridea, &.,” ‘ Quart.
Journ. Mier. Sci.,’ xxxiv, 1893.

8a. Goopricu, HE. 8.—‘‘ Note on a New Oligochete,” ‘ Zool. Anz.,’ xv,
1892.

4. Hessz, Dr. R.—“ Beitrage zur Kenntniss des Baues der Enchytreiden,”
‘ Zeitschr. f. wiss. Zool.,’ vol. lvii, 1894.

5. Stotc, Dr. A.—‘‘ Monografie Ceskych Tubificidu,” ‘Abh. k. Bohm.
Ges.,’ 1888.

6. Vespovsky, Prorrssor F.—‘‘Sur une nouvelle Planaire terrestre,”
‘ Revue Biol. du Nord de la France,’ 2 ann., 1890.

ON THE STRUCTURE OF VERMICULUS PILOSUS. 265

EXPLANATION OF PLATES 26—28,

Illustrating Mr. E. S. Goodrich’s paper, ‘‘ On the Structure of
Vermiculus pilosus.”

List oF REFERENCE LETTERS.

a. sp. s. Anterior sperm-sac. 6.2. Main body of the nephridium. ér.
Brain. 4%. v. Blood-vessel. 0. 2. Body-wall. ¢. Cuticle. c¢. a. Ciliated
ampulla. c.c. Ceelomic corpuscle. c. ep. Ceelomic epithelium. . ext. Canal
to the exterior. ci. Cilia. ci. Clitellum. . m. Circular muscles. . 2. Cell
of the nephridium. c. p. Ciliated process. cal. Colom. d. v. Dorsal longi-
tudinal vessel. ep. Epidermis. ep. cl. Epidermis of clitellum, ep. spth. In-
ternal lining of the spermatheca. ex. ci. External cilia. #. Flame-like cilia.
g.¢. Ganglionic cell. iz¢. Intestine. 7. d.s. Left dorsal bundle of sete.
1. d. v. Lateral dorsal vessel. 7. 7. Lateral line. 7. m. Longitudinal muscles.
l. v. s. Left ventral bundle of sete. J. v. v. Lateral ventral vessel. m. Mouth.
m. l. v. v. Middle lateral ventral vessel. m.x. c. Muscles on the nerve-cord.
m.n.c. 1. Lateral strand of muscles. m. z. c. m. Median strand of muscles.
m. p. sp. Median pore of the spermiducal chamber. m.s. Muscles of the
septum. #.c. Nerve-cord. x./. Nephridiostome. zp. c. Nephridial canal.
mph. p. Nephridiopore. «@s. (sophagus. ov. Ovary. ovd. Oviduct. ovs.
Ovisac. pel. Palpocil or sense-hair. p. /. v. v. posterior lateral ventral vessel.
ph. Pharynx. pr. Prostomium. p.s.c. Post-septal canal. p. s. m. Post-
septal mass. jp. sp. s. Posterior sperm-sac. 7. v. s. Right ventral body of
sete. set. Seta. sept. Septum. sp. p. Spermiducal pore. sp. ch. Median
spermiducal chamber. sp. f. Spermiducal funnel. spth. Spermatheca. spth.
p. Spermathecal pore. ¢. Testis. v. Valve. v. ov. Vessel to ovisac. 2». v.
Ventral longitudinal vessel. w. of sp. Wall of sperm-duct. 1. of v. Wall of
blood-vessel.

Fic. 1.—Adult Vermiculus pilosus, drawn natural size.

Fic. 2.—View of the anterior part of the body, slightly twisted so as to
display the dorsal and ventral sete and the median genital apertures.

Fie. 3.—Enlarged view of a seta. Z. apoch. 4 mm., oc. 4, camera.

Fie. 4.—Small portion of the body-wall, drawn from the living to show the
covering of hair-like processes. Z. apoch. 4 mm., oc. 4.

Fic. 5.—Semi-diagrammatic view of the vascular system from the dorsal
surface ; from the living.

Fic. 6.—View of the brain and anterior part of the alimentary canal; from
the living.

266 E. S. GOODRICH.

Fie. 7.—Portion of a transverse section of the body-wall in the region of
the lateral line. Z. apoch. 4 mm., oc. 4, camera.

Fie, 8.—Longitudinal section through the longitudinal dorsal vessel, passing
rather to the side, so that the lumen does not appear continuous. Z., ob. D,
oc. 4, camera.

Fie. 9.—Portion of the vascular system, showing the manner of communi-
cation between the lateral dorsal and the lateral ventral vessels ; from the
living.

Fic. 10.—Small portion of the dorsal blood system in the region of the
10th and 11th segments, greatly enlarged to show the valves; from the living.
Z., oil imm. 2 mm., 4 oc.

Fie. 11.—Diagram of the three positions taken up by the valves: a’, when
the flow is forwards; 4’, to stop the flow backwards; and c’, when they are
forced back.

Fie. 12.—Diagram representing the genital organs in situ; from the
living and sections.

Fic. 13.—Diagram representing the development of the sperm-sacs and
ovisac in an immature specimen, as seen in a solid longitudinal section ; chiefly
from sections.

Fic. 14.—Transverse section of the nerve-cord passing through a septum.
Z., ob. D., oc. 3, camera.

Fig. 15.—Transverse section through a ganglionic region of the nerve-
cord. Z., ob. D, oc. 3, camera.

Fic. 16.—Transverse section of the nerve-cord between two ganglia. Z.,
ob. D, oc. 3, camera.

Fic. 17.—Longitudinal section of the thick region of the sperm-duct. Z.,
ob. D, oc. 4, camera.

Fie. 18.—Longitudinal section of the narrow region of the sperm-duct.
Z., ob. D, oc. 4, camera.

Fic. 19.—Section through the wall of the spermatheca. Z. apoch. 4 mm.,
oc, 4.

Fic. 20.—Transverse section of an adult worm passing through the median
spermathecal pore. Z., ob. A, oc. 3, camera.

Fic. 21.—Tranverse section of a young worm passing through the sperma-
thece. It has been necessary to combine two sections to show this clearly.
Z., ob. A, oc. 3, camera.

Fic. 22.—Transverse section of an adult worm passing through the median
spermiducal chamber and its pore. Z., ob. A., oc. 38, camera.

Fic. 23.—Transverse section of a young worm in the same region, showing
the absence of a median chamber. It has been necessary to combine two
sections. Z., ob. A, oc. 3, camera,

ON THE STRUCTURE OF VERMICULUS PILOSUS. 267

Fic. 24.—Longitudinal section through the median spermiducal chamber.

Fic. 25.—General view of a nephridium, seen from below ; from the living.
Z., ob. D, oc. 4.

Fic. 26.—View of the nephridial funnel from below; optical section from
the living. Z. apoch. 4 mm., oc. 8.

Fic. 27.—View of the nephridial funnel from the side; optical section from
the living.

Fie. 28.—View of the post-septal mass of the nephridium, reaching through
the septum to the funnel; from the living. So far diagrammatic that the
whole course of the canal is represented, although it could not be followed .
out for certain in all its parts. Z., ob. D, oc. 8.

Fic. 29.—Section through the nephridiostome and post-septal canal, show-
ing the nuclei of the funnel cells. Z. apoch. 4 mm. oc. 4, camera.

Fie. 30.—Small portion of section through the body of the nephridium,
showing the canal and the nuclei of the nephridial cells. Z. apoch. 4 mm.,
oc. 4, camera.

Fie. 31.—Small portion of the living nephridium, showing the ciliated
ampulle. Z., oil imm. 2 mm. oc. 4.

Fic, 32.—Main body of the nephridium attached to the ventral longitu-
dinal blood-vessel; from the living. Z. apoch. 4 mm., oc. 4.

eiril i e's

i ¢
i
‘
{ ?
é
'
:
‘
’
*
|
ra, a
[*
‘
nm: “
Vira,

. =
>
—
—_ «¢
'
,
|

-.
rs
—_
a f '
—
*
é

aot

eit

‘
L_—
ad
!
'

> 9

Par

MOUTH-PARTS OF THE CYPRIS-STAGE OF BALANUS. 269

On the Mouth-parts of the Cypris-stage of
Balanus.

By

Theo. T. Groom, F.Z.S.,
Late Scholar of St. John’s College, Cambridge.

With Plate 29.

Tuer buccal mass of all ordinary Cirripedes consists, as is
well known, of three pairs of jaws, together with the labrum
and palps. The form and disposition of these mouth-organs
in the various genera of the Thoracica has been well investi-
gated by Darwin, and the differences observed between the
different forms in respect to these parts have been utilised for
purposes of classification. The morphological significance of
the jaws has, however, never been satisfactorily ascertained,
and the current views are all practically based upon supposi-
tions, the actual development having been traced in no case.

Darwin (Nos. 1 and 2) regarded the frontal filaments seen
in the larva as the first pair of appendages, and taking the
eyes to mark the first segment of the head, and believing that
the prehensile antennules arose within the fronto-lateral horns,
and that the three pairs of jaws arose in front of the Nauplius
appendages, regarded the mandibles of the adult as the appen-
dages of the fourth segment of the typical Crustacean, and the
two pairs of maxillz as the fifth and sixth respectively.

The eyes, however, are now no longer regarded as modified
appendages ; the frontal filaments, too, have been more cor-
rectly regarded as sense-organs, and the fronto-lateral horns
have been proved by Claus and others to be glandular processes
of the carapace.

270 THEO. T. GROOM.

Krohn, Willemoes-Suhm, and Lang (Nos. 3, 7, and 9) have
proved that the prehensile antennules of the Cypris-stage in
the Thoracica arise from the first pair of Nauplius appendages
(antennules), and Delage (No. 10) has shown the same for
Sacculina.

With reference to the fate of the remaining Nauplius
appendages, Metschnikoff, Willemoes-Suhm, and Lang (Nos. 8,
7, and 9) believed that both pairs were lost, and the first-
mentioned supposed that the mandibles and two pairs of
maxillz were all new structures formed inside a fourth pair of
appendages seen behind the third pair of Nauplius appendages.

Claus, on the other hand, supposed (No. 8) from the analogy
of other Crustacea that the mandibles of the Cypris-stage arose
from the third pair of Nauplius appendages, the latter not
being lost like the second pair, but greatly reduced in size.
This observer did not, however, succeed in proving his point,
and I shall attempt in the following remarks to show that this
view is the correct one.

The Cypris-stage, as is well known, possesses a well-defined
buccal mass. This, on account of its concealment within the
carapace, its small size, and delicate nature, as well as the
close packing of the component parts, is difficult to make out
satisfactorily.

Darwin found in the Cypris-stage of Lepas australis
“all the masticatory organs of a Cirripede in an immature
condition.”

Pagenstecher (No. 4) describes the mouth-organs of the
Cypris-stage of Lepas pectinata after fixation as consisting
of imperfect lobes and papille without bristles or teeth.

Willemoes-Suhm (No. 7) did not succeed in satisfactorily
separating the mouth-parts in Lepas fascicularis, but
describes the buccal mass as consisting of “three parts all
very rudimentary.”

Claus (No. 6) also describes three pairs of gnathites in an
undetermined Cypris-stage. These had the form of simple
outgrowths; the mandible was largest and connected with the
labrum by a finger-shaped palp referable equally to the upper

MOUTH-PARTS OF THE CYPRIS-STAGE OF BALANUS. 271

lip or to the mandible; the two hinder pairs Claus describes
as giving the impression of a single pair of appendages.

Of the two species of Balanus the Cypris-stage of which I
have carefully examined, B. perforatus, on account of its
minute size, is much less favourable for purposes of investiga-
tion than B. balanoides. Repeated dissections of the pupa
of the latter form have enabled me to form a tolerably good
idea of the constitution of the buccal mass.

Viewed from behind (fig. 3), a pair of appendages (mz?.),
clearly identical in position and relations with the lower lip, or
second pair of maxille of the adult, is recognisable. These
are attached on each side behind along an oblique straight
line, and are somewhat broader at the base than at the apex.
Like the second maxille of the adult, they are closely applied
to one another in the middle line. Seen in side-view (fig. 8)
they are somewhat lanceolate in shape, the apex generally
projecting somewhat forwards ; the hinder margin is convex,
while the anterior is somewhat sinuous. Like the remaining
mouth-parts, and like those of Lepas, as described by Pagen-
stecher, they are simple lobes devoid of teeth or bristles. At
the junction of these appendages with the first pair of maxillz
may be seen a conical chitinous process (figs. 1, 3, 4, 6).

The first maxille (mz!.) are situated externally to the
second, and project somewhat behind the latter (fig. 1, cf. also
fig. 11). They are similarly convex externally, but are simply
concave on the inner side ; they lie in planes inclined outwards
at a greater angle than the second maxille (cf. fig. 11). Like
the remaining mouth-parts they are broadest externally,
narrowing towards the centre. They do not attain quite the
vertical height of the second maxille (fig. 1),

The mandibles are large, and occupy the outer part of the
buccal mass (figs. 1—3), being found externally and anteriorly
to the first pair of maxille (figs. 1—3; cf. also fig. 11). The
main portion of the mandible passes distally into an inwardly-
curved process with a somewhat truncated end, giving the appen-
dage much the form of the adult mandible or first maxilla,
though, as already remarked, no teeth or hairs are present,

Paresh THEO. T. GROOM.

Anteriorly the basal portion of the mandible gives off a some-
what inwardly-curved palp-like process (figs. 1—5, palp.). This
has all the appearance of being an integral portion of the
mandible, and is filled by a mass of developing muscle con-
tinuous with that of the rest of the appendage, and running,
together with the muscle of the two pairs of maxille, to the
anterior part of the thorax (fig. 1). The anterior margin of
the basal portion of the mandible (figs. 1 and 5) is apparently
very short, though this is difficult to ascertain definitely, since
in teased preparations the buccal mass commonly tears away
from the head immediately in front of the palp; it is, thus,
easy to see why Claus regarded the palp as belonging equally
to the mandible or labrum. The evidence, too, afforded by the
muscle-supply must be taken with caution, as Nussbaum (No.
11) describes a muscle supplying the mandible in the adult
Conchoderma as having its origin in the palp. The evidence,
however, afforded by the immature Cypris-stage to be described
immediately appears to clearly show that the palp belongs to
the mandible.

The three pairs of gnathites, according to Claus, are situated
beneath a labrum; the labrum, however, as a definite promi-
nence, is exceedingly small at this stage in Balanus, and
projects as a very minute lobe between the distal ends of the
palps. The site of the future labrum is, nevertheless, recog-
nisable as a broad area in front of and between the mandibles
and their palps. The jaws are directed ventrally, so that this
area is to be described as situated anteriorly to, rather than
below the jaws.

Some stages of an undetermined species of Balanus ob-
tained from Messrs. Sinel and Hornell in Jersey in the spring
of the present year (1894) were specially valuable as throwing
light on the question of the fate of the Nauplius appendages,
and on the homologies of the jaws of the adult. The antennules
of the Nauplius, as already pointed out, have long been known
to give rise to the prehensile antennules of the Cypris-stage,
and these, according to Darwin, can often be recognised in the
adult. The fate of the antenne of the Nauplius has not been

MOUTH-PARTS OF THE OCYPRIS-STAGE OF BALANUS. 273

traced. Fig. 12 shows the position of the base of the antenna
and its socket in the last Nauplius-stage of the Balanus in
question. The socket is seen to lie at the side of the labrum,
just behind that of the antennule ; the innermost part of the
base of the mandible projects as a pivot, which fits into a slight
indentation in the side of the labrum. The origin of the
mandibles is well behind the mouth and attachment of the
labrum. As the Nauplius prepares for the moult which trans-
forms it into the Cypris-stage the body contracts, and many of
the structures belonging to this latter stage can be seen beneath
the cuticle. If the larva be now carefully dissected out of the
cuticle, or be examined just after the moult, it shows peculiar
features in which it approaches the Nauplius-condition, and
which are not found in the perfect Cypris-stage. Figs. 9 and
10 show the larva in this stage. The edges of the carapace
have not yet met in the mid-ventral line, and allow good views
of the ventral surface of the animal. The prehensile antennules
can be seen at the anterior end of the labrum, which, in most
cases, is still recognisable, often as a well-developed structure.
The sides of the labrum are indented, as in the Nauplius-stage,
by the remnants of the antennz, which are now undergoing
histolysis and absorption, but are easily recognisable. Behind
these can be seen the mouth-parts of the Cypris-stage already
well-formed. These consist of the mandibles with their palps,
and the first and second pairs of maxille. In the more advanced
stages the labrum and antenne are small (fig. 11) or have
practically disappeared, so that the larva has attained the con-
dition (so far as these organs are concerned) seen in the
perfect Cypris-stage. A series of stages were obtained showing
the gradual reduction of these parts. A highly important
feature is that the mandibles, clearly forming with their palps
a single appendage, can be detected on the site of the mandible
of the Nauplius, behind the antennz and labrum. The labrum
shows in some cases more or iess clear indications of the
median and lateral lobes found in the Nauplius at the distal
extremity. It is clear, then, that the palps can neither be
regarded as equivalent to the antenne, nor to the lateral lobes

274 THEO. T. GROOM.

of the labrum of the Nauplius, but belong to the mandibles.
Judging by analogy the palp will represent the ramus of this
appendage, the mandible proper being the gnathobase.

Of the two pairs of maxill, the first pair are indicated after
the first moult undergone by the Nauplius by a row of bristles
which has been termed the extra-maxillary are (No. 12). Ata
later stage this becomes a small foliaceous appendage (fig. 13)
provided with a number of bristles, and already correctly
regarded by Claus as an early condition of the “ outer
maxilla’? (No. 8). Inside this the first maxilla of the Cypris-
stage clearly arises (fig. 13). The second pair of maxille
appear later,simultaneously with the six pairsof cirri; they arise,
as described by Claus, just in front of the first pair of the latter,
and are clearly serially homologous with the thoracic appen-
dages. Further details as to the origin of the two pairs of
maxille will be given on a future occasion.

It may be regarded as tolerably certain from what has been
said above that :

(1) The antennz of the Nauplius become definitely lost
with the moult resulting in the production of the Cypris-
stage.

(2) The biramous mandibles of the Nauplius become reduced
at the same time to the small mandibles, the ramus being
probably preserved in the form of the small palp.

(3) The first pair of maxille arise behind the mandibles, and
at a later date, as a small pair of foliaceous appendages.

(4) The second pair of maxille arise still later just in front
of the first pair of thoracic legs (cirri).

The mandibles, first maxille and second maxille are accord-
ingly developed consecutively in the order named. They
cannot be regarded as parts of a single or fourth pair of
appendages as Metschnikoff maintains, neither can the two
pairs of maxille be regarded as outer and inner parts of a
single pair of appendages as Claus suggests. The mandibles
of the Nauplius are not lost as many authors have maintained,
but give rise, as Claus supposes, to the mandibles of the
Cypris-stage. The jaws represent, then, the third, fourth,

MOUTH-PARTS OF THE OYPRIS-STAGE OF BALANUS. 275

and fifth pairs of appendages of the typical Crustacean
series.

This conclusion is quite in harmony with that of Nussbaum,
as deduced from his beautiful anatomical investigations of the
adult. Cirripede, “ Die Anordnung der Musculatur spricht
dafiir, dass jeder der drei Kiefer aus einem Beinpaare hervor-
gegangen ist” (No. 11).

In the fate of the mandibles of the Nauplius and in the
constitution of the head of the adult the Cirripedia thus con-
form to what is recognised as typical among other Crustacea,

BIBLIOGRAPHY.

1. C. Darwin.—‘A Monograph of the Sub-class Cirripedia,’ vol. i,
“ Lepadide,” 1851. ;

2. C. Darwin.—‘A Monograph of the Sub-class Cirripedia,’ vol. ii,
‘** Balanide, Verrucide, &.,” 1854.

3. A. Kréun.—“ Beobachtungen iiber die Entwickelung der Cirripedien,”
‘Arch. f. Naturgesch.,’ vol. xxv, 1859.

4. A. PacEnstEcHER.—“ Beitrage zur Anatomie und Entwickelungsges-
chichte von Lepas pectinata,” ‘ Zeitschrift f. wiss. Zool.,’ vol. xiii,
1868.

5. E. Metscunrxorr.—“ Ueber die. Entwickelung von Balanus bala-
noides,” ‘ Sitzungsb. der Versamml. Deutsch. Naturf. zu Hannover,’
1865.

. C. Craus.—‘ Die Cypris-ahnliche Larva der Cirripedien und ihre Ver-
wandlung in das festsitzende Thier.,’ Marburg and Leipzig, 1869.

fon)

7. R. von WitLemors-Sunm.— On the Development of Lepas fascicu-
laris and the Archizoéa of Cirripedia,” ‘Phil. Traus. Roy. Soc.,’
vol. elxvi, 1876.

8. C. Cuaus.— Untersuchungen zur Erforschung der genealogischen
Grundlage des Crustaceen-systems,’ Wien, 1876.

9. Arnotp Lanc.—‘‘ Ueber die Metamorphose der Nauplius-Larven von
Balanus,” ‘Mittheilungen der Aargauischen Naturforschenden Gesell-
schaft,’ vol. i, 1863—1877.

10. Yves Dezace.—* Evolution de la Sacculine,” ‘ Archives de Zoologie
expérimentale,’ vol. ii, 1884.

11. M. Nusspaum.—‘Anatomische Studien au Californischen Cirripedien,
Bonn, 1890.

12. T. T. Groom.—“ On the Early Development of Cirripedia,”’ ‘Phil,
Trans. Roy. Soc.,’ vol. clxxxv, 1894.

276 THEO. T. GROOM.

EXPLANATION OF PLATE 29,

Illustrating Mr. Theo. T. Groom’s paper, ‘On the Mouth-
parts of the Cypris-stage of Balanus.,”

Fie. 1.—Side view of buccal mass of Cypris-stage of Balanus balanoides.
Mandible. Mandible with (palp) its palp. mz. and mz. First and second
maxilla, showing chitinous process between their bases. duce. m. musi.
Muscles supplying buccal mass. Thoracic muscles. Muscles filling up the first
segment of the thorax.

Fic. 2.—View of same from in front. Letters as in Fig. 1. Also, adductor.
Adductor muscle of carapace.

Fic. 3.—View of same from behind.

Fie. 4.—Similar view, the form of the appendages being shown by trans-
parency.

Fie. 5.—Mandible and its palp.

Fic. 6.—View of first maxilla from outside.

Fic. 7.—View of same from side.

Fig. 8.—View of second maxille from side.

Fie. 9.—Side view of an immature Cypris-stage of an undetermined species
of Balanus from Jersey. adductor. Adductor muscle of carapace. anf}.
Antennule. azé?. Antenna. caudal app. Caudal appendage. compd. eye.
Compound eye. /r/. aperture. Aperture of fronto-lateral gland. ma. First
maxille. ma. Second maxille. palp. Palp of mandible.

Fic. 10.—Ventral view of similar larva. Lettering as in Fig. 9. Also,
ant', muscles. Muscles to antennule. cp. edge. Edge of carapace.

Fig. 11.—Ventral view of mouth-organs of a larva in a somewhat more
advanced stage. The labrum is now greatly reduced, and the sites of the
antenne barely recognisable.

Fie. 12.—View of labrum and bases of antennules, and antenne with
their sockets (ant'. sock., ant. sock.) of the last Nauplius-stage of a species
of Balanus from Jersey. mandible sock. Socket of mandible.

Fie. 13.—View of the developing maxille (mz'. and mz*.) of the Cypris-
stage, as seen ina similar Nauplius. The small setigerous first maxille are
external, while the two pairs of pupal maxille are underneath the cuticle;
the first pair of these has the tip slightly withdrawn from the external
maxillz, inside which they were formed,

STUDY OF COCCIDIA MET WITH IN MICE. 277

A Study of Coccidia met with in Mice.
By

J. Jackson Clarke, M.B.Lond.
With Plate 30.

Nearty all that has been published on Coccidia of mice is
contained in the writings of Kimer! and Th. Smith.? Ludwig
Pfeiffer,® in reference to what these authors have said on this
subject, concludes: ‘“‘ Unsern frithern Auseinandersetzung
nach, handelt es sich hier um das Schwarmercystenstadium
einer Coccidie: das dauercystenstadium ist unbekannt.” In
order to supply data which may fill the gap thus indicated, I
venture to submit the following observations.

A white mouse, kept in a place previously occupied by
rabbits, was found dead. Dissection revealed the presence of
large numbers of Coccidia in every part of the alimentary
tract beyond the cardiac opening of the stomach. There was
no other abnormal feature. Sections of the intestine showed
an altered condition of the epithelium, but although decom-
position had not begun when the tissue was fixed I could not,
apart from the encapsuled parasites collected in the lumen of
the gut, have been certain that epithelial infection by Sporozoa
was present, so soon after death do some of the intracellular
protozoa lose their characters.

The contents of the intestine were spread on a clean slide,

1 Kimer, ‘ Ueber die Hi- oder Kugelformigen sogenannten Psorospermien,
&e.,’ Wiirzburg, 1870.

2 Th. Smith, Washington, ‘Journal of Comp. Med. and Surg.,’ 1889.

3 L. Pfeiffer, ‘ Protozoen als Krankheitserriger,’ 1891, p. 57.

VOL. 37, PART 3,—NEW SER. T

278 J. JACKSON CLARKE.

placed on damp blotting-paper in a Petri’s dish,! and kept at
room temperature (May—June).

The ripe free parasites were uniformly much smaller (aver-
aging 18 x 13 mw) than the Coccidium oviforme of the
rabbit (averaging 32 x 22 mu). They were also of a more
rounded shape. When first examined the capsules of the free
parasites were completely filled by granular protoplasm, having
the usual dense, round, central body ; they were examined on
successive days for a fortnight, and I may briefly indicate the
results by saying that with the modification in size previously
alluded to they underwent the same changes as does C. ovi-
forme of the rabbit when placed under similar circumstances.
Thus on the sixth day most of the parasites had subdivided
(fig. 2) into four granular spheres, of which some possessed a
clear oval corpuscle placed as far as possible from the point of
meeting of the segments (fig. 3). On the eighth day the sub-
divisions (sporogonia) had assumed an oval form, and some pre-
sented a capsule and a differentiation into spores and granular
residual matter (nucléus de reliquat). Many of the spores
showed the dumb-bell form described by Leuckart in C. ovi-
forme of the rabbit (see fig. 5), and in one or two cases I was
able to see the division of the C-shaped body into two comma-
shaped spores, as described by Balbiani. This arrangement of
lasting spores is, as in the case of the rabbit’s Coccidium, often
departed from, and appearances such as are depicted in fig. 6
occur frequently.

In order to have material for histological study, and to test
the time required for the manifestation of the disease, a young
healthy mouse, whose faeces were found to be free from Coc-
cidia, was fed with bread-sop containing some of the material
taken from the Petri dish after twenty-two days’ exposure
to air.”

1 This consists of two shallow glass dishes, one of which is larger than the
other, and when inverted forms a cover which excludes dust from the smaller
one.

2 T have given the days on which the various appearances were first noted,
and I do not wish to convey an impression that on any given day all the

STUDY OF COCCIDIA MET WITH IN MICE. 279

No change was noticed in the animal until six days after the
parasites were administered. On this day its coat was noticed
to be rough, the abdomen was distended, and the thighs some-
what drawn up towards the belly, so that the animal’s gait
was stiff. The stools were softer than normal. The next
day blood! was noticed about the animal’s anus, and this
sanguineous discharge contained coccidia of the same dimen-
sions as those described above. On the seventh day the animal
was distinctly ill and suffering, so it was killed. After death
the stomach was found to be distended, the contents consisting
chiefly of parasites, many of which showed the formation of
sickle-spores* as shown in fig. 8. The large and small intes-
tine contained practically little beside ripe Coccidia. The
stomach, intestines, and portions of the liver and kidneys were
fixed in saturated solution of corrosive sublimate and hardened
in the usual way. After embedding in paraffin, sections were
cut and stained in Ehrlich’s acid hematoxylin and eosin
(Gribler’s wasserléslich).

Histological Examination.

In the stomach the glands of the cardiac end contained free
Coccidia, many of which were encapsuled and filled with small
swarm-spores as shown in fig. 8; but the most striking feature
consisted in masses of minute bodies exactly resembling the
swarm-spores within such capsules, but stained and lying free on
the surface of the mucous membrane, and distending the ducts

coccidia presented the same appearance. On the contrary, up to nineteen
days after the parasites were placed in the moist chamber, nearly all the
phases could be observed at the same time; but on the date last named all
the parasites were subdivided into sporocysts, each of which comprised two
spores and a nucleus de réliquat, as the residual body has been termed by
Aimé Schneider.

1 This phenomenon is of interest in connection with an observation recorded
by Hess (‘Schweitz Archiv fiir Thier.,’ Ziirich, 1892, vol. xxxiv, p. 125), to
the effect that the “Rothe Ruhr” of cattle is associated with the presence of
a coccidium.

2 Compare R. Pfeiffer, ‘ Beitrage zur Protozoon-Forschung,’ Berlin, 1892.

280 J. JACKSON CLARKE.

of the glands. In the latter situation the accumulated spores,
devoid of any containing capsule, had an appearance which
recalled the well-known Sarcosporidia—a resemblance which
will be referred to again in a subsequent part of this paper.
That these small free bodies were swarm-spores derived from
the coccidia was shown conclusively, I think, by a close exa-
mination of the epithelial cells against which they lay. Many
of the cross-sections of the pyloric glands had the appearance
shown in fig. 9, where some of the lining cells contain one or
several minute bodies, many of the same size and appearance
as the small free bodies lying in the lumen of the gland.

From the minute intracellular parasites which lay in the
cell-protoplasm a gradation of forms could be traced up to the
encapsuled swarm-sporing parasites; but in the stomach the
full-sized intracellular parasites were very few in number
compared with those about to be described in the intestine.
This was probably due to the majority of the parasites having
subdivided into swarm-spores without the formation of a cap-
sule, a process which occurs in C. oviforme in the acute
disease, and which will be described immediately in the small
intestine in the mouse.

Small intestine.—Sections taken from various parts all
presented similar features. The epithelial cells of the glands
of Lieberkiihn were almost every one infected, and in the
majority of them the parasites were spherical with numerous
“corps albuminoides de réserve,’ and a small central body
which stained with eosin, and in the lumen of the gut lay the
encapsuled free parasites.

Sections of the small intestine showed that there the
epithelial cells were as much infested as in the stomach, but
the parasites presented a somewhat different aspect, as is
shown in fig. 10, which represents a cross-section of one of
the crypts. The epithelial cells show signs of having pro-
liferated, being in places three deep instead of forming a
single layer, and the sickle-shaped swarm-spores are wanting.
Most of the intracellular parasites are in the shape of granular
spheres, the granules staining with eosin; but some, such as

STUDY OF COCCIDIA MET WITH IN MICE. 281

the one marked a, have larger granules! staining deeply with
hematoxylin. One parasite (6) has surrounding it a thick
capsule, and is in the same phase as that shown in fig. 1. The
granules and the central body took the eosin and not the
hematoxylin of the stain. Other parasites were larger than
this, but still devoid of a capsule (c), and represented, I be-
lieve, an intermediate phase between the common form (/) and
the phase represented by the bodies d and e, which measured
respectively 34 x 19 and 35 x 22. The marginal part is
represented as seen in optical section, and is marked by a close-
set series of short rods which stained a deep blue with hema-
toxylin. The lighter dots in the bodies represent the surface
rods seen out of focus, and the series of darker dots in their
inner parts are represented as seen in focus, though they are
placed on the surface of the parasites. I have met with para-
sites in this phase in the liver and intestine of rabbits suffering
from acute coccidial disease, and as I have nowhere seen a full
account of this particular phase, I may introduce two figures
(11 and 12) taken from sections of a rabbit’s liver. These
short rods I believe to be of the nature of chromosomes, and
the parasites in this phase to be almost wholly nuclear in con-
stitution. I once believed them to be wholly nuclear, until
lately, at the suggestion of Professor Marcus Hartog (Cork),
I looked for a layer of protoplasm, which I have found to be
present, but so thin is this layer that it is difficult to make
out. Some of the modes of subdivision of this phase of the
parasite are illustrated in fig. 10, d, and figs. 11 and 12.

The large intestine contained innumerable coccidia both
within the epithelial cells and free, but I could find none in
the phase of sickle-formation as in the stomach, nor were there
any with peripheral rods as in the small intestine. The liver
and kidneys showed no abnormal features.

I have not yet been able to procure Eimer’s original paper,
but it is freely abstracted by Leuckart.* I have been able to

1 These granules, like those staining with eosin, are in some cases “corps
albuminoides,” i. e. stored food material.

2 Leuckart, ‘ Parasites of Man,’ Hoyle’s translation, 1886, pp. 197 and
219.

282 J. JACKSON CLARKE.

conclude that the parasites of Eimer (Eimeria, A. Schneider)
are the same as those described in this paper. The chief
grounds for this conclusion consist in the round form Eimer
refers to in many of his parasites, and to their dimeusions,
18H xX 18 and 26 x 16, approaching more nearly to the
parasites described above than to C. oviforme in the rabbit,
where the intestinal parasites average, I have found, 32 u x
22 p.

In order to test the transmutability of C. oviforme into
Eimeria, I fed two young white mice whose excrement I had
found to be free from coccidia with food mixed with the dung
of a rabbit containing spore-ripe C. oviforme. The result
of these experiments was negative.

A third experiment gave a positive result. A very old white
mouse, whose feces had been found to contain no coccidia,
was given food mixed with some material which had been taken
from the cecum of a rabbit, and kept for two months in a
moist chamber. There were many spore-ripe coccidia in this
material. Five days after the administration of the parasites
the mouse died, and its alimentary tract was found to be
filled with coccidia like those in the other mice referred to
above.

I may now conclude these observations by indicating the
conclusions to which they point.

Conclusions.

1. That the Sporozoa described above as they were found in
white mice belong to the Coccidiidea, and, like C. oviforme,
to the Disporez.

2. That they are probably identical with the Sporozoa de-
scribed by Eimer in the intestines of mice, and named
Eimeria by Schneider.

3. That Eimeria is probably only a variety of C. ovi-
forme, and may be but a modification of C. oviforme
(although two out of three experiments made with a view of
solving the point gave negative results) determined by the

STUDY OF COCCIDIA MET WITH IN MICE. 283

smaller dimensions of the epithelial cells of the intestine of
the mouse as compared with those of the rabbit.

4. That the appearance of large numbers of swarm-spores in
the gastric glands of the mouse is very similar to that pre-
sented by the Sarcosporidia, and suggests that the latter is
but one phase of a Sporozoon which may have in other phases
a form resembling that of the Coccidia. A similar conclusion
to this has been arrived at by L. Pfeiffer as the result of com-
paring Klossia with the Sarcosporidia.

October 3rd, 1894.

EXPLANATION OF PLATE 30,

Illustrating Mr. J. Jackson Clarke’s paper, “‘ A Study of
Coccidia met with in Mice.”

Fic. 1.—Ripe Coccidium, as seen with 4, 0. i. immediately after removal
from the body of the mouse.

Fig. 2.—Coccidium after six days in the moist chamber.

Fic. 3.—The same after the same length of time.

Fic. 4.—The same after eight days.

Fie. 5.—The same, 10th day.

Fic. 6.—The same, 12th day.

Fie. 7.—Section of a duct of a cardiac gland (mouse’s stomach).

Fie. §—Encapsuled parasite subdivided into sickle-shaped swarm-spores,
examined fresh from the mouse’s stomach.

Fic. 9.—Cross-section of duct of a pyloric gland.

Fic. 10.—Cross-section of a gland of the small intestine of a mouse, show-
ing two large naked intracellular parasites with close-set peripheral bars of
chromatin. The darker points on the surface of the parasite appear to mark
the commencement of subdivision.

Fic. 11.—Naked and free coccidium from a section of a rabbit’s liver,
showing the mode of subdivision of a naked parasite with peripheral chromatin
bars.

Fie. 12.—The same, showing another phase of subdivision.

fi Neit 4 a tlan

Al dametees MEA id | 2yi e FAt eee) on peti

TAL an Paden a

ay ie iAMiy

nh bn ) Loe :
> eonaye, ohhh gue i ?

cee dickies

i it x oF op 8Vty nee @

’ 4

vs .
ike Ch Qe
yor il a 4S etry we \' :

sifu aaa pde:! 44 Sr

tr Le ve) ' re
, ae

Faé { ‘ f
j | i
Vid } Aare h
“+
Py t ¢ " i
|
P
: i
i / j
j
a 1
FP
y +
fi z
f
. i
!
aL apy
‘ # *
- ih soa! POW Ban tty
f ps
i” i Japs h Moe Oe! ee 4

Se lle wo) yee r
Shaye vi a nee AJL 7

ees abana ve Seba Tee itn Ae ae

ivi ut) a ae Wr=4 ce ott aa — oe
any iieae © a

baie es nek my

; = ms
~~ Va 4

OBSERVATIONS ON VARIOUS SPOROZOA. 285

Observations on Various Sporozoa,
By

J. Jackson Clarke, M.B.Lond.

With Plates 31, 32, and 33.

Certain problems propounded in the last few years demand
for their solution a closer study of the intimate structure of the
sporozoa than has hitherto been found necessary. The varia-
tions of nuclear form presented by these organisms call for
especially close examination. For some of the higher members
of the group this work has been well begun by Wolters,!
who examined Clepsidrina blattarum, Monocystis
magna and agilis in Lumbricus agricola, and Klossia
in the kidney of snails. The description of the nuclear changes
in the common gregarines of the earthworm is particularly
complete, and since I have been able to confirm many of
Wolters’s observations it may be of interest if they are briefly
reviewed here. He found in Lumbricus agricola both
Monocystis magna and agilis constantly present, and
almost to the exclusion of other species. In both, at every
stage, a distinct nucleus was present. In the former it was
relatively large and oval, in the latter of a rounded form.
Under compression the nuclear membrane ruptured, but the
contents did not escape, and when the compression was removed
the nucleus tended to return to its primitive form. Thus the
nucleoplasm appeared to be of a solid structure. In M. agilis
this was also observed. The nuclear membrane was found to

1 Max Wolters, ‘Die Conjugation und Sporenbildung bei Gregarinen,”
‘ Archiv fiir mikroskop. Anat.,’ p. 99, 1891.

286 J. JACKSON CLARKE.

be strong and sharply defined, and in the young parasite there
was a Single large nucleolus. Later the outline of the nucleus
had become irregular and the nucleoli more numerous. In M.
agilis, and in one instance in M. magna, Wolters encountered
what he has named the flame-nucleus, a condition in which the
nuclear membrane has disappeared, and the nuclear substance
is prolonged at various points into the protoplasm of the body
of the protozoon. As soon as the nucleolus has broken up into
subdivisions the parasites are ripe for conjugation. It was
found that after the syzygium was formed the nucleus of each
parasite moved to the periphery, became elongated, and soon
exhibited a typical nuclear spindle with the chromatin now
massed together at the middle of the spindle. The chromo-
somes were very small. The division of each nucleus took
place, and half of each nucleus was extruded as a polar body.
Meanwhile the surfaces by which the parasites adhered to each
other became altered in such a way that instead of the
sharply marked line of division previously seen throughout a
complete series of sections of a syzygium, there was at one part
of the applied surfaces a communication through which the two
parasites fused together. In each parasite the polar body was
extruded on the surface opposite this communication, towards
which, after the polar bodies had been formed, the two nuclei
moved, and having reached the spot at the same time they
fused together. After this a nuclear spindle was to be seen in
each half of the syzygium, and could be distinguished, by its
position close to the area of communication, from the spindles
which were concerned with the formation of the polar bodies.
Thus it appeared that the conjoined nuclei had undergone
division, and that the daughter nuclei were again subdividing.
The resulting two nuclei moved towards the periphery in each
half of the syzygium, and there formed two spindles. This
Wolters found to be the case in complete series of sections.
These spindles were smaller than those of the polar bodies.
By repeated subdivision these peripheral spindles and resulting
nuclei increased in numbers and became surrounded by proto-
plasm constituting the sporogonia, which arranged themselves

OBSERVATIONS ON VARIOUS SPOROZOA. 287

peripherally, and also in spaces extending from the periphery
into the remains of the body substance of the parasite.

The single nucleus of the sporogonia subdivided into eight,
and meanwhile the sporogonia became surrounded by capsules
constituting sporocysts (pseudo-navicelle), the substance of
which split up into eight crescentic spores, each of which
contained a nucleus. Such are the chief results obtained by
Wolters with regard to the Monocystides, and the thorough-
ness of the work and the beautiful drawings by Nussbaum, at
whose instigation the work was undertaken, go far to establish
the conclusions arrived at.

I will now detail some of the features I have obtained by
examining the seminal vesicles of Lumbricus agricola,
taken in the month of May. I have not had the opportunity
of making such complete serial sections as Wolters, and in so
far my criticism must be incomplete; still, the observations
recorded below may be found of some interest. In this place
it is advisable to describe the methods employed. Wolters
found, and I have had the same experience, that Flemming’s
fluid did not give good results with gregarines. Wolters’s
results are chiefly based on the examination of sections of
material fixed in saturated solution of picric acid. I have
employed a method more commonly adopted, and which I have
found most satisfactory, not only for gregarines, but for
Coccidium oviforme and for animal tissues in general.
Small portions of the tissue are placed for twenty-four hours
in Foa’s reagent, i.e. a mixture of equal parts of a saturated
solution of corrosive sublimate in normal saline solution and
a 5 per cent. solution of bichromate of potassium or Miiller’s
fluid. Then the material is transferred for twenty-four hours
to running water, and afterwards placed on successive days in
30, 60, and 90 per cent. alcohol. After that they are placed
in absolute alcohol, and after saturation with chloroform are
embedded in paraffin, care being taken that the bath does not
reach a temperature higher than 50°C. ‘The sections were
cut with a Minot’s microtome, and fixed on the slide with
albumen and glycerine. After the usual process they were

288 J. JACKSON CLARKE.

stained with Ehrlich’s acid hematoxylin diluted with distilled
water, and when they had assumed a brownish pink colour were
transferred to a bath of tepid tap water and left for at least two
hours. Then for two or three minutes they were stained with
a solution! of Gribler’s water-soluble eosin, dehydrated,
cleared by xylol, and mounted in the usual way.

With regard to Wolters’s description of polar bodies, I can
only say that structures which are explicable only as of
this nature are of frequent occurrence in M. agilis. They
resemble flattened nuclei, and are placed beyond the surface
of the parasite and lie between it and the capsule. A nucleus
is usually to be seen within the parasite close to such bodies,
and frequently remains of a spindle can be made out passing
from the nucleus peripherally towards the polar body, as with
Wolters I am inclined to regard it.

I have not been able to give sufficient time to the investiga-
tion to place me in a position to criticise Wolters’s description
of the fusion of the nuclei after extrusion of the polar bodies.
I have been able to confirm Wolters’s view of the origin of
sporogonia in the main, but some modifications are, I think,
required of the process as described by Wolters, who would
appear to say that every nuclear division after the fusion and
re-separation of the original nuclei proceeds by regular mitosis.
Against this, such appearances as I have sketched in Pl. 31,
fig. 1, may be objected. The drawing represents a syzygium of
M. agilis. It was surrounded by a connective-tissue capsule.
About the middle of each half of the syzygium is a mass (a) which
I think can only be regarded as nuclear. These masses are
composed of fine granules, most of which are coloured purple?
by hematoxylin. Amongst these are coarser granules, which
give with the same reagent a deep blue reaction characteristic
of chromatin in both animal and vegetable cells. These chro-

1 This was obtained by dropping a few drops of a strong alcoholic solution
into a watch-glass filled with distilled water.

? This is Ehrlich’s metachromatic reaction. ‘‘ Metachromatisch d. h. in
Einer dem angewoéhnten Farbtone abweichenden Niiance farben,” Ehrlich,
‘Gesammelte Mittheilungen,’ 1891, p. 2.

OBSERVATIONS ON VARIOUS SPOROZOA. 289

matic granules are arranged in lines (0) which radiate out into
the substance of the parasite, in the body of which numerous
similar granules can be seen, and they are often joined together
by achromatic filaments. Besides these granules numerous
typical spindles (c) and nuclei are to be seen, placed for the
most part at the periphery of each part of the syzygium. I
regard the granules as a phase of mitosis, and the purple
granules of the nuclear bodies as chromatin in a modification
previous to that in which karyokinetic activity begins. Wolters
seems at one time to have held a similar opinion to mine in
reference to the radiating strings of granules, and to have relin-
quished it, I think, on insufficient grounds. ‘‘Um diese
Spindeln sah ich bei priiparaten welche durch Flemmingsche
ldsung abgetodte waren, viele stark farbende Kornchen in der
Substanz vertheilt, ebenso hier und da, auch weit ab von den
Spindeln, in den Syzygiten. Ich war geneigt dieselben als
chromatische Substanz anzusprechen. Spiatere Untersuchun-
gen an Hoden, die ich mit Umgehung dieser Lisung abtédtete
und hartete, zeigten nichts davon, sodass ich von meiner
ansicht zuriick gekommen bin, ohne eine befriedigende Erk-
larung dieser K6rnchen geben zu kénnen.” In a second
drawing, Pl. 31, fig. 2, of part of the periphery of another
couple of M. agilis I have represented some of these
granules (a) on a larger scale. In this case some of the deep
blue granules were surrounded by material which was coloured
by the eosin of the stain. On comparing these with the sporo-
gonia (¢c) lying at the periphery a close similarity of struc-
ture was observed, the body of the sporogonium staining in
the same tone with eosin as the material investing the chro-
matic granules. The nuclei of the sporogonia all stained
of the same deep blue as the granules, and they presented a
great variety of form. In some mitotic processes seemed to
have begun, and the same holds good for the sporogonia of
Mon. magna, as is shown at (d) in the lower of the two in
fig.3. Iam able to confirm Wolters’s description of the forma-
tion of sporocysts and spores. Whilst speaking of sporocysts I
may mention that I have encountered in the seminal vesicles

290 . J. JACKSON CLARKE.

of Lumbricus agricola prismatic pseudo-navicelle exactly
like those described by Bosanquet! in the body-cavity.

The observations made by Wolters on Clepsidrina blat-
tarum are, unfortunately, scanty. Much remains to be done
with regard to the earlier phases, both of the poly- and the
mono-cystides. More importance attaches to Wolters’s
description of certain phases of Klossia helicina. This
parasite is of great interest from the position it holds be-
tween the Gregarines and the Coccidia. From the fact that
conjugation has not been observed, I, like L. Pfeiffer,” should
be inclined to place it with the latter group. First described
by Kloss* of Frankfort, it has since been described by A.
Schneider and by L. Pfeiffer (loc. cit.). To the latter author
I am indebted for many beautiful preparations of the parasite
in Helix hortensis and Succinea Pfeifferi. Pfeiffer has
arrived at some interesting conclusions based on the study of
this parasite. Of these the more important are—lst, the
phenomenon of multiple * infection, as many as fifteen parasites
being found within a single epithelial cell; 2nd, that when
multiple infection occurs only one of the parasites reached
maturity; and 3rd, that the size attained by the parasite is
determined by the size of the epithelial cells of the kidney of
the species of snail infested by the parasite, though in all cases
the sporogonia have the same dimensions, the number of
sporogonia varying with the size of the parasite from which
they are derived.

The general features of Klossia I have never seen better
than in some common grey slugs which I examined in July,
1892. The slugs were found in a hollow in the rocks below
the falls of the river Shin, in Sutherland. With them were

1 W. C. Bosanquet, ‘Quart. Journ. Micr. Sci.,’ 1894, No. 148, p. 421,
fig. 19.
* L. Pfeiffer, ‘Protozoen als Krankheitserriger,’ 1891, p. 72.
3 Hermann Kloss, ‘ Senkenbergische Abhandlungen,’ vol. i, 1855-6.
4 LL. Pfeiffer compares this with what occurs in the Sarcosporidia, the micro-
sporidia in Coccidium salamandre, and in the Coccidia of the kidneys of
he goose and the dog.

OBSERVATIONS ON VARIOUS SPOROZOA. 291

numerous examples of Helix hortensis. The kidneys of the
snails and the slugs were all alike infested by the sporozoa,
but only in the slugs did I find swarm-sporing side by side
with the ordinary mode of reproduction by sporocysts. One
of these parasites, as seen in a teased preparation of a slug’s
kidney, is shown in fig. 4. The nucleus (@) was large and oval,
and showed a single large nucleolus (4). In fig. 5 is repre-
sented a cell (a) as seen in a section, and containing a parasite
subdivided into four sporocysts (c) (others were present out of
focus), each of which contains six crescentic spores (d). In
some of these a single nucleus could be made out. The spores
were very large, averaging 12, in length. All the sections of
the kidneys of several slugs showed a marked infection, and
in most of the sections swarm-sporing was well seen. Fig. 6
shows an example of this. A much hypertrophied cell (a)
contains a large sporing parasite, and six smaller parasites (e).
The sickles (4) are very large, 20uin length. Some of them
are undergoing farther subdivision.! The capsule of the para-
site has ruptured, and at the point of rupture are some free
sickles, which also are undergoing farther subdivision (c). The
appearance of swarm-sporing in a fresh teasing is shown in
fig. 7, where some detached sickles (4) are present. The colour
reactions in these sections were not good, probably owing to
the slugs having in the first instance been placed in Scotch
whiskey, so that they will not serve as a basis for comparison
with Wolters’s descriptious ; but since I have been able to find
in Coccidium oviforme all this author encountered in
Klossia, and also many additional features, I will pass to the
consideration of this more familiar parasite. For examination
I chose a highly infected liver in which the lesions were still
in process of evolution. Fig. 8 shows an average appearance
of a portion of the epithelial lining of a cyst as big as a small
pea. All the larger parasites show signs of nuclear activity.
The chromatin, in nearly every instance, gave a typical deep
blue reaction to acid hematoxylin. The most abundant form

1 The large sickles formed in swarm-sporing would thus appear to have the
equivalents of sporogonia, not of spores.

292 J. JACKSON CLARKE.

in which Coccidium oviforme usually presents itself in
sections is the spherical granular body represented in fig. 9 to
the left. The granules (4) stain slightly with eosin, but remain
transparent. They have, no doubt, the same signification as
the Gregarina corpuscles! in the higher Sporozoa, i.e. they
serve as stored food. The ‘‘ dauerform” (L. Pfeiffer) of the
parasite is represented in the same drawing to the right. In
it the granules have disappeared, and the nuclear body (a),
which is round in the spherical-granular phase (a’), is oval
in the encapsuled parasite. In neither case, however, does
the nucleus give the reaction of chromatin, but is stained by
the eosin. One other phase of the parasite may be mentioned
in passing. This is a small, dense, spherical body, devoid
alike of granules and of nuclear body, and staining throughout
without eosin. Sometimes a parasite possessing a thick oval
capsule is found to have broken up into sickle-shaped swarm-
spores, but the rule is that when the parasite multiplies whilst
still within the body of the host, it has either no capsule at
all or only the delicate so-called primordial capsule. It is to
the changes which lead to subdivision of the parasite within
the host that I wish to direct attention. R. Pfeiffer? first de-
scribed swarm-sporing, but chiefly in fresh specimens, and
without any detailed account of nuclear processes. LL. Pfeiffer
(loc. cit., p. 45) described distinct hematoxylin-stained nuclei
in the process of swarm-sporing, and says, ‘‘ Kingehendere
Untersuchungen sind hier noch sehr nothig, dar der mehr
oder weniger akute Krankheitsverlauf von Massgebenden
einfluss ist auf die Vermehrungsweise des zugehdrigen Para-
siten.”

Among the parasites in actively extending lesions of the
rabbit’s liver some present a distinct ‘‘ geflammte Kern,” like
that shown in fig. 10 (ce). Such nuclei take only the eosin of the
stain, or at most a slight tinge of purple at their edge. Most

1 The chemical nature of these bodies is not yet determined. They dissolve
n alkalies and mineral acids, and are not fat, nor do they contain lime. See
Max Wolters, loc. cit.

2 R, Pfeiffer, ‘ Beitrage zur Protozoenforschung,’ Berlin, 1891.

OBSERVATIONS ON VARIOUS SPOROZOA. 293

of the nuclei which are preparing for subdivision show distinct
radiating processes, which stain deep blue with acid hematoxy-
lin. Sucha nucleus is shown in fig. 11 (a). The food-granules
in such parasites have become diminished in numbers, and in
some cases stain more deeply with eosin. The next stage is
shown in fig. 12, where the nucleus has subdivided into two
parts with the formation of a typical spindle. Wolters was
unable to observe spindle-formation in Klossia, but from the
identity of some of the phases of nuclear structure in C.
oviforme, and in Klossia as described by him, it is probable
that in both as also in Gregarines, spindle-formation takes
place. In this way two or three (fig. 1, a and a’) nuclei are
formed. Fig. 13 shows the subdivision of a separated portion
of the nucleus. <A typical spindle (a) is present, and the chromo-
somes are like those of the Gregarines, extremely small. In
many instances the subdivision of the nucleus appears to take
place more rapidly, particles of nuclear matter being detached
to the periphery along single achromatic filaments. The
appearance results in a structure almost identical with that
figured by Wolters in Klossia (loc. cit., pl. vii, figs. 10 and
11). When the peripheral nuclei are large the resulting
subdivisions of the parasite may be termed sporogonia, for
they undergo farther subdivision in the formation of sickles
(see fig. 8, 5). When the peripheral nuclei are small, sickles
are formed immediately, as a comparison of figs. 14 and 15
suggests. Fig. 14 shows within a capsule (c) the optical sec-
tion of a parasite with a central nuclear mass (@) and small
peripheral nuclei (0). Fig. 15 shows a collection of sickles (4)
within a cell (a). Sometimes the peripheral arrangement of
chromatin takes the form of a zone of fine granules, as is shown
in fig. 16,a@. The peculiar behaviour of chromatin in Sporozoa
may be illustrated by the fact that sometimes capsules are met
with containing sickles stained onJy with eosin, though in per-
fectly similar capsules close to them the sickles have single
nuclei well stained with hematoxylin. With regard to the
phase of Coccidium oviforme, marked by the presence of
peripheral bars of chromatin which I mentioned in a previous
VOL. 37, PART 3.—NEW SER. U

294 J. JACKSON CLARKE.

paper, I may add that it arises as a modification of the peri-
pheral distribution of chromatin already described, but here the
particles are rod-shaped, and at certain stages are connected
with a central nuclear mass by a beautiful arrangement of
achromatic lines, as shown in optical section in fig. 17, where
within the host-cell (a) the central nuclear mass (8) is con-
nected by achromatic filaments with nuclear rods (c) at the
periphery. When once this system of peripheral rods is con-
stituted, the subdivision of the parasite by in-dippings of the
surface would appear to be the rule. Fig. 18 shows in a sur-
face view an example of sucha parasite in which subdivision has
proceeded to some extent. Some of the parasites with peri-
pheral rods are of considerable size, and show evidence of
having changed their form by active movement. The arrange-
ment of the deep blue rods is then often most complex ; this is
shown in fig. 19, where the large parasite (a) touches the base-
ment membrane on one side of a dip between two papille, and
on the other the nucleus of a cell, so suggesting a previous
conjugation of two parasites. It also has two dark bodies on
its surface near its base. These bodies have an appearance
quite similar to the polar bodies referred to in gregarines.
The general arrangement of chromatin in the swarm-sporing
process in Coccid. oviforme thus agrees with Wolters’s de-
scription of Klossia, and also with what L. Pfeiffer (loc. cit.,
p. 31) observed in coccidia: “Rasch folgen auf diese
anflockerung des primaren Kernes eine Reihe von Kerntheil-
ungen wie sie die Bilder in fig. 12 wiedergeben. Die
zahlreiche Tochterkerne liegen an dem Mantel der Parasiten
Kugel dicht an und oft in schéner geometrischer Anordnung.”
As far as I am aware, the description of nuclear spindles in
coccidia is here given for the first time. What may be the
more minute features of the nuclear figures, i.e. with regard to
the attraction-spheres, remains yet to be shown. The different
behaviour of the parasites with peripheral rods of protoplasm
as compared with such as the one shown, fig. 14, may perhaps
prove to throw light on direct nuclear division described by
Arnold, as compared with the more commonly observed

OBSERVATIONS ON VARIOUS SPOROZOA. 295

indirect division of cells. During the subdivision of parasites
such as the one shown in fig. 18, when acid hematoxylin is
used alone, bright crimson particles of chromatin aresometimes
visible in the interior of the segments.

The process of nuclear division is not limited to the larger
parasites. Some of the smaller (6—8 ) intra-cellular coccidia
present nuclei which are dividing indirectly, see Pl. 32, fig.
21,a. The chromatin is, as a rule, not arranged into chro-
mosomes which can be separately distinguished, but is arranged
in what under a magnification of 1000 diameters appear to be
masses of material which stain purple or blue with hematoxylin.
Such small parasites have no large food-granules, and in propor-
tion as the nuclei multiply the cytoplasm and the chromatin
increase in amount, giving rise to bodies such as those shown
a, P32; figs 21,6, and in’ Pl-31; fig: 195, 6:

Achromatic filaments are distinctly recognisable in many of
the parasites in this modification. The usual termination of
the process is the formation of sporogonia which contain
several minute nuclear masses embedded in eosin-staining
protoplasm. The sporogonia subdivide into sickles of the
ordinary form. Not infrequently, however, sickles are formed
directly, i. e. without a sporogonium stage. Again, the growth
of this non-granular form of coccidia may pass beyond the
average size (fig. 19, a), and the chromatin becomes minutely
subdivided and arranged at the periphery, thus producing the
phase marked by peripheral rods, as shown in Pl. 82, fig. 21, d.
Returning to the larger granular coccidia, I would here ex-
plain that regular mitotic figures, such as Pl. 31, fig. 12, are
extremely few in number. Not so infrequent are typical but
slightly irregular spindles, in which all the chromatin of the
original nucleus is concerned. Examples are given in Pl. 382,
fig. 20, a, 6, and z. The whole of the chromatin may become
active without the formation of a spindle, as shown in figs. c,
k, andl. In the parasite represented in fig. & the peripheral
collection of chromatin (2) gave a metachromatic (crimson)
reaction, whilst the other (1) extremity stained dark blue. In
the greater number of the parasites the main mass of the nucleus

296 J. JACKSON CLARKE.

remains unchanged for a time, the first indication of karyo-
kinetic activity being the separation of hzmatoxylin-stained ~
chromatin particles joined by achromatic filaments (figs. d, e,
and h). When this is the case the unchanged part of the
nucleus stains but faintly with hematoxylin alone, and when
eosin after staining is used assumes a red colour. Sometimes
the whole of the nucleus stains deep blue with hematoxylin
at first without any appearance of nuclear filaments. In such
cases the nucleus has the appearance shown in fig. f.
Irregular mitoses are abundant. Such are shown in figs.
f and g. In such mitoses the first portions separated from
the main nucleus frequently give the crimson reaction to
hematoxylin. Nothing is more striking than the growth of
chromatin in coccidia. In the final multinucleate condition
the coccidia possess more than a hundred times the amount of
chromatin they had at the commencement of the process.

Once more reverting to the parasites with peripheral chro-
matin rods, it should be observed that in a few cases, which can
only be found by patient search, the rods are replaced by fine
granules of chromatin, as shown in fig. m. Others of the
parasites with peripheral rods appear to break up into an
immense number of extremely minute sickles without previous
subdivision into segments. Finally, with regard to the seg-
mentation of the parasites in the phase marked by peripheral
rods of chromatin, it is sometimes seen that the subdivisions
possess sickle-shaped outlines like the one seen in optical
section in fig. ». It is to be observed, however, that such
crescentic segments are not homologous with single sickles,
but that each peripheral chromatin rod is the potential nucleus
of a spore.

I would now turn to mention, and it can be but too briefly,
L. Pfeiffer’s recent work on the Myxo-, Sarco-, and Micro-
sporidia.t In these groups Pfeiffer has added much to bio-

1 See L. Pfeiffer, ‘Protozoen als Krankheitserriger,’ 1891, and ‘ Unter-
suchungen tiber der Krebs,’ 1891.

For a general account of the Sporozoa and the groups included, together

with figures of most of the important forms, see Lankester’s ‘ Zoological
Articles’ (A. and C, Black), article ‘ Protozoa.”

OBSERVATIONS ON VARIOUS SPOROZOA. 297

logical knowledge, and has closely studied the effects of the
parasites on their hosts,—or, in other words, their pathological
effects. Since Dr. Pfeiffer has most kindly given me many
beautiful preparations of Sporozoa belonging to this and other
groups, I have been able to study, and, I may add, to confirm
his results.

With regard to the Myxosporidia, the description of the
epidemics they have caused from time to time in some of the
rivers of Germany forms a most interesting part of Dr.
Pfeiffer’s work, and is worthy of the close attention of
pisciculturists. Barbel, pike, and perch were chiefly attacked.
“The sick barbel present a striking appearance from the
presence of discoloured tumours in the skin, and of deep
crateriform ulcers on the head, the hinder part of the body,
and the tail: the ulcers have a widely infiltrated base.” Fig.
22 shows part of one of Dr. L. Pfeiffer’s sections of a myxo-
sporidial tumour of a barbel. The growth is alveolar in
structure, the alveolar walls being composed of fibrous tissue.
The contents of the alveoli consist solely of parasites. In
the upper part of the figure is a portion of the periphery of a
reticulated parasite from which sporocysts have separated ; at
the edge of the parasite is anuclear spindle. The thread-cells
of the sporocysts are stained deeply. Some of the sporocysts
are devoid of definite characters; whether they are young
forms, or residua after the escape of the single ameeboid spores,
is a question I have not been able to determine. LL. Pfeiffer
shows that these tumours begin as an infection of striped
muscle-fibres, and some of the preparations demonstrate this
point most distinctly, the young parasites in all stages of
existence being visible within muscle-fibres at the periphery
of the tumour.

Writing of the effects of Sarcosporidia, Leuckart has said,
« Although the tubes occasionally occur in immense numbers
close to one another, so that the flesh looks as if half of it
consisted of psorosperm tubes, yet they seem usually to cause
no special uneasiness. In many cases, however, the phenomena
of paraplegia, retarded respiration, and even suffocation, are

298 J. JACKSON CLARKE.

observed as associated with the presence of the tubes, and may
with some probability be referred to this cause.’’!

L. Pfeiffer has demonstrated the nature of these fatal cases,
showing that they are the result of the escape of the sickles
from the muscle-fibre in which they are developed, and their
entrance into the surrounding muscle-fibres, in which they
appear as groups of minute round cells, and growing to a
certain size constitute a new Miescher’s tube, which, unless the
host acquires a greater resisting power, again ruptures and
liberates its swarm-spores.

I have been able, both in Dr. Pfeiffer’s preparations and in
others of my own, to confirm these observations. Pfeiffer has
further shown that an emulsion of Sarcosporidia injected into
rabbits causes an intense inflammatory reaction and toxic
phenomena, and this is in keeping with what is seen in the
progressive infection of muscle referred to above. The free
spores seem to have a marked attraction for leucocytes (che-
miotaxis), and thus the process has been termed by Pfeiffer
* myositis sarcosporidica.”

When this progressive infection is limited to a certain region,
instead of inflammatory changes tumour formation is seen.
Thus in horses L. Pfeiffer has found that tumours included
by Kolliker in a group termed ‘‘ Muskelknospen” are
determined by sarcosporidia. Some of Dr. Pfeiffer’s
sections show this in, I think, a most conclusive manner.
The growth possesses a stroma of fibro-cellular tissue, in
which three zones may be described: an inner containing
dense foci of degenerated material surrounded by giant-cells,
&c.; a middle zone containing numerous Meischer’s tubes,
from which the sickle-spores are escaping ; and an outer zone
containing many muscle-fibres infected by Sarcosporidia.

One effect of a chronic progressive infection by Sarcosporidia
has been most exquisitely shown by the same observer. On
the oesophagus of sheep, white cysts, of which the largest are
as big as horse-beans, are sometimes encountered. The larger
(older) cysts are provided with a fibrous capsule. Dr. Pfeiffer’s

1 Leuckart, Hoyle’s (1886) translation, p. 202,

OBSERVATIONS ON VARIOUS SPOROZOA. 299

sections show that the capsule is formed by a reactive in-
flammatory process, and that between its fibres lie compressed
muscle-fibres (fig. 23, a). Those muscle-fibres which are
embedded in the deeper, i.e. older part of the capsule, are
distended by young sarcosporidia, which present a definite
nucleus and a clear protoplasm (fig. 28, 4). Next to the
capsule come large spaces (fig. 23, c), filled with sickles. The
spaces result from the distension of muscle-fibres and the
consequent separation of the fibrous fasciculi of the capsule.
It may be noted that some of the sickles in the spaces close to
the capsule are subdivided into segments. I may add that
whilst the sickles contain particles of chromatin throughout
the whole of their substance, the youngest parasites have a
definite nucleus and a clear protoplasm. IL. Pfeiffer has thus
shown that progressive inflammatory changes, cyst- and
tumour-formation may be determined by Sarcosporidia.

Finally Dr. Pfeiffer’s sections of the muscles of frogs show
that like the Myxo- and Sarco-sporidia, the Microsporidia occur
within muscle-fibres. Fig. 22 shows the remains of a muscle-
fibre containing some of these parasites. One group of spores
are not yet liberated from the parent cell, and present a
central nuclear spot and an unstained peripheral region. The
free spores are minute (2 m) bodies, most of them slightly
curved, so that they present the strongest possible resemblance
to a collection of vibrios.

I cannot close this article without expressing my warm
thanks to Dr. L. Pfeiffer for the liberal manner in which he has
answered my request for specimens and material.

Lonpon; December 27th, 1894.

300 J. JACKSON CLARKE.

EXPLANATION OF PLATES 31—38,

Illustrating Mr. J. Jackson Clarke’s “ Observations on Various
Sporozoa.””

Fic. 1.—Section through a syzygium of Monocystis agilis, x 500
diams. From the nuclear masses, a, in the central part of each half of the syzy-
gium numerous moniliform threads, 4, of chromatin are directed towards the
periphery, where are many small nuclei and nuclear spindles, ¢.

Fic. 2.—Section through part of the periphery of a syzygium of Mono-
cystis agilis, x 1000. From left to right are (1) the connective-tissue
capsule, a, (2) sporogonia, ¢, (3) part of the body of the parasite with gregarina
corpuscles, nuclear spindles, 6, and particles of chromatin, a, round some of
which a covering of protoplasm has collected.

Fie. 3.—Two sporogonia of Monocystis magna, x 1000 diams. The
nucleus of the lower of the two shows achromatic filaments, 4, joining the par-
ticles of chromatin. The body, a, of the structure stained with eosin.

Fic. 4.—A free example of Klossia from the grey slug, x 1000 diams,
Fresh specimen. Nucleus, a, nucleolus, 4. x 1000.

Fic. 5.—An epithelial cell in the kidney of a grey slug containing a parasite
subdivided into sporocysts. Some of the sickles contain a single nucleus:
There are no residual bodies. x 1000 diams. From a section.

Fic. 6.—A cell, a, from the kidney of a slug containing a parasite in the pro-
cess of swarm-sporing. There are six smaller parasites, e, in the protoplasm
of the cell. 4. Large sickle-shaped sporogonia. c. The same dividing. d.
Undivided central part. x 1000 diams.

Fic. 7.—A free Klossia showing swarm-sporing with some sickles, 4, de-
tached. Fresh teasing. x 750 diams.

Fie. 8.—Part of the epithelial lining of a coccidial cystadenoma of a rabbit’s
liver. a, a’. Parasites with two and three nuclei respectively. 4. A parasite
subdivided into sporogonia. x 1000 diams.

Fie. 9.—A round granular and an oval encapsuled coccidium. a, a’.
Nuclear body (red). 4. Cytoplasm. c. Definitive capsule. 1000 diams.

Fic. 10.—An intra-cellular coccidium with a ‘Geflammte-Kern” at a.
x 1000 diams.

Fie. 11.—<An intra-cellular coccidium, the nucleus, a, of which shows radi-
ating bars of chromatin. x 1000 diams.

OBSERVATIONS ON VARIOUS SPOROZOA. 301

Fic. 12.—Intra-cellular coccidium, the nucleus, ec, of which is dividing by
regular mitosis. 4. Cell-protoplasm with food-granules. a. Host-cell. x
1000 diams.

Fic. 13.—A coccidium which shows, besides the main nucleus, a nuclear
spindle. x 1000 diams.

Fie. 14.—An encapsuled coccidium (capsule, ¢c) with peripheral particles of
chromatin, 4, which are connected with a central nuclear mass, a, by achro-
matic filaments. Xx 1000 diams.

Fig. 15.—An epithelial cell, a, of the rabbit’s liver filled with nucleated
sickle-spores, 4. 1000 diams.

Fic. 16.—An intra-cellular coccidium with numerous minute chromatin
granules at a, the periphery.

Fie. 17.—An intra-cellular (host-cell, 2) coccidium showing close-set peri-
pheral rods of chromatin (¢) connected with a central nuclear mass (2) by
achromatic filaments. x 1000 diams.

Fie. 18.—A coccidium with peripheral chromatin rods seen in a surface
view, and showing the mode of subdivision. x 1000 diams.

Fic. 19.—Part of the epithelial lining of the rabbit’s cystadenoma showing
large free parasite with peripheral rods and granules of chromatin, and at
za multinucleated parasite. Xx 1000 diams. The free parasite is seen in
surface view.

Fie, 20.—a. Granular coccidium with a slightly irregular nuclear spindle.

6. A parasite with a nucleus similar to that shown in Fig. a, but with
granules of chromatin at the left half of the spindle.

ec. Granular coccidium, the nucleus of which has become altered so that it
presents a collection of granules of chromatin.

d. Granular coccidium, of which the outer part of the nucleus is separated
in an irregular spindle (1), whilst the central part remains unchanged (2).

e. A similar condition to that shown in Fig. d.

f. and g. Irregular mitoses.

h. A condition similar to that shown in Figs. d and e. The central
unaltered part (2) of the nucleus stained with eosin; the peripheral part (1)
(irregular spindle) stained with hematoxylin.

z. Irregular spindle.

Jj. Elongated nucleus stained densely with hematoxylin.

&. Irregular mitosis. Left extremity of the chromatin band stained blue,
right extremity crimson. Acid hematoxylin alone.

i. Irregular mitosis, granules stained purple.

m. A large parasite in a phase equivalent to that with peripheral arrange-
ment of rods, the latter replaced by granules of chromatin.

a. Large parasite with peripheral rods undergoing segmentation, one of
the segments crescent-shaped. Optical section.

302 J. JACKSON CLARKE.

21.—A portion of one of the papillary processes of a coccidial tumour seen
in section. a, Young parasites with division of nuclei.. 4. Larger but still
not granular coccidia with dividing nuclei. c. Capillary vessels and basement
membrane.

Fic. 22.—Section of a myxosporidial muscle-tumour of a barbel. From a
preparation made by Dr. L. Pfeiffer. x 1000 diams.

Fie. 23.—Section of a sarcosporidial cyst from a sheep’s esophagus. From
a preparation made by Dr. L. Pfeiffer. » 1000 diams.

Fic. 24.—Microsporidia in frog’s muscle. From a preparation made by
Dr. L. Pfeiffer. x 1000 diams.

A REVISION OF THE BRANCHIOSTOMID. 303

A Revision of the Genera and Species of the
Branchiostomide.

By

J. W. Kirkaldy.
(From the Laboratory of the Linacre Professor, Oxford.)

With Plates 34 and 35.

THE opportunity of examining a number of species allied to
the well-known Cephalochordate Amphioxus lanceolatus
was afforded to me by Professor Lankester, who had for some
time collected material for the purpose of revising the genus,
and suggested that I should go over the specimens in his pos-
session, and prepare a series of drawings showing the chief
characteristics of the species. The drawings have been executed
by Mr. Bayzand from rough ones prepared by me.!

The material examined by me is as follows:

1. Various collections from the Zoological Station, Naples
(A. lanceolatus).

2. Hight specimens from Ostend sent to Professor Lankester
by Professor van Beneden of Liége (A. lanceolatus).

3. A collection of about one hundred and fifty specimens
from the South Australian coast, presented to the Oxford
Museum by Professor Baldwin Spencer of Melbourne, Victoria
(H. Bassanum).

4, A collection of about ninety specimens from Torres
Straits, lent to Professor Lankester by Professor Haddon of
Dublin, by whom they were collected (H. cultellum).

1 The present memoir must be taken as superseding a note published by
me in the ‘ Reports of the British Association,’ 1894, Oxford meeting.

3804 J. W. KIRKALDY.

5. Four specimens from Ceylon collected by Mr. Haly,
and lent to Professor Lankester by Professor Haddon (H.
cingalense).

6. Eight specimens from the coast of Brazil, presented to
Professor Lankester by Professor van Beneden of Liége
(A. caribzus), by whom they were collected.

7. Twelve specimens from the Bahamas presented to Pro-
fessor Lankester by Professor Alexander Agassiz, by whom
they were collected (A. Lucayanum).

8. Five specimens from Ceylon, forwarded to Professor
Lankester by Professor Good Brown, of the National Museum,
Washington, U.S.A. (H. cingalense).

9. Ten specimens from the coast of California forwarded to
Professor Lankester by Professor Good Brown. Cooper’s type-
specimen of A. californiensis was among these.

10. In addition to the above-mentioned material which I
have been able to study, and in many cases to examine, by
means of transverse sections in Professor Lankester’s labora-
tory at Oxford, I have been allowed by Dr. Giinther to examine
the specimens of another species (A. Belcheri), of which
only five exist, in the British Museum.

I am desired by Professor Lankester to express his great
obligations to Professors Spencer, Haddon, van Beneden,
Agassiz, and Good Brown for their kindness in communicating
their specimens to him.

For the nomenclature of the various structures of the
Branchiostomide already in use the reader is referred to Pro-
fessor Lankester’s memoir (No. 7), and to the recently pub-
lished work by Dr. Arthur Willey (No. 15).

The characters to which I have, at Professor Lankester’s
request, given attention, and have found to be important for
the purpose of dividing the Branchiostomide into genera, sub-
genera, and species, are as follows. I enumerate the characters
approximately in the order of importance which it has been
found necessary to assign to them.

1. Przoral cirrhi uniformly connected to one another by a
low web, or per contra separated into two ventrad and two

A REVISION OF THE BRANCHIOSTOMIDA. 305

laterad groups, the ventrad groups distinguished by a very high
web.

2. A caudal expansion of the median dorsal and ventral fin-
ridge present or absent.

3. An unsegmented “urostyloid ” caudal region, one seventh
the length of the whole animal, present, or per contra
only a minute projection of the notochord beyond the last
myotome.

4, A single, or per contra a double post-atrioporal cecum
to the atrial chamber.

5. Only the right, or per contra both metapleura continuous
with the ventral portion of the snout or rostral fin.

6. The preoral gustatory groove (groove of Hatschek) super-
ficial and shallow, or per contra strongly marked.

7. Fin-ray spaces present or per contra absent in the median
ventral fin.

8. Fin-rays developed in the fin-ray spaces of the median
ventral fin, or per contra absent.

9. Right metapleur terminating behind the atriopore simi-
larly to the left metapleur; median ventral fin unconnected
with metapleura; or per contra right metapleur continued
without break into the median ventral fin.
thus, its posterior continuation.

10. Number of tentacles to the oral sphincter (intra-buccal
cirrhi) twelve, ten, or sixteen.

11. Gonads developed on both sides of the body, or per
contra only on one side.

12. Number of myotomes throughout the body ; number of
myotomes between snout and atriopore, atriopore and anus,

[N.B.—In counting the myotomes I have reckoned that
coincident with the atriopore as belonging to the anterior or
preatrioporal group ; and similarly I have reckoned the post-
anal group as commencing behind the posterior margin of the
anus. In many specimens the anus is elongated antero-
posteriorly, so as to extend under as many as three separate
myotomes. |

13. Proportional height of the median fin.

306 J. W. KIRKALDY.

14. Form of the anterior (rostral) and posterior (caudal)
expansions of the median fin.
15. Number and myotomic position of the gonad pouches.

The valuable diagnostic character afforded by the total
number of the myotomes, and of the number in the three
groups known as preatrioporal, preanal, and post-anal, was
first made use of by Sundevall and later adopted by Giinther
(No.5), whose essay is up to the present time the most important
on the taxonomy of the Branchiostomide. During the past
year, however, a very important memoir has appeared by
Andrews (No. 1) on a Branchiostomid from the neighbourhood
of the Bahamas, which presented so many novel features as
compared with the forms already known that it was necessary to
place it in a new genus, to which Andrews gave the unfortunate
name Asymmetron. Before the appearance of Andrews’
memoir Dr. Arthur Willey had, at Professor Lankester’s
suggestion, undertaken the examination of Professor Haddon’s
collection of Branchiostomide from Torres Straits, and the
result was a short memoir (No. 14) in which some important
characters of B. cultellum, Peters (of which the entire col-
lection consisted), are pointed out; and the significant sugges-
tion is made that the median ventral fin of the Branchiosto-
mide is not truly a median structure, but is the continuation
of the right metapleuron. The two papers of Willey and of
Andrews made it desirable to re-examine as far as possible all
the species of Branchiostomide, in order to ascertain whether
the peculiar features as to the unilateral character of the
gonads, continuity of metapleur and ventral median fin, absence
of ventral fin-rays in the one case and of ventral fin-ray
chambers in the other, described by these authors, obtain in
other forms which had not been looked at by their original
describers with these questions in mind. The nature of my
work is thus explained. Among the characters made use of,
I have added to those introduced into consideration successively
by Sundevall, Giinther, Willey, and Andrews, only one, viz.
the number of the cirrhi or tentacles lying on the inner face of

A REVISION OF THE BRANCHIOSTOMIDA. 8307

the oral sphincter (intra-buccal cirrhi), and this I have not
determined in every species. Possibly other more skilful
observers may be able to render some of the structures exa-
mined by me available for specific and generic characterisation,
but my own attempts to make use of the position of the atrio-
celomic funnels, of the skeleton of the preoral cirrhi, of the
form of the branchial bars (in section), and of the disposition
of Miiller’s renal papille, were unsuccessful.

The drawings given of B. Belcheri, Bassanum, cali-
forniense, and caribeum are the first which have been
published of those species.

The Branchiostomide are the sole family of known forms
comprised in the branch Cephalochorda of the phylum Ver-
tebrata. Retaining the name Vertebrata for the great phylum
to which some recent writers have proposed to apply the name
Chordata, Professor lLankester recognises three distinct
branches or lines of descent within that phylum, for which in
1877 (§ Quart. Journ. Mier. Sci.,’ vol. xxvii, p. 450) he proposed
the names Urochorda (the Tunicates), Cephalochorda, and
Craniata. Tothese three diverging branches Bateson has pro-
posed to add a fourth, the Hemichorda, to comprise the forms
known as Balanoglossus. We may begin the systematic cha-
racterisation of the Branchiostomide by a definition of the
Cephalochorda.

Branch CEPHALOCHORDA, Lankester, 1877.

Lertocarpil, Miller. ‘Abhandl. k. Akad. Urss.,’ Berlin, 1844,
p. 204,

Myetozoa, J. Geoff. St.-Hilaire. 1852.

Acranta, Haeckel. ‘Gen. Morphol.,’ 1866.

CIRRHOSTOMI, ‘Lectures on Compar. Anatomy,’

= En or { 1846.

Entomocrania, Huxley. ‘Proc. Zool. Soc.,’ London, 1876, p. 58.

CerHaLocHorDA, Hatchett Jackson. ‘Forms of Animal Life,’

1888, p. 437.

Vertebrata exhibiting the distinctive vertebrate combination

308 J. W. KIRKALDY.

of the four characters indicated by the terms notochord, pha-
ryngeal gill-slits, tubular myelon, and myelonic eye.

The notochord is large and unconstricted, is continued
through the region of the head, and projects beyond the first
myotome into the snout or rostrum. The longitudinal muscu-
lature of the body-wall is divided by membranous septa into a
series of well-marked myotomes. There is no marked enlarge-
ment of the anterior region of the myelon as brain, but a slight
dilatation of its cavity anteriorly. The myelon terminates
anteriorly short of the termination of the notochord ; there is
no special protective skeleton (cranium) for the anterior por-
tion of the myelon. The vascular system is devoid of a spe-
cialised “heart.” Minute tubular nephridia, arranged serially,
are present. Existing forms exhibit a deep-seated asymmetry,
masked by a secondary superficial symmetry ; they are of small
size, and probably are degenerate representatives of symmetrical
bilateral ancestors of more fully elaborated organisation.

Family (unica) BrRaNncHIOsTOMID#.

AmPuioxip#, Gray. ‘Synopsis Brit. Mus.,’ 1842, p. 150.

AmpHIoxINI, Miller. ‘Abhandl. k. Akad.,’ Berlin, 1844.

BRANCHIOSTOMID#, Bonaparte. ‘Cat. metodico Pesci Kurop.,’ 1846.

AmpuioxorpeE!, Bleeker. ‘Enum. sp. Piscium Archip. Ind.,’ 1859.

Crrrostomi, Giinther. ‘Cat. Fishes Brit. Mus.,’ vol. viii, p. 513,
1870.

Cephalochorda of minute size, and laterally compressed
elongate form—acutely terminated, both anteriorly and pos-
teriorly. The integument is raised dorsally into a median
plate or ridge forming the dorsal fin, which is uninterrupted
throughout the length of the entire body. Two similar plates
or outstanding ridges—the metapleura—exist on either side of
the anterior two thirds of the body, which is bluntly triangular
in section: the superior angle supports the median fin; the
infero-lateral angles support the two lateral plates or meta-
pleura. A median ventral fin is present in the hinder third
of the body; in some species it is continuous with the right
metapleur, which is also continued in front of the mouth to

A REVISION OF THE BRANCHIOSTOMIDA. 309

join the ventral portion of the rostral fin. The left metapleur is
not continued forward into the rostrum, nor more than a short
distance behind the atriopore. The gill-slits are numerous
and developed on both sides of the pharynx: they do not in
the adult present any numerical relation to the metamerism
of the myotomes ; they open into what is in the embryo a
median ventral groove, which subsequently becomes closed in
as a canal extending over the anterior two thirds of the ventral
wall, and opening by a pore (the atriopore) posteriorly. By
later enlargement this canal expands into a considerable cavity
(the atrium), separating a lateral region of the body on either
side (the epipleura) from the axial region; by similar growth
it is extended posteriorly within the body as the single or double
post-atrioporal czecum or ceca of the atrium. The gonads are
developed in either or both of the epipleura. A preoral
muscular hood is developed in front of the mouth, provided
along its circular margin with numerous tentacles, supported
by a cartilaginoid skeleton; there is one median unpaired
tentacle in the median ventral line, and from ten to twenty
(according to age and species) on either side.

A single circular area of pigment—the eye-spot—is developed
on the inner surface of the median anterior termination of the
hollow myelon. A single laterally placed olfactory pit may be
present on the left side of the snout, consisting simply of a
tubular depression of the epidermis by which it is brought into
continuity with a slight upgrowth of the terminal wall of the
myelon. The auus is placed on the left side of the ventral fin.

A forwardly projecting cecum (the hepatic cecum) is
developed on the right side from the alimentary canal at the
point where the pharyngeal perforations cease.

The oral sphincter (buccal apparatus) is provided with a series
of long sensory intra-buccal cirrhi, which project backwards
into the pharynx.

The gonads are developed in a numerous series of ccelomic
pouches, corresponding in number and position to the myo-
tomes of the mid-region of the body. There are no genital
ducts.

VOL, 37, PART 3,—NEW SER, x

310 J. W. KIRKALDY.

The mid-ventral surface in front of the atriopore is traversed
by numerous longitudinal pleats, which permit of the expansion
and contraction of this area so as to alter very considerably the
volume of the viscera and cavities of the body.

A cartilage-like tissue is developed as skeletal support for
the preoral cirrhi, and for the median and lateral (meta-
pleural) fin-plates. In the median dorsal fin, and usually, but
uot always, in the median veutral fin, this tissue takes the form
of separate fin-rays, each contained in a lymph-space or fin-ray
chamber; in the metapleura it forms a continuous unseg-
mented band with a related canal-like lymph-space.

I find it desirable to recognise in the family Branchiosto-
mide two genera—Branchiostoma and Asymmetron. The
genus Branchiostoma is divided by me into sub-genera—Am-
phioxus and Heteropleuron.

Genus I.—BRANCHIOSTOMA.

BrRANcHIostomA, Costa. 1834.
Amputioxvs, Yarrell. 1836.

Preoral tentacles or cirrhi, forming a single series united
to one another by a uniformly low intertentacular membrane.

A median ventral tentaculum impar present, the rest sym-
metrical.

Dorsal and ventral median! fins expanded in the caudal
region to form a lancet-shaped so-called “ caudal” fin, within
which the axis of the body terminates.

Fin-ray chambers, with or without enclosed fin-rays, present
in the preanal portion of the ventral fin.

Infra-rostral fin continuous with the right metapleur; the
left metapleur dying out before reaching the infra-rostral fin.

Atrial chamber produced behind the atriopore into a single
tapering cecum, reaching as far as the anus.

Preoral tentacles provided with numerous sensory papillz.

1 Although morphologically the ventral median fin is possibly a portion of
the right metapleur, it is convenient to retain the name “ventral median
fin” for that continuation of the right metapleuron which lies between atrio-
pore and tail terminus in the mid-ventral line,

A REVISION OF THE BRANCHIOSTOMIDA. SEE

Gustatory groove of Hatschek (on the right side of the
roof of the proral hood) shallow.

Sub-genus 1.—AmpuHioxus.

Ampuioxvs, Yarrell.

Both metapleura terminate immediately behind the atriopore,
and overlap the median ventral fin. The ventral fin-ray
chambers form a single median series, each containing a pair
of fin-rays, excepting the more anterior and more posterior
chambers.

Gonad pouches developed on both right and left epipleura.

1. Amphioxus lanceolatus, Pl. 34, fig. 1.

LiMaX LANCEOLATUS, Pallas. ‘Specil. Zool.,’ x, p. 19.

BRANCHIOSTOMA LUBRICUM, Costa. ‘Cenni Zool.,’? Napoli, 1834,
p. 49.

- Fe 35 ‘ Frammenti Anatomia Compa-

rata,’ fas. i, Napoli, 18438.

AMFPHIOXUS LANCEOLATOS, Yarrell. ‘ British Fishes,’ 1836, p. 462.

BRANCHIOSTOMA LANCEOLATUM, Gray. ‘Cat. Brit. Mus. Fish.,’
vol. vii, p. 150, 1851.

Myotomes, maximum number 62, minimum number 58,
most frequent 60 (J. W. K. 50 specimens); myotome formula
(przeatrioporal, preeanal, post-anal) 86, 15, 10 (cf. Lankester) ;
35, 14, 12 (cf. Andrews); 35, 14, 11 (J. W. K,).

Dorsal fin of moderate height, namely, one seventh of the
height from the crest of the fin to the free edge of the metapleur
at the mid-point of the animal’s length.

Snout or rostral fin small and pointed, not marked off from
the dorsal fin. Caudal expansion of the dorsal and ventral fin
long and lancet-shaped with sharp angles.

Ventral fin with a variable number of median fin-ray
chambers, in each of which are two paired fin-rays (excepting
those at either end of the series). Thirty-four to forty-one
pairs of ventral fin-rays present.

Oral sphincter placed in a line drawn vertically from the
apex of the anterior angle of the seventh myotome.

312 J. W. KIRKALDY.

Intra-buccal tentacles of the oral sphincter are twelve in
number. Oral hood large; tentacles ofthe oral hoodtwenty-one
in small specimens to forty-one in the largest specimens.
Olfactory pit present.

Gonad pouches twenty-three to twenty-nine on the right side,
and twenty-one to twenty-eight on the left side (in fifty speci-
mens, J. W. K.); usually twenty-six on each side.

Average length of fifty specimens (sent from Naples as well-
grown) 4°8 cm. (J. W. K.) ; maximum length 5:8 em.

Distribution.—Mediterranean Sea, English Channel,
North Sea, coast of Norway.

2. Amphioxus californiensis, Pl. 34, fig. 4.

BRANCHIOSTOMA CALIFORNIENSE, Cooper. ‘Nat. Wealth of Cali-
fornia,’ 1868, p. 498.

Myotomes, maximum number 73, minimum number 69,
most usual 71 (in ten specimens, J. W. K.). Myotome formula
45,17, 9(J. W.K.) ; 44, 19, 8 (J. W.K.) ; 44, 16, 9 (Cooper’s
type-specimen, J. W. K.).

Dorsal fin of moderate height; series of dorsal fin-ray
chambers extending from the eye-spot to the anal myotome.
Rostral fin very small. Caudal expansions of the dorsal and
ventral fins long and shallow. Ventral fin with fin-ray
chambers and paired fin-rays. Anterior extremity of notochord
dipping downwards. Oral hood and circlet of tentacles
relatively small in size. Gonad pouches, thirty-one right and
thirty-one left.

Oral sphincter underlies the apex of myotome 4, Maximum
length of ten specimens 7:4 cm. (J. W. K.).

Distribution.—Coast of California.

Remarks.—This species comes nearesttothe B. elongatum
of Sundevall, the specimens of which are lost. It may be
recognised by the relatively small size of the cephalic region ;
the oral sphincter is placed far forward (myotome 4), and the
anus is placed far backward (myotome 62).

A REVISION OF THE BRANCHIOSTOMIDA. 313

3. Amphioxus caribeus, Pl. 34, fig. 5.

BRANCHIOSTOMA CARIBZUM, Sundevall. ‘Ofvers. vet. Akad. Forsk.,’
vol. x, 1853, p. 12.

Myotomes, maximum number 61, minimum number 59,
most usual number 60 (in eight specimens, J. W. K.). Myo-
tome formula 37, 15,9; 38, 14, 7; 37, 15, 8 (most usual).

Dorsal fin low—one eighth of the height from the crest of
the fin to the free edge of the metapleur at mid-region of body.
Fin-ray chambers commence in front of the eye-spot, and
extend to myotome 55. Rostral fin marked off from the
dorsal fin by a shallow notch, and terminating bluntly in
front. Caudal fin small and shallow. Ventral fin low, with
fin-ray chambers and paired fin-rays. Oral sphincter underlies
the apex of myotome 5, and is provided with twelve intra-
buccal cirrhi. Przoral tentacles and gonad pouches as in A.
lanceolatus.

Maximum length among eight specimens 4 cm. (J. W. K.);
according to Sundevall 5:1 cm.

Distribution.—East coast of the United States and of
South America; West Indies.

Remarks.—This species stands very near to A. lanceo-
latus. It is distinguished from that form chiefly by the
slight development of the caudal fin and the shortness of the
post-anal region.

4, Amphioxus Belcheri, Pl. 35, fig. 8.
BRANCHIOSTOMA BELCHERI, Gray. ‘Proc. Zool. Soc.,’ 1847, p. 35.

Myotomes, maximum number 65, minimum number 68 (in
four specimens examined in the British Museum (J. W. K.).
Myotome formula 37, 14, 14 = 65 (according to Ginther) ;
38, 17, 10.= 65 (J. W. K,) ; 37, 16, 10 = 63 (J. W. K.).

Rostral fin well marked, and separated from the dorsal fin by
a depression. Notochord dipping downward before its anterior
termination. Inferior lobe of the caudal fin relatively large.

314 J. W. KIRKALDY.

Other characters apparently as in A. caribeus and A.
lanceolatus,
Maximum length among four specimens 5 cm. (J. W. K.).
Distribution.—Coast of Borneo (Sir Edward Belcher) ;
Prince of Wales Island, Torres Straits (Dr. Coppinger).
Remarks.—The only known specimens of this species are
preserved in the British Museum, Natural History.

Sub-genus 2.—HETEROPLEURON.
HETEROPLEURON, J. W. K., sub-genus nov.

The left metapleur terminates immediately behind the atrio-
pore; the right is directly continued without interruption
into the median ventral fin.

Ventral fin chambers with or without fin-rays.

Gonad pouches limited to a single series, which is developed
on the right epipleur.

Intra-buccal cirrhi sixteen in number.

1. Heteropleuron Bassanum, Pl. 34, fig. 6.

Brancuiostoma BassanuM, Giinther. ‘ Report Zool. Coll., H.M.S.
* Alert,” ? 1884, p. 31.

Myotomes, maximum number 78, minimum number 70,
usual number 75 (in 50 specimens, J. W. K.).

Myotome formula 43, 16,12; 44, 14,17; 45,17, 15; 40,
16, 14 (usual).

Dorsal fin shallow. Rostral fin large, and marked off from
the dorsal fin by a dip. Caudal fin long and low, the superior
and inferior angles entirely removed (as compared with A.
lanceolatus). Ventral fin with fin-ray chambers and paired
fin-rays. Preeoral cirrhi from 31 to 33 (in specimens examined).
Olfactory pit present. Oral sphincter underlies apex of seventh
myotome. Intra-buccal cirrhi sixteen in number. Gonads
26—31, only present on the right epipleur.

Maximum length of 50 specimens 4°3 cm. (J. W. K.).

Distribution.—Bass’s Straits, Australia.

Remarks.—The specimens of H. Bassanum sent by Pro-

A REVISION OF THE BRANCHIOSTOMIDA. "315

fessor Baldwin Spencer were in an excellent state of preservation.
They all show a greater compression of the body from side to
side, 1.e. less thickness, than do similar specimens of Amphi-
oxus lanceolatus.

Although the gonad pouches develop on the right epipleur,
they extend when ripening across the mediau line of the body,
so as to be visible through the left epipleur.

2. Heteropleuron cingalense, n. sp., Pl. 35, fig. 7.

Myotomes, maximum number 64, minimum number 61
(in eight specimens, J. W. K.).

Myotome formula 39, 17, 6; 39,17, 8; 39, 16,8; 38,17, 8.
_ Dorsal fin low (one eighth of height from crest of fin to edge
of metapleur at mid-body). Rostral fin small, not marked off
from dorsal fin. Ventral fin with fin-ray chambers and paired
fin-rays. Oral sphincter underlies the apex of the fourth myo-
tome. ‘Intra-buccal cirrhi not determined. Gonad pouches
twenty-five, only present on the right epipleur. |

Maximum length of eight specimens 3 cm. (J. W. K.).

Distribution.—Coast of Ceylon.

Remarks.—Four of the specimens examined by me were
collected by Mr. Haly, and supplied to Professor Lankester by
Professor Haddon. Three of these gave the myotome formule
39, 17,8; one (that figured) gave 39,17, 6. Five specimens
were lent to Professor Lankester by Professor Good Brown
from the United States National Museum. Of these, one
appears not to belong to this species at all, and, in fact, belongs
to the sub-genus Amphioxus. The remaining four presented
the myotome formula 39, 16, 8; 38, 17,8; and 37, 15, 9.
From these facts I think it is clear that there is a Cingalese
Heteropleuron distinct from H. Bassanum, from which it
differs chiefly in its smaller number of myotomes. Whether
the specimen of Amphioxus included amongst the specimens
sent from the United States National Museum indicates a
distinct Cingalese species of the sub-genus Amphioxus, or is
referable to A. Belcheri, I prefer to leave an open question.

316 J. W. KIRKALDY.

3. Heteropleuron cultellum, Pl. 34, fig. 2.

EpPIGONICTHYS CULTELLUS, Peters. ‘Monatsbericht der k. Preuss.
Akad. der Wiss.,’ Berlin,
1876, p. 322.
BRANCHIOSTOMA CULTELLUM, Giinther. ‘ Report Zool. Coll. H.M.S.
* Alert,” 1884, p. 32.
Willey. ‘Quart. Journ. Micros. Sci.,’
vol. xxxv, 1894, p. 361.

33 33

Myotomes, maximum number 56, minimum number 50,
usual number 52.

Myotome formula 82, 10,10; 82, 12, 8; 33, 10, 10; 34,
11, 10; 32, 10, 10 (usual number in thirty specimens).

Dorsal fin of great height, especially in the anterior region,
where it is more than one third of the total height from fin
crest to metapleur edge. The rostral fin is short and deep,
and is not marked off by a notch from the dorsal fin. The
ventral fin presents fin-ray chambers, which do not, however,
ever contain fin-rays, either single or double. (Willey first
observed this, and I can confirm him.)

The caudal fin is lancet-shaped, but not strongly marked.

The oral sphincter underlies the angle of myotome 6; there
are sixteen intra-buccal cirrhi. The preoral tentacles are from
forty-one to forty-three in number (in the specimens examined).
The notochord is slightly depressed in the rostral region, and,
instead of tapering to its anterior extremity, is expanded to
form a club-like termination. Posteriorly the notochord pro-
jects beyond the last myotome more than it does in other
species of Heteropleuron or Amphioxus, and terminates
bluntly. The gonad pouches are from seventeen to twenty
in number, forming a series on the right epipleur. The
maximum length shown by specimens in Haddon’s collection
was 3°5 cm. (Willey).

Distribution.—Torres Straits, Australia.

Remarks.—This species differs more from the other species
of Heteropleuron than any of them do from one another, and
differs in features which remove it further than they are from

A REVISION OF THE BRANCHIOSTOMID. 317

the sub-genus Amphioxus. It would be almost justifiable to
place H. cultellum in a distinct sub-genus on account of the
absence of fin-rays from the ventral fin-ray chambers, the
great depth of the dorsal fin, the swollen anterior knob-like
termination of the notochord, and the considerable tract of
terminal notochord projecting posteriorly beyond the last
myotome.

Dr. Arthur Willey (loc. cit.) was the first to draw attention
to the unilateral character of the gonadic pouches in H.
cultellum, and, in fact, to all the points here noted, excepting
the number of the intra-buccal cirrhi. Dr. Willey states that
the preoral tentacles are devoid of the projecting sensory
papillz which occur in all species of Amphioxus and Hetero-
pleuron. I find in well-preserved specimens where the epi-
thelium is still present that the sensory papille are clearly
developed.

Dr. Willey was unable to find an olfactory pit in this
species, and I have not found one.

Genus II].—AsyMMETRON.

AsymMeEtRON, Andrews. ‘Johns Hopkins University Circulars,’ June, 1893,
vol. xii, p. 104.

Przoral tentacles grouped into ventrad and laterad series by
the presence of a very high intertentacular membrane uniting
the tentacles of the two ventrad groups, the lateral series having
a low intertentacular membrane like that of the whole series
in Branchiostoma: a median free ventral “‘ tentaculum impar ”
between the two ventrad groups of high-webbed tentacles.

Dorsal and ventral median fins expanded some distance in
front of the caudal extremity, and contracted again along the
terminal seventh of the body, so as to leave a narrow caudal or
urostyloid process, and no “caudal” fin. Myotomes not
developed in the urostyloid process, which is, however, traversed
by the notochord and nerve-cord. No fin-ray chambers or fin-
rays present in the ventral median fin.

Right metapleur continued without break to join the ventral

318 J. W. KIRKALDY.

median fin (as in Heteropleuron); left metapleur dying out
immediately behind the atriopore.

Infra-rostral fin in direct continuity with both the right and
the left metapleura. Atrial chamber extending behind the
atriopore as two laterally paired ceca. Przeoral tentacles devoid
of sensory papille. Gustatory groove of Hatschek (on the right
side of the roof of the przoral hood) very deep and pit-like.

Species unica.—A. Lucayanvm, Pl. 34, fig. 3.

AsymMETRON Lucayanum, Andrews. ‘Johns Hopkins Univ. Studies,’
1893, p. 213.

Myotomes, maximum number 69, minimum number 63,
usual number 65—66.

Myotome formula 43, 8,12; 44,9, 13; 45,10, 14; 46, 10,
12; usual formula 44, 9, 13.

Dorsal fin of moderate height (one seventh of length from
crest of fin to edge of metapleur at mid-body). Rostral fin
very slightly developed, either above or below the notochordal
rostrum, which, however, is of unusually great length. Ventral
fin devoid of fin-ray chambers, and of fin-rays ; a few irregular

spaces apparently represent the fin-ray lymph-spaces of other
‘genera. Oral sphincter underlies the angle of myotome 8.
Ten intra-buccal cirrhi are present. From twenty-one to
twenty-nine proral tentacles were counted. The nine most
ventrally placed are separated into two groups of four by
a single free median ventrad tentacle; the two sets of four
ventrad tentacles right and left of the tentaculum impar pre-
sent a high intertentacular membrane, by which tentacle 2
(counting the impar as tentacle 1) is united to tentacle 3, 3 to
4, and 4 to 5: the remaining tentacles on either side corre-
sponding to the numbers 6, 7, 8, 9,10, &c., are free from one
another excepting for a low basal connection as in other
Branchiostomide. All the preoral tentacles are smooth and
destitute of projecting sensory villi or papilla. Gonad pouches
twenty-six to twenty-nine, in a single series on the right

epipleur.

A REVISION OF THE BRANCHIOSTOMID”®. 319

Maximum length observed in 12 specimens 1:9 cm.
(J. W..K.).

Distribution.—Off the Bahamas, pelagic (?).

Remarks.—Several interesting points in the structure of
this species are detailed in Mr. Andrews’ account (loc. cit.).
Attention may especially be drawn to the following in addition
to those which have been indicated as generic and specific
characters. The anus occupies a nearly median position, the
ventral fin being here deflected to the right, occupying thus
more nearly its true morphological position as right metapleur.
The first pair of nerves arise below the eye-spot; in place of the
second pair there is a single nerve, which, with the first pair,
supplies the rostrum.

In young specimens the urostyle-like process is not developed,
and there is a terminal caudal fin.

The terminal branches of the rostral nerves are furnished in
all Branchiostomidz with cellular end-organs; these are of
especially large size in A. Lucayanum.

Incerte Sedis.
1. AmpHioxus ELoNGATUS, Sundevall. ‘Ofvers. vet. Akad. Forhand.,’
1852, p. 147.
BRANCHIOSTOMA ELONGATUM, Sundevall. ‘ Ofvers. vet. Akad. For-
hand.,’ 1853, p. 12.

Myotomes 79. Formula 49, 18,12. Dorsal finlow. Caudal
fin small. Oral cirrhi wanting. No eye-spot. Length6 cm. |

Distribution.—Coast of Peru.

Remarks.—This species is described as above by Sundevall,
but the description does not enable us to determine whether
the species belongs to the sub-genus Amphioxus or Hetero-
pleuron, or to the genus Asymmetron, or to a distinct genus.
The apparent absence of oral tentacles is very probably due to
alcoholic shrinking.

Eigenmann (8) records the capture of a number of Branchio-
stoma from San Diego Bay, California, which he is disposed
to refer to this species, but from his description it appears
probable that they belong to the species A. californiensis,

320 J. W. KIRKALDY.

2. BRANCHIOSTOMA PELAGICUM, Giinther. ‘Challenger Reports,’ vol.
xxxi, 1889, p. 43.

Myotomes 67. Myotome formula 36, 16 (?), 15.

Dorsal fin low ; caudal fin paddle-shaped ; ventral fin without
fin-rays ; oral cirrhi wanting ; notochord projects beyond the
myotomes posteriorly. Nerve-cord with eye-spot ; gonads
in two series, twenty-six (?) on either side.

Length 1 cm.

Distribution.—A single specimen taken in the tow-net
near Honolulu.

Remarks.—Of this species only the single specimen above
described is known. After Dr. Giinther’s description and
figure were published it was examined by Professor Lankester
by means of transverse sections, but the state of preservation
was such as to render any satisfactory observations impossible,
The specimen had been stained strongly in carmine before
it came into Dr. Giinther’s hands, and mounted under a
cover-glass which had greatly compressed it.

TABULAR ENUMERATION OF THE GENERA AND SPECIES OF
BRANCHIOSTOMIDA.

Genus I].—Brancuiostoma, Costa.
Sub-genus Amphioxus.

A. lanceolatus, Yarrell (Pallas).
A. californiensis, Cooper.

A. caribeus, Sundevall.

A. Belcheri, Gray.

Sub-genus Heteropleuron.

1. H. Bassanum Giinther.
2. H. cingalense, J. W. Kirkaldy.
3. H. cultellum, Peters.

Genus IJ.—Asymmetron, Andrews.
1. A. Lucayanum, Andrews.

A REVISION OF THE BRANCHIOSTOMIDA, 321

INCERT SEDIs.

1. Branchiostoma elongatum, Sundevall.
2. Branchiostoma pelagicum, Gunther.

Note.—Whilst the three species of Heteropleuron seem to
be distinctly characterised, it is very queStionable whether
more than two species of the sub-genus Amphioxus should be
recognised, namely, A. lanceolatus (including A. caribeus
and A. Belcheri) and A. californiensis.—E. Ray
LANKESTER.

BIBLIOGRAPHY.

1. AnpREews, E. A‘ Johns Hopkins Univ. Studies,’ vol. v, 1893, p. 213:
“ An Undescribed Acraniate, Asymmetron Lucayanum.”

2. Bovert, Ta.— Uber die Bildungsstatte der Geschlechtsdrusen und die

Entstehung der Genital kannern beim Amphioxus,” ‘ Anat, Anz.,’ vii,
1892, pp. 170—181.

8. Ercenmany, C. H.—‘ Ann. Nat.’ 1892, p. 70: ‘“‘Amphioxus in San
Diego Bay (? B. californiense).”

4. Gi, J.—‘ U.S. Nat. Mus.,’ v, 1882, p. 515: “ Notes on Leptocardians.”

5. GintHer, A.‘ Report, H.M.S. “ Alert,”’ 1884.

6. LancErnans, P.—‘ Archiv fiir Micr. Anat.,’ xii, 1876, p. 290 (lanceo-

latum).

7. Lanxester, E. R.—‘ Quarterly Journal Micros. Sci.,’ xxix, 1889, p. 365

(lanceolatum).

8. Morren, C.—‘ Bull. Acad. Roy. Belge, 1875, xxxix, 2, p. 312 (struc-

ture of notochord, caribeum).

9. Miurr, J.—‘ Konigl. Akad. der wissenschaften,’ Berlin, 1842, p. 79.
10. Peters, W.— Monatsbericht Akad.,’ Berlin, 1876, p. 322 (cultellum).
11. DE QuatreFaces, A.—‘ Ann. Sci. Nat.’ (8), iv, 1845, p. 197 (lanceo-

latum).
12. Rozen, W.— Morph. Jahrb.,’ ii, 1876, p. 87 (lanceolatum).

13. SuxpEvarL.— Ofversigt k. vet. Akad, Zérhand.? 1852, p. 147; 1853,
p: il.

322 J. W. KIRKALDY.

14. Witter, A.—‘ Quarterly Journal Micros. Sci.,’ 1894, p. 361: “ Report
on Professor Haddon’s Collection of Amphioxus from Torres Straits
(cultellum).”

15. Wittzy, A.—‘Amphioxus and the Ancestry of the Vertebrates,’
Macmillan, London, 1894.

EXPLANATION OF PLATES 34 & 35,

Illustrating Miss J. W. Kirkaldy’s paper, “A Revision of the
Genera and Species of the Branchiostomide.”

All the figures are drawn of the same absolute size, so as to facilitate a
ready comparison of the proportions of each species. A line placed above
each pair of figures gives the usual length of adult specimens of the species
indicated. With the exception of fig. 1, which is copied from originals drawn
from living specimens at Naples by Professor Lankester, the drawings have
been made from spirit-preserved specimens. In the views from the ventral
surface the preoral hood has been represented as expanded, as it would
probably appear in life, but in the side views (excepting in fig. 1) the actual
condition of contraction of the hood seen in the spirit specimens has been
more closely followed, and accordingly the side view is not in this region pre-
cisely coincident with the ventral view. The ventral mid-surface is repre-
sented as distended as it is in life, so that in the profile views it projects
below the metapleur. It would be as easy as it is desirable to have drawings
from the life of the South Australian Heteropleuron Bassanum and of
Amphioxus californiensis.

Fie. 1.—Ventral and profile views of Amphioxus lanceolatus, Pallas,
copied from Lankester (No. 7) with the addition of a tentaculum impar to the
oral hood. ‘The animal is represented as seen in life in a basin of water lying
on its back (upper figure) or on its side (lower figure), with the preoral hood
and tentacles expanded, the atrial chamber distended, and the atriopore widely
open.

Fic. 2.—Ventral and profile views of Heteropleuron cultellum,
Peters. Preoral hood contracted in the profile view.

Fic. 3.—Ventral and profile views of Asymmetron Lucayanum,
Andrews,

A REVISION OF THE BRANCHIOSTOMIDA. 323

Fie. 4.—Ventral and profile views of Amphioxus californiensis,
Cooper. In the profile view the preoral hood is represented as much con-
tracted as it is in spirit specimens.

Fie. 5.—Ventral and profile views of Amphioxus caribeus, Sundevall.
In the profile view the preoral hood is incompletely expanded.

Fic. 6.—Ventral and profile views of Heteropleuron Bassanum,
Giinther.

Fie. 7.—Ventral and profile views of Heteropleuron cingalense,
Kirkaldy. Preeoral hood in the profile view only partially expanded.

Fic. 8.—Ventral and profile views of Amphioxus Belcheri, Gray. In
the profile view the preoral hood is not nearly so fully expanded as in
the ventral view.

‘
wereeonie veruolropeeal
ail 4 ised * ae ia fab imay
RE wid || Lath Vee rai
Gere) tay a Ch ae Th te
’
| ‘
‘| '
: q
‘2A
a 7 ad
mils
a iy +
v_ =
: 7 - :
I
a a as
aan 2
—_
he
(
_ ° 3 7 7 . : +
7 i» P - a
a 7 7 s ie 7 _ «

ore re CPOE, & 2:1

vir iemiijen teeiinith +
Se) ee
—_ e

ber DONE ean yt. Aa Mariette

cer Ale ei tae La

Wamaeehie “Ph tUaiotutelh la

PMdal i

ered) clligay, han Tay a '
Wea) wh i Hits: feta a
iy pay ive ln i
iA fei ‘iy 4 itt weary WH iA ull
i
*
v
a ‘- -
ls he a
SS

tine hy STi “a

acpi? fie dey eT a 7
t. CaS 2h een
canhkeds Ayqe ata) Gan

re eniilovd ide TaN «0.00
Liviu belGopiey al als Mu) taal

init, align tae len “yt wie
Re)

a
c

THEORY OF THE EMBRYONIC PHASE OF ONTOGENY. 9825

Sedgwick’s Theory of the Embryonic Phase of
Ontogeny as an aid to Phylogenetic Theory.

By

E. W. MacBride, B.A.,
Fellow of St. John’s College, Cambridge ; Demonstrator in Animal Morpho-
logy to the University of Cambridge.

In a recent number! of this Journal there appeared a paper
by Mr. Adam Sedgwick on the significance of the embryonic
phase in development, which embodies a principle which, if
true, seems to me well fitted to throw light on some obscure
problems in morphology.

It is not the object of the present essay to discuss the cor-
rectness of Mr. Sedgwick’s views, but rather, assuming them
to be true, to point out some of their consequences.

These views may be briefly stated as follows. Making a
broad survey of the facts of ontogeny, we find that there are
two main types or phases of development—the larval and the
embryonic. In the former case the immature organism pur-
sues a free life, engaging in the struggle for existence; in the
latter case the developing animal is shut off from the influence
of external conditions, either inside an egg-membrane or in
the uterus of the mother; but in both cases it is relieved from
the necessity of having to seek its own living, since nourish-
ment is provided for it either in the shape of food-yolk or fluid
nourishment exuded from the uterine walls.

In many cases the whole course of the ontogeny of an animal

1 “On the Law of Development commonly known as von Baer’s Law; and
on the Significance of Ancestral Rudiments in Embryonic Development,”

Quart. Journ. Mier. Sci.,’ April, 1894.
VoL. 37, PART 3.—NEW SER. ¥

326 E. W. MACBRIDE.

is embryonic, but in every case larval development is preceded
by a longer or shorter period of embryonic development.

The whole interest of the science of Embryology lies, of
course, in the fact that features observed in both types of de-
velopment seem inexplicable except on the assumption that
they are reminiscences of structures possessed by the ances-
tors of the animals in whose development they appear. Such
traces of the history of the race are to be found in the vast
majority of larve ; in embryos they are likewise to be found,
though here they are less prominent, as is seen by comparing
the development of two allied forms, in one of which the larval
type prevails, and in the other the embryonic. Now Mr. Sedg-
wick’s theory of the relation of the two types to one another is
that that portion of embryonic development in which ancestral
features are observable represents a larval stage passed over
inside the uterus or egg-membrane and modified in con-
sequence. Thus the chick during the first four or five days
of its existence is to be regarded as an immensely modified
larva.

If this view be true it follows that, however modified the
record of ancestral history contained in the larval development
may be, the embryonic record of the same history can never
rise above it in value.

It was until lately customary to assume, explicitly or im-
plicitly, that there was an inherent tendency for the ontogeny
of the individual to be a summarised repetition of the phylo-
geny of the race. In proof of this statement we may adduce
Balfour, who in his ‘Text-book of Comparative Embryology’
(vol. ii, p. 298), says, “ Unless secondary changes intervened
this record [of ancestral history] would be complete ;” and
Bateson,! in his discussion of the ancestry of the Chordata,
commits himself to a similar position. That there can be no
such general tendency is, however, shown by the fact that in

1 «The Ancestry of the Chordata,” W. Bateson, ‘Quart. Journ, Micr.
Sci.,’ 1886. ‘Development within an egg-shell as involving a less compli-
cated struggle with enviromental forces, is less subject to variation than that
in the open sea, and consequently is more likely to preserve ancestral features.”

THEORY OF THE EMBRYONIC PHASE OF ONTOGENY. 327

bud ontogeny there is no trace of anything which can be inter-
preted as ancestral structures, and that some most striking
and recent changes, such as the loss of limbs in snakes or the
reduction of the toes of the ostrich to two, are not recorded in
embryology, i.e. the organ concerned shows from its inception
the adult arrangement.

How then does the theory we have adopted account for the
retention of ancestral characters by larve ?

So far as we can judge by comparative anatomy, the stimuli
to evolution (in the sense of change of structure) have been
two, viz. (1) change of environment and habits, and (2)
increased or decreased demands on the working of certain
organs. As we therefore pass along a series of genetically
connected animals, we should find, pari passu with the
environment and the functional demands of the organism, the
structure changing. If these stimuli commenced to act from
the beginning of free life, then each individual adult in the
chain would show from the beginning the modified structure
belonging to it; but if these stimuli were deferred in their
operation till the animal had attained a certain size, then what
was before a uniform life-history would become differentiated
into two periods—a larval during which the ancestral habits
were retained and the structures corresponding to them, and
an adult in which new habits were assumed and structure
correspondingly modified.

An illustration will make this clear: if young flat-fish when
they emerge from the egg were at once to adopt the adult
mode of life, then that most interesting larval stage, in which
they are bilaterally symmetrical, would be missed out in their
development.

Thus we see as arace of animals progressed from point to
point in evolution, it would tend to develop a trail of larval
stages, each grade of development surmounted being re-
presented by a new larval stage intercalated in the ontogeny.
This process, however, could not go on indefinitely; there
would soon arise the tendency for the earlier larval stages to
be passed over whilst still in the egg-membrane, and so a

328 BE. W. MACBRIDE.

portion of the development would become embryonic, and so
subjected to the various modifying influences which are con-
nected with this type of ontogeny. Therefore it follows, as
the first important deduction from Sedgwick’s theory, that in
seeking to obtain a basis for phylogeny, most importance must
be attached to animals which show long larval histories.

Balfour with his usual sagacity has, so to speak, instinctively
anticipated this conclusion. Although he points out that “ the
favourable variations which may occur in the free larva are
much less limited than those which can occur in the foetus,”
he says that there is “a powerful counterbalancing influence
tending toward the preservation of ancestral characters, in that
larvee are compelled at all stages of their growth to retain in a
functional state such systems of organs at any rate as are
essential for a free and independent existence” (‘ Comp. Emb.,’
vol. 11, p. 299).

The objection, alluded to in Balfour’s statement, that larvee
as well as adults have been subjected to the modifying influ-
ences of their environment, will readily occur to most minds.
Let us consider whether it is possible to approximately estimate
the nature and amount of such influences; and, first of all, let
us consider what is meant by secondary larve.

Balfour imagined that secondary larval forms might be pro-
duced by a diminution in the food yolk, and consequent earlier
commencement of free existence (loc. cit., p. 300). There is
no evidence to suggest that such a change has ever taken
place ; all the facts point in a contrary direction. We shall
see that food yolk produces the most diverse distortions of
development; the developmental processes of free larve are,
on the other hand, remarkably uniform. Secondary larve
must be regarded as having arisen owing to the young adults
having taken to a new mode of life; the best instances of this
are perhaps the aquatic larvee of the may-flies and dragon-flies.
We have the strongest reason for believing that the immediate
ancestors of insects were terrestrial animals, aud the aquatic
larvee mentioned show their secondary character by the fact
that their respiratory organs are modified from organs adapted

THEORY OF THE EMBRYONIC PHASE OF ONTOGENY. 3829

to air breathing. Their “ tracheal” gills, instead of like all
other gills bringing the blood into close proximity to the water,
bring their blood first into contact with air contained in a
system of closed tubes, and then this air into contact with the
water.

In the case of ordinary larve, the probability of modifica-
tions due to adaptation to the environment cannot be denied.
If, however, Sedgwick’s hypothesis is correct that “ larval
history is constructed out of ancestral stages,” or, in other
words, that the larva retains ancestral characters because it
retains the ancestral mode of life, then the environment has
remained to a large extent constant (at any rate in the com-
monest case, that of pelagic larva), and the changes they are
likely to have undergone, instead of being, as Balfour supposed,
unlimited, will be comparatively few in number.

Of these changes reduction in size is the most important.
The passage to the adult state is often accompanied by the loss
of larval organs, and great changes in those which are retained,
necessitating in some cases the complete destruction of their
constituent cells, and their reconstruction from rudiments
which have retained the embryonic condition (histolysis). It
is, therefore, clearly to the advantage of the larva to grow no
larger than necessary before it undergoes metamorphosis.
Correlated with this loss of size is the frequent disappearance
of all traces of segmentation, since this is probably to be
regarded as essentially the same phenomenon as vegetative
reproduction, only held in check by the individuality of the
whole. Metameric series of organs are represented only by
those members which are absolutely necessary. Another
change which larve are prone to undergo, is the acquisition of
transparency. What results this carries in its train will be men-
tioned below. Finally, the occurrence of long spines is a wide-
spread phenomenon, though what their precise use is it would
be rash to surmise. Possibly they are of a protective nature.

Let us.now apply these principles in a concrete case, for
example the larve of the Crustacea.

The characteristic larva of the Entomostraca is the well-

300 E. W. MACBRIDE.

known Nauplius, which shows no signs of segmentation. We
have however from comparative anatomy, strong reason for
believing that the ancestor of the Crustacea was segmented,
and that it was probably related to the Polycheta. How is
this apparent contradiction to be explained? We answer that
the Nauplius retains the ancestral habits of Crustacea, and
with this a certain necessary amount of ancestral structure ;
but it has diminished in size and external segmentation has
disappeared. Since in the ancestor locomotion and prehension
of food were effected by one or two pairs of anterior append-
ages only, we have these alone represented in the Nauplius,
though the ancestor doubtless possessed in addition a series of
segments bearing undifferentiated parapodia-like appendages.
The complete disappearance of these is a mark of the high
specialisation of the larva; if we compare the various
families of the Entomostraca with one another, we find that
in the primitive group of the Branchipoda, the Nauplius shows
indications of a postoral segmentation; whereas in the highly
specialised Cirripedes and Ostracods we get a specialised
Nanplius. In the former case this is brought about by the
outgrowth of great spines, in the latter by the precocious
appearance of the adult bivalve shell, and in neither instance
is there a trace of segmentation.

The larva of the Malacostraca the Zoza, has been a great
puzzle to morphologists. It is quite impossible to regard
some of the peculiar features, such as the suppression of the
thoracic segments and their appendages, as ancestral, and the
question has been raised by Claus,! whether it has any
phylogenetic significance at all.

Applying Sedgwick’s principle, we explain the Zoza as
representing a later ancestral stage than the Nauplius, in
which some of the Nauplius appendages had become ex-
clusively masticatory and others exclusively tactile in function;
the main locomotor function had been, so to speak, passed on
to the two or three pairs of maxillipedes, which are

1 ZL. Claus, “Zur Kenntniss d. Malakostrakenlarven,” ‘Wirz. Naturw.
Zeitschrift,’ 1861.

THEORY OF THE EMBRYONIC PHASE OF ONTOGENY. 38381

always large and biramous, whilst at the same time some of the
most posterior segments have been modified to form a powerful
jointed “ tail.’ The thorax retained its primitive character,
and is accordingly suppressed in the larva; though here in
comparing the Zowe of the various groups we meet with a
precise parallel to the case of the Nauplii. The Zoza of the un-
differentiated Schizopod possesses only one pair of maxillipedes,
and even they are short and somewhat foliaceous, but it shows a
distinct segmentation of both thorax and abdomen. Still more
instructive are the larve of the lower families of the Decapods,
the Sergestide and Penzide. Taking Peneus for example, we
find that it escapes from the egg-membrane as a Nauplius: it
gradually changes to a Zozea with two pairs of maxillipedes and
the thorax distinctly segmented and with rudimentary append-
ages ; this passes into a form with thorax well developed and
all its appendages biramous—the so-called Mysis-stage, closely
resembling the adult Schizopod,—and from this it passes to the
adult state. On the other hand, in the highly specialised
Brachyura we find a highly specialised Zoza, in which the
thoracic segments are totally suppressed and the thorax pro-
longed into great spines, the Mysis-stage is dropped but a new
“ Megalopa”’-stage is introduced, which strongly recalls the
Macrura, and may be taken to indicate a Macrurous ancestor.
The existence of the Megalopa and Mysis stages, the signifi-
cance of which is obvious, affords the strongest reason for
maintaining the ancestral significance of the Nauplius and Zoza
stages; in doing so one merely follows the universal rule of
science, 1. e. reasoning from the known to the unknown.
Turning now to the embryonic type of development, let us
examine the causes which are likely to modify a course of
development which is primitively larval. First we must
discriminate between various kinds of embryonic development.
There is, in the first place, the type in which the organism is
confined within the egg-membrane and supplied with nutriment
by means of yolk stored up in its cells. Secondly, we have
cases in which the embryo, still remaining in the egg-mem-
brane, is retained in the body of the mother, the egg being

332 E. W. MACBRIDE.

closely applied to the uterine wall, from which nourishment is
obtained, and the yolk having consequently in large measure
disappeared. Thirdly—a much rarer case,—a number of eggs
are enclosed together in a capsule and only one develops; the
eggs destined to be eaten being known as yolk cells. As aless
extreme case we have the eggs all developing up to a certain
stage, but only a few surviving. This condition is seen in
Prosobranch Mollusca. Lastly, we may mention those cases
in which the uterus or other brood-pouch of the mother is
used, so to speak, as a nursery for the larve, the embryos
escaping from the egg-membrane, and passing the earlier part
of their existence as free-swimming organisms inside the
brood-pouch.

Taking the first case, which is by far the commonest, the
disturbances of development which are found in it are due to
two main causes—yolk and the egg-membrane. It is owing
to the cramping influence of the latter that external differentia-
tion of: form is to a large extent lost. The gastrula of
Asterina gibbosa, for instance, is almost spherical, con-
trasting thus with the common form of Echinoderm gastrula, .
which is more or less elongated. Where the egg is enclosed in
a roomy capsule, on the other hand, as in the Pulmonata, this
is less frequently the case; for instance, we have the velum of
Limneus and Planorbis. Mere disuse will not suffice to
account for the disappearance of external organs, as in this
case all traces of ancestral history ought to disappear in internal
as well as external organs, and this is not the case.

The presence of food yolk exercises the most distorting
influence on development. To Lankester’ is due the credit of
firstlaying emphasis on this. In treating of the development of
Mollusca he points out that the question whether the endoderm
is represented by many or few cells, and whether, consequently,
these are invaginated to form the gut, or whether the ectoderm
grows over them, is entirely determined by the amount of yolk
present. Balfour, who had almost at the same time instituted a

1 “On the Invaginate Planula or Diploblastic Phase of Paludina vivi-
para,” E, Ray Lankester, ‘ Quart. Journ. Micr. Sci.,’ vol. xv, 1875.

THEORY OF THE EMBRYONIC STAGE OF ONTOGENY. 3833

similar comparison between the segmentation of the eggs! of
vertebrates, subsequently put forward the thesis,? based on a
comprehensive survey of the facts of embryology, that the
rapidity of segmentation of a given region of the ovum is in-
versely proportional to the amount of yolk contained in it.

The general effect of the presence of yolk, therefore, when
massed specially in the endodermic end of the ovum, is to
impede cell division, and render processes of development
which depend on folding (e. g. invagination) impossible.

There is, however, another manner in which yolk can be
accumulated, and that is in the more central portion of the
ovum, instead of at one end. This is characteristic of the
Arthropoda. When it is comparatively moderate in quantity,
as in the case of Lucifer, segmentation and invagination can
proceed normally, though the number of cells composing the
blastosphere is small. When it is somewhat greater in quantity,
as in Branchipus, segmentation at first proceeds normally, but
soon the inner yolky ends of the blastomeres fail to be governed
by the rapidly increasing nuclei, and segmentation only affects
the outer layer of the egg, the inner ends of the first formed
blastomeres fusing together to form a central yolky mass. In
most Crustacea the yolk is so large in quantity that only
superficial segmentation is possible from the beginning. In-
vagination of this outer layer toform the gut still occurs in some
cases, the yolky mass being pushed before it ; but, since the yolk
is eventually absorbed by the endodermic cells, even this soon
ceases to be possible, and we reach eventually a condition in
which thesegmentation and first processes of development recall
to a certain extent those found in telolecithal eggs when the yolk
increases to such an extent as to prevent segmentation at the
endodermic pole at all (meroblastic eggs). In the scorpions and
insects segmentation in its earlier stages is totally suppressed,
and represented merely by the multiplication of nuclei; and in
the later stages segmentation only occurs where developing
organs require it, and thus a mimetic meroblastic segmenta-
tion is produced.

1 «Quart. Journ. Micr. Sci.,’ 1875. 2 “Comp. Emb.,’ vol. i, p. 121.

334 E. W. MACBRIDE.

The second type of embryonic development, viz. that in
which the egg is applied to the uterine wall, is characterised
by the reduction of the food yolk, so that the segmentation
reverts to the total type. Sharply marked traces, however, of
the former presence of yolk remain. The gorging of the en-
doderm with yolk has rendered the archenteron functionless,
and as it still remains functionless when the plan of absorbing
nutriment from the uterus through the general surface is
introduced it is deferred in development, and in fact the
destiny of the first products of segmentation is totally different
in Mammalia from what it is, for example, in Echinoderms.

In the third type of embryonic development, in which the
ovum is enclosed in a capsule with a number of yolk cells, the
most weird changes are produced in development. We read
of a complete separation and subsequent reunion of the blasto-
meres for instance; this type is almost confined to the
Platyhelminths, and is alluded to here only for the sake of
completeness, and to show how few traces of ancestral history
the development of these animals affords. One family only,
the Dendroceela, lay their eggs singly, and in this case we have
a large amount of food yolk present; yet Platyhelminths
have bulked largely in many phylogenetic speculations,

Lastly, we have those few cases in which the developing
animal escapes from the egg-membrane, but remains in the
uterus or brood-pouch. In these cases we have comparatively
little interference with the normal course of development. The
early stages occur in a perfectly regular manner, and we have,
in fact, free-swimming larvee within the brood-pouch; it is
only in the later stages that they commence to absorb fluid
from its walls. It is necessary to emphasise this type of
development, though it is comparatively rare (Brachiopods,
Paludina, and Amphiura squamata) because if it is con-
founded with the foregoing types, its totally different character-
istics would seem quite inexplicable.

I ought perhaps to mention that the earlier stages of
Amphiura squamata are described as being abnormal by

THEORY OF THE EMBRYONIC STAGE OF ONTOGENY. 385

Russo,! but neither the figures nor the methods of this author
are calculated to inspire much confidence. These earlier stages
are very difficult to obtain, but I have strong reason to suspect
from those which I have seen, that the earlier development
follows the ordinary Echincderm type.

Having thus rapidly reviewed the principal disturbing
factors in embryonic development, we can employ our know-
ledge in attacking one of the most vexed questions in mor-
phology, viz., the significance of the mesoderm and its contained
cavity the celom.

Is the former to be regarded as a differentiated portion of
the gut-wall, and the latter as a portion of the enteric cayity,
or is the ceelom to be regarded as a mere enlargement of the
cavity of the gonad as Hatschek? has suggested ?

In all Annelids and all Mollusca (Paludina and Cephalopods
excepted) the mesoderm first appears as two symmetrically
situated large cells—the primary mesoblasts. In Paludina,
Echinodermata, Sagitta, Brachipoda, and Amphioxus, it arises
as one or more pouches of the gut. Now, leaving out of
account Anthropods, Vertebrates and Cephalopods, where the
development has been complicated by the enormous amount
of yolk present, we find that of the other groups the Echino-
derms have by far the most prolonged larval development.
They are unique amongst the Celomata in the fact that the
blastosphere is a free-swimming larva, and that consequently the
development of both endoderm and mesoderm takes place during
their free-swimming life. Here, then, we may on our hypothesis
expect to find ancestral structure preserved, and here we find that
the coelom is developmentally a part of the archenteric cavity.

No Annelid or Molluscan larvee commence free life so early ;
most of them may be ruled out at once for the purposes of
this comparison, since the disturbing presence of yolk shows
itself plainly in the fact that the endoderm is represented by
a few large spheres, and the production of a pouch has become

1 Achille Rosso, “Embryologia d’ Amphiura squamata,” ‘ Rendiconti

della Societa Reale di Napoli,’ tome vii (2nd series).
2 «Lehrbuch der Zoologie,’ 1891, B. Hatschek,

336 E. W. MACBRIDE.

an impossibility. A few Annelid eggs have, however, very
little yolk, and the larvee commences a free life in the gastrula
stage.| Even here, however, if the blastosphere be compared
with that of an Echinoderm, one is struck at once by the com-
paratively small number of cells it has, and one understands
why the mesodermal rudiment should be represented by a
single cell. The small number of cells is doubtless due to the
comparatively large quantity of yolk, even if it be fairly
uniformly distributed. This is probably, however, not the
only reason why the celomic gut pouch is not found in the
Annelid larva. If we compare Echinoderm larve with one
another, we find that the blastocele or segmentation cavity
and the ccelom vary inversely with regard to one another.
Thus in the creeping larva of Asterina the coelom is very
spacious, and the blastoccele reduced to a mere slit; in the
pelagic larva of Asterias, on the other hand, the blastoccele is
exceedingly large, and the ceelom has the form of two narrow
tubes the lumen of which is in parts occluded. A similar
comparison can be made between the ordinary Tornaria larva
of Balanoglossus and Bateson’s larva. The reason of this
difference is not far to seek. It is to the over-development of
the blastoccele with its contained jelly that pelagic larve owe
that transparency which is so invaluable to them; hence the
great development of the blastoccele in pelagic larve and con-
sequent feeble development of the ccelom.

Now whatever may be the functions of the ccelom in
Echinoderms, in Annelids its main functions are excretion, and
the production of the sexual cells. Of these the first is
performed in the Trochophore (the characteristic Annelid larva)
by the so-called protonephridium, and the second has, of
course, no place in larval economy. Hence, if we regard the
Trochophore as bearing somewhat the same relation to the
Echinoderm larva as the Zowa of the crab does to that of
Penzeus, we see why the ccelom should have been entirely sup-
pressed and the mesoderm represented only by a few large cells.

1 Compare figures given in Korschelt and Heider’s ‘Lehrbuch der Ver-
gleichende Embryologie.’

THEORY OF THE EMBRYONIC STAGE OF ONTOGENY. 3837

Thus the ccelom appears to be, phylogenetically, simply a differ-
entiated portion of the archenteron ; when the lumen of the
latter is small, and its walls are composed of only a few large
cells, the mesodermic walls of the coelom are represented by a
single large cell on each side.

A little consideration will throw light on the reason why
what we see to be probably the primitive mode of development
should appear in Paludina Brachiopods and Amphioxus. In
all these cases the yolk is exceedingly small in quantity and
uniformly distributed; in the first two cases we have a
pseudo-embryonic development—the fourth type mentioned
above,—and of course inthis case there are no pelagic conditions
to require a suppression of the celom. Amphioxus has a long
larval history. Sagitta, on the other hand, pursues a true
embryonic development within the egg-membrane ; but the
yolk appears to be quite uniformly distributed, and hence its
primitive character.

A second vexed question which naturally follows directly on
that of the origin of the mesoderm, is the origin of the
endoderm, and consequently of the gastrula itself. Echinoderm
development suggests the idea that the ancestral form of
metazoon was a sphere of ciliated cells, and that the archenteron
arose through the specialisation of a portion of the surface of
the sphere to fulfil digestive functions and its invagination
into the anterior. This is the view adopted by Korschelt
and Heider ; and it is, of course, the famous gastrzea hypothesis
of Haeckel.? On the other hand, contrary opinions have been
put forward by Metschnikoff,? Lankester,* and Sedgwick.®

1 ¢Lehrbuch der Vergleichende Embryologie,’ vol. i, p. 81. Heider
showed experimentally that carmine granules were swept by the cilia to the
posterior end.

? Haeckel, ‘Studien der Gastraea Theory,’ Jena, 1877.

3 Hl. Metschnikoff, ‘“ Spongiologische Studien,” ‘ Zeit. fiir wiss. Zool.,’
Bd. xxxii, 1879.

4 HK. Ray Lankester, ‘‘ Notes on Embryology and Classification,” ‘ Quart.
Journ. Mier. Sci.,’ vol. xvii, 1877.

® A. Sedgwick, “The Development of the Cape Species of Peripatus,
pt. iii,”’ * Quart. Journ. Micr. Sci.,’ vol. xxvii, 1887.

338 E. W. MACBRIDE.

Metschnikoff starts, like Haeckel, from the blastosphere or
blastula as an ancestral form; he supposed it, however, to
have become filled up by cells wandering in from the periphery.
In the midst of these a digestive cavity was later developed,
and finally the mouth was formed by the specialisation of the
area through which food was taken in. Lankester, starting
from the same form, supposes the inner ends of the cells of the
blastula to have become differentiated so as to be specially
digestive in function, and later that they became separated off
as a special layer. The cavity of the blastosphere was thus the
digestive cavity, and food at first taken in over the whole surface,
was later taken in only at one point, and thus a mouth was formed.

Sedgwick, on the other hand, is inclined to start from a
protozoon in which cell territories were non-existent, though
many nuclei were present. He supposes that the gut originated
as a digestive vacuole, and that the nuclei acquired a definite
arrangement with regard to this vacuole and other organs, and
thus tissues were constituted. Cell territories, in so far as they
exist in the adult, he regards as due to secondary rearrangement
of the protoplasm.

We have already pointed out that Echinoderm development
tells strongly in favour of the view supported by Haeckel and
Korschelt and Heider, and that Echinoderm development,
from its almost exclusively larval character, is of the very
greatest importance in deciding such a question. Its evidence
is by no means solitary ; the statement may be made that, in
all Coelomata without exception, when the yolk is feebly
developed and evenly distributed we find the embryo pass
through a blastula stage which is converted into a gastrula
by invagination (cf. Leucifer among Crustacea, Polygordius
and Serpula and many others in Annelids, Paludina and
Chiton in Mollusca, Amphioxus and Cyclostomes in Verte-
brata, &c.). The groups which constitute the chief support
of Metschnikoff’s theory are Sponges and Celenterata. We
may leave the first entirely out of account, as it is quite pos-
sible that they constitute a distinct phylum to the rest of the
Metazoa. In many Celenterates we start from a blasto-

THEORY OF THE EMBRYONIC STAGE OF ONTOGENY. 3839

sphere, which is in some cases a_ free-swimming larva.
This blastospere, however, becomes filled up by cells which
wander in from the external layer; in most cases this
seems to take place from one end of the somewhat elongated
blastosphere, and in Aurelia this process is replaced by in-
vagination. Now the important point to notice in these
larvee is that during their free-swimming life the gut is func-
tionless; and this accounts for the fact that it is represented
by a solid rudiment. A precise parallel to the difference
between the endoderm of the Celenterate and Echinoderm
larve may be found amongst the larve of the Ectoproct
Polyzoa. In the pelagic larva of Membranipora (Cyphonautes),
which has a long free-swimming life, we find a perfectly well-
developed gut with mouth and anus; on the other hand, in
the larva of Alcyonidium we have a stomach of yolky cells,
an almost occluded cesophagus, and no intestine, whilst in
that of Bugula the whole mesoderm and endoderm is repre-
sented by a solid mass of cells. These larve are developed
from yolky eggs and take in no nutriment during their free
life. I hold, therefore, that Heider! is perfectly justified in
his statement that the ancestor of the Ccelenterata was “a
ciliated, oval, free-swimming form, in which by invagination
at the posterior end an archenteron was formed.”

Lankester’s view finds its chief support in the development
of Geryonia. In fact, this form is the only known one in
which such a process as he supposes to have taken place in
the blastula ancestor appears in the ontogeny. Are there any
reasons for regarding the development of Geryonia as specially
primitive? I think we may fairly say none; but that on the
contrary it shows manifest signs of its secondary character ;
the egg is yolky, and the development proceeds directly to the
medusa form, the hydra form being suppressed. The most con-
clusive argument, however, against Lankester’s hypothesis is
that on his assumption the cavity of the blastosphere is identical
with the cavity of the future gut. Now all recent investigations
have gone to show that the blastoccele is the rudiment of the

1 «Tehrbuch der Ver. Emb.,’ vol. i, p. 81.

34.0 E. W. MACBRIDE.

blood system, and has no connection with the gut whatever. In
many Celenterates the stage of the hollow blastosphere is
missed out, and segmentation results in a “ morula” of which
the external layer is the ectoderm and the rest endoderm. I
think we must imagine that in the development of Geryonia
the shortening process has gone one step further, and that as
a result of segmentation we reach at once the stage of the
hollowed-out planula.

Sedgwick’s hypothesis was suggested from a study of the
embryos of Peripatus capensis. Their developmental his-
tory is, however, the very last place where one ought to seek
for indications of the ancestral meaning of the earlier stages
—at any rate if Sedgwick’s own hypothesis as to the significance
of the embryonic phase be correct. All species of Peripatus
so far as is at present known are oviparous; in Peripatus
Nove-Zelanie, however, the eggs are large and yolky, and
the development conforms to the ordinary centrolecithal type
so characteristic of Arthropods—the peculiarities of which we
have described above. In Peripatus capensis nutriment is
supplied by the wall of the oviduct, and the yolk has in large
measure disappeared, at any rate its more solid portions ; but
the development still bears the impress of centrolecithal seg-
mentation, i.e., in the imperfect definition of the blastomeres.
It is obvious one might with equal justice expect to find
information as to the character of the ancestor of Metazoa in
the eggs of mammals.

Let us now briefly rehearse the conclusions to which the
foregoing discussions seem to point. The earliest well-marked
larval stage which we have discovered is the blastula—a sphere
of uniformly ciliated cells. This “ animal Volvox,” as Huxley!
calls it, may be regarded as a protozoon colony, not in the
sense of consisting of independent units any more than does
Volvox, but rather in the sense of being built up by the repeti-
tion of a unit asa result of what Lankester® calls ‘ eumero-
genesis,” just as is the colony of a Hydromedusan. At first all

1 «Anatomy of Invertebrates,’ p. 678.
2 Art. “ Hydrozoa,” ‘ Encycl. Brit.’

THEORY OF THE EMBRYONIC STAGE OF ONTOGENY. 341

elements in the blastophere were alike in structure and func-
tion. Later, however, coincidently with its acquiring the
capacity for moving in a definite direction, a change would take
place: the form first became elongated, and it is interesting to
observe that the free-swimming blastulas of both HEchino-
cyamus and Eudendrium have this form; then the cells at the
posterior end, being least favorably situated with regard to
promoting the locomotion of the colony, and best situated for
seizing food particles, since they are in a kind of backwater
from the eddies produced by the ciliary motion of the rest,
would become specially digestive ; increase in their number
could only take place coincidently with invagination if the
form of the colony were to be preserved and at the same time
the digestive cells were to remain in contact with the sur-
rounding medium—and thus we have the archenteron formed.
The cells at the anterior end, on the contrary, are in the best
position for receiving stimuli from the outer world; and here
we should expect the first sense-organ to appear, and it is just
at this spot that we find the larval sense-organ of Conatula
with its associated nervous tissue, and the still more primitive
sense-organ of the Echinocyamus larva, this latter consisting
of a thickened patch of ectoderm bearing stiff cilia, which take
no part in locomotion. In the same place the apical plate or
larval brain of the Trochophore is found, also bearing cilia or
more probably sense hairs.

Metschnikoff’s! great objection to regarding invagination as
the primitive method of forming the endoderm was that the
blastopore sometimes became the mouth and sometimes the
anus. Sedgwick’s suggestion, however, that mouth and anus
were differentiated from a slit-like blastopore, seems to answer
this difficulty. That a slit-like opening can be represented
by two independent perforations is shown by Echinoderm
development. Thus in Holothurians the larval mouth by a
shift of position becomes the adult ; in Asterids and Echinids,
on the other hand, it is represented by a totally new perfora-

+ “Vergleichend Embryologische Studien. (3) Uber die Gastrula
Hiniger Metazoen,” ‘ Zeit. fiir wiss. Zool.,’ Bd. xxxvii.

VOL. 37, PART 3,—NEW SER, Z

3842 E. W. MAOCBRIDE.

tion. No one can suppose that the ancestral form gave up
its old mouth and developed a new one; the change was one of
size and relative position only. We must assume that the
original mouth was a wide one, and that part is utilised by the
larva and aborted in the adult.

The ccelom, as we have seen, arose as a specialised portion
of the gut.

It is to be observed that the history we have just sketched
is in accordance with that rule which seems to hold in all cases
where we can by means of comparative anatomy show with
reasonable probability that evolution has occurred, viz. that
new organs never arise de novo, but by the differentiation of
older organs. This rule seems, however, to me to be violated
by supposing that either archenteron or ceelom arose as a
split in a solid mass of cells. The history affords also an
explanation of that rigorous separation of primary and secondary
body cavities, the blastoccele or hemoccele, and the celom,
which all recent research has tended to emphasise. The first
is, in fact, morphologically inside and the second outside the
primitive blastosphere.

Lastly, the conception of the primitive metazoon as a colony
of Protozoa is in accordance with that repetition of similar
parts on which Bateson! has laid so much stress as one of the
most marked characteristics of living things. We should
recall also the high individuality acquired by colonies of
Siphonophora, Polyzoa, and Ascidians.

1 * Materials for the Study of Variation,’ W. Bateson, Cambridge, 1890.

THE ANATOMY OF ALCYONIUM DIGITATUM. 343

The Anatomy of Alcyonium digitatum.
By

Sydney J. Hickson, M.A., D.Sc.,

Beyer Professor of Zoology in the Owens College, Manchester ; Fellow of

Downing College, Cambridge.
With Plates 36—39. If

InDEX oF CONTENTS.

PAGE
PREFACE 344
Section I.—History oF ouR KNOWLEDGE OF THE ANATOMY OF
ALCYONIUM . 344
Section IJ.—Gernrrat Form anp Naturat History oF THE Cotony 349
Shape—Size—Colour—Expansion and contraction of the polyps
—Contraction of the colony as a whole—Sexuality—Geo-
graphical and bathymetrical distribution—Hnglish species
of Aleyonium.
Section I[].—GrnERAaL ANATOMY i 254
The composition of the colony—The tudiethal polyps ce the
colony—The appearance of the extensible portion—Size—
Tentacles — Mouth— Disc — Stomodseum — Mesenteries—
Ventral mesenterial filaments— Dorsal mesenterial filaments
—Gonads—Communication between individual polyps—
The endodermic canals—Distribution of the spicules.
Section [V.—Minute Anatomy . 363
The eta aects sc almcdeniaeoMesenter al fila-
ments — Mesoglaea— Spicules — Endoderm — Ovaries and
testes—The buds.
Section V.—Norss oN THE CIRCULATION OF THE FLUIDS IN THE
Cotony anp THE DIGESTION 380

VOL. 37, PART 4.—NEW SER. AA

344 SYDNEY J. HICKSON.

PREFACE.

Tue work that is recorded in these pages has occupied a
considerable portion of my time for the past two years, and
was entirely conducted in the morphological laboratory at
the University of Cambridge and the Marine Biological Asso-
ciation’s laboratory at Plymouth. In some respects it is not
so complete as I had hoped it would be. My original inten-
tion was to publish with the account of the anatomy of the
species a description of its development and spermatogenesis,
but unforeseen difficulties presented themselves, and I have
decided to postpone the publication of the results I have already
obtained until they are more complete. The development of
Alcyonium can only be studied in the months of December and
January, when the investigations are frequently interrupted or
delayed by storms and rough weather. I have experienced
very great difficulty in keeping the young embryos alive when
I have been successful in effecting artificial fertilisation of the
ova.

If I am successful this winter in obtaining the stages in
development that I have not hitherto obtained, I hope to pub-
lish an account of the maturation and fertilisation of the ovum
and the development in the course of next year.

As I felt that the determination ofthe chemical constitution
of the homogeneous substance composing the bulk of the
mesogloea required considerable skill and time, which I did not
possess myself, I entrusted this portion of the work to Mr.
W.L. Brown, of St. John’s College, Cambridge, and his paper,
which accompanies this, is a valuable addition to our knowledge
of the chemical constitution of the Celenterate tissues.

Section I.—History or INVESTIGATIONS.

The common “ dead men’s fingers” of our coasts was con-
sidered by the ancients to be one of those substances formed
by the churning of the sea waves. It received its name
Alcyonium probably from Alcyone, daughter of Molus and

THE ANATOMY OF ALCYONIUM DIGITATUM. 345

Enarete, who threw herself into the sea for grief on the death
of her husband, Ceyx, in a shipwreck.

In the middle of the last century a great controversy arose
as to the nature of the so-called zoophytes, of which Aleyonium
was one. Peysonnel (1727), de Jussieu (1742), Reaumur
(1742), Ellis (1754), and others, all maintained that they were
animals, notwithstanding the vigorous opposition of the great
Linnzus, who considered them to be partly animal and partly
vegetable in nature, of Baster and others. It is not necessary,
however, to give an account of this controversy here, as the
reader will find a most interesting account of it in the first
edition of Johnston’s ‘ Zoophytes.’ In the course of the con-
troversy, however, de Jussieu paid a visit to Honfleur to study
these forms in the living state, and the result was (19) that he
produced for the first time a description and some fairly good
figures of the polyps of Aleyonium digitatum. He says,
‘‘ Nous avons soin de tremper dans nos bocaux une branche
de chacun des ces plantes en particulier, et nous fumes sur-
pris au premier aspect d’apercevoir, sans secours d’aucun in-
strument, de petits insectes qui avoient chacun pour loge une
des petites cellules formées dans le tissu de ce qui nous parois-
soit la feuille d’une plante.”

It is to be noted that in this memoir de Jussieu uses the
term “polype:” “J’appelle et dans la suite j’appellerai en
général polype, une famille d’insectes de la nature des vers
plus ou moins longs, dans les uns desquels la téte et dans les
autres le corps sont ou environnez ou parsemez de cornes, qui
servent aux uns connue de mains pour prendre des choses dont
ils font la nourriture, et outre cet ouvrage tiennent encore lieu
aux autres de pieds pour se mouvoir.”

It is difficult to say, however, whether he or Reaumur was
the first to apply the word polype to the Celenterate individual.
The following passage in Reaumur’s great work (84) is interest-
ing as showing the reason for its adoption. Writing of Hydra
he says, “‘ De concert avec M. de Jussieu qui en avoit observé
aux environs de Paris et fait dessinez une espéce du méme
genre, mais plus grande et d’une autre couleur je leur imposai

346 SYDNEY J. HICKSON.

le nom de polype parceque leurs cornes nous parurent analogues
aux bras d’animal de mer qui est en possession de ce nom.”

The term was subsequently adopted by Trembley (86), and
used by him in his classical work on ‘ Hydra.’

At the commencement of the present century, however,
the name Alcyonium was used to include a considerable number
of marine fleshy, gelatinous, and spongy zoophytes, which are
now known to belong to widely different groups of the animal
kingdom.

The description given by Ellis (8) of the genus is as follows:

* Animal, plant forma crescens.

“‘Stirps fixa, carnosa, gelatinosa, spongiosa, vel coriacea ;
epidermide cellulosa poris stellatis s. osculis pertusa, hydras
tentaculatas oviparas exserentibus.”

Baster (1) describes the genus quite briefly as follows:

“Die Alcyonia zyn zagte dog vaste zeegewassen die als een
middelsoort tussen de Kruidige (Herbacea) en Hoornagtige
(Keratophyta) entmaken.”

Included in the genus, as defined by Ellis, were a large
number of Ascidians, sponges, and probably also some Bryozoa
and Hydroids. In the systems of Miller (1776) (81), Fabricius
(1780) (10), Berkenhout (1795) (2), Turton (1809), Bose
(1807) (3), and others, we find the same conglomeration
of animals included under the name Alcyonium.

It must not be supposed, however, that because these natu-
ralists included all these different forms of animals under the
one generic name they were necessarily ignorant of more easily
discerned characters of Aleyonium digitatum; it was rather
their want of knowledge of the anatomy of the Ascidians,
Bryozoa, &c., which led to the confusion.

Ellis’s (7) description of his figure runs as follows :—“‘ A
piece of Alcyonium manus marina cut perpendicularly
through the middle to show that it is formed of tubes, which
branch out into others, each ending on the surface in a starry
opening of eight rays; in each of these openings is a polype-
like figure or sucker with eight claws, fastened to the inside of
the tube at its lower part by eight fine, tender filaments, by

THE ANATOMY OF ALCYONIUM DIGITATUM. 347

which it can raise or sink itself at pleasure in its tube; all
these tubes that compose this Aleyonium are connected to-
gether by minute reticulated fibres; these enclose a stiff
gelatinous substance which seems to be the flesh of this com-
pound animal, and these fibres, with their enclosed con-
tents, to be the muscles; for by the excretion of these it assists
in opening or closing the stars on the surface, while the suckers
or polype-like figures are pushing themselves out in search of
food, or when they are retreating to secure themselves from
danger.”

In 1816 Savigny clearly pointed out that the species
Alcyonium digitatum, A. exos, and A. arborea, in the
possession of eight pinnate tentacles, must not be classed with
the Ascidians, the anatomy of which he described in some
detail. Alecyonium exos was probably the Alcyonium
palmatum of the Mediterranean, and Alcyonium arborea
the Paragorgia of Norway.

In Cuvier’s ‘ Regne Animale’ it is stated that in Aleyonium,
as in the Pennatulez, we observe polyps with eight denticu-
lated tentacles, and a stomach prolonged into the ovaries ; but
the first detailed description of an Alcyonid was that given by
Milne Edwards in 1835 (9).

This distinguished naturalist obtained some specimens of
Paralcyonium (?), to which he applied the generic name
** Aleyonide,” at Cape Matifou, in Algeria, and he gave to the
Académie des Sciences a description of the anatomy, which,
considering the simplicity of his methods, is remarkably accu-
rate and detailed. He pointed out the individuality of the
polyps, and that each of them is capable of independent
action, although acting in response to stimuli with its
neighbours. He cut off a single polyp, opened it longitudi-
nally, and discovered the stomodeum, with its free opening
into the celenteron, the eight mesenteries with their attach-
ments, the ovaries, and the cavities in the tentacles. With
regard to the structure of the colony, he pointed out that tubes
(loges) running down into the colony are really the continua-
tion of the polyps, “‘ Il suffit d’un examen superficiel pour se

848 SYDNEY J. HICKSON.

convaincre que ces loges ne sont autre chose que la continua-
tion du corps des polypes eux-mémes;” and asserted that the
polyps do not retreat into their tubes like the Serpule, &c.,
do, but into themselves by asort of invagination. He noticed,
moreover, the mesenterial filaments, which he called “les
organes intestinaux,” and added with extraordinary perspi-
cuity that they ought to be considered organs of secretion,
“analogues aux canaux biliaires des insectes.”

In more recent times the only account we have of the
anatomy of Aleyonium digitatum is that given by Carl
Vogt and Jung in their ‘ Lehrbuch’ (87) ; but it contains so
many blunders, and the illustrations are so exceedingly diffi-
cult of interpretation, that itis really of very little use to the
student of zoology.

Von Koch’s (21) account of the anatomy of Aleyonium
palmatum contains some points of interest, but as it is very
brief and but imperfectly illustrated it is not of much practical
value.

It was the absence of any good account of this common
British Aleyonarian which first induced me to make a careful
study of some of the details of its anatomy; but I soon found
that I must defer the investigation of many details until an
opportunity presented itself of studying the living animal at
the sea-side. I found, however, that many points presented
greater difficulties than I anticipated, and I have been obliged
to delay the publication of my results until I felt quite certain
by repeated observation and experiment of their accuracy.
Just before the close of last year I learned that the late Prof.
Milnes Marshall was preparing for publication a description
of the anatomy of Aleyonium palmatum, and the last letter
I received from him, only a few weeks before the lamentable
accident which caused his death, was on the subject of the
formation of the buds in Alcyonium. With characteristic
generosity he proposed to postpone the publication of his work
until my paper was finished, and invited me to see his pre-
parations before I went to press. ;

Since his death his brother has forwarded to me the few

THE ANATOMY OF ALCYONIUM DIGITATUM. 349

rough notes and drawings on this subject that were found
among his papers, for which kindness I am glad to have the
opportunity of expressing my warmest thanks. The notes are
only those made in the course of his work in the laboratory,
and were not prepared in any kind of way for publication, but
whatever help or suggestion I may have gained from them I
have expressed in the text; and as the few drawings that were
finished are those of another species of the genus to that I
have myself investigated, I have not incorporated them at all
into this memoir.

Section II].—Gernerat Form anp Naturat History or THE
Cotony.

Alcyonium digitatum in its younger stages occurs in
the form of small incrusting masses on shells, weeds, worm-
tubes, &c.! As it grows older and larger it becomes dome-
shaped, then spherical and cylindrical, and finally branches
into five to eight blunt processes which become slightly
flattened in one plane when the colony reaches its largest size.
In the branched condition it is commonly called ‘“‘cow’s
paps,” ‘‘dead men’s toes,” or “dead men’s fingers,” all of
which terms are more or less appropriate to its lifeless, sodden,
pulpy appearance. |

The largest colonies I have measured are eight inches long
from the base to the tip of the longest branch, and six inches in
breadth.

The colour of Alcyonium varies very considerably with the
locality. The specimens caught in deep water off the Eddy-
stone Lighthouse are almost invariably of a pale pinkish colour
when alive, but they rapidly bleach just before death. Those
living in shallower water are quite white. Johnston states
that he occasionally met with it of a reddish-orange colour,
and I have myself dredged up yellow varieties off the west
coast of Scotland, and have seen similar specimens from the
Shetlands, the Bristol Channel, and elsewhere.

1 The youngest specimen I have found consists of a single polyp with one
bud sprouting from its side.

350 SYDNEY J. HICKSON.

The yellow variety owes its colour to a pale yellow tint in
the spicules, which is retained after the idee is dead and
preserved in spirit.

When first brought on deck from the dredge or trawl, the
colonies are of a soft, flabby consistency, with the extensible
portion of the polyps retracted, but still visible as transparent
circles on the surface. If a colony be allowed to remain out
of water for some little while the transparent circles become
smaller and smaller, until at last only a star-shaped depression
is left to indicate the position of the polyps. At the same
time the colony itself contracts considerably until it reaches
a size not much more than two thirds of its original bulk.

When the colonies are placed in a sea-water aquarium they
soon regain their former size by the absorption of water, and
in a few hours the polyps expand. The colonies, however,
do not, if they remain in a healthy condition, continue ex-
panded indefinitely, but are seen to contract and remain
contracted for some time at intervals. These periodic con-
tractions are quite independent of light, but as a rule seem to
be synchronous with tide intervals,—that is to say, they occur
twice in every twenty-four hours.

My experiments at Plymouth, two years ago, proved that
for the first three or four days after capture they contracted
regularly twice in every twenty-four hours, but after that their
times of contraction became irregular.

A number of experiments were tried with some Alcyoniums
placed in a tank in which an artificial tide, rising and falling
only once in twenty-four hours, was introduced, and at the
end of a fortnight I found that the two colonies which re-
mained healthy were contracting regularly only once in twenty-
four hours.

My observations were not sufficiently satisfactory to enable
me to draw any general conclusions from them, but they seem
to indicate that there is a normal rhythmic contraction of the
polyps of Alcyonium corresponding with some state of the tides
—probably low water, and that a new rhythm may be induced
by an artificial tide of different duration from the normal one.

THE ANATOMY OF ALCYONIUM DIGITATUM. 351

The matter seems to me to be worthy of further investiga-
tion, and it is to be hoped that some one who is able to remain
at the Plymouth laboratory for some months will repeat the
experiments.

The sexual reproduction of Aleyonium digitatum occurs
on our southern coasts during the months of December and
January. The colonies are invariably dicecious. Although I
have opened several hundred colonies, I have never found a
single instance of ova and spermatozoa occurring together.

Karly in March the gonads may be seen as minute swellings
on the ventral and lateral mesenteries. These gradually get
larger and larger as the year progresses, but it is not until
August or September that any difference can be seen between
the ova and the sperm-sacs. The ova then acquire a pale
yellow tinge, which becomes redder and redder as the winter
approaches. The gonads do not reach their full size until
December.

The spawning takes place in an aquarium very slowly, the
ova being shot out of the mouth, one by one, at considerable
intervals, and it may occur either by night or during the day-
time. Whether this is also the case in the natural conditions
I am, of course, unable to say; but colonies may be found at
the end of December in all conditions between those densely
packed with ova or sperm-sacs and those completely shotten.

Early in January the number of pregnant colonies becomes
smaller and smaller, until in the last week of that month
scarcely a single ovum can be found.

The extraordinary long period (nine months) that elapses
from the first appearance of the gonads to the time they reach
maturity is particularly noteworthy. I can offer no satisfac-
tory explanation of it.

I may say that there are no external indications of the sex
of a colony of Aleyonium. In shape, size, and colour the male
and female are the same. Nor is there any probability that,
as in some Hydrozoa, they are protandrous or protogynous,
for small colonies not more than an inch in height may be
either males or females.

352 SYDNEY J. HICKSON.

The time of year at which the spawning takes place is also
remarkable.

According to Wilson, the spawning of the Pennatulid
Renilla and the Gorgonid Leptogorgia takes place in the
summer months. Moreover, in the case of Renilla, “ the eggs
are always laid at very nearly the same hour of the day,
viz. 6 a.m.”

The following statistics as to the breeding times of certain
Alcyonaria at Naples are given by Lo Biancho (28), the
numbers referring to the months.

“Alcyonium palmatum.—Con uova mature 1i—iii, de-
posizione ix—x.

*“Gorgonia Cavolinii.—Con uova mature e deposizione
di larve v—vi.

“Pennatula phosphorea.—Uova mature nei polipi xi.

“Pennatula rubra.—EHsemplari con uova iil.”

It would seem, then, from these facts that the breeding
season of different Aleyonaria, and even species of Aleyonium,
varies very considerably.

There is not sufficient evidence at present to enable us to
give a very exact account of the distribution of Aleyouium
digitatum. It extends from between tide-marks to a depth
of forty or fifty fathoms, and does not, I believe, extend into
deep water.

Professor Herdman informs me that Alecyonium digi-
tatum is very common in the Firths of Forth and Clyde, and
at various places in the Irish Sea. It is also common between -
tide-marks at Hilbre Island (Cheshire), Puffin Island, and at
the south end of the Isle of Man.

Canon Norman informs me that he has dredged the species
in the Bergen and Hardanger fjords in Norway, and believes
that it occurs much further north.

As the determination of the species of Aleyonium is a matter
that requires some care, all that can be said at present is that
there is no evidence that it occurs outside the area of the North
European coasts.

There are only two species of the genus Aleyonium found in

THE ANATOMY OF ALOCYONIUM DIGITATUM. 353

the British area, namely A. digitatum and A. glomeratum
(Pl. 36, figs. 1 and 2). The yellow variety of A. digi-
tatum referred to above cannot be separated as a distinct
species from the white variety. There are no features in the
mode of growth, character of the spicules, or general anatomy
which distinguish the two varieties, the only difference being
the yellowish tint in the spicules of the yellow one.

Alecyonium glomeratum was first discovered by Hassall
in 1841, and he referred it to Miller’s species Alcyonium
rubrum; but on being told by a Mr. Macgillivray that it
differed from A. rubrum, gave it the name Alcyonidium
glomeratum (14). (Throughout his descriptions Hassall uses
the word Alcyonidium in place of Aleyonium.) The species is
defined as follows—‘‘ Polypidoms massive, of no very definite
outline; colour a deep uniform red, the shade of which ap-
proaches to vermilion.”

In the second edition of Johnston’s ‘ Zoophytes,’ vol. i,
p. 178, there is a long quotation from Couch on A. glomera-
tum, who says, ‘‘The colour externally is of a deep blood-
colour, and internally is but slightly lighter. The lobes are
very numerous, and divide nearly as low down as the base.
The spicula are numerous and irregularly arranged ; they are
linear, elongate, pointed at both extremities, with uneven or
granular spaces between; sometimes they are simple, and at
others united into K-shaped bodies, and occasionally wanting
one or other of its members, forming an imperfect K.

When Gray examined the species in 1865 (12) he made it
into a new genus, giving it the name Rhodophyton Couchii.
The description of the genus is as follows:—‘‘ Coral flesh
cellular, covered with a hard calcareous coat, contracted at the
base, expanded above, and divided into several oblong lobes or
branches, covered with short cylindrical tubes with a circular
mouth. Polyps half retractile, forming when contracted a
white tubular termination to the cells. The more developed
cells of the polyps, especially those at the end of the lobes,
are longitudinally grooved.”

“This genus,” he continues, “differs from the typical

35 4 SYDNEY J. HICKSON.

Alcyonia or Lobularie, taking A. digitatum for the type, in
the outer surface being covered with a continuous crustaceous
coat, and in each of the polyps being enclosed in a distinct
tubular sheath projecting from the general surface. It differs
from all the Aleyonia in the polyps being only half retractile.”

I have examined a couple of specimens sent to me from the
Marine Biological Association’s laboratory at Plymouth, and I
am able to give a figure of the spicules (fig. 5). The descriptions
given by both Couch and Gray are fairly accurate, but, as I
shall point out presently, the condition of retraction of the
polyps is not a character upon which too much reliance can
be placed for purposes of classification, as it is one which
depends so largely upon the means employed to kill and pre-
serve the colony.

A. glomeratum (Hassall), then, differs from A. digitatum
(Linn.), in that the lobes are more pointed and more deeply
divided, the spicules are reddish in colour, and there are no
dumb-bell-shaped spicules.

Section IIJ.—Gernerat ANATOMY.

The colony of Aleyonium digitatum is composed of a
number of polyps which are fused together for the greater part
of their length, but have each a free extensible and retractile
portion which bears the mouth and tentacles. The mesoglea
of the larger portion of the body-wall of each polyp is con-
siderably thickened, and in direct connection with that of its
neighbours, so that no limits can be drawn in this region
between one polyp and another.

In most of the text-books of zoology it is customary to
restrict the term “ polyp” to the free extensible portions of
the individuals, and to give the name ‘‘ ccenosare”’ to the part
of the colony which is composed of their fused basal portions.
It cannot be too strongly insisted upon that this is an erroneous
conception of the structure of Aleyonium. ‘The dorsal mesen-
terial filaments of each primary polyp extend to the base of
the colony, the ventral mesenteries are visible as ridge-like

THE ANATOMY OF ALOYONIUM DIGITATUM. 35D

projections for a long distance down the ceelentera of the fused
portion of the polyps, and the endoderm shows no important
modification in that region. Consequently the coelenteron of
the fused portion of each polyp is no more “colonial” in
character than that of the free extensible portion. It is, in
fact, simply a part of the individual polyp.

A comparison with the genus Tubipora may render this
matter clear. In this genus the polyps remain separate from
one another for the greater part of their course, being con-
nected together only by the horizontal platforms and the basal
stolon. The only part of the colony that is, strictly speaking,
“ colonial,” or, in other words, common to more than one
polyp, are the stolon at the base and the horizontal platforms.
In these structures there runs a number of branching canals,
lined by endoderm, which establish a communication between
the polyps, and form a basis from which new polyps originate
as buds. In the case of the colonial Hydrozoa it is often a
matter of difficulty to determine where the ccelenteron of a
polyp ends and the ccenosarcal canal begins, and consequently
the distinction between polyp and ceenosarc is indefinite. But
in the case of the Alcyonarian there is no such difficulty. The
celenteron of a polyp of Tubipora, for example, can be distin-
guished from a canal in a horizontal platform at once by the
fact that it contains, either complete or in a shrivelled condi-
tion, the dorsal mesenteries and mesenterial filaments.

In Tubipora, then, there can be no doubt that the ccenosarc
is composed of the horizontal platforms and basal stolon, and
by these structures only.

In the case of Alcyonium the distinction between the polyps
and cceenosare cannot be defined. In consequence of the enor-
mous development of the mesoglea in this genus, the parts
corresponding to the polyp walls and horizontal platforms and
basal stolon of Tubipora are all fused together, and conse-
quently it is impossible to say where the polyp wall ends and
coenosarc begins.

The colony, then, is mainly composed of a number of polyps
partially fused together. Some of these—the primary polyps

356 SYDNEY J. HICKSON.

—extend from the surface right down to the base; but others
—the secondary polyps, which are formed during the later
stages of growth as buds from endodermic canals proceeding
from the celentera of the primary polyps—terminate at various
depths in the substance of the colony, according to the stage
of growth at which they were formed.

My first impression on examining the anatomy of the colony
was that the celentera of the secondary polyps fused directly
with those of the primaries, but I am now convinced that this
impression was an erroneous one. The secondary polyps
terminate in the mesoglea, but as the terminal end is usually
connected with neighbouring ccelentera by short open canals,
they cannot be said to terminate blindly (PI. 37, fig. 8).

When a thoroughly contracted specimen of Alcyonium is
examined the surface may be seen to be covered with a number
of rounded prominences, each bearing in its centre a star-
shaped depression. These prominences indicate the position
of the polyps, and are on an average 1‘5 mm. in diameter.
In a few instances I have measured prominences as much as
2 mm. in diameter, and several young buds are very much
smaller than the average dimensions given above.

On the upper free parts of the colony the polyps are very
closely packed, so that in the contracted condition there is
very little space between the protuberances; but on the sides
near the base of the colony the protuberances are frequently
separated by considerable intervals. In a specimen I obtained
some years ago off the coast of North Wales the base of the
colony is spread out like a membrane over a Serpula tube,
and in the outer portions, where the base is thinnest, the polyp
protuberances are as much as 1—3 mm. distant from one
another.

In a fully expanded colony the base of the extensible portion
of each polyp averages 2 mm. in diameter, but it is not so easy
to give the exact length from the mouth to the surface of the
colony, on account of the extreme sensibility of the polyps
during life.

It is quite probable that in the fully expanded condition, in

THE ANATOMY OF ALCYONIUM DIGITATUM. 357
their natural habitat, they stretch themselves out much further
than they do under any circumstances in the tanks of an
aquarium. All that can be said, then, about this measurement
is that, of the polyps expanded in the tanks of an aquarium,
the distance from the mouth to the surface of the colony does
not exceed 5 mm.’ Similarly it is impossible to give any
reliable figures as to the exact length of the tentacles, but, from
observations made upon numerous living polyps with a simple
magnifying glass, my impression is that, when fully expanded,
they are normally about three quarters the height of the
extended portion of the polyp, that is 3 mm.

Tn living Alcyoniums, as they are observed in an aquarium,
the tentacles are more often contracted than not, so that the
crown has the appearance of an eight-pointed star, as it is
figured in pl. xxxiv,! fig. 3, of the second edition of Johnson’s
‘ Zoophytes,’ and they will remain in this contracted condition
for a very long time together. It is only on exceptional occa-
sions that the tentacles expand themselves completely, so as
to show clearly their characteristic pinnate Alcyonarian appear-
ance.

When the colony is killed, expanded by sudden immersion
in Lo Biancho’s No. 2 chrom-acetic solution, and then trans-
ferred to methylated spirit for preservation, the zooids usually
lose their transparency, and become opaque, excepting at their
lower extremities, where the body-wall usually remains trans-
lucent. There is a good deal of difference, however, in the
degree to which they lose their transparency when preserved.
In some young colonies the body-walls remain perfectly trans-
parent, only the tentacles and stomodzum being opaque; and
occasionally the same transparency may be seen in the polyps
of the larger and older colonies.

In Alcyonium palmatum of the Mediterranean the ex-
panded zooids, when preserved, are as a rule transparent ; this
is due to the fact that the body-wall is much thinner than it
is in the British species.

‘ In the MS. notes on the anatomy of A. palmatum, by the late Professor
Marshall, I find 10 mm. given for this measurement.

358 SYDNEY J. HICKSON.

The degree to which the extensible portions of the polyps
contract on death is sometimes used as a character for the de-
termination of specific characters, but although it may be used
at times with advantage, it is one which should be used with
great caution. The group of spicules, which may be seen in
fig. 6 to extend from the surface of the colony a short distance
into the base of the extended portion, may be much more con-
siderable in some specimens than in others, and may afford a
physical obstruction to the complete retraction of the polyps.
I have observed, too, that when a colony has remained in
imperfectly aérated water the polyps will die expanded, as sea-
anemones do; and a colony thus moribund when placed in
spirit will present a large number of expanded polyps. The
condition of the retraction of the polyps of a spirit specimen
of an Alcyonarian may depend, then, not only on anatomical
characters, but also on the condition of vitality, or want of it,
at the time the colony was preserved. These facts should not be
ignored, as they too frequently have been, in the determination
of species and the descriptions of new ones.

Colour.—All the extended polyps of Aleyonium digita-
tum I have examined are perfectly white and transparent
when alive; but in the specimens obtained in Provence the
spicules of the Halskranse (crown) are, according to Vogt and
Jung, sometimes reddish in colour.

Tentacles.—The eight tentacles of Aleyonium digita-
tum are equal in size, and in the fully expanded condition
stand at right angles to the long axis of the polyps. They are
provided with a number of pinne on each side, which varies
from twelve to twenty-two.

It is not possible to count with accuracy the number of
pinne in contracted or preserved specimens. Some of the
proximal pinne are capable of being so completely withdrawn
that they are not visible except in the fully expanded condi-
tion. My numbers are given from the careful observation of
a great many fully expanded polyps in the Plymouth la-
boratory.

In his description of the tentacles of A. palmatum, von

THE ANATOMY OF ALCYONIUM DIGITATUM. 359

Koch (21) says, ‘ The tentacles possess twelve to fourteen pairs
of pinne (Fiederpaare).” In our English species it is quite
impossible to recognise the pinne as pairs. Not only do we
find an uneven number of pinnz on the two sides of many of
the tentacles, but a minute examination of the free end of the
tentacle usually shows a small pinna on one side without any
corresponding one on the other (see fig. 10).

Carl Vogt and Jung’s description (37) of the tentacles is as
follows :—“ Bei missiger Ausdehnung bieten sie die Form
langen Blumenblatter mit sageformigen Randern, aber in
gewissen Fallen verlangern sich alle Theile dermaassen, dass
die Fihler Hirschhérnern ahneln, welche auf beiden Seiten
mit sehr langen Zacken besetzt sind.”

I have never seen this “ antler” appearance of the tentacles
except in spirit specimens, and I am inclined to think it is
entirely due to the action of the preservative fluid, for in life
the tentacles appear to be perfectly regular (fig. 11).

The tentacles of fully expanded polyps when examined alive
with a low power of the microscope vary very considerably in
appearance. In Pl. 37, fig. 11, I have drawn the crown
of a polyp I observed in a recently caught specimen which
expanded in the bucket of sea water in which it was placed
immediately after its removal from thesea. The tentacles were
extremely extended, and at the base of each of them there was
a bullate swelling, on the side of which there were no pinne.
From time to time a sudden and simultaneous movement of
contraction would occur, caused perhaps by a slight shaking of
the table, and the tentacles would assume a more or less rigid
appearance, with the bullate portion very much extended, but
the terminal portions and the pinne very considerably short-
ened (fig. 6). These contractions lasted only a few seconds,
and then the tentacles extended themselves again and resumed
the condition represented in fig. 11.

It is but rarely, however, that it is possible to observe the
polyps so fully extended as this. A condition such as that
figured in fig. 10, in which there is apparently no bullate por-
tion, is not unfrequently seen, and it is quite possible to kill a

VOL. 37, PART 4,—NEW SER. BB

360 SYDNEY J. HICKSON.

certain number of polyps in this condition by the chrom-
acetic acid method.

Under the ordinary, somewhat unfavorable, conditions of life
in an aquarium the polyps do not usually expand more fully
than those represented in pl. xxxiv of Johnston’s ‘ Zoophytes,’
the body-walls being fully extended, but the tentacle lightly
contracted, forming an eight-pointed star on the summit of
the polyp (Pl. 37, fig. 9).

The condition shown in fig. 10, in which there is no bullate
basal swelling to the tentacles, may be called the first stage of
contraction.

The condition just described may be called the second stage
of contraction (fig. 9).

In the third stage (fig. 7,2’) the tentacles become bent on
themselves, and a ring-shaped collar folds over the crown,
partially hiding it.

In this condition the polyp contracts down to the level of
the surface of the colony. If they contract rapidly in response
to strong stimuli the body-wall is often very considerably dis-
tended; but if they contract slowly, either naturally or in con-
sequence of something distasteful in the water, they withdraw
below the surface of the colony without any alteration in their
diameter.

When the crown has withdrawn below the surface of the
colony the basal portion of the body-wall of the extensible
portion of the polyp closes over the hole, the stomodeum
becomes folded like a concertina (fig. 7,4’, St.), and the tentacles
become tightly drawn down.

In the last stage of contraction (fig. 7, 5’) the transparent
basal portion of the extensible portion of the polyp is drawn
down below the surface, and the aperture is tightly closed by
the opaque and densely spiculated crust of the colony, and at
the same time the tentacles, disc, and stomodzeum become so
tightly pressed together that it is difficult in sections to make
out their outlines.

Each tentacle is, when fully expanded, provided with a wide
lumen, as in all Aleyonaria. This lumen, which is a prolonga-

THE ANATOMY OF ALCYONIUM DIGITATUM. 361

tion of the intermesenterial space of the cclenteron, is con-
tinued into each of the pinne. In Alcyonium palmatum
the ectoderm of the pinne is thickened in regular patches, which
contain batteries of nematocysts. I have seen similar patches
on the pinne of our English species in favorable preparations.

The mouth varies in appearance very considerably. When
fortunate enough to be able to focus the microscope quite
vertically down on to a fully expanded polyp, the mouth is seen
to be a slit-shaped opening on a shallow ridge in the centre of
the flat oral disc. At one end it has a crescentic widening,
which indicates the position of the siphonoglyphe lower down
in the stomodeum. When the polyp is slightly retracted, as
is usually the case, the true mouth cannot be seen, since the
oral disc is depressed in a funnel-shaped manner, forming, as it
were, a secondary mouth, which is oval in outline (fig. 6).

The stomodeum of Alcyonium digitatum consists of a
short, somewhat flattened tube about 1 mm. in length, opening
to the exterior by a slit-shaped mouth, and into the ccelenteron
by an oval aperture.

On the ventral side there is a siphonoglyphe extending
through two thirds of the aboral end (fig. 14). The epi-
thelium lining the siphonoglyphe is considerably thicker than
it is in other parts of the stomodzum (see Section IV), and
provided with numerous long flagella.

The epithelium of the stomodeum other than that of the
siphonoglyphe is thrown into a series of six longitudinal folds,
which are rather more pronounced at the oral than at the
aboral end (Pl. 38, fig. 12).

In this respect the stomodzum of the polyps of A. digita-
tum show a marked contrast with that of the polyps of A.
palmatum. In all the sections of the polyps of this species
that I have examined the ectoderm of the stomodzum is quite
smooth, as is represented in pl. 1, fig. 2, of my former paper (16).

The mesenteries of Alcyonium digitatum present no
features that require special comment. As in all Alcyonarians,
they bear on their ventral faces the large retractor muscles.
The details of their histology are given in the next section.

362 SYDNEY J. HICKSON.

The mesenterial filaments may be seen through the trans-
parent body-wall of a fully expanded living polyp. The single
pair of dorsal filaments may be readily distinguished from the
other six by proceeding in almost straight lines from the edge
of the stomodzum down into the depths of the celenteron ;
and if they be traced further in sections of preserved speci-
mens, they may be seen to continue right down to the base of
the polyp,—that is to say, in the case of the primary polyps to
the base of the colony, and in the case of the secondary polyps
to that part of the endodermal canal system from which they,
as buds, arose (Pl. 37, fig. 6).

The four lateral and twoventral mesenterial filaments are very
short, and even in the most fully expanded polyps present a very
sinuous outline. It is difficult to say what their exact length is,
as I have found it impossible to measure them accurately in
the living polyp; but judging from the length of the tentacles,
and other measurements that I have obtained accurately, I
should say that they are never more than 2—2‘5 mm. long.

The gonads are spherical bodies, reaching in their mature
condition a diametrical measuremeut of aboutO‘5mm. When
ripe the ova are of an orange-red colour, the sperm-sacs milky
white. They are covered by a coat of endoderm continuous
with the endoderm of the lateral and ventral mesenteries, to
which they are attached by short stalks. They are invariably
attached either to the side or edge of the free portion of the
mesenteries below the termination of the mesenterial filaments,
and never extend for a distance of more than 10 mm.

The gonads do not occur, so far as my observations go, on
the dorsal pair of mesenteries.

As stated above, the ceelentera of the different polyps com-
posing the Alcyonium colony do not communicate with one
another directly, but running in the mesoglea between the
cavities there are canals, lined by endoderm, with a wide
lumen, by means of which the water can flow from the ceelen-
teron of one polyp into that of its neighbours. In the peri-
pheral portions these canals are fairly numerous, but they are
very scanty in the deeper parts of the colony, being almost

THE ANATOMY OF ALCYONIUM DIGITATUM. 363

confined to the basal portions of the secondary polyps and
the adjacent portions of their neighbours (fig. 8).

In addition to these canals, the mesogloea contains solid
rods or rows of cells, isolated cells and spicules (fig. 26, &c.).

The skeleton of Alcyonium is entirely composed of isolated
spicules of calcium carbonate.

The spicules occur only in the mesoglea, and are far more
numerous in the periphery, where they are densely crowded,
than in the deeper parts of the colony.

When an expanded polyp is examined with the microscope,
a group of spicules may be seen to extend some little distance
up the walls of the base of the extensible portion between the
mesenteries (fig. 6). In some polyps a few scattered spicules
may extend right up to the slight constriction just below the
base of the crown of tentacles. Immediately above this con-
striction there is almost invariably found a ring of scattered
spicules, which sends a radiating row a certain distance along
the aboral side of the tentacles.

When the crown is examined from above a ring of spicules
may be seen surrounding the mouth, and these send off lines
of scattered spicules along the sides of the oral surface of each
tentacle (fig. 10).

Section 1V.—Minure Anatomy.

The ectoderm of the general surface of the colony of
Alcyonium is very liable to become detached, either by friction
in the bottle in which the animal is preserved or during the
process of decalcification, and it is only in sections of speci-
mens that have been very carefully treated that the ectoderm
can be seen at all.

It consists of a number of columnar, spindle-shaped and
flask-shaped cells, 0°02 mm. long, connected at their outer
borders, but free from one another for the greater part of their
course. At the base of the epithelium there are a few spherical
interstitial cells of different sizes, and here and there may be
seen a cell which, like a ganglion cell (fig. 17), is irregularly
star-shaped.

364 SYDNEY J. HICKSON.

I have examined the ectoderm very carefully,to determine
whether it is ciliated or not; but neither in the living animal
nor in my best preparations can I find any trace of cilia on
any of the cells, either on the general surface of the body, the
extensible portion of the body-wall, the tentacles, or the oral
disc.

This is not inconsistent with the work of other observers,
for in no other work on the minute anatomy of Alcyonarians
can I find the external ectoderm! described as being ciliated.
This is a noteworthy point, because in the Actinie, according
to the Hertwigs and other observers, the ectoderm is ciliated.
In the Hydrozoa the ectoderm is never ciliated, so that in this
respect the Alcyonaria apparently agree with the Hydrozoa
and differ from the sea-anemones.

It is difficult to determine with certainty the origin of the
cells that give rise to the spicules; but, for many reasons, I
am inclined to agree with von Koch’s (23) results on Gorgonia
and Clavularia, and attribute them entirely to the ectoderm.
Among the interstitial cells of this tissue one frequently finds
large spherical cells which lie beneath their neighbours, and
cells very similar to these may be seen isolated in the subjacent
mesogloea. Moreover large cavities may be seen in isolated
cells in the same region, which, in all probability, contained a
spicule before decalcification (fig. 17, sp. c.).

Attention has already been directed to the fact that the
spicules are far more numerous at the periphery than in the
deeper parts of the colony. This suggests very forcibly that
the spicules are only formed at the periphery, and that with
the growth of the mesogloea they become more and more sepa-
rated from one another. It is not likely that during the growth
of the colony they are dissolved.

1 Von Koch (28) says that in the Alcyonaria the ectoderm-cells are
provided with cilia, which often disappear in old colonies; but he gives
no figure to illustrate this ciliation. Kolliker (24) says that the ectoderm
of Pennatulids shows ciliation, but he was unable to determine its extent
in the specimens he examined, and Prof. Marshall did not describe nor
figure it.

THE ANATOMY OF ALCYONIUM DIGITATUM. 365

The ectoderm is almost invariably covered by a coat of
transparent mucus, in which grains of sand, minute alg, and
other foreign bodies occur. This mucous coat is probably
formed from the secretions of some of the flask-shaped cells.

With the possible exception of an occasional star-shaped
interstitial cell I can find no trace of a “ Nervenschicht” in
the general ectoderm of the colony, nor can I find true mus-
cular processes to any of the cells, nor any cnidoblasts or
nematocysts.

The ectoderm-cells of the tentacles are very similar in
general appearance to those of the general surface of the colony,
but they are not quite so long,i.e. about 0:015 mm. They
are prolonged below, however, into well-marked muscular pro-
cesses, and between them may be seen numerous cnidoblasts and
nematocysts. A nerve-sheath is very distinctly present in the
form of minute star-shaped and round ganglion-cells connected
with a plexus of very delicate anastomosing fibrils (fig. 22).

The ectoderm of the oral disc and body-wall is not so easy
to investigate, as it becomes detached or partly so with the
slightest friction. It seems to vary very considerably in thick-
ness according to the condition of expansion of the polyp,
but in other respects it does not differ in any very marked
characters from that of the general surface of the polyp.

The ectoderm lining the stomodzum is composed of closely
packed columnar cells, 0°017 mm. in length, provided with short
cilia (fig. 14), asin Pennatulids, Clavularia, and A. palmatum
(Wilson). The epithelium of the siphonoglyphe differs from
that of the other parts of the stomodzum in that the cells are
narrower, somewhat clearer, considerably deeper (0°025 mm.
in length), and bear long (0°025 mm.) whip-like flagella
(fig. 29).

The nematocysts of Aleyonium are extremely small (0°0075
mm.), and all of one kind. They may easily be seen by
snipping off a tentacle of an expanded polyp and examining it
with a microscope, when, on the addition of a little acid, they
may be seen to shoot out their threads, and occasionally get
loose from the ectoderm and float off freely in the surrounding

366 SYDNEY J. HICKSON.

water. They seem to swell considerably in dilute acetic acid,
so that some of the free nematocysts measured as much as
0-015 mm. in length, but I do not think they are ever as large
as that in their natural position inthe ectoderm. The filament
is long, easily broken, and is not ae, so far as I could
discover, with any barbs.

In sections of preserved specimens the nematocysts are not
very easily observed, as they are usually shot and lost during
the process of killing the colony in an expanded condition ; but
they may be demonstrated by staining sections of colonies
killed contracted in spirit, or any other preservative, in eosin
and hematoxylin. By this method the nematocysts are stained
blue and the ectoderm-cells pink.

The extremely small size of the nematocysts of Aleyonium
is not by any means exceptional in the case of the Alcyonarians.
Moseley stated that in Sarcophyton there are no nematocysts,
and the late Professor Marshall and his brother (29) were un-
able to find them in Virgularia mirabilis. In Pennatulids
generally, according to Kolliker (24), they are exceedingly
minute, and their presence could not always be determined in
spirit specimens.

Wilson (38) says that in the endodermic mesenterial filaments
of Paraleyonium there are “scattered irregularly through the
filaments minute nettle capsules. They are remarkable for their
very small size, being smaller than the nuclei of the endoderm-
cells. They have an oval form, and each contains a spirally
coiled filament. In the minuteness and rarity of the nettle
capsules the mesenterial filaments of the Alcyonaria differ from
those of the Actinians, and it seems possible that in the former
group they are to be regarded as rudimentary organs.”

The figure given by this author of the nematocysts of the
endodermic filaments on Paraleyonium is the only figure yet
published of the nematocysts of Alcyonarians.

The Mesenterial filaments.—I have nothing new toadd
to Wilson’s excellent account of the structure of the dorsal
mesenterial filaments.

The dorsal mesenterial filament consists of columnar ciliated

THE ANATOMY OF ALCYONIUM DIGITATUM. 367

cells, “arranged so as to form a long solid band on the edge of
the septum (i.e. mesentery). As seen in sections this band
has a bilobed form, or, in other words, a longitudinal groove
runs along its middle” (fig. 18). I have never seen the Y-shaped
appearance of this groove in section which Wilson describes in
Funiculina, but in several sections I have seen the lobes folded
over so as to form an almost complete canal (fig. 16).

The cells of the groove are in appearance and in detail
similar to those of the stomodeum.

As regards the ventral mesenterial filaments my observations
do not coincide in all respects with Wilson’s figures and descrip-
tion of these filamentsin Paraleyonium. Like Wilson, I recog-
nise that there are essential and very important differences
between these filaments and the dorsal ones. There is no
median groove, there is no sharp line of demarcation between
the cells of the filament and the neighbouring endoderm-cells
of the mesenterial epithelium, and the cells are not all of the
same kind, some staining very much more deeply than others
(fig. 19). Two distinct kinds of cells may be made out in good
preparations with a high power : first, large non-ciliated gland-
cells, which stain deeply with hematoxylin; and second,
elongated columnar cells filled with numerous minute granules.

The mesoglea of Alcyonium is composed of a thick homo-
geneous substance containing (1) endodermic canals ; (2) solid
cords of endoderm; (3) minute isolated cells connected with
one another and with the endoderm by fine anastomosing fibrils ;
(4) spicules of calcium carbonate.

(1) The endoderm canals—that is to say, cords of endo-
derm containing a lumen which communicates with the
celenteron of one or more of the polyps—are chiefly found in
the superficial parts of the colony. They consist of a single
layer of endoderm-cells embracing a narrow lumen, which
becomes obliterated in some of the branches (figs. 13, 17).
The canals then are continuous with—

(2) The solid cords of endoderm which are- found
throughout the mesoglea of the colony, and may be easily
studied by osmic acid or methyl blue preparations of fresh sec-

3868 SYDNEY J. HICKSON.

tions. The cords are in some cases fairly compact, resembling
a canal in all respects except the presence of a lumen; but in
others the cells are only loosely connected with one another,
become elongated or star-shaped (fig. 26), giving off fine fibrils
at their angles. There may be only a single row of oval or
cubical cells, or in some cases the row may be drawn out into
a chain of elongated spindle-shaped cells (fig. 27).

(When studying the mesoglea of Aleyonium last December
I noticed a number of oral bodies lying in and among the cells
[fig. 36, gr.] of the endodermic cords. They may be readily
distinguished from the endoderm-cells by their dark but homo-
geneous appearance. In one or two instances I have suc-
ceeded in making out a somewhat irregular body in the centre,
which may be a nucleus. It is difficult to say with any degree
of certainty what these bodies are, but it is possible that they
may be some form of parasitic sporozoon.)

(3) The isolated mesogloea cells may be seen in osmic acid
and other preparations of fresh sections. They are star-shaped
or bipolar, and are connected with one another by fine fibrils,
which branch and anastomose in the homogeneous mesoglea
(fig. 26).

These cells are connected not only with one another, but
also with some of the cells of the endoderm cords, and fine
fibrils from them may also be seen passing into the endoderm
lining the celentera.

This account of the histology of the mesoglcea differs in many respects from
that given by Carl Vogt and Jung. These authors describe numerous small
canals on the growing branches of the colony which they call the ‘ Nahr-
candle.” These end blindly at the periphery, and are continuous with the
ceelentera. These ‘“ Nahreanile” are probably the same as the “endoderm
canals” of my description.

In addition to these, the authors describe in the substance of the ecenen-
chym (mesogloea) a network of canals, “ Sammelcanile,” and capillaries. It
is noteworthy that in none of the figures to which reference is made is any
lumen drawn in the “ Sammelcanale” or capillaries, nor is there any indication
in the text that the authors ever saw any lumina in these so-called canals.
They are probably the same, then, as the solid endoderm cords of my account,
but quite erroneously described. The network of fine fibrils has not been
previously described, and was overlooked by Vogt and Jung.

THE ANATOMY OF ALCYONIUM DIGITATUM. 369

I have searched in vain for anything corresponding to the muscular fibres
of Vogt and Yung’s account, and I am quite convinced that neither in Al-
cyonium digitatum nor in A. palmatum are there any muscular fibres in
the mesogloea. It is quite possible that some of the cells lying in the meso-
gloea may be to a certain extent contractile, but they are quite different in
appearance from the muscular processes of the endoderm-cells lining the ccelen-
tera, which will be described later on.

The chemical characters of the homogeneous substance which
forms the bulk of the mesogloea are described in a paper by
Brown, which appears in this number of this Journal.

Spicules.—The spicules of Alcyonium digitatum
present so many varieties of form, that it is not possible, without
writing a long treatise on this subject, to do more than give a
description of a few examples.

The general impression obtained by examining a slide of
spicules, made by boiling a small branch of Aleyonium
digitatum in potash, is that no two spicules in the field ot
the microscope are alike.

The majority, however, are somewhat like a dumb-bell in
shape (fig. 3, a, 6, c), with numerous irregular and blunt
projections at each end, and about 0°1 mm. in length. In
addition to these there is a considerable number of larger
(0°:2—O0'3) spicules, which are branched (d) or shaped like a
capital K, or simply crosses with numerous blunt projections.
A few spicules may be seen scattered about among the more
characteristic forms which are quite irregular in shape, some
with a central plate-like centre from which spring a few short
branches (fig. 3, f), some with two or three long arms and
one or two short ones, some very small ones like spheres
covered with irregular tuberosities, and others of many
different shapes and sizes.

In all preparations such as the one mentioned, in which the
tentacles, the disc, &c., of the polyps are boiled down with
the other part of the colony, there may be seen a certain
number of long unbranched lancet-shaped spicules (fig. 42, e)
covered with irregular tuberosities. They vary very much in
shape, some being like a thick pin (without its head), others

370 SYDNEY J. HICKSON.

lancet- or spindle-shaped, and others again slightly curved like
a boomerang. These long unbranched spicules occur chiefly in
the tentacles and disc of the polyps. They do occur in other
parts of the colony, but I do not remember to have seen any
dumb-bell-shaped spicules in the extensible portion of the
polyps.

When one examines a similar preparation of the spicules of
the yellow variety of Aleyonium digitatum, one notices
precisely the same forms in very much the same proportions.
In fact, if it were not for the faint yellow colour which they
all possess, it would be impossible to distinguish them from the
spicules of the white variety (fig. 4).

The figures were drawn by Mr. Wilson, from preparations
in my collection of the spicules of the yellow (fig. 4) and white
(fig. 3) variety respectively ; the specimens were selected by
him quite at random from a very large number on each slide.

A preparation of the spicules of the other English species,
Alcyonium glomeratum, presents many features of
marked difference from those of the yellow and white varieties
of A. digitatum. In the first place, the colour is of various
shades from pale pink to blood-red. The majority of the
spicules are elongated needles and spindles, and there is an
entire absence of the small dumb-bell-shaped forms, very few
K’s and crosses, and there are several club-shaped forms
(fig. 5), which I have never seen in any preparations of
A. digitatum.

While, then, I have found it impossible to distinguish the
preparations of the spicules of the varieties of A. digitatum
by any other characters than the colour, a preparation of the
spicules of A. glomeratum can be easily and immediately
recognised by the numerous characters given above. It is
true that many of the elongated and spindle-shaped spicules of
A. glomeratum are almost exactly the same shape as the
spicules of the tentacles and disc of the polyps of A. digi-
tatum, but the clubs are peculiar to it, the dumb-bell absent,
and the K’s and crosses very rare.

Alcyonium digitatum is not a favorable form to take

THE ANATOMY OF ALCYONIUM DIGITATUM. 371

for the study of the development of the spicules, as it is a
matter of very great difficulty to make a thin section of the
surface of the colony before decalcification.

We have two very different accounts of the development of
these structures in Alcyonaria. According to von Koch
(20) the spicules of Clavularia prolifera are formed
within and continue to grow within a cell which is given off
from the ectoderm, the young spicules being always enclosed
in a nucleated protoplasmic sheath. In my decalcified sections
I have seen these sheaths (fig. 17, sp. c.), which, I imagine,
contained young spicules, and I have also seen cells
containing a small vacuole lying close under the ectoderm.
I believe, therefore, that von Koch’s account is substantially
correct.

Schneider (35), on the other hand, if I understand his
description and figures correctly, believes that in Aleyonium
acaule one or more cells gradually assume the shape of the
future spicule, and then become calcified. I have seen nothing
in my preparations to support this view.

The Nervous System.—In longitudinal sections of the
tentacles a number of minute cells may be seen lying between
the epithelium and the layer of muscular fibres. These cells
are spindle-shaped, triangular, or star-shaped, and are con-
tinuous with a fine plexus of fibrils, some of which have a
beaded appearance (Pl. 38, fig. 22 n. p.).

Whether this system of cells and fibrils is really nervous in
nature and function has not been determined experimentally,
but it is undoubtedly homologous with the layer called
*Nervenschicht ” by the Hertwigs in the ectoderm of the
Actinie.

A similar plexus of cells and fibrils may be seen in the
endoderm, when suitably prepared sections are made which
cut the celenteric tubes longitudinally and slightly obliquely.

I have represented in fig. 83 a portion of a tube that has
been cut somewhat obliquely, so that in the upper part of the
section the endoderm-cells may be seen, but in the lower part
these have been shaved off, and only the circular muscular

Bp SYDNEY J. HICKSON.

fibrils remain. If the part marked with an asterisk be
examined with a high power, the plexus represented in fig. 35
will be visible.?

We have evidence, then, of the existence of a nervous plexus
in both the ectoderm and endoderm of Alcyonium. It has not
been found possible at present to trace it in the ectoderm and
endoderm of all parts of the body.

The occurrence of a fine plexus of cells and fibrils in the
mesogloea was noted in a previous publication (15); and
although, owing to technical difficulties, this mesogloeal nervous
plexus has not been proved to be connected with the nervous
plexus of the ectoderm, it is extremely probable that such a
connection does exist. There is no difficulty in tracing its
connection with the endodermic nerve plexus.

The nerve plexus of the mesogloea is represented in fig. 26.
The figure was drawn from a section through a part of the
colony about one inch from the surface. The difficulty of
tracing it nearer to the ectoderm is due to the fact that the
only satisfactory preparations showing the plexus are made by
cutting sections of the living colony as thin as possible with
the free hand, and treating them immediately with osmic or
pyrolignic acid. The superficial regions of the colony are,
however, so densely crowded with spicules that it is impossible
to get them in very thin sections. Sections of decalcified
specimens which have been made with a microtome do not
show the nerve plexus very satisfactorily.

Muscular System.—The walls of the ccelenteron of
Aleyonium are provided with a complete sheath of endodermic
circular muscular fibres (fig. 33). This sheath is probably
used during life for causing the expansion of the polyps. Any
contraction of the circular fibres would cause an increased
water-pressure in the cwlentera and force the polyps out; but
when the colony is disturbed or artificially wounded, and the
great mesenteric retractor muscles are contracted, it forces

1 Herdman (17) has described a number of fusiform, polygonal, and trian-
gular cells in the endoderm of the mesenteries of Sarcodictyon.

THE ANATOMY OF ALCYONIUM DIGITATUM. 373

water out of the tubes, and causes a considerable contraction
of the whole colony.

If one takes an expanded colony out of the water one notices
that, in the first place, the crown and free portion of the
polyps are withdrawn ; but when these are completely concealed
the colony undergoes still further contraction in size, and water
may be seen to be oozing from openings in the centre of the
star-shaped tubercles that mark the position of the polyps.
This contraction may continue until the colony is reduced in
size by about one third of its original bulk, and it is then a
great deal harder to the touch than it was when fully expanded.
Specimens of Alecyonium in museums and laboratories that
have been killed slowly by immersion in weak spirit are always
very much contracted in this way, and they offer a marked
contrast in consistency to the soft spongy specimens which
have been killed suddenly by Lo Biancho’s No. 2 chrom-acetic
acid solution.

A similar contraction of the colony may be brought about by
rolling the specimen about in a bucket of sea water.

This circular muscular sheath may act, then, either as a
protractor of the polyps, or, in extreme cases, as a constrictor
of the colony.

The great retractor muscles of the polyps are situated, as in
all Alcyonarians, on the ventral faces of the mesenteries. In
the regions where the muscles lie, the mesoglea projects in the
form of simple or branched lamelliform folds. The muscle-
fibres are situated on these folds (fig. 21), and are covered by
an epithelium of endoderm-cells.

In most of my sections it is not easy to distinguish between
the muscle-fibres and the processes of mesogloea, as both of them
are perfectly homogeneous; but a double stain obtained by
hematoxylin and eosin differentiates the two tissues most beauti-
fully, the mesogloea staining blue and the muscle-fibres pink.

The muscles run longitudinally on the mesenteries, but
approach nearer and nearer to the stomodeum in their passage
from below upwards (fig. 14), but their eventual insertion is on
the disc close to the mouth.

374 SYDNEY J. HICKSON.

On the opposite face of the mesenteries the muscular
processes of the endoderm-cells run transversely to the long
axis of the polyps. These may be seen quite clearly in any
good series of sections taken longitudinally through an expanded
polyp. They constitute what has been called the protractor
muscular system of the polyps (figs. 14 and 21, P. msc.).

In the disc there is a circular band of muscle-fibres situated
in the ectoderm, and from this band there run fibres along
the borders of the inner surface of each tentacle. These fibres
undoubtedly act as the retractors of the tentacles (fig. 10).

The Endoderm.—The method adopted for examining the
minute structure of the endoderm of Alcyonium was to place
some thin strips of a colony into } per cent. osmic acid for
thirty minutes, then wash in distilled water and transfer to
picro-carmine for thirty minutes. The walls of the ccelenteric
canals were then scraped with a fine scalpel, and the endoderm
thus removed teased up with needles in glycerine. A modifi-
cation of this method, which yielded some good results, was to
substitute a weak solution of gentian violet for the picro-
carmine.

The cells lining the ceelenteric canals are by these methods
seen to be very similar in their general characters to the myo-
epithelial cells (fig. 28).

The protoplasmic portion contains a number of coarse
granules, which stain brown with osmic, and have strong
affinities for gentian violet. They are usually irregular in out-
line, but, owing to their very minute size, it is not possible to
assert that they have crystalline shapes. The granules are
probably of the same nature as the nutritive spheres which
occur in some of the endoderm-cells of the Hydrozoa.

A few cells present one or more large clear vacuoles.

The nucleus is round, and stains faintly in picro-carmine.
A careful and prolonged examination of these nuclei with the
highest powers at my disposal failed to reveal anything beyond
a clear homogeneous structure. I have never seen any stages
in the division of the nucleus.

Most of the cells show a single very delicate flagellum on

THE ANATOMY OF ALCYONIUM DIGITATUM. 375

their free ends. Itis possible that those cells (such as fig. 28,
a, ¢,0,n) which do not show a flagellum may have lost it during
the preparation and maceration of the specimen.

The muscular portion of the fibre is homogeneous and trans-
parent. It is usually branched at the extremities. Although
most of the cells possess a muscular portion, there are un-
doubtedly some cells in the endoderm which do not. The cell
represented in fig. 28 fis a type of cell very frequently met
with in these maceration preparations. It possesses two long
protoplasmic processes.

A few cnidoblasts may be seen in such preparations of the
endoderm-cells, and are represented in fig. 28, A and k;
but it is very probable that they have been swallowed with the
prey, and are not of endodermic origin.

Although the study of macerated preparations reveals many
points of great interest which cannot be seen satisfactorily in
any other way, it has the disadvantage of giving no evidence of
the true relationship of the cells to the organs of the body.

The endoderm is not precisely the same in all parts of the
celenteron, and consequently it is necessary to supplement our
study of macerated specimens by a study of carefully prepared
sections through different parts of the body.

In the lower part of the ceelenteric tube the endoderm has a
very vacuolated appearance. ‘This is due, not to the occurrence
of intra-cellular vacuoles, but to great gaps between the endo-
derm-cells.

The individual cells are thinner than they are in other parts
of the body, and loosely connected together in groups which
are separated from one another by considerable rounded
spaces.

The character of the endoderm in this region is very similar
to that described by Lankester as occurring in the proximal
third of the gastric tube of Limnocodium (27).

If sections be taken through the upper parts of the tube in
the region of the ventral mesenterial filaments, the endoderm-
cells (fig. 23) are seen to be much more closely set, and to con-
tain larger granules.

VOL. 37, PART 4.—NEW SER. Gc

376 SYDNEY J. HICKSON.

A few round interstitial cells may be seen in the endoderm
in nearly all good sections through this region. These cells
are of different sizes, and probably replace and supplement
the other endoderm-cells during the animal’s life.

The endoderm covering the mesenteries is composed of
short cubical cells which become flattened and fusiform over
the muscular ridges on the ventral faces of the mesenteries ;
but the general appearance of this part of the endoderm, as well
as that just described in the last paragraph, varies very much
with the condition of contraction or expansion of the polyps.
When the polyps are fully expanded, it is a thin flat layer of
cells ; but when they are contracted the cells are long, thin, and
so crowded together that, in sections which do not pass through
it quite vertically, it has the appearance of consisting of two
or three layers.

The sheath of endoderm which covers the gonads and the
endoderm forming the six mesenterial filaments are not affected
in this way by the condition of the polyps. In the former case
the cells are cubical in shape, finely granular in consistency,
without very distinct walls, and without muscular processes
(fig. 25).

The sections of partially retracted tentacles that I have made
show that the endoderm in that condition is a compact tissue
composed of short cubical cells with round nuclei. I can find
no muscular processes connected with any of the cells, and I
am convinced, after a prolonged study of these cells, that they
are not present. Each cell is very finely granular. There are
no nematocysts. The lumen of the tentacles is wide, and it is
continued for a short distance into each of the pinne (fig. 10).
The endoderm of the pinne is similar to that lining the axial
cavity of the tentacles.

The sexual cells first make their appearance in the thickened
border of the ventral mesenterial filaments. In both sexes
they occur as a mass of polygonal cells (fig. 37) covered by a
single layered sheath of endoderm. There can be no doubt
whatever of their endodermic origin.

Odgenesis.—If a series of sections be made through a

THE ANATOMY OF ALCYONIUM DIGITATUM. 377

female colony collected in the month of April or May, several
of the mesenteries may be seen to bear close to their free
margins a group of small, clear, round or polygonal cells. In
the smaller groups the cells are all of approximately the same
size, but in some of the larger groups a few of the cells are
very much larger than the others, and contain two instead of
one nucleus (fig. 87). An examination of such a group with an
immersion lens shows that the nuclei are of two different
kinds, and take up the staining reagents differently. Some
present a well-marked nuclear membrane, and a number of
highly refracting chromatin granules giving the characteristic
appearance of a resting nucleus ; the others are perfectly homo-
geneous, staining deeply in hematoxylin, and presenting no
granules, nucleoli, or other structure.

In the cases of the larger cells which present two nuclei
one of them is invariably of the former description, and the
other of the latter.

The explanation of these facts seems to be that some of the
young ova—namely, those containing the homogeneous nuclei
—cease to grow, and divide at an early stage in their develop-
ment, and these become absorbed by the others which con-
tinue to grow.

At the end of May, and in the months of June and July,
ovaries may be found on the mesenteries containing one or two
ova very much larger in size than the others, and in some of
these young ova may be seen partially disintegrated (fig. 38).

There can be no doubt, then, that some of the young ova
formed in the ovary are swallowed up by the larger ones;
others, however, simply degenerate, and their shrivelled
remains may be seen between the more mature ova.

During the later months of the year the ova continue to grow,
and assuming a light yellow colour in August, they reach their
full size and deep red colour at the end of November or the
beginning of December.

The ripe ova in the follicles of a female Alcyonium in
December are about 0°5 mm. in diameter, but I have noticed
a very considerable variation in the actual size of the ova

378 SYDNEY J. HICKSON.

when spawned. The germinal vesicle of such an ovum is
extremely large, its diameter being about a quarter of that of
the ovum, and is situated quite at its edge. In some cases a
large germinal spot may be found, but more frequently it is
absent. The vesicle contains a dense network of fibrils and a
considerable number of granules, which stain very deeply in
carmine, hematoxylin, Léffler’s blue, and other stains (fig 40).

The protoplasm of the ovum is filled with a larger number
of spherical or irregular globules and granules of yolk, which
turn perfectly black when treated with any reagent which con-
tains osmic acid.

Spermatogenesis.—The very young testis of Alcyonium
cannot be distinguished from the very young ovary. It con-
sists simply of a number of small round cells on the mesen-
teries covered by a layer of endoderm-cells. It soon shows a
difference in that the cell outlines completely disappear, and
the testis has then the appearance of a dense crowd of nuclei *
(fig. 42).

The nuclei increase immensely in numbers during the
summer months, but, although I have made many series of
sections, I have not satisfied myself of their mode of division.
At present I have not been able to discover any signs of
karyokinesis. | When the testis has reached a certain size,
about 0:1 mm. in diameter, a space free from nuclei filled with
an irregular coagulum makes its appearance in the centre
(fig. 43), The coagulum shrinks somewhat in the later months
of the year and disappears entirely in the ripe testis, which is
simply filled with spermatozoa.

The spermatozoa when mature consist of head with a cone-
shaped anterior end, followed by a spherical body and a long
flexible tail (fig. 45 a).

On making a teased preparation of a ripe testis, a number
of spermatozoa may be seen bearing just behind the head a
clear vesicular body, similar to that described and figured by
Pictet (33) in Strongylocentrotus and other Invertebrates.

The details of the formation of the ripe spermatozoon from
the spermatids I have not at present been able to follow.

THE ANATOMY OF ALCYONIUM DIGITATUM. 379

Gemmation.—The buds of Aleyonium digitatum arise
in the canals near the surface of the colony, which run in the
mesogloea between the celentera of the polyps. There seems
to be no definite law governing their appearance, as they
appear quite irregularly between the older polyps in all stages
of the growth of the colony. On referring to Marshall’s
manuscript noteson thegemmation of Aleyonium palmatum
I find a similar remark with regard to that species. The first
rudiment of the bud is a diverticulum from the canal in the
direction of the surface of the colony. The cells lining this
diverticulum proliferate rapidly, and it is usually found to be
almost filled up with loose cells. The diverticulum increases
in size and becomes pouched. In the older buds there are
undoubtedly eight of these pouches, but I have not been able
to follow the order of their formation (fig. 46). The intervals
between the pouches become the mesenteries of the polyp.
At the time of the formation (fig. 46) of the pouches an
invagination of the superficial ectoderm immediately above the
bud rudiment. This ectodermic invagination sinks down, and
opens eventually into the centre of the endodermic diverticulum
from the canal (fig. 47).

The later stages are very much more difficult to follow, as it
is extremely difficult to obtain sections that run exactly
parallel with the central axis of the young polyp. In very
young colonies—and these are the most favorable for the study
of gemmation—the axis of the bud rudiment is by no means
parallel with the axis of the neighbouring polyps, and conse-
quently it is impossible to orient the colony before embedding ;
and in older colonies the young buds occur at such rare
intervals that hundreds of slides of sections may be hunted
through without furnishing abud in such a conditionas will lead
to any important results. There can be no doubt, however,
that the epithelium of the stomodzum is formed from the cells
of the ectodermic invagination.

As in Kenilla, so in Alcyonium, the two dorsal mesenterial
filaments are the first to be developed, and for a long time
apparently these filaments are the only ones to be found in the

380 SYDNEY J. HICKSON.

buds. The epithelium of the central portionof these filaments
is precisely the same as that of the stomodeum, the cells being
smaller, and staining more deeply in borax carmine than the
endoderm-cells, so that there can be little doubt that Wilson
(88) is correct in saying that (in Funiculina) the dorsal
mesenterial filaments are of ectodermic origin.

The tentacles do not appear until quite late in the develop-
ment of the bud,—not,in fact, until after the formation of the
ventral mesenterial filaments, which arise as thickenings of the
endoderm at the edge of the mesenteries.

Section V.—NoreE ON THE CIRCULATION OF THE FLUIDS IN
THE PoLyPs.

In a communication I made to the Royal Society in 1882, I
pointed out that the action of the cilia in the siphonoglyphe of
Aleyonium was to create a current of water flowing from the
mouth into the celenteron. Almost simultaneously, Wilson
discovered that the cilia of the dorsal mesenterial filaments
produce a current flowing in the opposite direction.

During the month of December I examined these currents
again with great care, and I can now confirm not only my own
results, but also those of Wilson.

When the polyps of Aleyonium digitatum are fully
expanded, the long cilia of the siphonoglyphe are in active
movement, causing a swift current of water to flow into the
celenteron on to the six endodermic—that is to say, ventral—
mesenteries. A reverse current from below upwards towards
the mouth is produced by the cilia of the two long dorsal or
ectodermic mesenterial filaments.

The current produced is a tolerably swift one, and must
cause a rapid distribution of any soluble products of the diges-
tion caused by secretion of the gland-cells of the mesenterial
filaments, and there can consequently be little doubt that the
filaments are not alone the absorbents of the food materials in
solution. These must be carried down by the ciliary currents
to the lower parts of the celenteric tube, and absorbed by the
endoderm in any region of the tube.

THE ANATOMY OF ALCYONIUM DIGITATUM. 381

This appears to be the case from the following considera-
tion:—The dorsal mesenterial filaments of the primary, 1.e.
the oldest, polyps of the colony extend right down to its base ;
and, so far as I could observe, there is no diminution of the
vigour of the current produced by them in the lower regions
of the tube. If there is a current of water in one direction at
the side of a tube which is closed at one end, it is obvious
that there must be a current in the reverse direction on the
opposite side. The endoderm of the lower parts of the tube,
although provided with large intercellular spaces as described
above, is nevertheless living and active. (In Tubipora and
Clavularia viridis the endoderm dies in the lower ends of
the tubes.) I have no doubt, then, that the endoderm of all
parts of the ceelenteron can and does absorb nourishment in
the fluid form.

There is no evidence that particles of food are taken up in
the solid form by the endoderm of Alcyonium. I have not
been able to observe any solid particles of food other than
the nutritive spheres in any of the endoderm-cells I have
observed, either in sections or in teased preparations. I have
seen no diatoms or other unicellular alge or fragments of
animals embedded in the endoderm. Particles of carmine
suspended in the water are, however, readily seized by these
cells wherever they occur; and in some cases I have seen
them embedded in the cells of the mesoglea, indicating that
they may be handed on from cell to cell throughout the
system.

LITERATURE.

ae

BastEerR, Jos.—‘ Natuurkundige Uitspanningen,’ eerste deel, p. 26, 1762.

i)

. Berxennout, J.—‘ Synopsis of the Natural History of Great Britain
and Ireland,’ i, p. 212, 1795.

8. Bosc, L. A. G.—‘ Histoire naturelle des Vers,’ iii, p. 151, 1827.

4. Coucn, J—‘ A Cornish Fauna,’ vol. iii, p. 58, pl. xiii, 1, 1844.

5. Danreissen, D. C.—‘ Norske Nord-Havns’ Expedition, Aleyonida,’ 1887.
6. Ex.is, J.—‘ Essai sur |’ Histoire naturelle des Corallines, &c.,’ 1756.

382 SYDNEY J. HICKSON.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.
24.

25.

26.

27.

. Exits, J.— An Account of the Sea-pen, &e.,” ‘Phil. Trans.,’ liii.

. Exus, J., and Sotanprr, D.—‘ The Natural History of many Curious

and Uncommon Zoophytes,’ 1786, pl. i, fig. 7.

. Mityz-Epwarps.—“ Recherches anatomiques, physiologiques, et zoolo-

giques sur les polypes,” ‘Ann. Sci. Nat.,’ 2°, iv, p. 321.

. Fasricius, O.—‘ Fauna Greenlandica,’ 1780.
. Fremine, J.—‘ A History of British Animals,’ 1828.
. Gray, J. E.— Notice on the Rhodophyton, a New Genus of the Alcyo-

niade found on the Coast of Cornwall,’ ‘Proc. Zool. Soc.,’ 1865,
p. 105.

Hassatu.—* Supplement to a Catalogue of British Zoophytes,”’ ‘Ann.
Nat. Hist.,’ vii, p. 285, 1841.

Hassatt.—“ Remarks on Three Species of Marine Zoophytes,” ‘ Ann.
Nat. Hist.,’ xi, p. 112, 1848.

Hicxson, 8. J.—‘‘ Some Preliminary Notes on the Habits and Anatomy
of Aleyonium digitatum,” ‘Proc. Camb. Phil. Soc.,’ vii, p. 7.
Hicxson, 8. J.—‘* The Ciliated Groove in the Stomodeeum of the Alcyo-

narians,” ‘ Phil. Trans.,’ pt. ili, 1883.

Herpmay, W. A.—‘“‘On the Structure of Sarcodictyon,” ‘Proc. Roy.
Soc. Edin.,’ viii, p. 31.

Jounston, G.—‘ A History of the British Zoophytes,’ 1838.

Jussieu, B. pe.—‘‘ Examen des quelques productions marins qui ont été
mises au nombre des plantes, et qui sont l’ouvrage d’une sorte d’In-
sectes de Mer,” ‘ Mémoires de l’Académie Royale des Sciences,’ 1742.

Kocu, G. von.—‘ Anatomie der Clavularia prolifera,” ‘Morph.
Jahrb.,’ vii, p. 484.

Kocu, G. von.—‘‘ Die Alcyonacea des Golfes von Neapel,”’ ‘ M. T. Zool.
Stat. Neapel,’ ix.

Kocu, G. von.— Kleinere Mittheilungen tiber Anthozoen,” ‘Morph.
Jahrb.,’ xvi.

Kocu, G. von.—‘ Die Gorgonide des Golfes von Neapel,’ 1887.

Koiuiker, A.—‘ Anatomisch-systematische Beschreibung der Alcyo-
narien,’ 1872.

Lamarck.—‘ Systéme des Animaux sans Vertébres,’ ii, 413; 2nd ed.,
ii, 631.

Lamovurovx.—‘ Histoire des Polypiers coralligenes flexibles, vulgaire-
ment nommés Zoophytes,’ 1816, 336, pl. xii, fig. 4; pl. xiii.

Lanxuster, BE, R.—< On the Intra-cellular Digestion and Endoderm of
Limnocodium,” ‘ Quart. Journ. Mier. Sci.,’ xxi, p. 119.

THE ANATOMY OF ALCYONIUM DIGITATUM. 383

28. Lo Biancno Satvarore.— Notizie biologiche riguardanti special-
mente il periodo di maturita sessuale degli animali del Golfo di Napoli,”
‘ Mitt. z. Stat. Neap.,’ viii, 386.

29. Marswatt, A. M. and W. O.—‘ Report on the Pennatulida collected in
the “Oban” Dredging Expedition,’ 1882.

830. MoseLry, H. N.—“ Report on Corals,” ‘ ‘ Challenger’ Monographs,’
vol. ii.

31. Miuter, O, F.—‘ Zoologie Danice,’ 1776, 255, No. 3078.

32. Pattas.—‘ Elenchus Zoophytorum.’ 3

33. Pictret, C.—“ Recherches sur la spermatogenése chez quelques Inverté-
brés de la Méditerranée,” ‘ Mitt. z. Stat. Neap.,’ x, p. 75.

34, Reaumur, R.—‘ Mémoires pour servir a |’ Histoire des Insectes,’ 1742,
vol. vi, p. 55, preface.

85. ScuneipeER, K. C, M.—* Hinige histologische Befunde an Coelenteraten,”’
‘Jen. Zeits.,’ xxvii, p. 440.

36. TRemBLEY, A.—‘ Mémoires pour servir a |’Histoire d’un genre de Polypes
d’eau douce, a bras en forme de cornes,’ Leiden, 1744.

87. Voet, C., and June, E.—‘ Lehrbuch der praktischen bean
Anatomie,’ Brunswick.

38. Witson, E. B.—“ The Development of Kenilla,” ‘ Phil. Trans,’ 1883,
Part iii.

39. Witson, E. B.—“‘ The Mesenterial Filaments of the Alcyonaria,” ‘ Mitt.
z. Stat. Neap.,’ v.

EXPLANATION OF PLATES 36—39,

Illustrating Mr. Sydney J. Hickson’s paper on ‘‘The Anatomy
of Aleyonium digitatum,”

Lettering used throughout.

D. Dorsal intermesenterial space. Het. Ectoderm. nd. Endoderm. can,
Canal in mesoglea. c.g. Coagulum in spermagem. Dm/f. Dorsal mesen-
terial filament. fc. Flagellate cells of ventral mesenterial filaments. g.c.
Gland cells of ventral mesenterial filaments. gox. Gonad. gon. 2. Ovary.
gon. d. Testis or spermagem. gr. Parasitic sporozoon? g.v. Germinal

384 SYDNEY J. HICKSON.

vesicle. 7.c. Interstitial cells of the ectoderm. m. Muscles. mes. Mesogleea.
mf. Mesenterial filaments. xem. Nematocyst. w.p. Nerve plexus. P. Msc.
Protractor muscles. &. Msc. Retractor muscles. S87. Siphonoglyphe. sp.
Spicules. sp.c. Spicule forming cells. S¢. Stomodxum. sp.g. Spermagem.
t. Tentacles. V. Ventral intermesenterial space. v. m./. Ventral mesenterial
filaments. y.s. Yolk spaces in the ovum.

PLATE 36.

Fic. 1—A specimen of Alcyonium digitatum in the Cambridge
Museum.

Fic. 2.—A specimen of Aleyonium glomeratum from the Marine Bio-
logical Association’s laboratory at Plymouth. Both of these illustrations were
drawn from spirit specimens, and are for the purpose of illustrating the dif-
ferences of external form of the two species as may be seen in ordinary
museum specimens.

Fie. 3.—Spicules of Aleyonium digitatum (white or pink varieties)
x 240.

Fie. 4.—Spicules of A. digitatum (yellow variety) x 240.

Fic. 5.—Spicules of A. glomeratum xX 240.

PLATE 37.

Fic. 6.—A fully expanded polyp drawn from a living specimen. ‘The bases
of the tentacles are extended in a bullate manner, and the pinne are partly
withdrawn. The stomodeum (S¢.) and the six short and two long mesenterial
filaments may be seen through the transparent body-wall. At the base of the
crown of tentacles, and at the base of the extensible portion of the polyp, the
body-wall is more opaque, and contains clusters of spicules.

Fic. 7.—Vertical section through a portion of a colony (semi-diagrammatic)
to show the polyps in different stages of retraction, the different sizes of the
polyps, the canal system, the distribution of the spicules, &c. The mesen-
terial filaments are represented as being in the same plane as the tentacles,
which they are not, for the sake of illustrating certain points in the arrange-
ment of the structure. The following points are illustrated in this figure :—
1’ represents a fully expanded polyp. 2’ represents a partially retracted polyp
in which the tentacles are contracted and folded over towards the centre of
the disc ; the body-wall forms a circular fold over the crown of tentacles. 3!
represents a polyp in which the tentacles and stomodeum have sunk to the
level of the general surface of the colony. 4’ represents a polyp in which these
organs have sunk below the level of the general surface of the colony. 5!
represents a polyp which is completely retracted, the thin body-wall of the
extensible portion having sunk below the surface, and the densely spiculated

THE ANATOMY OF ALCYONIUM DIGITATUM. 385

mesoglcea closed over it. It will be noticed that the long dorsal mesenterial
filaments are all on one side, i.e. the axial (Marshall) of the celentera. The
spicules in the colony are much more numerous at the periphery than in the
deeper parts of the colony. The celentera of neighbouring polyps do not
actually run into one another, but they are in communication with one another
by means of canals.

Fie. 8.—Diagram of a section through a portion of a colony to show that
the secondary polyp cavities (11) do not run into the primaries (1), but are in
communication with them by means of canals. The peripheral canals, spicules,
and other details are omitted.

Fie. 9.—The dise of a polyp in the second stage of retraction. The ten-
tacles are contracted, but have not yet begun to bend over.

Fic. 10,—The disc of a polyp in the first stage of retraction. This drawing
was made from a preserved and stained preparation of an expanded polyp, and
shows the arrangement of the muscles (m.) on the disc, and the spicules of the
peristome and of two of the tentacles. These muscles are all ectodermic.

Fie, 11.—A drawing of the dise ofa living and fully expanded polyp, showing
the bullate base of the tentacles and the shape of the mouth.

PLATE 38.

Fie. 12.—Transverse section through an expanded polyp (x 130) in the
region of the stomodeum, showing the folds in the epithelium lining the stomo-
deeum (not present in the expanded polyp of A. palmatum), the siphono-
glyphe, the arrangement of the muscles on the mesenteries, &c.

Fie. 13.—Vertical section through a retracted polyp. On the left-hand
side the section passes through the side of a tentacle, and through one of the
long dorsal mesenterial filaments ; on the right-hand side it passes through the
centre of the tentacle and the siphonoglyphe. ‘The stomodzum is seen to be
folded on itself.

Fie. 14.—Vertical section through the extensible portion of a polyp (x 70),
showing the mesentery bearing the transverse protractor muscles, the longi-
tudinal retractor muscles, one of the ventral mesenterial filaments, and the
siphonoglyphe.

Fic. 15.—Transverse section through one of the pinne of a tentacle, show-
ing the ectoderm with its nematocysts and the lumen in the endoderm.

Fic. 16.—Transverse section through one of the dorsal mesenterial fila-
ments.

Fie. 17.—Vertical section through the surface of a decalcified colony of
Alcyonium, showing the ectoderm with its covering of slime, the mesogloa
containing the endodermal canals, the empty spaces which contained the

386 SYDNEY J. HICKSON.

spicules before they were dissolved away in acid, and the spicule-forming
cells (sp. c.).

Fie. 18.—Transverse section through a dorsal mesenterial filament and a
portion of the mesentery. The groove in this case lies freely open instead of
being nearly closed, as it is in fig. 16.

Fic. 19.—Transverse section through one of the ventral mesenterial fila-
ments showing the gland-cells (g.¢.) and flagellate cells (f.c.).

Fic. 20.—Transverse section of an expanded polyp (x 130) below the level
of the stomodzum, to show in section the six endodermic mesenterial fila-
ments (Vm/f.) and the two ectodermic mesenterial filaments (Vm/.).

Fic. 21.—Transverse section through a portion of a mesentery to show the
great retractor muscle-fibres (2. msc.) on their branched supports of mesoglcea
(coloured darker in the figure), and the very delicate protractor muscles (p.
msc.) on the opposite face of the mesentery.

Fie. 22.—Longitudinal section of a tentacle in a partially retracted condi-
tion such as that shown above in Fig.9. The ectoderm (ecé.) is shown con-
taining several nematocysts, and the ectodermic nerve plexus (z.p.).

Fic. 23.—A small piece of the endoderm taken from the polyp tube a little
way below the termination of the ventral mesenterial filaments to show the
intercellular spaces.

Fig. 24.—The same taken considerably lower down, showing that the inter-
cellular spaces are much larger and the endoderm-cells much more shrivelled
in appearance.

Fic. 25,—A small portion of the endoderm covering a gonad, showing that
in that region no intercellular spaces can be seen.

PLATE 39.

Fic. 26.—A section of a portion of the mesoglea and endoderm of a polyp
cavity taken about an inch below the surface of the colony, cut with the free
hand from the living animal, stained in weak methyl blue, and subsequently
treated with pyrolignic acid. In the mesoglcea there may be seen a number
of very small stellate cells connected with one another by fine branching pro-
toplasmic processes. In addition to these there are a number of spherical,
spindle-shaped, or elongated cells collected together into strings or clumps,
and connected at intervals with the plexus of fibres from the stellate cells.
The plexus of fibrils may be seen to be connected with a plexus in the lower
layer of the endoderm.

Fic. 27.—A string of cells in the mesogloea from another preparation.

Fic. 28.—A number of cells obtained by scraping the walls of a polyp tube
and treating with osmic acid and picro-carmine. @., /., /.,0., and x. are all typical

THE ANATOMY OF ALCYONIUM DIGITATUM. 387

myo-epithelial cells; gy. probably represents a group of epithelial cells from the
dorsal mesenterial filament; /. is a enidoblast, but it is not possible to assert
that it belongs to the endoderm, as they were not seen in situ in the endo-
derm in any of the sections.

Fie. 29.—A portion of the epithelium of the siphonoglyphe.

Fic. 30.—Two nematocysts after they have been exploded by the action of
dilute acetic acid.

Fic. 31.—Two cnidoblasts found in a teased preparation of the ectoderm of
a fresh tentacle.

Fic. 32.—Two unexploded nematocysts found in the same preparation.

Fie. 33.—A section cut very obliquely through a cclenteron of a polyp
some distance below the surface of the colony. In the upper portion the pro-
toplasmic portion of the myo-epithelial cells are cut across; in the lower
portion, marked by an asterisk, the circular arrangement of the muscular fibres
of these cells may be seen. Outside the limits of the tube the mesoglea may
be seen with its strings and clumps of cells, stellate cells and fibres.

Fic. 34.—The appearance of the protoplasmic portion of the myo-epithelium
as seen in such a section as that figured in 33, showing the large, irregular,
intercellular spaces.

Fie. 35.—The circular muscle-fibres of the endoderm lining the ccelenteron,
such as are represented at the point marked with an asterisk in Fig. 33, but
more highly magnified, showing the endodermic nerve plexus.

Fic. 36.—A string of cells from the mesoglcea (osmic acid preparation)
showing intra-cellular bodies (gr.) of doubtful meaning, but possibly parasitic
sporozoa.

Fic. 37.—Transverse section through a mesentery in March, showing a young
ovary (yor. 2). Some of the young ova have a single nucleus containing a
few chromatin granules in it ; others have a single nucleus which stains deeply
and is homogeneous ; others again have two nuclei, one of each kind.

Fic. 38.—Transverse section taken through a mesentery later in the year,
showing an older ovary. One of the ova (ov’.) has reached a considerable
size ; another (ov"’.) contains two ova partly disentegrated.

Fie. 39.—Section through an ovum in the month of September, x 125 dia-
meters, showing the nucleus (g. v.) containing a single large nucleolus situated
at a short distance from the periphery.

Fic. 40.—Section through an ovum in the month of December, x 125 dia-
meters. The nucleus (g. v.) contains a number of chromatin granules, but no
large nucleolus. It is situated quite at the periphery.

Fic. 41.—Section through a portion of an ovum in December after treat-
ment with alcohol, more highly magnified, showing the spaces occupied by
the yolk.

388 SYDNEY J. HICKSON.

Fic. 42.—Section of a mesentery in the early spring, showing a young
spermagem (sp. g.).

Fie. 43.—Section through a spermagem in the autumn, showing a central
coagulum and a crowd of densely packed nuclei in a matrix at the periphery.

Fic. 44.—Section of a portion of the above showing the nuclei more highly
magnified.

Fie. 45.—A series of stages of the final development of the spermatozoa as
seen in a preparation of a broken testis in December. a. A ripe spermatozoon.
b.,¢.,d. Stages in the formation of the tail. £ A cell containing four nuclei
occasionally met with in these preparations.

Fic. 46.—A lobed diverticulum from an endodermal canal of Alcyonium
meeting an ectodermic invagination to form a bud.

Fic. 47.—A later stage, when the endodermal diverticulum has joined the
ectodermal invagination, but the mouth has not broken through. ‘The dorsal
mesenterial filaments have been formed.

Fic. 48.—A still later stage, less highly magnified, when the mouth has

been formed. The connection with a neighbouring fully-formed polyp is
shown. The tentacles have not yet been developed.

CONSTITUTION OF MESOGL@A OF ALCYONIUM DIGITATUM. 3889

Note on the Chemical Constitution of the
Mesoglea of Aleyonium digitatum.

By

W. Langdon Brown, B.A.,
Late Hutchinson Research Student, St. John’s College, Cambridge.

[From the Physiological Laboratory, Cambridge. ]

THE main organic constituent of the mesoglea in Alcy-
onium digitatum would appear to belong to that class of
bodies described by Kruckenberg! as hyalogens, so widely found
among Invertebrate skeletal structures. Hyalogen is charac-
terised by its insolubility, and its conversion by various re-
agents (e.g. 5 per cent. solution of caustic soda) into a soluble
substance, hyalin. Hyalin is practically a mucin, yielding on
decomposition a proteid-like body and a carbohydrate.

I. The following experiments show that mucin is readily
obtainable from the mesogloea ; whether, in life, any of it is
present as such, or whether it is derived from a hyalogen by
the necessary preliminary treatment, it is, of course, difficult
to say.

Fresh specimens, freed as far as possible from cellular layers,
were washed well with distilled water and minced. They were
then extracted with a 5 per cent. salt solution to remove any
globulins, and left under lime water for seventy-two hours. The
filtrate was then treated with acetic acid. A white precipitate
was obtained, which, on standing for some hours, settled ; it was
then filtered off, and washed with water acidulated with acetic
acid. It was again dissolved in lime water and re-precipitated.
The collected precipitate was boiled with 2 per cent. sulphuric

1 Kruckenberg, ‘ Zeit. f. Biol.,’ xxii.

390 W. LANGDON BROWN.

acid for about twenty minutes; on neutralisation a reducing
sugar and a peptone-like body were obtained. A portion of the
original colony, boiled direct with sulphuric acid, yielded a
reducing sugar.

Throughout the investigation the endeavour was made to
confirm and locate the results obtained from chemical analysis
in bulk by the application of micro-chemical tests. For the
micro-chemical detection of mucin, Waymouth Reid recom-
mends thionin, as used by Hoyer,! which he considers to be
thoroughly diagnostic. It appears probable, however, that
mucin gives the ruddy purple reaction with this dye in common
with other basophilic substances. Nevertheless this reaction
serves to locate the position of the mucin or hyalin which had
been extracted. Sections of specimens hardened in corrosive
sublimate were cut and stained as recommended by Reid.
The endoderm appeared blue, the mesogloea a ruddy purple.
This latter coloration was also seen in cover-slip films pre-
pared from the precipitated substance extracted as above.

Hence we conclude that the mesogloea yields a substance
which resembles mucin—

(i) In its solubilities.
(ii) In its decomposition products.
(iii) In its micro-chemical reaction.

II. After treatment of the mesogleea with dilute acids, how-
ever, lime water or baryta water is capable of dissolving a
much larger proportion. Prolonged boiling or treatment with
superheated water (120° C.) will effect the same change. Lime
water or baryta water then leaves but a small granular
residue, whereas the main mass of the mesoglea is left by
these reagents when repeatedly applied to the fresh material.
Treatment for a few days with comparatively dilute (e. g. 50
per cent.) alcohol greatly impairs the solubility.

The filtrate from this lime water extract will on decomposi-
tion yield a reducing sugar and a proteid substance.

This recalls the condition of affairs met with in the edible

1 Waymouth Reid, ‘Journal of Physiology,’ vol. xiv. Hoyer, ‘Arch. f,
Mikr, Anat.,’ Bd. xxxvi.

CONSTITUTION OF MESOGL@A OF ALCYONIUM DIGITATUM. 391

bird’s-nest which Kruckenberg (op. cit.) regards as composed
of a hyalogen-neossidin, giving rise on solution to a hyalin-
neossin. The mesoglea further resembles the reactions of
the edible bird’s-nest as described by Green! in the absence of
a cellulose reaction, and in the fact that on boiling with dilute
sulphuric acid a pinkish coloration, which subsequently
darkens, is seen.

Another feature frequently shown by hyalins, in which that
under consideration agrees, is the absence of Millon’s reaction.
An application of this test micro-chemically to mesogloea shows
us that the cells alone stain red. This indicates that the pro-
teid extracted by 5 per cent. salt solution (which presents all
the characteristic features of a globulin) is located in the cells,
and not in the mesoglea ; it need not then detain us.

The micro-chemistry of hyalogen presents a slight difference
from that of mucin. Pieces of mesogloea which yielded no
further mucin to lime water were decalcified, and cut with the
freezing microtome. Stained with thionin by Reid’s method,
the purple coloration of the mesoglea only appears under
artificial light, recalling the conditions of the basophilic reaction
with methylene blue. The endoderm-cells appear blue both
with natural and artificial light.

We shall not err greatly, then, if we describe the organic
constituent of the mesogloea as mainly a hyalogen, probably
mixed with some amount of mucin. The hyalin obtained
from the hyalogen is precipitated on saturation with ammonium
sulphate as a gummy mass.

III. As stated above, after extraction of previously acidified
mesogloea with lime or baryta water a small amount of a some-
what granular residue is left. Examined under the microscope,
it is seen to consist of clumps of globules which yield the
xanthoproteic reaction. They recall in some measure the
description given by Mall? of the appearances of elastin acted
upon by acids or alkalies. The substance, however, is not
elastin, as is seen by its having no affinity for Victoria blue or

1 J. R. Green, ‘ Journal of Physiology,’ vol. vi.
2 Mall, ‘ Anat. Anzeig.,’ iii Jahrg.
VOL. 37, PART 4,—NEW SER. DD

392 W. LANGDON BROWN.

safranin, and in its resistance to prolonged tryptic action. It
swells up, however, under the action of this enzyme. It is
markedly insoluble, showing no diminution in bulk on addition
of weak or strong acids and alkalies, cold or boiling. The only
exception to this is concentrated sulphuric acid, which causes
it to swell up, and after boiling for a short time to dissolve with
an accompanying pink coloration, soon passing toalight brown.
No leucin, peptone, or sugar could be demonstrated as a
product of this action.

These reactions remind us somewhat of the insoluble residue
described by Schafer! as obtained from the entostemite of
Limulus. The total amount, however, is but slight.

An artificial pancreatic digestion of mesogloea yields the
results which might be anticipated from the foregoing. The
residue was composed of antialbumid (soluble in 1 per cent.
caustic soda) calcareous matter (readily soluble in dilute acetic
acid) and the granular substance just referred to. The filtrate
yielded albumoses, peptones, a reducing sugar, crystals of
leucin and of tyrosin—the last presumably from the cells,
since in them alone is Millon’s reaction successful.

LV. In investigating a material like the mesogloea two ques-
tions naturally suggest themselves: does it contain gelatine?
and does it contain nucleo-albumen? (a) Does the mesoglea
contain gelatine? In some experiments on this point, thin
slices of mesogloea were treated with distilled water under
pressure at 120° C. for over two hours; in others, portions of it
were minced and boiled with distilled water for 14 hours. The
subsequent procedure was that described by Young? for retiform
tissue. The extract was filtered, while boiling, through a hot
filter into alcohol. The precipitate which formed was collected
after standing for some hours, dissolved in distilled water,
boiled, and filtered. The filtrate was concentrated to a very
small bulk. In all cases this did not “jelly” perceptibly.
Hence we conclude—

1 Schifer, foot-enote to Ray Lankester on ‘“ Skeleto-trophic Tissues,”

‘@..J. M. 8:,’ vol. axa;
2 R. A. Young, ‘ Journal of Physiology,’ vol. xiii.

CONSTITUTION OF MESOGL@A OF ALCYONIUM DIGITATUM. 395

The mesogloea does not contain gelatine. Hence. also, it
cannot contain chondrin-like bodies. The popular term
“jelly” applied to this substance appears to have no basis in
chemical fact.

(6) Does the mesogloea contain nucleo-albumen? The recog-
nition of this body in several non-cellular tissues suggests its
presence here. To test this point a colony was stripped of its
external layers and treated by Halliburton’s method for extract-
ing nucleo-albumen.! None could be detected.

Thanks to the method recently introduced by Lilienfeld and
Monti,” the examination of this question micro-chemically is
also feasible. Specimens hardened in osmic acid were cut with
the freezing microtome, then washed thoroughly and placed
in a solution of ammonium molybdate. After being washed
for a few seconds in a mixture of ether (9 parts) and water
(1 part), they were put into a 20 per cent. ethereal solution of
pyrogallic acid. The cells in such specimens were seen under
the microscope to be stained black, but the mesogloea was not.
Hence we conclude—

The mesogloea does not contain nucleo-albumen.

V. Conclusions:

(i) The mesoglea of Aleyonium digitatum is chiefly
composed of a hyalogen.

(ii) Prior to the conversion of the hyalogen into hyalin the
mesoglea wil] yield a mucin.

(iii) It also contains a small amount of an insoluble albu-
minoid body, whose nature was not determined.

(iv) It does not contain gelatine or nucleo-albumen.

1 Halliburton, ‘Brit. Assoc. Reports,’ 1888 ; ‘Journal of Physiology,’
vol, xiii.

2 Lilienfeld and Monti, ‘Zeit. f. physiol. Chem.,’ Bd. xvu. See also
Gourlay, ‘ Journal of Physiology,’ vol. xvi.

6 : 4
. a: e
>.
bay, rey ot) BD ~ PDAS Wh Ve) | WET ee
a8 ips - iag his ig ye note LA ho .:
mi ne) ilfps | a fs praia (} A iy ay) picks age
iii braiit a Pi) is ark ooh t OF B.S F it ¥

. | = me
—".t oP. Caintidlaeaionn Ty joy, eo pone aaah 4
art) fark) 3 Jee Ses) Se ie ad AO eRe
PS eC ee war tate. pi

eee yi Leica Wesel it vend Pipi marin :
ietieioh wh biie awe os@pdls celal
fe Diesels Lb ty Vidensi! ult i La iu) 4466
ol ee Cee Ten ae Ope thep a silks wild * pi

7 - - “st
era re yh) Oe Sgr 5) ie Ae eA} Ss ISS ol BTEE

‘ oe wane Ty yiasl > limape LOPS vy fork SR
A Biche — sini / eiaidulanun oi ulmanise Tne sae
ae Hhans( Gurr Gpeowtt ¥ Tig hat, B) els OE ral nel.
Te) Lee rd a a Serpe OE eat dts. Maa wii 4
Ment Chobe sim jal peg cca | age Lah L <,i4ea als

‘
19) Ai we Ai)" Y dd nih ype Leagan jem tt adios ong a lb
‘i Stay Sole ee
Jules ts Sa SCRE Hote tie fone amen oN
— ; a sieG > p
Svitireh Grins ih Pulte eA Need Sou
: : eu’ F aia
CUP Nev WED Pow sl Ud te nie ot Re ideal veoh
. WVHA lipin Chie
pid alike Foliar) nis see OA ter Be |
Nive
auido or outages hae

inhi a bt SUR iN ta eetay wt. sib 4

Riss

rT lished’, aolaouh” » 2) Vth opt Sh

7}
CR a a * qiWonk, “Nk iis
| ite Selatan “a

ees —
Hy ; eo! er —. bg 7
pig _ an +e
1 , he pes
> yon a

A STUDY OF METAMERISM. 395

A Study of Metamerism.
By
T. H. Morgan, Ph.D.,
Associate Professor of Biology, Bryn Mawr College, U.S.A.

With Plates 40—43.

TABLE OF CoNTENTS.

PAGE

I. Typical Forms of Modification in Annelids : - : . 395
II. Variation in the Position of the Reproductive Organs. . 403
III. Abnormalities in Front of or including the 15th Metamere . 407
IV. Study of Embryos . : ‘ : : ; 5 . 418
V. Abnormalities at the Posterior End : : : : : . 421
VI. Modification of Internal Structures . : : : : ~ 425
VII. Study of Polychetous Annelids and Leeches . : : . 428
VIII. Summary. : : - : : . 431
IX. Modifications in Ancens of Arairopods : : : : . 435
X. Abnormal Metamerism of Locust . ; : : : . 437
XI. Study of Colour-bands of Echinoderms . : : : . 439
XII. Regeneration in Earthworms . ; : : : ‘ . 445
XIII. General Conclusions . ‘ ; F : ; : : . 460

I. Typrcat Forms or MoptrFicaTion.

Tuat there were occasionally to be found in the Annelids
irregularities in the serial repetition of the rings seems to have
been known to several of the earlier writers on descriptive and
systematic zoology. The fact did not attract more than a
passing attention, and such irregularities were relegated to
that waste-heap of abnormalities from which subsequent in-
vestigation has often drawn valuable material.

Simultaneously in 1892 two articles appeared dealing with

396 T. H. MORGAN,

these abnormal conditions, one by C. J. Cori (18), and the
other by the present writer (88). Both writers pointed out
the general interest attached to these modifications, and their
importance for an interpretation of the general problem of
metamerism. Subsequently a third paper appeared (10), re-
cording the presence of similar abnormalities in many groups
of Annelids, but without making any attempt to solve the
problem itself.

My own paper was only a preliminary notice, and until the
present time I have not had an opportunity of fully describing
the material that I had at that time already accumulated,
studied, and drawn. The present paper attempts to give a
full consideration of the facts only touched on before, and to
extend over a wider field the conclusions reached.

The modification of the rings, segments, or metameres of
the Annelid fall into two general classes,—not, however,
sharply separated from one another.

The first of these I shall speak of as the compound metamere,
in contradistinction to the normal or simple metamere. In
the earlier: paper the shorter term “ split metamere” was
used, although it was there pointed out that the term was
misleading.

The other type of modification may be spoken of as the
spiral metamere or spiral modification, or briefly as the spiral.

The simplest case of the compound metamere is represented
by the diagram on Pl. 40, fig. I, a,3,c. Fig. I, a, is a dorsal
view of a compound segment. The segment is double on the
right side, normal on the left. Fig. I, B, is the same, turned
over so as to be seen from below, showing a similar doubling
below on the same (right) side of the body. Fig. I, c, repre-
sents the compound segment as opened along the “split.” It
is conceived to be transparent, and the dotted line to represent
the lower surface outlines. The general appearance is that a
metamere has been split into two on one side of the body, but
has retained on the other side its normal structure. The
question at once arises, are we dealing here with a case of
division of a metamere on one side of the body, or is it a case

A STUDY OF METAMERISM. 397

of three half-metameres united together? It is one of the
aims of the paper to answer this question.

To the other class of modifications belong the spirals, one of
the simplest of whichis shown in fig. VII, a, B, c. The spiral
is produced by a combination of two half-compound metameres,
i.e. one metamere is “split”? below, but the same one is
normal above. The next metamere is split above and on the
same side of the body, but not below. The posterior arm of
the anterior half-compound metamere is continuous at the
side with the anterior arm of the posterior (upper) half-com-
pound metamere. The same fact may be stated in a different
and perhaps a clearer way. A half-metamere on the right
side of the body unites in the mid-line below with a metamere
in front of it, and above with a metamere behind it, fig.
VII, c. We have formed as a result of this arrangement one
more half-metamere on the right side of the body than on the
left side, as is best shown by the construction in fig. VII, c.

We may proceed to study more in detail the many kinds of
modification that group themselves around these two types.
We shall find that nearly all of the geometrical combinations
conceivable are to be found in the worms themselves.

Category 1, fig. I, a, B, c.—This form has already been
described as the type. It is equally common on the right and
left sides of the body.

Category 11, fig. II, a, B, c (above, below, and reconstructed,
as seen from above).—In this form one (above) of the lines of
the split turns to one side (posteriorly) to fuse with the septal
line. This leaves, as shown in the reconstruction, a free end
forming a modification of the compound metamere. This
modification may appear either on the right or left side of
the body, and the free end may occur above or below, anteriorly
or posteriorly.

Category 111, fig. III, a, B, c.—By a simple shifting of the
lines in the last form we pass to the third category. Here
both of the lines of the extra half-segment turn towards the
same septal division, and we have formed an intercalated half-
metamere,

398 T, H. MORGAN.

The variations of this category are formed by the lines
turning both anteriorly or both posteriorly, and on the right
or left side of the body. In reality the half-segment is simply
intercalated between two perfect segments ; and when we speak
of the lines turning either forward or backward we only imply
that the intercalated piece has encroached more on the one or
the other segment.

Category rv, fig. IV, a, B, c.—Here three half-segments of
one side unite with one half-segment of the opposite side, so
that the right side appears to be “ split”? by two incomplete
divisions into three half-segments. This may be called a
double compound metamere. Variations arise owing to the
depth to which the incomplete lines (splits) pass towards the
middle line. A combination of this form (really only a ha/f-
double compound) with others is shown in Pl. 40, fig. XIV.
This same modification may be carried a step further, and in
one case I found four half-metameres of one side united to
one of the opposite. See Pl. 42, fig. 47.

Category v, fig. V, a, B, c.—Here on the upper side appear
two eompound segments—one on the right side, the other
on the left. On the lower surface, B, two perfect metameres
are found. By tracing the division lines between these seg-
ments from the lower to the upper side we see the result is
brought about by the failure of these two septal lines to meet
above ; each turns somewhat to one side. The reconstruction
shows that a middle connective is left to unite the two seg-
ments above. This is brought about by the fusion above of
the middle of the right posterior segment with the left anterior.

The variations are few. The incomplete meeting may take
place either above or below, and the lines turn anteriorly or
posteriorly.

Category vi, fig. VI, a, B, c.—The arrangement found here
inay be produced from fig. v by extending the free ends of the
segment lines till the one meets the segment in front, and the
other the one behind. A reconstruction of such a figure
gives the spiral shown in fig. VI,c. This spiral form terminates

A STUDY OF METAMERISM. 399

both anteriorly and posteriorly in free ends. The two meta-
meres involved take two complete turns around the body.

Variations of this (like those of category v) readily suggest
themselves.

Category vit, fig. VII, a, 8, c.—This spiral form has been
already described as a type of the spiral.

Category vitt, fig. VIII, a, 8, c.—On the upper side are two
half-compound segments opening in opposite directions (a).
On the lower side three normal segments are present (8).
The reconstruction (c) shows that the posterior arm of the
first half-compound segment takes a complete turn around the
body, and then unites with the second half-compound segment
as its anterior arm. The spiral involves three segments.
The middle of the three, instead of forming its proper union
above, unites above on the right side with the segment
anterior to it, and on the left side and above with the segment
posterior to it.

The variations of this form consist in having the union
above or below, on the right and on the left, or on the left
and on the right side.

Category 1x, fig. IX, a, 8B, c.—This form is merely a longer
spiral than that of category vir. The spiral turns one and a
half times around the body, beginning and ending as in vi.
Three segments on one side of the body correspond to four on
the opposite.

This process may theoretically be continued through an
indefinite number of segments, but generally a few turns
suffice to bring the spiral to a close. The most extreme case
that I have found involved twelve and a half turns to the
spiral (Pl. 41, fig. 26).

Category x, fig. X, a, B, c.—This spiral is an extension of
that of category vi11. It involves a greater number of meta-
meres, and in consequence the spiral is longer. Two turns
are taken around the body.

In both categories 1x and x all the variations found possible
for vi1t and 1x are here applicable.

400 T. H. MORGAN.

Fig. XI, a, B, shows a further extension of vit, with three
turns.

There are several combinations of the preceding categories
that are possible, and some of these I have found. It is not
particularly important to discuss these possibilities, as they all
reduce back to forms already described. As an example,
however, one such is given in fig. XII, a, B, c; the reconstruc-
tion, c, sufficiently explains the conditions. It will be noted
that five half-segments of one side correspond to three of the
opposite.

Whenever a double half-compound metamere is introduced
a more complicated form of spiral results (see fig. XIV, and
Pl. 42, fig. 47, at x1). This causes a double spiral, i.e. two
spirals, to take a parallel course, as shown in fig. XIV. Oneor
both of the spirals may end in a half-compound metamere—
both so end in fig. XIII. This reconstruction will, I think,
sufficiently explain the conditions, so that a further description
would be superfluous.

The double spiral may be formed in other ways. Two suc-
cessive compound metameres may be introduced in such a way
that a new spiral is started along with the one already present.
Such cases are shown in fig. XIII, and in fig. 47, Pl. 42, at
x’, Several variations of these combinations suggest them-
selves, and several have been found amongst the worms, but
the discussion of these variations would not add materially to
the former cases, and may be omitted. Even a triple spiral
was found in one case, as shown in fig. 47 at x*. It was of
short duration, owing to the introduction of compound meta-
meres in such a way that the spirals were quickly absorbed.

Lastly, the series of compound metameres, double and
triple spirals shown in fig. 47, Pl. 42, were all drawn from
the same worm. We find 134 half-metameres on the right
side and 118 half-metameres on the left. In all there were
fifty-two perfect rings, and the numbers in the plate between
consecutive groups of compound metameres, spirals, &c., in-
dicate the number of perfect rings in that locality. An exa-
mination of this reconstruction will show how far it is possible

A STUDY OF METAMERISM. 401

for the variations—abnormalities—to be present, and the worm
still be able to reach sexual maturity. In another worm con-
taining many abnormalities there were 131 half-metameres on
the right side of the body and 189 on the left side. In all
there were seventy-nine normal rings. The number of half-
rings in these worms is noticeably very large as compared with
the normal number. It should also be noticed that the
abnormalities are not confined to any one part of the body, -
but scattered throughout the whole length of the worm.

We may now study the relative proportion in which
the different modifications, described above, are found. Ina
lot of 318 worms 218 were found to be normal externally with
respect to the metameres, and 100 abnormal. This is in the
proportion of 1 to 2°18, or in general one worm of every three
was abnormal.

In the same worm there often occurred more than a single
abnormality. Thus in the above 100—

With. . 1 abnormality : ‘ : 65
- : . 2abnormalities . : ? 16

er 3 eo as : 2 : 10
» more than 3 os ‘ = : 9
100

An examination of the material shows that there is practi-
cally no difference in the number of compound metameres on
the right as compared with the left side of the body.

The examination shows, however, that a much larger
number of “‘ splits” are present on the dorsal surface of the
body than on the ventral surface. In a compound metamere
we have a split above for one below, but in such spirals as
shown in figs. vri1, x, &c., we have two above and none below,
and such spirals have their “splits” on the dorsal surface in
the greater number of cases. I regard this as important, as it
gives a clue to the solution of the problem.

An examination of the material shows that there are thirty-
two cases of spirals beginning and ending above (category vi1r),
and only one case of a spiral beginning and ending below.

402 T, H. MORGAN.

There were ten cases of failure of lines to meet above as in
categories v and vi. No cases below.

There were twenty-five cases of spirals beginning above or
below, and ending on the same side of body—below or above.
Amongst these twenty-five cases there were fifteen that began
above and ended below, and ten that began below and ended
above. If we compare the data of the thirty-two cases with
the twenty-five cases we find that we have ninety-nine cases
where the split was on the dorsal surface, and twenty-five
cases where the split appeared on the ventral surface.’ That
is to say, there were nearly four times as many cases of false
unions above as below.

If we arrange the different forms of abnormalities found
under the categories given in the preceding section, we find
the relative proportion of the abnormalities in the 100 abnor-
mal worms to be as follows. In this count no difference is
made as to whether an abnormality occurred once or twice on
the same worm. It is based on the results of all abnormalities
found present.

Category I, II . : ‘ : ; 42
- inf = 5 : : : 1
3” Wo VI. ° . . . 10
Vito: : : : . 20 }
” 96
{ se Tie Fis ; : 5 : 6
% VILE. 5 ; : : 16
| uf dee # < 5 5 ‘ s} oe
=f xI, &e. : ‘ 4 ; 8
Mixed forms too irregular for classification . a

1 These numbers were derived from the data given above as follows:

32 (spirals beginning and ending above) x2 = 64
US Gees “3 above and ending below). —— fi |)
TO > ss below and ending above) . wg
10 (cases where lines failed to meet above) . = 10

99
15 (spirals beginning above and ending below) . = 15
OCs pS below and ending above) . = 10

25

A STUDY OF METAMERISM. 403

II. Vartations 1n THE Position oF THE REPRODUCTIVE
Organs.

With the hope of throwing additional light on the modifica-
tions in the arrangement of the metameres, I was led to
examine carefully those cases where the false arrangement
occurred in the segments anterior to the fifteenth. Here
there are definite landmarks, viz. on the outside are the
openings of the vasa deferentia, oviducts, and seminal re-
ceptacle, and on the inside are the organs belonging respectively
to these openings. The openings of the vasa deferentia on
the fifteenth segment are the only ones of these openings that
are readily and with certainty to be made out, so that I have
confined my study largely to the modifications in the position
of these openings. When we come to study the internal
arrangement of the organs, then the position of ovaries and
seminal receptacles will also be considered.

In utilising the fifteenth segment as a point of departure, it
was first necessary to determine whether this was in reality a
fixed point. In a lot of 799 worms, twenty were found with
the vasa deferentia opening abnormally. In twelve cases of
the twenty the openings were on the same segment, but this
was not their normal segment (fifteenth). In six cases the
openings (right and left) were not on the same segment. In
two cases the openings were doubled on one side or the other.
To summarise:

A. Openings on the same segment, but not on the 15th . 12
B. Right and left openings, not on the same segment <6
c. Openings doubled on one side . : ; ae

A. The following table records the openings of the vasa
deferentia in the first category.

No. of Worms.
E vasa deferentia opened on 10th metamere.
1 EP) 3) llth »
2 ” » 12th »
4 2 2”? 14th 2
4 » » 16th ,,
12

404. T. H. MORGAN.

In those cases in which the openings of the vasa deferentia
occur on a segment anterior to the 15th metamere, we may
be dealing with a case of incomplete regeneration of the ante-
rior metameres. In fact, in the first case given there was
evidence for believing that the anterior end had just regene-
rated, and even the second case gave slight evidence of the
same thing. That all of the cases can be explained in this
way is, I think, highly improbable. We find, as the table
shows, the greatest variation on each side of the 15th meta-
mere, pointing in favour of individual egg variation rather
than regeneration. ‘This view is also strongly supported by
an examination of the internal conditions. Again, those cases
recorded above in c and p show conclusively that the openings
of the vasa deferentia may shift, and in these cases regenera-
tion is largely out of the question.

In order, however, to satisfy myself completely, i.e. experi-
mentally, on this point, I tried by (artificially) cutting off the
anterior metameres of a large number of worms to determine
what proportion of these regenerated the full number of
anterior metameres. The results are given in a later section.

B. There were six cases out of the twenty in which the
openings of the vasa deferentia were in different metameres of
the same worm. These are figured in Pl. 42, figs. 48—58.

In the first case, fig. 48, the 15th metamere has on its right
side (left of figure seen from below) an opening of the vas
deferens. The left side of the 15th has no opening, but on
the 16th metamere of that side occurs the opening of the left
vas deferens. One such case is recorded.

In fig. 49 another variation is found. Here the 15th meta-
mere has on the left side of the body its proper opening, but
there is no opening on the right side of that segment. On
the 16th metamere, however, and on the right side, the right
vas deferens opens. Four such cases are recorded.

In fig. 50 we find that the right side of the 15th metamere
has an opening, but not the left side. On the 14th metamere
we find the opening of the left vas deferens. One such case is
recorded.

A STUDY OF METAMERISM. 405

There can be no doubt but that an examination of a larger
number of worms would show other combinations, but these
are sufficient to show that variations in the position of the
vasa deferentia do occur.

c. Two worms in this lot showed a doubling of the openings
of the vasa deferentia on one side.

In fig. 51 we find on the 15th metamere a pair of openings,
right and left; but in addition to these we find on the left side
of the 16th segment another opening of a vas deferens.

In fig. 52 we find the 15th metamere with its pair of
openings, the right somewhat enlarged. On the left side of
the 14th metamere another opening is present, as shown in
the figure.

Dissections were made of nearly all of the worms recorded
in categories a, B,c. There is always the possibility of mis-
takes when only surface views are examined, but fortunately
in all cases recorded the evidence from internal structure has
supported the conclusion from the external appearance. More-
over additional data of considerable interest has, I think, been
gained.

We may examine first the results of dissections of worms
in which the paired openings of the vasa deferentia are on
some other segment than the 15th.

No. 1. Openings of the vasa deferentia on the 10th meta-
mere, fig. 53. The ovaries were present two metameres in
front of the vasa defereutia openings, i.e. the ovaries were in
the 8th metamere. All of the other parts of the reproductive
system were present and normally developed. The brain
appeared in the Ist metamere.

No. 2. Openings of the vasa deferentia on the 11th meta-
mere. Only one ovary was found in the 9th metamere (left
side) ; the other probably lost in the process of dissection.

No. 3. Openings of the vasa deferentia on the 12th meta-
mere. ‘The ovaries were found in the 10th metamere, fig. 54.
The first segment seemed incompletely regenerated.

No. 4. Openings of the vasa deferentia on the 14th
metamere, fig. 55. The ovaries in the 12th metamere.

406 T. H. MORGAN.

Seminal receptacles in the 8th and 9th metameres. In
another worm similarly modified the ovaries were also present
in the 12th metamere.

No. 5. Openings of the vasa deferentia on the 16th meta-
mere. The ovaries were present in the 14th metamere.

We may next examine those cases where the openings of
the vasa deferentia are, right and left, on consecutive meta-
meres.

No. 6. Three worms were dissected in which the right vas
deferens opened on the 15th metamere, and the left on the
16th, figs. 56 and 57 (projections from above). In two of the
three worms the ovaries showed a similar variation, appearing
on the 13th (right) and 14th (left) metameres,—that is, each
two segments in front of its corresponding vas deferens. In
the third worm both ovaries appeared in the same metamere,
viz. the 13th.

No. 7. Two cases recorded with the left vas deferens
opening on the 15th metamere, and the right on the 16th.
The ovaries varied correspondingly, and were found in the 13th
(left) and the 14th (right) metameres (see fig. 58). The seminal
receptacles showed a similar displacement, and were found in
the 9th and 10th (left) and 10th and 11th (right) metameres.

No. 8. The worm drawn in fig. 52 in which the left vas
deferens is doubled (14th and 15th metameres) was dissected,
and a corresponding doubling of the ovaries of the same side
was found.

No. 9. In another case of doubling the vasa deferentia of
the right side opened on the 16th and 17th metameres; that
of the left side was single and appeared on the 16th metamere
(fig. 59). The ovaries varied also, appearing in the 13th and
14th metameres of the right side, and in the 14th of the left
side. The variation here is uot the same in the ovaries and
vasa deferentia on the right side, for if the ovaries had
varied to correspond they would have appeared in the 14th
and 15th metameres of the right side.

No. 10. The paired openings of the vasa deferentia are
found on the 15th metamere. In addition there is another

A STUDY OF METAMERISM. 4.07

opening on the left side of the 16th. The ovaries are found
in the 13th metamere (fig. 60, a).

No. 11. The paired openings of the vasa deferentia are
found on the 15th metamere. On the left side of the 14th
there is an additional opening. One pair of ovaries are found
in the 13th (fig. 60, B).

No. 12. One case of doubling is recorded for L. terres-
tris (the preceding cases were all for L. foetidus) (fig. 74,
a, 74, 8B). The opening of the right vas deferens is doubled, and
opens both on the 15th and 16th metameres. The left only
appears on the 15th. Dissection showed both ovaries present
in the same segment and not doubled. The dissection shows
also the course taken by the vasa deferentia (fig. 74,8); and
we find a single vas deferens on the right side that supplies
both openings of that side.

I regret that on account of the difficulty in dissecting out
the vasa deferentia in L. foetida I did not get further evi-
dence as to the method of doubling of the external openings in
those worms.

III. ABNORMALITIES IN FRONT OF OR INCLUDING THE 15TH
SoMITE.

In a lot of 799 worms (L. feetidus), twenty, as recorded in
the preceding section, showed abnormalities in the position of
the openings of the vasa deferentia. In this same lot, after
the removal of these twenty (leaving 779), there were thirty-
two worms with abnormalities in front of or including the
15th metamere. This is in the proportion of 1 : 24.

In 1893 I again collected material in order to obtain a
larger number of abnormalities in the anterior ends of the
worms. From 957 worms eighteen were found showing modifi-
cations of the anterior metameres. This is in the proportion
of 1 : 53.

The material from these two sources, with a few additions
from other lots of worms, has been brought together and
figured in Pls. 41 and 42.

VOL. 37, PART 4,—NEW SER. EE

408 T. H. MORGAN.

The modifications may be classified under the following
categories :
. Part of a metamere compressed.
. Failure of external septal lines to meet.
. Spirals in front of the 15th metamere.
. Spiral involving the 15th metamere.
. Results of the introduction of compound metameres.

A. Plate 41, figs. 1—3.

In fig. 1, 4 B, the 4th metamere is not fully developed on
the right side of the body. The ring is narrower, and has less
pigment on the right side. The vasa deferentia open on the
16th metamere, as shown in B.

Fig. 2, a B, shows another worm with the 2nd metamere
similarly modified. The vasa deferentia open on the 15th
metamere.

Fig. 3, a B, shows the 5th metamere reduced on the left
side. The vasa deferentia open on the 16th metamere.

Conclusion.—It is not a little surprising to find in two of
the three cases where an anterior metamere is partially unde-
veloped, that the openings of the vasa deferentia have shifted
one segment posteriorly. The data are too small to lead to any
definite conclusion.

B. Plate 41, figs. 4—11.

In fig. 4 we see that the line of division between the 9th and
10th metameres failed to meet exactly on the upper surface.
The vasa deferentia opened normally, if we count the 9th and
10th as true metameres.

In fig. 5 the line between the 5th and 6th metameres failed
to meet above. Vasa deferentia normal.

In fig. 6 the line between the 4th and 5th metameres failed
to meet. Vasa deferentia normal.

In fig. 7 the line between the 3rd and 4th metameres failed
to meet. Vasa deferentia normal.

In fig. 8 the line between the 13th and 14th metameres
failed to meet.

In fig. 9 the line between the 12th and 13th, and in fig. 10

HoOOwW Pe

A STUDY OF METAMERISM. 4.09

the line between the 11th and 12th. In these last three cases
all failures were above, and in all the vasa deferentia were
normal.

In fig. 11 the line between the Ist and 2nd failed to meet
below; in fig. 12 between 18th and 14th below. In fig. 13
the line between the 5th and 6th metameres failed to meet at
the sides. Both in figs. 11 and 13 the vasa deferentia were
normal, but in fig. 12 there were indications of an additional
opening on the left side of the 16th metamere.

Conclusion.—In nine out of the ten cases the position of
the openings of the vasa deferentia were not affected by the
failure to meet of metameres anterior to the openings, nor was
a shifting to be expected on a priori grounds, since the modi-
fication is clearly only a lack of perfect union between the
right and left half-metameres (fig. 13 perhaps excepted).

Fig. 12 is, then, a case of doubling of the openings of one
side, like those described in the last section, and can have
nothing to do with the lack of perfect union of the more
anterior segments.

It is noticeable that in seven out of these ten cases the
imperfect union is on the dorsal surface. In only two cases is
it found below, and in one at the side. We will return toa
consideration of the point at another time.

C. Figs. 14—19, 20—26.

Two cases are to be considered under this head: 1st, those
in which the spiral introduces no new half-metameres on
either side; 2ndly, those cases where a new half-metamere is
introduced in front of the 15th metamere, so that one half of
15 becomes united to 16 (15—16), and the other half that
remains is united to half of 14 (15—14).

lst case.—In fig. 14 a spiral is found involving the 7th,
8th, and 9th metameres. It belongs to category viit (Pl. 40,
fig. vi11). No new half-metameres are introduced, and the
vasa deferentia open on their normal (15th) metamere.

In fig. 15 a similar spiral involves the 12th, 13th, and 14th
metameres. Vasa deferentia as before, normal.

410 T, H. MORGAN.

In fig. 16 a similar spiral involves the llth, 12th, and 13th
metameres. Vasa deferentia normal.

In fig. 17 a spiral is present that involves the 12th, 13th,
and 14th metameres. The lower end of the figure (posterior
end of the spiral) turns to one side and ends in a point. The
openings of the vasa deferentia were not apparent on the
surface, but dissection showed the ovaries were in the 138th
metamere, that is within the spiral. Presumably, then, the
vasa deferentia were also in their proper metamere.

In fig. 18 a longer spiral involves the 6th, 7th, 8th, and
9th metameres. The spiral belongs to category x (Pl. 40,
fig. x). This form does not introduce any extra half-meta-
meres, and we find the vasa deferentia normal in position.

In fig. 19 a similar (reversed) spiral is found. It includes
the llth, 12th, 13th, and 14th metameres. The vasa defe-
rentia open normally.

Conclusion.—In these six cases where a spiral is intro-
duced in front of the 15th metamere the position of the
openings of the vasa deferentia is not affected. No better
demonstration is needed to show that the spiral is due simply
to a rearrangement of the segments involving their union
across the middle line. It will also be noticed that in all of
these cases the false union is made on the dorsal side of the
worm.

2nd case.—In fig. 20, a, B, c, are drawn the upper and
lower surfaces of a worm, and (in c)'a reconstruction of the
spiral. The 2nd, 3rd, and 4th metameres are involved, and
an extra half-metamere (3rd) is introduced upon the left side
of the body. The modification belongs to category vit
(Pl. 40, fig. vir). We find corresponding to this intercalation
that the openings of the vasa deferentia are affected, aud
appear on different but consecutive segments. Each, however,
opens on its own 15th half-metamere. In other words, the
15th half-metameres still develop their proper reproductive
openings, and in consequence of an interpolated half-metamere
in the spiral anteriorly the two openings do not fall into the
same metamere,

A STUDY OF METAMERISM. 411

In fig. 21, a, B, c, a somewhat similar spiral involves the
9th, 10th, and 11th metameres, and also introduces an extra
half-metamere on the left side. The external openings of the
vasa deferentia were not found, but dissection showed that the
ovaries were present in consecutive segments, i.e. each in the
13th half-metamere of its side. The explanation is the same
as in the last case.

In fig. 22, a, B, c, we find a spiral belonging to category
1x (PI. 40, fig. rx). An extra half-segment is introduced.
Nevertheless the openings of the vasa deferentia seemed both
present on the same ring (metamere). This metamere is made
up of the 15th half-metamere of one side and the 16th half-
metamere of the other. In fig. 23, a, B, c, D, are shown two
modifications in the anterior end. The first of these is brought
about by an incomplete lateral union (c) of the line between
the 5th and Gth metameres. More posteriorly a spiral is intro-
duced that involves the 13th and 14th metameres. An extra
half-metamere, as shown in D, is introduced on the left side.
The external openings of the vasa deferentia were not seen.
Dissection showed that only one ovary was developed, and that
on the right side (or at least only one was found). This was
in its normal half-metamere. The seminal receptacles were in
their normal segments (9th and 10th), which showed that the
more anterior modification (failure of lines to meet) did not
affect subsequent metameres.

In fig. 24, a, B, c, a short spiral involving the 5th and 6th
metameres is present, and introduces an extra half-metamere
on the left side. External openings of vasa deferentia were
not found. Dissection showed that the ovaries were present
in the same segment (13th of one side, 14th of other side). I
could not fully decide whether this case was due to a true in-
tercalation of a half-segment or to an overgrowth of the middle
line by the right half 6th metamere, due to an incomplete
dorsal growth of the left half 5th metamere. The latter inter-
pretation would be more in accordance with the presence of
the ovaries in the 18th—14th metamere, which would then be
only the 13th metamere. Moreover, as this is the only case

412 T, H. MORGAN.

recorded of such a modification, I think the latter interpreta-
tion is probably correct.

In fig. 25, a,B,c, arather complicated spiral involves the 10th,
11th, 12th, and 13th metameres,as showninc. An additional
half-segment is introduced on the left side. Nevertheless both
of the openings of the vasa deferentia appear on the same seg-
ment (15th—14th).

Conclusions.—The number of worms recorded is too small
to give any conclusive data. In two cases (figs. 20, 21), where
an extra half-segment is introduced, each opening of the vas
deferens is on its normal half-segment, and therefore on con-
secutive rings.

In two other cases (figs. 22, 25) the openings of the vasa
deferentia occur on the same ring (14th—15th in one case,
15th—16th in the other). In a third case only one ovary was
found (fig. 23), and in a fourth the nature of the spiral was
doubtful.

D. Figs. 26—82.

Two kinds of modification belong to this category: Ist,
those where the spiral passes through the 15th metamere ;
and 2nd, those cases where the 15th metamere begins or ends
a spiral.

lst Case.—In fig. 26 is shown a dorsal view of a worm with
a spiral involving the 11th to the 24th metameres. The spiral
winds around the body thirteen times. The number of half-
segments is not increased by the spiral. Looking at the
lower surface of the worm the vasa deferentia are found on
the 15th metamere, and are in no way affected by the fact
that the metamere is part of a long spiral, and not a closed
ring.

In fig. 27, A, B, C, a spiral involves the 10th to the 16th meta-
meres. Looked at from below, B, the vasa deferentia are seen
on the 15th metamere. The reconstruction in c shows that
the 15th metamere is a part of the spiral, and still carries the
openings of the vasa deferentia.

In fig. 28 a spiral is found involving the 13th to the 24th

A STUDY OF METAMERISM. 413

metameres. The openings of the vasa deferentia were not
seen on the outside, but dissection showed the ovaries to be
present in the 13th metamere.

Conclusion.—In these three cases where the spiral in-
volves the 15th metamere the reproductive organs appear in
their normal halfmetameres. Im all of these cases the false
union of metameres that produces the spiral is above. The
lower surfaces of the half-metameres join one another as under
normal condition,—that is, the 14th below and on the right
side joins the 14th, &c. On the dorsal side the reverse is
true,—that is, the 14th joins the 15th half-metamere, &c. As
the reproductive organs are in the lower part of the segment,
we would expect to find them still in their normal segment
(and not in any way modified), and this we do find in all cases
recorded.

2nd Case.—In fig. 29, a,B,c, a spiral is drawn involving the
14th, 15th, and 16th metameres. The spiral begins in the
14th half compound metamere (14—14). The openings of the
vasa deferentia are present in their normal half-metameres.

In fig. 30, a,B,c, a spiral involves the 15th, 16th, and 17th
metameres. Here, as in the last case, the openings of the vasa
deferentia are not affected.

In fig. 31, a,B, a similar spiral involves the 15th, 16th, 17th,
and 18th metameres. Here a half compound metamere
(15—+5), beginning the spiral, carries the openings of the vasa
deferentia on their normal 15th metamere.

In fig. 32, A,B, a spiral involves the 15th, 16th, and 17th
metameres. The openings of the vasa deferentia occur on
their normal metamere, which forms the anterior limb of the
half compound metamere (15—+5) that begins the spiral.

Conclusion.—In none of these last cases does the presence
of a half compound metamere involving the 15th metamere
affect the openings of the vasa deferentia.

General Conclusion from Category D.—Nearly every
case points unmistakably to the conclusion that, however
falsely the segments may unite in front of the 15th metamere,
or so that the 15th metamere is involved in a spiral or half

414 T, H. MORGAN.

compound metamere, still the openings of the vasa deferentia
appear on their 15th half-metamere.

It is also very noticeable that in by far the larger majority
of cases the false unions of the metameres are on the dorsal
surface.

EK. Figs. 33—46.

The results recorded in the preceding section are, as has
been said, fully in accord with the statement that, however
varied the union of the half-metameres across the median line
may be, still the openings of the vasa deferentia occur in most
cases on their normal (15th) half-metamere.

In the present category (£), where we should hope to find
this same relation to hold, the matter stands otherwise. In
only one case, viz. in that first given, does the anticipated
result follow. Many of the cases are unintelligible, or
nearly so.

I used in my earlier paper the first figure referred to above
to illustrate the explanation offered as to the value of the
compound metamere (there called the split metamere). I still
believe that it does this, but I was not then aware how much
in the minority, as far as numbers go, this modification really
is. I fear that by picking out this one case as an illustration
in my short preliminary communication I exaggerated the
value of the modifications as furnishing evidence of my view.
I was careful to state only that this was “one of the most in-
structive cases,. .. . and gives us a clue by which to interpret
the split metamere.” This statement I still think is true, but
by far the weightier evidence for my conclusion is now fur-
nished by the cases recorded in the preceding sections rather
than in this one.

Pl. 41, fig. 33, 4,8, shows the anterior end of a worm having
the 10th metamere compound (10—+1%) ; there is, therefore,
one more metamere on one side of the body (the left) than on
the other. Correspondingly the openings of the vasa de-
ferentia are not on the same metamere, but on consecutive |
ones. Each opening is on the 15th half-metamere of its side.
Each half-metamere has developed its normal structures, but

A STUDY OF METAMERISM. 415

owing to the shifting of the metameres, due to the introduc-
tion of an extra half-metamere, the halves of the 15th meta-
mere do not unite with one another.

In fig. 34 two modifications of the metameres are introduced
in front of the 15th metamere. Anteriorly the line between
the 4th and 5th metameres fails to meet above. The 7th
metamere is compound, so that an extra half-metamere is
introduced on one side (the left); nevertheless the two vasa
deferentia open on the same segment (16—15).

In fig. 35, a,B,c, we find again two separate modifications.
A small half of the 3rd metamere forms above a union with
half of the same segment on the right side, and below it is
wedged in between the 2nd and 4th metameres. Again, further
back a half-segment is introduced on the left side between the
8th and 10th metameres (or, if the more anterior extra half be
counted, between the 9th and 10th). The openings of the vasa
deferentia were not made out, but the ovaries occurred in
consecutive segments, i. e. on the 13th half-metamere of each
side if the more anterior half-metamere (3rd) is not counted.
If this was counted, then the ovaries would lie in the 14th
half-metamere of the left side, and in the 13th half-metamere
of the right side.

In fig. 36, a,B,c, a spiral! is introduced that involves the 11th,
12th, and 13th metameres, as shown in c. An extra half-
metamere (12th) is introduced on the right side. The open-
ings of the vasa deferentia were not seen, but dissection showed
that the ovaries are in consecutive segments. In the left
side the ovary occurred in the 14th half-metamere, and on the
right side in the 16th half-metamere.

In fig. 37, a, B, the 12th metamere is compound, and we find
that the vasa deferentia open on correspondingly shifted meta-
meres. But, besides this, the openings of the right side of the
body have doubled, so that the 15th—16th metamere has two
openings, right and left.

In fig. 38, a,B,c (lateral), a half-metamere is introduced
between metameres 3 and 4. The openings of the vasa

! This case belongs to a preceding category, D.

416 T. H. MORGAN.

deferentia are on the same segment. (It seems to me very
doubtful whether the surface line of the 4th metamere is due
to a half-metamere introduced.)

In fig. 89, a,B,c, a compound metamere occurs on the 15th
segment (15—15). The opening of the right vasa deferentia
alone was visible on the exterior. The ovaries were found in
the 13th metamere.

In fig. 40, a,B, we find a spiral involving the 8th, 9th, and
10th metameres, introducing an extra half-metamere on the
right side. The vasa deferentia open, nevertheless, on the
same metamere (15th—16th). Another modification starts
at the 15th metamere, and involves a few more posterior
segments.

In fig. 41, a,B,c, we find four compound metameres (see c),
followed by a spiral that involves segments immediately in
front of the region of the opening of the vasa deferentia.
Fifteen half-metameres of the right side correspond to twenty
of the left side. The openings of the vasa deferentia appear
on the same ring, and this ring is made up of the 16th half-
metamere of one side and the 2lst half-metamere of the
other.

In fig. 42, 4,B,c, we find a compound metamere (3—3), fol-
lowed by a complicated spiral (c) that involves the 14th to the
21st metameres. The openings of the vasa deferentia were not
seen on the outside. Unfortunately I have no records of the
dissection of this worm.

In fig. 43, a, B, c, we find two spirals and a compound meta-
mere in the anterior region. The openings of the vasa de-
ferentia occur in the last spiral. The opening on the left side
is in the 18th half-metamere, and on the right side in the 16th
metamere.

In fig. 44, a, B, a compound metamere (6—#) is first
found, followed by a spiral belonging to category vii, Pl. 40,
fig. vit, and this followed by another compound metamere
(18—15). Three extra half-metameres are introduced on
the right side. The openings of the vasa deferentia are

1 Belongs to another category, i.e. D.

A STUDY OF METAMERISM. 417

doubled on both sides. On the left side the openings are
found on the 15th and 16th half-metameres. On the right
side of the body the openings appear on the 20th and 2lst
half-metameres.

In fig. 45, a,B,c, we find the 6th metamere somewhat reduced
on the left side. We have an intercalated half-segment (15th)
on the right side, followed by a spiral involving the 15th to the
19th metameres. The latter introduces an extra half-metamere
on the right side. The openings of the vasa deferentia were
not found. Dissection showed the ovaries in the 18th half-
metamere of the left side, and in the 21st half-metamere of
the right side (in the half-metamere that follows the 18th on
the other side).

In fig. 46, a,B,c, an extremely modified anterior end of a
worm is drawn. As the reconstruction shows, many extra
metameres are introduced on the left side of the body. In
the exterior of the 24th and 25th left half-metameres swollen
areas seem to point to openings of the vasa deferentia, but
none appear on the right side. In the dissection of this
worm no ovaries could be found, and no data throwing any
light on its “ make-up” were obtained.

Conclusion.—It is very difficult to reach any conclusion
in the face of so much conflicting evidence. It would seem
as though an intercalated segment might or might not affect
the position of the reproductive organs. If to get out of the
dilemma we assume that some of these cases are not due to
the intercalation of true half-metameres, but to unilateral
division of metameres, then other difficulties are encountered
as great or greater than those that appear on our first assump-
tion. It was only after the preceding pages were written and
the hopelessness of any explanation realised that a partial
solution—or what seems at first to be such—suggested itself
tome. Such cases as those drawn in figs. 41 and 46 might
be due to a regeneration in the adult of anterior meta-
meres. We see that this might offer an easy escape from
thedilemma. It seemed also to explain why the segments in some
of these worms that carry the openings of the vasa deferentia

418 T, H. MORGAN.

and the ovaries are so far from the anterior end. Here more
segments might have regenerated than were lost.

Fortunately we are dealing with an hypothesis that could
be tested by experiment. I at once set to work to determine
whether or not the same number of segments reappeared when
the anterior ends were cut off. I also wished to find out
whether a greater number of modifications than in the average
worms were thus produced. A later section gives the result
of these experiences.

IV. Stupy or Emsryos.

Capsules containing embryos of L. foetidus were collected,
and the embryos killed immediately on their emergence from
the cocoon. Out of 170 embryos there were twenty-five that
showed abnormalities in the arrangement of the rings. This
is in the proportion of 1 to 5:2. In other words, there are
fewer cases of abnormalities in young worms than amongst
adults. In the latter the proportion was one to two.

This apparent contradiction finds its explanation in the fact
that the adults are often found regenerating lost metameres,
and proportionately the number of abnormalities in these
newly formed parts is greater than in the embryos. This will
be taken up more fully in another section.

The number of metameres in the young worms that have
just left the capsules depends to some extent, as the following
table shows, upon the size of the young worms. ‘The size seems
to be in general connected with the number of embryos that
develop in the capsule. Amongst the worms in the same
capsule, however, there are variations in size, and we can only
make some such general statement as this, that where many
embryos are present in the same capsule they are each smaller
when they leave the cocoon than when fewer embryos are
present.

The following table indicates, in a rough way, the rela-
tion between the size of the embryos and the number of
metameres :

A STUDY OF METAMERISM.

Size of Worm.

419

No. 1 111 metameres Very large.
ene 102 ef Large.
3 100 : ms
awed: 85. a Small.
shat : ‘ 85 3 : F Bs
Sets ; : 78 . : ‘ a
ey 67 : Very small.

Numbers 6 and 7 came from a capsule that contained 12 worms.

The following table gives the record of twelve young worms
(including those of the preceding table), showing the number
of metameres present :

85
100
102

85
107
106
11
: : 78
9 ‘ : ; ; 67
Seal : 87
Pa . 85
, 12 97.

12 ) 1110
92°5

aonronrrFr wnr

I attempted to determine whether the young worms that
had abnormalities came in proportionally larger numbers from
individual capsules. Capsules were also examined to see
whether the number of embryos in a capsule was a factor in
the production of abnormalities. This examination ought also
to show whether the eggs from certain worms had a tendency
to produce abnormally joined metameres; if, for instance, in
all the capsules that contained a large number of embryos
many were abnormal, it would follow, with a great deal of pro-
bability, that the crowding (or smaller amount of food, &c.)
brought about (directly or indirectly) the result. If, on the
other hand, certain capsules, regardless of the number of

420 T. H. MORGAN.

worms present, showed a large percentage of abnormalities,
then the cause probably arose in the egg.

The following tables give the data that I have collected.
The number of recorded cases is smaller than I could have
wished, but the evidence is all in one direction and fairly con-
vincing. The capsules were isolated and the worms collected
as soon as they crawled out.

Number of embryos in a capsule. Number of abnormal worms.
No. 1 6 ‘ : ‘ 1
the 3 0
aye 6 2
si 6 0
San gs 12 il
5 (6 5 1
ar Oe 3 1
eae ss 4 0
ar Ae 12 1
” 10 5 ]
55 ake 12 1
” 12 8 if

Conclusion.—In none of the capsules is there a prepon-
derance of abnormalities, nor does there seem to be any rela-
tion between the number of worms in a capsule and the
presence or absence of wrongly united metameres.

We must conclude that neither of the possible causes sug-
gested above is active in producing the abnormal embryos.

We get no clue from these data as to any outside forces
producing the result ; and although we have not by any means
exhausted all the possibilities, still it seems not improbable
that in each particular case local internal conditions determine
the result.

All of the commoner forms of abnormalities recorded for the
adult worms are also found in the embryo. This applies to
abnormalities in the anterior part of the body as well as in
other parts. There are more abnormalities in the tail end
than in the head end.

A few typical abnormalities found in some of these embryos
(young worms just out of the capsule) are shown in PI. 42,

A STUDY OF METAMERISM. 421

figs. 75—78. In fig. 75, a,B, isdrawn a compound metamere ;
in fig. 76, 4, B, a spiral of category vir; in fig. 77 a spiral of
category vill; and in fig. 78 a longer spiral of category x.

A number of adult worms was examined, and the following
table shows the number of metameres inthem. Mature worms
were chosen as far as possible.

No: 101
ae 102
me 92
ea: 85 (perhaps regenerating).
nea oF
EP) 6 97
Pe iL 105
at KOL - ; ‘ : 95
Pree |S ie : é : 101
lO 88 (probably regenerating).
ro es 9
Or 85
513 eM
alin? 84.
mela : 95
15 ) 1416

94°4 average.

It will be seen, if we may judge by the figures given, that
few, if any, new segments are added to the worm after it leaves
the capsule.

Another count of 25 worms gave the following figures :—
99, 89, 85, 100, 90, 96, 95, 95, 97, 101, 101; 105, 101, 106,
88, 84, 88, 95, 100, 92, 100, 91, 105, 92, 96. These numbers
give an average of 99°6. If we combine this with the average
of the 15 cases given above (94°4) we get 97 for the average
number of segments of 40 worms.

V. ABNORMALITIES AT THE PostreRIon EnpD.

In one lot of 525 worms 40 were found showing newly
formed posterior ends due to regeneration.
In these 40 worms we find—

422 T. H. MORGAN.

A. Regenerated portion just beginning to show metameres. 4

B. Regenerated portion normal

a. Number of new metameres : : 6
Bs hiss - : : 6
c. Regenerated portion having one abnormality . 12
a. Number of new metameres ; : 19
b. a3 55 6
(p: Be e 30
ee, if 16
é. ss As 25
iy 0 2» 3
UE Bs $3 : q 4
h. 0 = 5 ; ll
a 39 33 4
je ~ oS : : (several)
k. 5 5 : : 15
l. as 33 : : 6
p. Regenerated portion with two abnormalities. 7 OF
(Number of uew metameres not recorded.)
E. Regenerated portion with three abnormalities . . 4
rv. Regenerated end with more than three abnormalities . 12

In p, £, and F the number of new segments was not counted,
but measurements were taken of the regenerated portions for
comparison with c, and were as follows:

c. D. E. F.
1 18 mm. 26 mm. 27 mm. 25 mm.
2 IBY Ay ilies 20). ¥s5 gore
3 10 ” 21 ” 9 oP) 25 3
4 (ae 16 8 ss 22 =,
5 Bes SS 4 ie" oo.
6 oh GE Sie es 18 mm. Hl ee
7 avs 6) 98 ,, 19; 5
8 ae 164 mm. i es
9 Aes 145;
10 2 ”? 13 ”
ll 2 5 12,
12 2 + As
12) 80 , 13\) 294 14)

65 mm, average length. 182 mm,

A STUDY OF METAMERISM. 423

These measurements show conclusively that the greater
number of new abnormalities occur not more frequently at the
beginning of the newly formed portion than throughout all the
later period of growth.

Summary.—In the 40 cases mentioned above 4 were not
well enough developed to furnish any data. In the36 that remain
there were only two that did not show any abnormality, and it
is noticeable that each of these had only formed at the time a
few new segments (six each). We must conclude from the
data that about 18 worms out of every 19 would show abnor-
malities in the regenerated posterior end.

In these same worms the number of abnormalities in the
region of the body anterior to the regenerated end was recorded.
This gives us data to show whether the irregularities are due to
inherited peculiarities of the tissue or to the conditions acting
during regeneration. In the 12 cases recorded in c 9 were
normal in front of the regenerated portion, and 3 showed ab-
normalities,—that is, as 1to 3. In the6 cases recorded in p, 5
were normal in front and | abnormal (1 to 5).

In the 4 cases in £ there were no abnormalities in front. In
the 12 cases recorded in r there were 2 cases unrecorded, and
the remaining 10 had no abnormalities in front.

To sum up, the evidence points unmistakably to the con-
clusion that the abnormalities found so frequently in regene-
rated portions are due to the conditions acting during regene-
tion (from within or from without), and in no way connected
with an hereditary. tendency to be more abnormal in one case
thaninanother. That is to say, the tissues of a worm that has
developed normally from the egg are just as apt to develop
irregularly in regenerating as are the tissues that have de-
veloped irregularities during embryonic growth. Heredity
seems to have nothing to do with causing the abnormalities.
In fact, one might say that the tissues inherit a strong tendency
to regenerate normal metameres, but the means at command
are so imperfect that abnormal results are frequent.

Will the large proportion of abnormalities present in regene-
rated worms, taken in connection with the number of worms found

VoL. 37, PART 4,—NEW SER. FF

424 T. H. MORGAN.

regenerating naturally, account for the difference between
adults and embryos? We have seen the number of worms
found showing regenerated (and regenerating) posterior ends
to be 1 in every 18 (1 to 12); therefore in a lot of 225 we
should expect to find 19 worms having new posterior ends. In
our lot of 225 worms we found 100 abnormal worms. If we
deduct from this 100 the number of those that should have
regenerated tails (19), we should have 81 cases remaining.
The proportion would then be 81 to 225, or 1 to 2°8 (approxi-
mately 1 to 8). We found, however, that the proportion of
abnormal to normal embryos in A. foetida was only 1 to 5.

The difference that is still found is probably due to several
causes, and these are not in the least of a hypothetical nature.
Firstly, after a time a new regenerated end cannot be distin-
guished from the rest of the body. Secondly, a small per-
centage of regeneration must also take place in the anterior
end. Thirdly, the difficulty of seeing the abnormalities in
the embryos is much greater than in the adult, and defects
may have occasionally escaped even a careful examination.
Fourthly, the data is drawn from too small a number of cases
to make an exact agreement very probable, even if it did exist.

In the light of these conditions I think the closeness of the
result is as near as could be expected.

A number of worms were examined in which the posterior
end of the body was regenerating, to see if any could be found
in which the tip of the “ tail” showed, in process of formation,
modifications of the typical arrangement. Several were found,
a few of which are drawn in PI. 43, figs. 79—81.

The first of these (fig. 79, 4,B,c) shows a compound metamere
in process of formation from the terminal piece or telson. On
the side that forms the “split” the telson is proportionately
longer than on the other.

In fig. 80, a, B, c, the end of the body is very much modified,
and a spiral is in process of formation. The reconstruction (c)
shows sufficiently the conditions present.

In fig. 81, a, B, c, another modification is present. Above and
on one side of the telson the outline of a metamere is seen, while

A STUDY OF METAMERISM., 425

below this line only extends a third across. Whether this
would ultimately form a complete ring, or whether we have
here the beginning of a spiral starting with a half-compound
metamere below, is uncertain.

Several vertical longitudinal sections were made through
these “ tails,” but no additional information was gained from
the sections. The irregularities in the positions of the rudi-
mentary body cavities was very apparent.

VI. Mopirications oF INTERNAL STRUCTURES.

Before any conclusion can be reached as to the value of the
irregularities seen on the surface, we must examine the condi-
tion of the internal organs, particularly the arrangement of
the septa.

We can formulate this general statement, that the ar-
rangement of the septa generally conforms to the
curves of the lines seen on the surface. This means
that the phenomena are deep-seated, and that all the struc-
tures of the body are involved in the new arrangement, and
not merely the external surface lines.

Exceptional cases are not uncommon, in which we find two
principal departures from the rule. First, the arrangement of
the septa does not always agree with the surface lines. Some-
times the arrangement of the septa is more perfect, and some-
times simply different. Secondly, the arrangement of the
somatic attachment of the septa is sometimes different from
the splanchnic attachment. Usually in this case the somatic
portions of the septa follow the lines found on the surface,
while the splanchnic attachment differs in its arrangement,
and is usually more irregular in its form.

If a worm having a compound metamere be opened we find
the septa arranged as shown in fig. 61, 4,B. One septum (s)
reaches only to the mid-dorsal line. If we remove the diges-
tive tract we find the same septum ending freely below, just
before reaching the middle line(s’). This half-septum is found
to correspond to the “split” in the compound metamere, and
lies between the two half-metameres of that side.

426 T. H. MORGAN.

The septum is attached on the somatic side to the body-wall
along the “split.” Centrally it is attached to one half of the
splanchnic wall of the digestive tract. We should find on
sectioning such a worm that the half-septum had a double
wall, and was like a true half of a septum in every respect. We
should see also, as a result of this arrangement, that the body
cavity of the single half-metamere is continuous in the mid-
ventrai and mid-dorsal lines with two half-metameres on the
opposite side.

The number of the nephridia corresponds to the number of
the half-metameres, as shown in fig. 62, so that there is one
more nephridium on one side than on the other.

The nervous system is often modified, as shown in fig. 62.
Here each of the half-metameres is seen to receive its full
number of nerves from the ventral cord. This is not always
the case, for, as shown in fig. 61, B, the double side gets only
the supply normal for a single half-metamere. In figs. 63
and 64, B, we see other irregularities in the nerve-supply.
This may be connected with the degree to which the half-
septum approaches to the mid-ventral line. When it falls far
short the nerve-supply is less than when it reaches the mid-
line. The nephridia, however, that lie laterally in the body are
invariably doubled. In fig. 64, a, the surface line between the
half-metameres of the compound metameres is not very
extensive. The figure represents the body-wall flattened out
(the worm had been previously opened). Fig. 64, B, shows
that the septum s is attached to the somatic wall over a
correspondingly small area.

The condition of the septa in the spiral modification is very
interesting. One of the simplest cases is shown in fig. 65, A,B.
The first figure shows a short spiral beginning and ending
above (category vit1). On opening the worm from below
(fig. 65, B) the septa are seen to follow the same arrangement,
both on the body-walls and on the intestine. There results a
single spirally winding septum beginning with the
half-septum of the compound metamere anteriorly,
and ending with another compound metamere poste-

A STUDY OF METAMERISM. 427

riorly. In other words, there is a continuous body
cavity (cclom) lying between the coils of the septum,
and this cavity is continuous from the anterior to the
posterior end of the spiral.

Fig. 66 shows a similar spiral, the only difference between
this and the last being that in the former (fig. 65) the anterior
end of the spiral abuts against the septum in front, while here
it ends freely.

Fig. 67, a,B,c, shows a similar but longer spiral. Opened
dorsally the septa are found as in B, and it will be seen that
they follow the course of the outer spiral. In the middle line
the septa over the intestine show a tendency to irregularity
(even more so than the figure shows). In c the septa are
drawn as seen from below. They run obliquely over the lower
wall of the digestive tract. In both B and c the figures were
drawn from preparations made by dissecting free the digestive
tract and its attached septa from the body-walls.

These examples will suffice for the regular forms, where
surface spirals and septal spirals agree; but this, as stated
above, is not always the case.

In fig. 68, 4B, we find on the surface a spiral shown in a;
that belongs to category x (Pl. 40, fig. x). Opening the worm
we find that the septa are also spirally arranged, but the spiral
is shorter than the surface spiral by one turn, although both
begin and end on the upper surface.

Fig. 69, a B, shows a parallel case, where the internal spiral
is shorter than the external. In this case the septal spiral is
shown only in its somatic attachment.

In fig. 70, a,B,c, we find a surface spiral like the last. We
find the septa arranged dorsally as shown in B, and on the
ventral as shown in c. It will be seen that the external
and internal (septal) spirals belong to different forms (cate-
gories). ‘The external belongs to category x1 (Pl. 40, fig. xr),
while the septal belongs to category 1x (extended), Pl. 40,
fig ix;

Fig. 71 shows a form in which the external line between two
metameres failed to meet in the mid-dorsal line. In dissection

428 T, H, MORGAN.

the arrangement of the septa was found to be normal ; that is
to say, the septum has met normally across the mid-dorsal line,
but the external lines have failed to do so.

In fig. 72, a B, a surface spiral is drawn, and in B the arrange-
ment of the septa beneath is shown. The septa shows an
abnormal arrangement, but this does not correspond to the
surface lines: it is more of an approach to the normal.

Fig. 73 is from a compound metamere but in which the
septa form a simple spiral making a little more than one turn
of the body.

It is not always easy to picture to one’s self the relation
existing between septal and surface lines when the two do
not correspond. In the simpler cases it is easily understood.
When two spirals are formed of different lengths, it is due to
the fact that the septa sometimes unite with one another
beneath the surface before the surface spiral is brought to an
end. That is, if both start together one may continue for a
longer time than the other. We must conclude, then, that
while the two usually vary together, yet they may vary inde-
pendently. 1 have dissected a far greater number of worms
than recorded above, and from that number have selected
those given as the most interesting examples to illustrate the
main points that concern us at present. Many other rela-
tions have suggested themselves, but I have only wished to go
into the subject as far as concerns the matter in hand.

VII. Stupy or Potycu#tous ANNELIDS AND LEECHEsS.

A large number of species of polychztous Annelids have
been examined, and many modifications in the arrangement
of the metameres have been found in them. From this number
I have chosen a very few for description, since a large number
of cases have been given in the papers by Cori and Buchanan.

The modifications that are figured here are all taken from
specimens of Amphinome.

In fig. 82, a B, we find above B a compound metamere with
the split on the left side. In the ventral side of the worm, a,
we find the left anterior half of the compound metamere

A STUDY OF METAMERISM. 429

wedged in ventrally between two segments, while the other
left half completes the ring. We have here evidently a modi-
fication of the compound metamere similar to category 11,
Pl. 40, fig. 11.

In Pl. 48, fig. 88, a, B, we see a metamere imperfectly deve-
loped on the right side. The dorsal parapodium and gill are
absent on this side, but the ventral parapodium is present, as
seen in side view in B. We have here probably a case of
incomplete development of a metamere of one side.

Another somewhat similar modification is shown in fig. 84,
AB. The first of these shows the ventral side, and the middle
segment of the three drawn is seen to be imperfectly deve-
loped. The ventral parapodium of the right side (left of
figure) is present, but the metamere does not reach as far
dorsally as the line of dorsal parapodia (see B). On the left
side of the body we find the metamere completely undeveloped.
We seem to have here a half-segment of the right side that
has not fully developed even on that side, but has overgrown
the mid-ventral line on the left side.

Fig. 85 shows a spiral beginning and ending on the dorsal
surface. Seven segments are involved, giving five turns to
the spiral.

The large number of abnormal forms found in Amphinome
is in part due no doubt to the frequent regeneration of por-
tions of the body that takes place. The segments are broad,
and for this reason it is surprising to find abnormal com-
bination of the metameres so frequent. In other poly-
chetous Annelids, where the metameres are very narrow, the
false unions and imperfections of the metameres are very
numerous.

Cori has figured several modifications, and I have found
similar ones that are exceedingly difficult to explain as simply
due to modifications of half-metameres. Such cases are
particularly common in polychetous Annelids.

The majority of these very abnormal forms are, I think, due
to regeneration rather than to egg-variation. Since the method
of regeneration is more irregular than the growth of the

430 T. H. MORGAN.

embryo from the egg, it becomes much more difficult to
explain on the metamere assumption these variations due
to regeneration.

A number of leeches were examined to see whether the
annuli that make up the broad metameres ever showed false
unions like those found in the earthworm’s metameres.
Such modifications are not rare.

In fig. 86, a B, the line between the second and third annuli
of the ninth metamere failed to extend above over to the left
side of the body. Again, in this same metamere we find that
the line between the 5th and 6th annuli turns forward to join
the line between the 4th and 5th, so that the last annulus of
this metamere forms a compound annulus with the first
annulus of the next metamere, producing a short spiral form.

A more complicated spiral is shown in fig. 87, 4,8. The
16th and 17th metameres are involved. The spiral com-
mencing in the 16th metamere runs over into the next meta-
mere, as the reconstruction B shows. The annuli are also
imperfectly joined at the sides, as shown in B. This last
result is due to a failure of the annuli to unite perfectly with
one another.

A third spiral-form, involving a double compound annulus,
is shown in fig. 88, 4, B. A half-compound annulus above
and on the right side forms a spiral with a half-double com-
pound annulus above and on the left side. The first half-
compound ring is the 3rd.

We see that the annuli of leeches show many of the same
sorts of false unions that are present in the metameres of the
earthworm, but in the leech this represents only a superficial
alteration.

The modifications in the leech would correspond to those
modifications occasionally found in the earthworm, where the
septa are normal but the surface markings abnormal in their
arrangement.

It will be noticed that in the figures given no new annuli
are introduced into the metameres, and only the method of
union of the annuli varies, This result may only be due tothe

A STUDY OF METAMERISM. 431

relatively few leeches examined, although the number was
sufficient to show that the modification involving additional
half-annuli, if it occurs, must be rare.

VIII. Summary.

In the papers of Cori (18) and myself (32) the same ex-
planation was offered as to the origin of the simpler cases of
abnormal unions. Cori said, “ Nun kann es sich ereigen, dass
in der einen K6rperhalfte wahrend der Entwickelungsperiode ein
Ursegment mehr gebildet wird, dem auf der Gegenseite kein
Ursegment entspricht. Auf diese Weise wird die Bildung jener
Falle von Schaltsegmenten verstiindlich, wie sie im vorherge-
henden anatomisch beschrieben worden. Betreffend das Ver-
haltnis eines Schaltsegmentes zu den iibrigen Metameren des
Korpers sind zwei Méglichkeiten vorhanden. Es kann sich ein
Schaltsegment vollstandig ausbilden, und kann von den anderen
Segmenten abgegrenzt bleiben oder es geht eine Verbindung
mit dem vor oder nach folgenden Metamer ein.’

My own statement was very similar:—‘‘* We know that in
the embryo the metameres are laid down right and left of the
middle line of the body in blocks of mesodermal tissue; that
under normal conditions these hollow blocks come to lie
exactly opposite (right and left of) one another, so that the
opposite pairs unite across the median dorsal and ventral lines.
If we conceive that the blocks are slightly displaced on one
side, or that two consecutive blocks of one side are smaller
than two of the opposite side, we may have, as a necessary
mechanical result of the relative position of the block, a split
metamere.” ‘To account for the spiral arrangement I said,
and I am still of the opinion that this is the true explanation,
“Tf we imagine one of the mesodermic blocks of one side to
be larger either above or below (but not both above and below),
so that above (let us say) it opens into two body-cavities of the
opposite side, while below it opens into but one, then we have
produced the conditions necessary to start the spiral. Each
of the consecutive blocks on the same side as the supposed
larger block will open below into its proper opposite block, but

432 T, H. MORGAN.

above (on account of the first displacement) into the one lying
in reality behind it. . . . The spirals when once started do not
run on continuously, but end after passing around the body
several times. The ending likewise finds its explanation in the
inequality of the blocks of opposite sides.” I did not in this
preliminary paper consider a point with respect to the ending
of these spirals that may be spoken of here.

If a spiral once started was dependent in order to end, on
another accidental displacement appearing, so that the spiral
is, as it were, satisfied, the chances are that most spirals would
reach a prodigious length before terminating. As a matter of
fact the spirals are generally short, the shorter the commoner.
This shows, I think, that the conditions, after a spiral has been
started, are of such a nature that the chances are the spiral will
soon end itself. It is easy to offer a formal explanation of
why this is so. The shifting of the blocks by the double union
established at the anterior end! will tend to displace the block
behind, so that two of one side will come to correspond to one
of the other.

In the preceding section of the paper there are in reality
two topics that have been discussed in close connection with
each other. The question of the shifting of the reproductive
openings was considered in connection with the origin of the
compound metameres and spirals. The reason for doing so
was based on the hope that the one relation might help in the
interpretation of the other.

Moreover, in respect to the abnormal position of the open-
ings of the reproductive organs, we have considered first those
cases where, in worms with normal metameres, the openings
have shifted or doubled ; and secondly, those cases where there
were both abnormal metameres and abnormal openings of the
reproductive organs in the same worm. The former of these
subjects has been already noticed by previous authors. Beddard
(5) in 1886 published some most interesting observations on
variations in the reproductive organs of Perionyx excavatus.

1 It is assumed that the spiral starts at the anterior end. But it is also
possible that the shorter spiral arrangement involves all the segments,

A STUDY OF METAMERISM. 433

In a lot of 489 individuals there were found seventeen that
showed such variations.

Benham (6) described in 189] a variation in the openings of
the reproductive organs of Lumbricus herculeus. The vasa
deferentia appeared on the 14th and 15th metameres.!

With respect to the abnormal unions of metameres, and the
relation borne to the position of the reproductive organs, there
is little to be added to the conclusions reached in the sections
dealing with abnormalities in front of or included in the 15th
metamere. ‘T'wo cases were there sharply separated,—those
cases where there is a half-metamere more on one side than on
the other (C, second case, E), and those cases where the spiral
does not introduce an additional half-metamere on one side
(B, C, first case, D).

The results from the latter cases are in full harmony with
the interpretation of the spiral stated above. In those cases
where an additional half-metamere is introduced there are
certain examples that seem to verify the same statement ; but
there are other cases, and these form the majority, where the
results will not bear this interpretation. We have, then, to
assume, in some of the cases recorded, either that the half-
metameres are not equivalent to the normal, or else that the
reproductive organs do not give unfailing evidence of the
change that has taken place. Anyone who has studied the
facts will, I think, agree with me that it is far more probable
that the reproductive organs have appeared on a wrong (?)
half-metamere than that the half-blocks are not equivalent.

This leads us to another side of the problem, as yet not
dealt with. It is natural to think of every half-block of one
side as having a half-block that belongs to it on the other
side; in other words, to look upon one of these blocks as
predestined for the other, and we marvel to find them
separated. I think there are no real grounds for such an

1 Bateson’s memoir, ‘ Materials for the Study of Variation,’ that has just
reached me, refers to numerous cases of abnormalities in the reproductive
organs of earthworms recorded in a paper by Michaelsen, ‘Jahrb, Hamb,
wiss. Anat.,’ vill, 1891. I have not had access to this paper.

434, T. H. MORGAN.

expectation. If the facts recorded in the preceding pages
indicate anything clearly, it is that the union of the blocks
across the middle line is the natural result of their size and
of their position. Under normal conditions the blocks are
so placed that they unite pair for pair. If the order is dis-
turbed they unite differently. This is brought out most
strikingly in that case where there were 134 half-blocks on
one side and 118 half-blocks on the other. Here throughout
the body there are very few half-segments that are united with
the one opposite bearing the same numerical relation from
before backwards.

In the light of these statements I think we ought not to
look to any one half-metamere as predestined to carry under
all conditions the openings of this or that organ. Rather
should we expect that those segments, which happen to get so
placed in the body of the worm that they correspond to the
normal position for a particular organ, will go ahead
and develop that organ. We might say that the position in
which an organ will develop is determined by a definite region
of the body irrespective of how that region has gotten into the
necessary position, provided this happened when the cells were
still undifferentiated. The question is too large to discuss
here, and the facts too meagre.

There is a liability to err if the statements made above in
regard to the methods of union of the metameres be applied
to all cases. All that I contend for is that the majority of cases
will bear this interpretation. Sooner or later one is sure to
come across individuals where it is impossible to see the appli-
cation of the theory. Particularly have I found this true in
the Polycheta. I have met such cases quite often in examin-
ing the regenerated anterior ends of worms where the head had
been artificially cut off very obliquely. Sometimes pieces are
left, where the new head joins the old body, that cannot be
interpreted as half-metameres. Whether all of these unusual
cases are the result of regeneration I am not prepared to say.
Many of them seem to be; others, so far as I know, may have
come in from the egg,

A STUDY OF METAMERISM. 435

The results that have, on the whole, been the most puzzling
to me, and which I cannot pretend to entirely explain, are
some of those cases where the internal and the external
spirals do not agree. It seems quite certain that in the far
larger number of cases the false unions are at bottom, the
result of imperfect joining of the mesodermic blocks, and that
the ectodermic grooves mould themselves on the internal
arrangement.

But in cases of disagreement it seems clear that the outer
grooves have to a certain extent an independent individuality,
and may behave differently from the spirals in the mesoderm.
When we find similar spirals in antenne, &c., it is not, after
all, so surprising that the outer body-wall of the worm should
show this independent variation.

IX. Mopirications 1n ANTENN#H OF ARTHROPODS.

The antennz of Arthropods are made up of a series of seg-
ments or rings. In some Arthropods the segments are long
and relatively few in number. In others the segments are
narrow rings and very numerous. The latter kind are more
likely to show variations in the arrangement. The antennz
and antennules of the lobster (Homarus vulgaris) I found
showed many variations. Perhaps the lobster’s antenne and
autennules may owe many of these false arrangements to the
fact that if they have been lost new ones will regenerate, but
that all the abnormalities come from this source is extremely
improbable.

Ihave found in the antenne of the lobster nearly all the sorts
of variation in the arrangement of the rings that are to be
found in the earthworm. A few of these variations are shown
in Pl. 43, figs. 90 to 95. Fig. 90, a,B, shows a compound ring,
Fig. 91, A,B, shows a double compound ring, where three half-
rings of one side are united to one of the other. Both ofthese
modifications are common, particularly the former, at the broad
base of the antenne.

In nearly all cases the ‘‘splits” run in from the sides.
Rarely a modification similar to that shown in fig. 92, a,B, is

436 T. H. MORGAN.

found. Here the “split” is found over the broad surface of
one side, and not over the other. ‘This condition is never
found in the falsely joined metameres of the Annelids, for even
in those cases where a metamere appears below, and not above,
it is also found to extend over and on one side (lateral) of the
body.

Fig. 95, a, B,c, shows a spiral formed by a union of three
compound rings above and one below, as shown in the re-
construction.

Fig. 93, 4,B,c, shows a longer and more complicated spiral.
In the middle of this spiral we find a “double spiral” developed.

Fig. 94, a,B,c, alsoshows a complicated spiral. These spirals
in the antennz are similar to the spirals found in the earth-
worm, and need no further description.

The antennules of the lobster also show the same modifica-
tions. Both the antennz and the antennules are flattened
from above downwards, and the splits extend as a rule on the
lateral sides of the appendages. ‘The basal joints of nearly all
antennz show imperfect rings. The greater number of abnor-
malities occur in the proximal portion of the antenne, and are
found in less and less frequency as far as the middle region.
They are rarely found in the distal ends of the antenne, and
it is important to notice that distally the rings become much
longer relatively to their breadth. Another interesting condi-
tion is found. By far the greater number of “ splits” occur on
the convex side of the antenne and autennules, particularly in
the latter. There results a larger number of half-rings on the
convex side. In one case there were ten more half-rings on
the convex side of one antennule than on the opposite side.

When we see that it is normal for the antennules to bend
outward, i.e. to turn their convex side towards the middle
line, it seems fair to draw the conclusion that the tendency of
the antennule to turn to one side is the direct or indirect
cause of the greater number of rings on the convex side.
There is no reason for believing that the greater number of
rings causes the turning, since the antenna is bent whether
more rings are present on one side or not. Processes that go

A STUDY OF METAMERISM. 437

on in the cells of the antenne before each moult will deter-
mine where the new lines of division between the segments
come in, and in some way the bent condition of the antenna
seems to effect to a large extent the formation of the new
lines.

I have also examined the development of the antenne in the
lobster. At the third (?) moult the terminal division of the
endopodite of the antenna contains within it a rod of cells, and
these show alternate constrictions to form the lines between
consecutive rings. ‘These lines of constriction are exceedingly
near to one another, since the segment containing the new series
is considerably shorter than the total number of segments that
ultimately develop from it at the next moult. It is easy to
see that under these cramped conditions the arrangement of
the division lines might easily be disturbed by local causes.

Au examination of the antenne of certain insects where
regeneration of the antenne of the adult is not probable (and
that in the larva hardly possible) shows that in those forms
where there are a large number of rings, and each ring very
short, variations exist in the arrangement, while in those
forms with long rings such variations are absent (or not found
commonly).

In the locust I have not seen any misformed rings in the
antenne (in as many as fifty individuals examined). In the
cockroach (Blatta), where the rings are narrow, compound seg-
ments and short spirals are frequently found both in the larvee
and the adults.

X. AspNnorMAL MetTAMERISM OF LocUwstT.

It seemed to me highly improbable that such specialised
metameric forms as the Crustacea and Insecta, in which the
metameres are so definitely limited in function, number, and
position, should show any variations comparable to those in
the earthworm. However, I had brought to me a locust
(‘grasshopper ”’) in which there was a half-segment more on
one side of the abdomen than on the other.

In fig. 89, a, is drawn a dorsal view of this locust with the

438 T. H. MORGAN.

wingsremoved. Corresponding to the 5th metamere of the abdo-
men we find a compound metamere, so that one half-metamere
of the left side corresponds to two half-metameres of the right
side. A figure of the abdomen from below is drawn in
fig. 89, B, and the abdomen seen from the right side is seen in
fig. 89, c. On account of the interpolated half-metamere the
abdomen is bent towards the left. The interpolated segment
reaches exactly to the median line above and below. The
abdomen ends normally, having all of the accessory structures
of the male perfect in all details.

Fig. 89, c, shows that each of the half-metameres on the
right side of the compound metamere bears stigmata, so that
there is one more of these on the right than on the left side of
the body, i.e. eight on the right (one more than normal) and
seven on the left.

If we assume each of the right halves of the compound
metamere equivalent to a true half-metamere of the series,
then each of the half-metameres following the compound
metamere is joined to a half-metamere that is not its normal
half. As a consequence, one new additional half-metamere
will have to be formed at the end of the series on the right
side to make a perfect ending to the abdomen. In so highly
modified a form as the locust this view seems very improbable.
An alternative view would be to suppose that a half-metamere
has been intercalated between the 4th and 5th, or between
the 5th and 6th, and that the intercalated half has united
across the middle line to the 5th in common with the normal
half of the 5th. We might think of this intercalation as due
to a half of one metamere dividing into two, or we may
suppose that when the ventral mesoblast broke up
into a series of blocks it formed on one side one more
block than on the other (one more than the normal).
If the last view be true—and it seems more probable than any
of the other suggestions—we see at once the futility of trying
to explain the conditions on an assumption of predestined
right and left half-metameres. When we recall that the whole
ventral plate of the insect becomes segmented at the same

A STUDY OF METAMERISM. 439

time, and when we recall the complicated modifications of the
terminal segments, it seems highly probable that posterior to
the compound metamere no shifting of the segments of one
side can have taken place, and that no one segment has divided,
but that the plate as a whole has broken up into a greater
number of half-metameres on one side than on the other.

XI. Srupy or THE CoLouR-BANDS OF ECHINODERMS.

In the summer of 1891, while in Jamaica at the Marine
Laboratory of the Johns Hopkins University, I began a study
of the colour-bands on the arms of the Ophiurians in the hope
of getting results that might help towards a solution of the
problem of metamerism. The work was laid aside for two
years, during which time the colour was so far removed from
the alcoholic specimens as to render renewed study unprofit-
able. My earlier results, although not carried very far, point
to certain conclusions that are not without interest in the
present connection. In several species of Ophiurians (“ brittle-
stars’) the arms are banded at regular intervals by pigment
of a different colour from that of the rest of the arm. Each
band of pigment is confined to a single segment—metamere—
of the arm. In fig. 96 three such pigmented segments are
shown. The colour of the band in this case is a rich orange-
red, while the rest of the arm is an olive-green. Along the
mid-ventral line there runs a longitudinal narrow line of
orange-red pigment not shown in the figure.

We find that a definite number of uncoloured segments
alternate with a coloured segment. Three uncoloured seg-
ments lie between the coloured segments, i.e. every fourth
segment is coloured. Occasionally, however, the regularity of
the arrangement is disturbed.

Fig. 97 shows a common variation. Instead of finding a
fourth coloured segment between the upper and lower coloured
segments of the figure, we find two consecutive segments
coloured, each over half its extent. The middle line of the
arm marks the extent to which each is pigmented, the first
half-band lying to the right and the more distal to the left.

VOL. 37, PART 4.—NEW SER. GG

4.4.0 T. H. MORGAN.

We see that the right and left sides of the arm vary independ-
ently, so that each forms its half-coloured band. In the
present case something has disturbed the regularity of the
process, so that the colour has appeared too soon on the right
side, as the figure shows.

In fig. 98 we have asomewhat similar case. Here the half-
bands are separated by an entire uncoloured segment. In
other cases that I have seen the half-bands may be even
farther separated by uncoloured segments.

From a study of these and similar variations we find that
the following relations exist. When a segment is only half
coloured on one side it is generally, though not invariably,
followed, sooner or later, by a segment coloured only on the
opposite side. We might say that the colour-ring had split,
and its two halves had appeared on different segments, so that
when we find a half-band on the right we would expect to find
its other half on the left, and vice versa.

When half-bands appear, one of the halves seems always to
come in too soon proximally. The other half then appears on
its normal segment or beyond it.

My material has been too limited to warrant any attempt at
further explanation of the phenomenon, and no doubt a more
extensive study would show exceptions to these statements
that would render an explanation still more difficult.

The three following tabulations of the colour-bands show the
main variations that comein. The x indicates a coloured ring.
The figures between these indicate the number of uncoloured
segments. The half-coloured bands are indicated by x L. } and
x } R., depending whether the colour is on the left or right
side of the arm as looked at from below.

441

MUETAMERISM.

AVSiLUDYS On

First REcoRD.

on >.< Gel >.< oe &6 Cel ><

xX XO Ko KO Kod

Xoo XO K OD KX OD X

Gel >< Ge Se Ge) >< Ge) >

Se Gel 5 Gey OS Gelb 2S Gr) Xe

on

on

ine)

4 Ge) > Gel > GP)

Gel >4 ar) >.4 Ge) >.

Xoo Xo KN

Fa
XR daaiacs KX oO Koo X
xX 4
x

mel ox Gel 9<Ge) > Gel < ine) 3K

xcs Xo KX oO X co xX co

ee pa
HAM ANG XK IW@dam XK oo
ees Hx
x x

pS
aa XN xX oo XN XK
x

Broken.

oo Ko K OD K OD K OD KX OD KX OD KO XK OD

c2 X 0 K 20 X od X

xX KX oO Ko XK OO

Xoo KX oo X od K oO

co KX oD K oO K OD

Broken.

Broken.

H. MORGAN.

T.

4.4.2

Seconp ReEcorp.

pe a Pe

xX om KO a KX WAIN GR IN INR KH KS XK OD wi IN Hi GR OK OD XK OO

ere eae ak ix
45% - x *
pa pa

X om KX od K a ANG NIN Ge ier XK OD XK CO
i re eS
x x x

a od
Koo KX OD XK ay AG AN ied GV ig rn KX OD X OD XK CD
a Sil ee
x x x

443

STUDY OF METAMERISM.

A

Tarrp REcoRpD.

S ec
KO WM KH KOO K OD Keo KOO KH Kod Karrinanninee | KA Ka Xa Xa
4 ie ox
x x
= e 8
ee MR

XD Xe0Xo2 Xe9 X29 Kemet ccs ets emar X | XA XH XK KAKA KA Xo Xe
lee eel oe ee x
x x

X™ XOX KD KD KN | Ka XA KA xX

Xm Xoo XK Xo KD KR | Xa XA XA XG

pa pa A
esl acl XK OD KK ay iG &K Alo XO KAKA KS XA XA KA XA XR Xa XRXGKR
4 Xen See OC x
Xx Xx x x

444, T, H. MORGAN.

A few of the more obvious relations that one finds in these
tables may be pointed out. In some cases exceptions to the
statements made above are found. At ain the third record,
last (fifth) column, we see that a half-colour band is present on
one side, but not on the other. And again this is seen at a
in the third record, fourth column.

In the third record, first column, we find in one region many
irregularities present, although taken altogether the same
number of half-rings are present on the right and left sides.

In the third record each arm was found to have a new distal
end that had regenerated. In this part every other segment
was coloured. Whether this was a permanent or only a tem-
porary coloration I cannot say.

It would be interesting to find out whether in the same in-
dividual similar variations showed a tendency to appear on
the different arms. Even the limited data of the three tables
give a slight amount of evidence in favour of such a view.

The regular arrangement of the colour-bands on every fourth
segment finds an interesting parallel in certain Annelids, where
the coloured rings bear a more or less definite relation to the
metameres of the body.

Andrews (1) makes the following statement in regard to the
arrangement of coloured bands in the polychetous Annelid
Procerea tardigrada, which belongs to the family Syl-
lidee:—‘‘ The female has a dark dorsal transverse band upon
Somites 3, 6, 8, 9, 18, 17, 21, 25, 27, 29, 32, 35, 38, 42, 46,
49, 51, 58, 56, 57, 70, 71, 74, 77. . . . The non-sexual form
has pigmented bands like those of the female, but arranged
according to a definite law or general rule, to which the bands
in the female conform also; bearing in mind that the female is
formed as a cut-off part of the non-sexual stage, separating
almost always just posterior to the thirteenth somite, and
hence having thirteen less somites than that stage. In 110
individuals carefully studied, only three had the bud formed
just posterior to the fourteenth somite; seventy-nine had an
evident bud just posterior to thirteenth somite.

“ Having tabulated the arrangement of the coloured bands in

A STUDY OF METAMERISM. 44.5

these 110 individuals, there results the general rule that the
bands occur upon the third and fourth somites, then upon
every other or alternate one up to and including the twelfth,
then (in the region of the bud) upon every fourth one up to
and including the twenty-fifth, then upon every fifth one up
to and including the forty-first, after which the exceptions
become so numerous that no rule is evident. The examina-
tion of so many cases shows a definite tendency to limitation
in the bands to certain somites in the anterior region, and a
greater and greater irregularity in the posterior region.”
After illustrating some cases of failure in the normal
arrangement of the bands, Andrews adds, “ These facts seem
sufficient to indicate that we have in this Syllid a marked
tendency to the acquirement of a regular metameric marking,
which, however, does not coincide with the metamerisation of
the somites, but tends to follow a special law best expressed
in the oldest part of the body in which certain alternating
coloured and not coloured somites are distinguishable—a series
of groups or combinations of somites thus following one
another.”

It is not without importance to find in the typically meta-
meric Annelids regular serial markings following definite laws
in each portion of the body. Groups of metameres seem
here to act as a unit. This case is certainly paralleled by the
colour-bands on the arms of the brittle-stars. The serial
repetition of the appendages of the Crustacea furnish examples
of somewhat similar regional variation. It is not improbable
that if we find an explanation for one set of phenomena we
will be able to explain them all.

XII. Regeneration 1n Eartoworms.

In the winter of 1887-8 I made a small number of experi-
ments to determine the extent of regeneration in the earth-
worm. Again, in the spring of 1892, another series of
experiments were started, but an accident spoiled the results.
In the past winter of 1893-41 made a more elaborate and sys-
tematic attempt to work out the same problem. I am much

4.4.6 T, H. MORGAN.

indebted to Miss Elizabeth Nichols, Fellow in Biology, Bryn
Mawr College, who began this study of regeneration with me.
Many of the early experiments were largely carried out by her ;
and later, when the work devolved on me, I profited much by
the results of the previous work.

There were several main problems that I wished to work
out. First, the extent to which the earthworm could regene-
rate; secondly, the number of new segments that would reap-
pear in the anterior end after the removal of a definite number ;
thirdly, the presence or absence of abnormalities in the re-
generated anterior segments.

Certain rough results were at first obtained, which showed
that when many segments were cut off only a few segments
replaced them. That is to say, there was no apparent connec-
tion between the number of segments that were cut off and the
number that regenerated. Four and rarely five new segments
came back.

I then set to work to determine what result would follow
when only a few segments were cut off, for obviously if four
came back when one, two, or three were cut off, the result
would appear as though the reproductive organs had all shifted
posteriorly. The tables below that are first given show the
results of these latter experiments.

One of the most conspicuous results was the great decrease
in size that the worms suffer during the period of regeneration.
When only a few segments were cut off the regeneration was
soon accomplished, and no great decrease in the size of the
body of the worm was obvious; but where many segments
were cut off, and regeneration only took place after several
months, or not at all, the body dwindled until it got to be less
than a half of its original size. I have not made a histological
study of these worms to determine what organs have suffered
most during the period.

The worms used were L.(orAllolobophora) fetidus, which
live in manure heaps. They were kept in ordinary flower-pots
filled with the manure in which the worms were found living.
The pots stood in about an inch of water, and each was covered

A STUDY OF METAMERISM. 44,7

with a glass plate. The pots stood in a large glass case in one
of the rooms of the laboratory. The temperature of the room
varied from about 70° F. during the daytime to about 50° F.
at night. Under the same conditions the normal worms kept
perfectly healthy.

TaBLe I.

Two segments cut off, 3/16, 94.
Killed, 4/28, ’94.

2 segments regenerated : ' . vas def. 15
2 + cp : : : eee
2 22 2? ° ° ° 5) 15
2 a 5 : : : a ko

Two segments cut off, 3/4, 94.
Killed, 4/28, ’94.
2. segments regenerated ‘ : . vas def. 15_
Two segments cut off, 3/16, ’94.
Killed, 5/5, 794.
2 segments regenerated. Sem. recept. 9, 10, 11 (normal)

Tase II.

Three segments cut off, 3/4, 94.
Killed, 4/28, 794.

3 segments regenerated : ° . vas def. 15
3 3 33 (and piece of 4)

3 os » : : “ ay LB
3 x 3 : ‘ ‘ fee lb
eee a : : ae
ar » (and small piece of 4)

2 5 - (with indications of athird) ,, 14
2

23 29 33 be)

Three segments cut off, 3/16, 794.
Killed, 4/28, 794.

3 segments regenerated : ; . vas def. 15
es om 3s A ‘ : : aa Lo
2 a .s (sem. recept. 8, 9, 10)

3 36 ‘5 (and little piece of 4) a 25

Three segments cut off, 3/16, ’94.
Killed, 5/5, 94.
2 segments regenerated (sem. recept. 8, 9, 10)
2 a % Os Pr ) vas def. 14.

448 T. H. MORGAN.

Taste ILI.

Four segments cut off, 3/16, 94.
Killed, 5/5, °94.

4 segments regenerated : ; . vas def. 15
3 ” 2 . . . ” 14
3 4 5 5 . : ceeds
3 ” ” Ge . . 3 14
Be 5 %9 : ; ; eg
3° ” ” . : : $5 14
3 ” FP) a 9 TA
3 ” ” . . . ” 14
2 ” ” (+ 1/3 of 3) ZA 5 i137
3 ” ” (sem. recept. 8, 9, 10)

ye ee = : A 3 sy ale

Four segments cut off, 3/4, 94.
Killed, 4/28, 94.

3 segments regenerated ‘ . vas def. 14
4, “ = (+ 4 of 5th) (8rd

segment imperfect) so be
4, mt a (a small piece of 4

had been left) . ae ls

TasLe IV.

Five segments cut off, 3/16, 794.
Killed, 5/5, 794.
4 segments regenerated 2 : . vas def. 14
3 ae os (+ 4 of 4th) ‘ 3 ke

Five segments cut off, 3/16, 94.
Killed, 4/28, 794.
4 segments regenerated.
3 33 ”
4 33 ”

Five segments cut off, 3/4, 94.
Killed, 4/28, ’94.
4 segments regenerated (+ 4 of 5th) . vas def. 15
eae a5 (+ 4 of 3rd) : a as

A STUDY OF METAMERISM. 449

TABLE V.

Attempts made to cut off 1 and 2 segments, 3/4 and 3/16, ’94.
Killed, 5/2, 794.

1 segment regenerated - ; . vas def. 14
1 ”» 2” C . ey) 15
1 Pe - (indications of a divi-

sion into 2) 2 ee
2 segments ms ope eel iy

Conclusions from Tables I—V.—In all cases recorded
(6 cases) where two segments were cut off two regenerated.

When three segments were cut off (14 cases) three grew
back in nine worms and two grew back in five worms.

In those worms (14) where four segments were cut off, four
segments regenerated in five worms, and three segments grew
back in eight worms, and two segments and a third! in one
worm. ;

When five segments were cut off in no case did five grow
back. In some of these worms a little more than five must
have been cut off, so that a small piece of the sixth was taken
off (see Nos. 2 and 6 and 7 of Table IV) and regenerated. In
four worms four segments grew back, in two worms three
segments grew back, and in one worm only two segments grew
back.

Taking these four tables together, we see that up to four
segments lost, and including this, there is a tendency for the
worm to reproduce the number lost. This was actually done
in all the cases where two segments were lost, in nine cases
out of fourteen where three were lost, and in five cases out of
fourteen where four were lost. More than four segments the
worm does not seem to be able to regenerate, as a rule.

The data given in Table V are not as accurate as in the pre-
ceding tables, because there was some confusion in the numbers
of the pots containing these worms. So far as the figures go,
they show that in three cases two segments must have been
cut off, that one individual then regenerated two, and two

1 The amputation was oblique, and cut off a part of the fifth segment.

450 T. H. MORGAN.

individuals regenerated only one segment each. In one case
one segment must have been cut off and one regenerated.

The results of these tables have a direct bearing on the con-
ditions found in adult worms, and recorded in a previous
section. Those worms in which both openings of the vasa
deferentia were found on a segment anterior to the fifteenth
may have been, in most cases, the result of the loss of ante-
rior segments, and a subsequent incomplete regeneration. We
cannot affirm that all cases were the result of the process,
because, in the light of other facts, it is not improbable that
such variations may have come from the embryo.

These records also show us that regeneration of the anterior
end will not account for any of those cases where the vasa
deferentia open on a metamere posterior to the fifteenth, for
in no cases were more segments generated than amputated.

What the result would have been in the cases recorded below
where a worm regenerated many segments, after amputation far
posterior to the fifteenth segment, I cannot tell. The worms
were not kept for a long enough time to determine whether
or not reproductive organs would ever have appeared.

In the preceding and following tables it will be noticed that
in many cases the position of the openings of the vasa deferentia
is recorded, and when these were not found the segments con-
taining the seminal receptacles (9—10—11 normally) are re-
corded. These landmarks were located after regeneration had
taken place. This served as a check for those cases where the
number of segments amputated had been previously recorded,
and in the other cases gave fairly accurate evidence as to the
number of segments that had been cut off.

In the next two tables,—the results of the first series of ex-
periments,—a large number of recorded segments were cut off
to find the limit of the power to regenerate anterior segments.

A STUDY OF METAMERISM. 451

TasLE VI.
Segments cut off, Oct. 12, 93.
No. cut off. No.of worms. Noy. 14— Jan. 1—
(1) 4 segments regenerating.
(1) 4 32 » S
10 (4) (1) 4 Bb + ame as Nov. 14
(1) Imperfect.
Jan. 80— May —
(1) aa (1) 6 segs. regen.
(1) ni38 ai
12 (5) ¥ GQ) see (1) 5 segs. regen.
(1) ves (1) 335,»
(1)? 5 segments regenerating.
Jan. 30— Apr. 14— May 18—
14 (5) (1) 5 seg. regenerating. bea aie
(1) Imperfect. (1)
(1)
(1) Dead.
(1)
Jan. 30— Apr. 14— May 18—
16 (4) (2) Alive. (1) Alive. (0) Alive.
Jan. 30— April 14—
19 (2) (2) Both regenerating. (1) Had regenerated 4 or
5 segments,
(1)?
Jan, 30—
19—24. (5) (1) Regenerating imperfectly.
Jan. 30— (0)
24. (5) (3) Not regenerating.
Dec. 2. Jan. 3— Jan. 30— Apr. 13—
26 (5) (3) (2) (2) (0)
27 (3) (2) (2) (1) (0)

452 T. H. MORGAN.

TaBLE VII.
Segments cut off, Jan. 12, 94.
No. cut off. No.of worms. Apr. 14— May 18—

19 (1) (0)

21 (1) (0)

22 (3) (2). (1) Regenerated (2) Very imperfectly

imperfectly. regenerated.

23 (1) (1) » (0)

24 (1) (0)

25 (1) (1) (1)

26 (1) (1) (0)

27 (1) (0)

The results from Tables VI and VII show that one cannot
say definitely, ‘here the power of regeneration ends.” The
figures show that some worms regenerate where others fail to
do so. It may be that the possibilities are different for different
worms, or that at the time of operation certain worms were in
better condition than others, or the external conditions (bac-
teria, &c.) may have been different in different cases.

The tables show that posterior to the twelfth segment the
power of regeneration rapidly decreases. Worms that have
lost more segments than this number may live for some time,
and heal up the wound, or even regenerate imperfectly. But
sooner or later the majority of these die. Occasionally re-
markable exceptions are found. In Table VI the fifth record
shows that a worm that had lost nineteen segments regenerated
four or five new ones, and in the next tables more remarkable
cases still will be recorded.

It is a tedious operation cutting off a definite number of
segments from a living worm. In the three following tables
the number of segments cut off was not counted at the time,
but could be calculated with approximate certainty after re-
generation by the position of the vasa deferentia or segments
containing the seminal receptacles, or, when the amputation
was behind these, something like an approximation could be
obtained by utilising the anterior end of the clitellum, or even
the posterior end of the body.

A STUDY OF METAMERISM. 458

Taste VIII.
Anterior segments cut off, 1/22, 94.
Killed, 4/18, 794.

No. of segments regenerated. Vasa def. Calculated No. cut off.
3 13 bee
3 (4) 13 B (3)
3 (4) 14 4 (4)
3 12 6
3 12 6
3 (irregular) 14 4
3 14 4
4 12 te
4 12 r
4 (irregularly united to body) 138 6
4 12 7
4 (3) 13 6 (3)
4 (2 + 3) 14 eee
5 15 5
Anterior segments cut off, 1/22, 94.
Killed, 5/5, ’94.
No. of segments regenerated. Vasa def. | Calculated No. cut off.
3 13 5 -
3 +4+4+9) 15 3 (+)
3 (with compound 1—4) 14(orl4—15) 4
3 (3 + ¢ irregular) 13 5 (+) -
3 (irregular) 12 6 v
4 12 re
4 12 a
4 14 55
4 12 ty
4 (2) 14 5 (4)
4 (2) 12 7 (3)
5 (4) 14 6 (2) ~

454, T. H. MORGAN.

Tanne IX.
Anterior segments cut off, 3/4, 794.
Killed, 4/28, 794.

No. of segments regenerated. Vasa def. Calculated No. cut off.
i 14 2
1 14 ae
1,(4) (vas def. double +4 — 14 lor 2 (?)
3 (3) 15 3 (2)
3
3 (3)
4
4
4
4 vas def. 13 6
4 14 5

The Tables VIII and IX add a large number of data to
those of the preceding tables. It will be noticed that there is
one definite case of five new segments coming in for five cut
off. There are also a number of irregular methods of union of
new and old parts, and several cases of imperfectly formed new
rings. The next table covers much the same ground as the
preceding, but is interesting because the worms had been kept
for a much longer time than those in the preceding table, and
because the last record shows a case where from thirty to forty
segments were in all probability cut off, and yet three and a half
segments had regenerated.

A STUDY OF METAMERISM. 455

TaBLEe X,

Anterior segments cut off, 10/18, 793.
Killed, 2/11, 794.

No. of segments regenerated. Sem. recept. Vas. def. Calculated
(Normal, 9 —10—11) No. cut off.
2 (2 + ¢ irregular) 5—6—7Z 11 6 (+)
3 7—8—9 13 5
3 4—5—6 -— 8
3 (+) (worm very small) -- 10 8 (2)
4, 5—6—7 — §
4 S$—9I—10 — 5
eI — 12 7
5. (irregularly joined) —6—7 (?) os 9 (?)
5 ~~: 11 9

5 (small piece)
5 (very small piece)

4. (F) = 13 6 G)
6 (+ 3) —7—8 = ‘10 (3)

3 (4) small worm, with only 65 old segments and old anus. 30—40
In addition two other worms, where union of new head and body was
too irregular to count even approximately the segments.

TaBLeE XI.

Anterior segments cut off, 11/15, 93.
Killed, 4/14, 794.

No. of segments regenerated. Sem. recept. Calculated No. cut off.
3 9—10—11 3,
5 6—7--8 8
4 — §
Not regenerated 10 in front of clitellum 15
3 ere 34, 19
a (mouth present) 9 a ne 16
Scarcely regenerated (small piece)
3 or 4 very irregular 9 m 3 16
3 imperfect segments 8 re ss Ly
4 (1st partially divided) vi = _ 8
4 A * above) 15 AS - 10
5% (1st imperfect) * 15 _ ss 10
15 or more regenerating 63 segments in front of anus 385—40

This table is interesting because it shows a large number of
WoL. 37, PART 4.—NEW SER, HH

456 T. H. MORGAN.

worms regenerating imperfectly a few segments in front of
the clitellum. Moreover the last case is particularly instruc-
tive, because here again 30—40 segments were probably cut
off. The number of worms that were operated on at the be-
ginning of this experiment was unfortunately not recorded ;
hence we do not know what the per cent. of mortality has
been.

TaBLE XII.

Anterior segments cut off, 11/18, ’93.
Killed, 1/21, 94.
No. of segments regenerated. Sem. recept. Vas. def. Calc. No. cut off.

2 —3—4 8 9

3 — 10 8

3 (united irregularly) 8—9—10 14 4

3(+ $44) Sie - oe)

4 (+ 4) 6—7 —8 12 7 (+)

4 (+ 4+ 4d irregularly joined) — 11 8 (+)

2 3—4—5 9 8

3 —4—5 — 9

3 (2) 45—6—7 = ‘Bice

a 7—8—9. 13 6°

4 (3) —6—7 — 8 (3)

3 13 in front of clitellum 12

3 (3) ll » ” 14

3 12 %3 23 13

3 (or four imperfect) just cE ” 23

a 17 23 2» 8

4 14 33 33 ll

5 16 33 > 9

5 14 ” ” ll

2 (or three very imperfect) 4 5 + 21

2 (or three very imperfect) 14 y » il
Very imperfectly healed 7 ” ” 18
Imperfectly healed 12 es » 13
Healed—evidence of few rings 9 s a 16
Beginning to regenerate 10 99 » _ 1

oy) 0 “15 ” ae ~ 15
Healed, not regenerated 15 ie a 10
Healed—regenerating (?) 5 ” » 20

8 je more segments. Rings

faint. Small piece } 56 old segments in seals of anus.

—

A STUDY OF METAMERISM. 457

In several instances in the preceding tables records are
found of worms that, in addition to the formation of new
segments, completed parts of segments accidentally cut off in
the operation. The following records are of worms in which
purposely very oblique amputation had been made.

Taste XIII.

Anterior segments obliquely amputated, 3/4, 794.
Killed, 4/28, 794.

§. segments completed (9—10—11)sem.recept. vas.def.—
9 aS », (dorsal) — esi Be
ll ” ” (lateral) = J 15
2 % 9—10—11) ,, Ee
4 = », (lat. vent.) +] newseg. (9—10—1]) 3 —
4 ” ” » +3newsegments — eA
6 me » (lateral) +3 53 —_ —
7Q@ 5 », (lateral) = i ae
7 s », (lateral) +2 ‘5, (8—9—10) 8 at

The last table shows very clearly that the power to com-
plete segments that have in part been removed is much
greater in the earthworm than the power to regenerate whole
segments. In the table we find as many as twelve segments
completed, and in a comparatively short time. Such
a number of segments is never or rarely regenerated when the
anterior segments are cut squarely off. /

The latter records of this same table show that when ante-
rior segments are entirely cut off, in addition to those cut
obliquely, that we have both a regeneration of those lost
(within a limit) and a rebuilding of those injured. Moreover,
although the results are too few to speak with entire confidence,
it would seem that the presence of segments completing them-
selves does not interfere with the formation of the full comple-
ment of new regenerated segments.

These facts, it seems to me, throw an interesting light on
the problem of regeneration. They are too few to warrant at
present any speculation. It is my intention to make a fuller
and more accurate study of these phenomena.

The obliquity of the cut ‘was lateral in most cases, in a few

458 T. H. MORGAN.

the slice was taken off the dorsal surface, and in others off the
ventral.

In a few cases no record was made. Whether or not the
injured rings complete themselves more readily at the side than
dorsally or ventrally cannot be determined from the table, for
the position of the parts cut off in the first instance was not
recorded.

In connection with the preceding experiments another was
carried on. ‘The anterior ends that were cut off were in many
cases kept alive to see what power of regeneration remained in
them. It became evident very soon that the length of the life
of the pieces was in a general way proportionate to their linear
length. Those with a few segments died in the course of a
week ; those with more lived longer, &c. No cases of survival
of as few segments as fifteen were ever found. The two fol-
lowing tables show to what extent longer pieces remained
alive, but with one exception (twenty-four anterior segments)
they all died after a time. It is surprising to find this to be
the case, for such pieces contain the mouth (so that the piece
may feed) and all of the important (?) organs of the body.

TaBLe XIV.
Record of anterior segments.
Date—Oct. 12. Nov. 14. Dec. 2. Jan. 4.
‘No. of anterior No. of
segments. worms.
12 (5) (1) (0)
a (5) (4) (2) (0)
16 (4) (0)
19 (2) (2) (1) (0)
19—24 (5) ~- (2) (0)
24 (5) (5) (4) (3) May 5 (0)
26 (5) (5) (3) (1) Apr. 26(1)
Jan. 12. Apr. 14. May 5. May 18.
19 (1) (0)
22 (3) (3) (3) (?) (0)
23 (1) (1) (0)
24 (1) Regenerated a half-inch.
26 (1) 0)

(
27 (1) (1) (0)

A STUDY OF METAMERISM. 459

The worms that had the anterior segments amputated, the
history of which is recorded in Tables I—XIII, were examined
from time to time during the period of regeneration. One
striking result was often apparent. When the amputation of
the anterior segments had taken place obliquely, the new seg-
ments grew out approximately at right angles to the cut
surface. It was not uncommon to find a new regenerating
anterior end with its axis inclined at as much as (or even more
than) 45° to the long axis of the body. I have no records as
to whether the same thing happens when the oblique surface
is above or below. Those referred to were from lateral oblique
sections.

This result is comparable to that of Barfurth on the regene-
ration of the tadpole’s tail. Here the new part appeared at
right angles to the cut surface, and subsequently swung round
into line.

To get the same result from the tail of the tadpole and the
head of an earthworm suggests that there is some fundamental
law of growth underlying both phenomena that should be
more extensively investigated.

It was my intention to examine the power of regeneration
in young worms for comparison with the adult, but only a few
experiments have been made, and there has not been sufficient
time to allow the completion of this side of the work.

Miss Adelene M. Fielde (14) has published a few fragmen-
tary notes on the power of regeneration in L. terrestris.
Pieces containing twenty to thirty segments from the posterior
end of the worm lived forty days, but did not regenerate at
either end. In these pieces new half-segments had been in-
serted, the authoress affirms, hecause she could not find any
such modifications (compound metameres) in other worms!
In nine worms five anterior segments were amputated. These
“wholly regenerated.” In ten worms five anterior and twenty
to thirty posterior. These were found regenerating.

4.60 T. H. MORGAN.

XIII. Generat Concivusions.

The solution of the problem of metamerism has often been
attempted by morphologists with varying success. It has
become more and more evident that the problem is a difficult
one, and I think the results have shown that the final solution
can only come with a better knowledge of the fundamental
relations of the parts of the body to one another, with an
acknowledgment of the imperfection of the phylogenetic
method, and with a better insight into ontogenetic laws.

Darwin said over thirty years ago, in the ‘ Origin of Species,’
“We need not here consider how the bodies of some animals
first became divided into a series of segments, or how they
became divided into right and left sides with corresponding
organs, for such questions are almost beyond investigation.”
The attempt of morphologists to solve even the simpler of
these phenomena, viz. metameric repetition, shows how true
Darwin’s words remain even to-day.

It may, therefore, seem doubtful whether anything will be
gained by a new analysis of the problem of metamerism, or by
a critical consideration of the numerous theories already
advanced. It would certainly be unwise to add any new
speculation to that already afloat. In the following pages no
new theory is offered, and I have attempted no more than a
consideration of those methods which have proved sterile or
erroneous as contrasted with the methods that have been
fruitful and suggestive.

During the last fifteen years the methods of the phylo-
genists have been applied to the solution of metameric repe-
tition. Comparative anatomy has been the point of departure ;
but speculation has leaped far beyond its legitimate boundaries.
The results have shown that the method of phylogenetic inter-
pretation is subjective rather than objective, and the conclu-
sions have given, therefore, at most a probable course of
evolution, and often only a conceivable process of transition.

It is only fair to say that in many cases the speculation has
been advanced tentatively, as suggestion rather than conclu-

A STUDY OF METAMERISM. 461

sion, and the admirable work on which the speculation has
been based has been in no degree vitiated by the attempts of
the authors to push their conclusions beyond the limits
deducible from their immediate results.

The Ceelenterates have formed the basis of Sedgwick’s theory
(34) of the origin of metamerism. The radial actinian has
been worked over into a bilateral metameric form, and all the
details of structure of the higher forms (celom, nephridia,
gill-slits, trachea, &c.) have been evolved from the hypothetical
actinian ancestor. Wilson (37) has supported the same view,
“but only as a suggestion for further investigation of the
facts.” The Turbellaria have been the starting-point of Lang
(25) and Meyer (29); but each author has described an
entirely different course of transition from the flat-worm to
the Annelid.

The Nemertian claims have been pressed forward by
Hubrecht (22) and Balfour (2), and hinted at by others.

The Echinoderms have proved refractory, yet Wagner! has
made out a possible phylogeny from these to metameric forms,
and Haeckel (15) reversing the process, made a star-fish out
of five fused Annelids.

The Enteropneusta, declared unsegmented by Bateson (3)
and segmented (32) by the present writer, have been believed,
nevertheless, by both authors to throw light upon the problem
of metamerism.

With this divergence of opinion it is not surprising to find
as great a divergence of method. In one case a radial form
has given the starting-point, in the others a bilateral form.
But the ways in which bilateral forms have been transformed
into the metameric form have been very different. The budding
theory has played perhaps the most conspicuous part. Nearly
all of the older writers looked upon segmented forms as
animal colonies—Quatrefages, Cuvier, Owen, Duges, Geoffroy-
St. Hilaire, Lacaze-Duthier, Herbert Spencer, and Perrier.
The same idea is often found in later speculation as well. In

1 Quoted on the authority of Meyer (29). I have not been able to find
the original,

4.62 T. H. MORGAN.

the budding theory a short bilateral form is supposed to have
elongated by a repetition of itself to form a long chain of
united individuals. Other authors—Hubrecht, Lang, Meyer
—have supposed metameric repetition to have appeared in a
long animal which split up secondarily. Hubrecht (22, 23)
has supposed that a long animal, such as a Nemertian, is con-
tinually in danger of injury from without, and the animal has
met this danger by acquiring a remarkable power of regenera-
tion! hose animals that had the main organs of the body
repeated were better able to regenerate any pieces that were
broken off; for such pieces would contain, in all probability,
all of the essential organs of the body.

Lang’s (23) description of the origin of metameric repetition
from the flat-worm, Gunda segmen tata, is too well known
to need rehearsal.

In this elongated form the repetition of the digestive diver-
ticula have been the centres around which the metameres have
been built up.

Meyer (29) has contended that the repetition of the meta-
meres is an expression of the alternate bendings of the sides of
the body of a free-swimming elongated worm. A pair of elon-
gated gonad-pouches have been broken up into a series of meta-
meric compartments, owing to the swimming movements of the
body, and around these as centres the metameres have evolved.

The attempts to find the solution of metamerism within the
metameric groups have been equally unsuccessful. The
simpler Annelids, such as Polygordius and Protodrilus, have
been interpreted as.archaic forms by Hatschek. This explana-
tion has been rejected by Kleinenberg, Eisig, and Meyer. The
lower Vertebrates have been equally difficult to interpret.
Lankester (28) found that the tail of Appendicularia showed
muscle-plates with corresponding ganglia in the nerve-cord.
The Ascidian larva has no similar structures. Brooks (8) has
argued that Appendicularia is a very old archaic form; while
Willey (89) has interpreted the tadpole larva of the Ascidians
as a secondary larval form derived from a fixed ancestor. By
inference, therefore, Appendicularia is a sexually mature larva,

A STUDY OF METAMERISM. 4.63

and its segmented tail either an antefact or a secondary
acquirement.

Bateson (8) has attempted to show that Balanoglossus is
related to the Chordata, and is unsegmented. I (82) have
defended the same supposed relationship, and described
Balanoglossus as segmented. Spengel (86) believes Balano-
glossus not to be related at all to the Chordata.

Amphioxus is a typically segmented form, but has given no
clue as to the origin of its metamerism.

By many morphologists (Semper, Dohrn, Van Wijhe, Minot,
&e.) the group Chordata is supposed to be derived directly
from segmented Annelids. But the Annelid-vertebral con-
nection has as many opponents as defenders, and is one of
the most illustrious results of the phylogenetic method.

It must be admitted, of course, that morphologists are
dealing with complex and difficult problems in attempting to
unravel the past connections between the larger phyla of the
animal kingdom, and that their speculations have been in
many cases ingenious working hypotheses; but the results
show, I think, that the method is easily carried too far, and
that, after many trials, it has not led us to any definite con-
clusion.

Ontogeny has been also fruitful in speculation. The cry
has been that Ontogeny tended to repeat Phylogeny, and
larval forms without end have been set up as archaic
remains.

The rise, culmination, and decline (?) of the Gastreea theory
illustrates in a concrete case the history of the ontogenetic
method ; and the Nauplius theory teaches a lesson of caution
that ought not to be forgotten. The ccelom theory, built up
on the splendid results of Agassiz, Metschnikoff, Kowalewski,
Lankester, Hatschek, and Hertwig, while a most suggestive
working hypothesis, has led to no settled conclusion; for the
well-ascertained fact that in many groups (Sagitta, Amphioxus,
Brachiopods, Echinoderms, &c.) gut-pouches give rise to the
celom has not led us to any decision as to whether we
have here an ontogenetic performance or a phylogenetic repe-

4.64, T. H. MORGAN.

tition. In the Annelids, for instance, where there is a typical
ceelom developed, there are no traces of gut-pouches, and
this has to be interpreted as a secondary loss. Hatschek has
suggestively remarked (15), ‘‘ The two mesodermal teloblasts
may correspond to the celom-sacs from which they were
derived by a reduction in the number of their cells. In fact,
we find this method of formation only in those cases where the
number of cells of the embryo is very small.”

Whatever be the conclusion reached as to the formation of
the ccelom, we shall still be far from a decision as to how and
when the ccelom repeated itself in the metameric forms.

The trochosphere of the Annelids, Mollusca, and Mollus-
coidea (?) has been utilised by Hatschek to support the whole
family tree of the higher Metazoa (the Vertebrates perhaps
excepted).

Kleinenberg has interpreted the trochosphere as a recapitula-
tion of ahydro-medusa. E. B. Wilson and others have rejected
the trochosphere ancestry, and believe the larva to be cceno-
genic. Whitman wrote in 1887, “In spite of volumes
devoted to the discussion of the subject, the larva of Poly-
gordius still remains a morphological puzzle.” -

The trochosphere theory got a strong support from Semper’s
discovery of Trochosphera equatorialis.’ Even Kor-
schelt and Heider, who, as a rule, are most circumspect in
accepting embryological speculation, wrote in 1890, ‘‘ Hochst
wahrscheinlich liegt in der Trochophora der Anneliden die
ontogenetische Recapitulation einer Stammform vor, welche
den Anneliden Mollusken und Molluscoiden gemeinsam war
und von der aus sich diese Thierstiimme als selbststiindige
Gruppen abzweigten.””

Recognising the relationship existing between the trocho-
sphere larvee of Mollusca and Annelida, embryologists have not
been satisfied to postulate only the archaic nature of the larva,

‘ Semper’s Rotifer was known long before Hatschek’s theory, and sug-
gested to me the name “trochosphere” for the larval form (‘ Devel. of

Lynneus,’ 1875), which some years latcr Hatschek adopted from me with
the change of the word to “ trochophore.’—KH. Ray LanKusTER,

A STUDY OF METAMERISM. 465

but have gone further, and postulated it as the ancestral adult.
They have been led to believe that such a minute few-celled
larva has evolved into the complicated segmented Annelid, and
into the equally complicated but unsegmented Mollusc. The
resemblances between the adult Mollusc and Annelid they
have been content to call “ parallel developments.”

For all the evidence we have at present we might satisfy the
facts just as well by assuming that a large many-celled unseg-
mented form stood as the bottom of these two groups, having
a trochosphere as its larval form. The Rotifers and related
forms would then be interpreted as arrested forms. The
one conclusion would be, I think, as justifiable and as
easily maintained as the other, and both impossible to
demonstrate.

In the face of so much conflicting embryological indecision,
it seems to me we have in reality arrived with certainty no
nearer to the solution of metamerism.

There have been only a few attempts to explain the meta-
meric repetition as the result of mechanical action. Hu-
brecht’s (22) explanation for the Nemertian is scarcely a me-
chanical explanation, since it presupposes a repetition of parts
and an ability to regenerate in the worm itself. Kennel’s (24)
ingenious attempt to interpret division of animals as the result
of external stimuli is scarcely a mechanical explanation.

His (19) has attempted to explain the metamerism of the
Vertebrate as an embryological phenomenon—as the result
of series of breaks occurring in the two lateral mesodermic
sheets of the embryo. Meyer’s (29) view, referred to above, is
distinctly a mechanical hypothesis. The repetition that he
assumes to have come into a long Nemerto-Turbellarian an-
cestor was the result of the movements of the body in swim-
ming. Many objections are easily raised against Meyer’s
romantic speculation, As Hatschek has pointed out, those
Annelids that are adapted for swimming are heteronomously
segmented, while Meyer’s hypothesis seems to demand first a
homonomously segmented form as the result of its own activity.
In the second place Meyer’s sketch starts off on Lamarckian

4.66 T. H. MORGAN.

principles. Although many naturalists still admit the prin-
ciple of use and disuse as a factor of organic evolution, perhaps
an equally large number reject the explanation. Hence, until
we get definite proof of the truth or falsity of any such theory,
it ought not to be used as the starting-point on which to build
up other or new theories.

Caldwell (11) has offered a brief and interesting attempt to
explain metameric repetition. So brief, indeed, is the presenta-
tion, and so obscurely worded, that I am not certain that I
have entirely understood the meaning of the author. It isa
mechanical theory par excellence. The theory assumes
that as early as the blastula! stage the endoderm, as well as
the mesoderm, is represented in cells or groups of cells at the
surface of the sphere. Now the endoderm may before inva-
gination get separated into two parts, owing to an early
elongation of the blastula, so that when gastrulation sets in it
may take place at two regions of the surface. One of the
regions may contain much more endoderm than another,
giving an oral or an anal gastrulation asa result. Similarly
the mesodermal “ Anlage” may be pulled apart and turn in with
the endoderm at one region or the other, or may even have
been left along the line where the two endodermal masses
separated.

In many cases the whole of the mesoderm may be turned
into the gastrula cavity with the endoderm, and subsequently
set itself free from the endoderm by one, two, or many gut-
pouches. When many gut-pouches arise, they mark the be-
ginning of metameric repetition. The reason for many pouches
appearing in some forms is to be explained as due to an early
elongation of the invaginated endoderm forming the archen-
teron, so that the mesodermal “ Anlagen” get pulled out and
broken apart.

If, as Caldwell supposed, there is pre-formation for the
mesoderm, there must be for all the other organs of the body,
and this the author admitted.

1 The author says planula, which makes his explanation obscure ; unless
he means technically a blastula.

A STUDY OF METAMERISM. 467

It is difficult to see how by simple elongation of a larva to
meet a supposed larval advantage, in the first place the
organs in the ectoderm should get separated into exactly the
same number as the number of gut-pouches, and in the second
place that all of these should correspond so as to give organs
repeated symmetrically in both ectoderm and mesoderm. I
cannot, I admit, resist the conviction that metameric repeti-
tion is far too difficult and fundamental a problem to be ex-
plained as the result of mechanically pulling out the supposed
Anlagen of all the organs of the body.

The phenomena of metameric repetition and apical growth
are closely associated together. Whether the latter stands in
any causal relation to the first cannot be definitely asserted, or
if it could we should have no means of determining whether
the relation is an ontogenetic or a phylogenetic one; whether
the apical elongation was established in the embryo or in the
adult (if, indeed, we can draw any line of any value between
the embryonic and adult appearance of any organs).

Nevertheless it is important to emphasise the fact of the
connection, for we find both in Vertebrates and Annelids the
two phenomena closely bound up together. Further, we have
the interesting case of the star-fishes and brittle-stars, where
at five radial points there is apical growth, and the five arms
are segmented. In the higher plants also there is a repetition
of similar parts (phytomeres) and apical growth.

Whether the repetition of the calcareous skeleton and tube-
feet of the arm of a star-fish is a repetition comparable to the
repetition in the Vertebrate and Annelid will depend largely
upon definition of terms. As to the fact of a symmetrical
repetition and the presence of apical growth there can be no
question, and that is the main point. There are no grounds
for assuming that the repetition of the parts of the star-fish
arm was ever connected with any attempt of the animal, in the
past, to reproduce itself at five equidistant points, nor would
such a suggestion be believed by anybody for a moment. The
method of growth in these arms, where there is a terminal
piece carrying the eyes and a subterminal growing region,

468 T. H. MORGAN.

is so similar to the method of elongation of the Annelid body
that even the most casual observer must be impressed by the
comparison. The greatest drawback to any attempt to refer
the two cases to a common method of growth (not phylo-
genetic, of course) would probably be met hy the statement
that the repetition of a metamere is something entirely
different from the repetition of the vertebral ridge in the star-
fish’s arm. If we can succeed in breaking down this conven-
tional and artificial definition of the value of metameric
repetition as compared with other repetitions of the body we
shall have made a step forward Iam confident. This question
will be more fully dealt with in the next section.

I have asked the opinion of eminent botanists on several
occasions as to the meaning of the repetition of the parts of
the higher plants, particularly the Phanerogams. So far as I
can learn, the botanical phylogenists have not had a much
better time of it than the zoological phylogenists. The former
have had the immense advantage of much paleontological ma-
terial, but, as I understand, even with this the great gaps come
just where the evidence is most wanted. I have not, however,
found any botanist who believed that in the past a single stem
and leaf or two leaves (phytomeres) represented the ancestral
plant which grew long by repeating itself, as the embryo
plant does at present.

Haeckel, in his ‘ Generelle Morphologie, 1866, used the
term “ promorphology ” to include the fundamental relations of
the parts of an animal to one another, in the same sense that
the relations of the axes of a crystal are the expression of its
form. The aim of promorphology, Haeckel said, is to deter-
mine the ideal fundamental form by a process of abstraction,
and to discover the natural laws according to which organic
matter develops its outer form. The relation of the parts, i.e.
the form, results with absolute necessity from the architectural
union of the constituent parts, in the same sense that an inor-
ganic crystalline form results from the union of crystalline
material, and from its relation to its environment.

Whether morphology will be ultimately driven to an inter-

A STUDY OF METAMERISM. 469

pretation of the symmetry of organic form as the expression of
physical laws of protoplasm, rather than due to slow adapta-
tion of an amorphous or irregular substance to its surround-
ings, is too large a question to attempt to discuss. Whatever
the explanation may be, there are certain well-established facts
in this connection that have, it seems to me, an important
bearing on metameric repetition.

The relation existing between bilateral and radial symmetry
is one of the most suggestive fields of promorphology. We
get from this source more suggestion as to a possible interpre-
tation of the facts of metameric repetition than from any other
source. It would lead too far to attempt anything like a full
discussion, but I may cite two cases that will serve as simple
illustrations of a large field of inquiry.

The radial symmetry of a sea-urchin is a very perfect type of
five-rayed structure. The test, made up of a mosaic work of
calcareous plates, is a marvellous piece of detailed fitting. Yet
there are several cases on record of individuals that have a
sixth ray (antimere) introduced. Each of the six antimeres
may be a perfect copy of the others, as well as of the normal.
The same condition is not uncommon in the star-fish and other
Echinoderms, but owing to the lack of a mosaic calcareous
skeleton the result is not so impressive.

Again, in other individuals one of the five antimeres may be
entirely or in part omitted, and yet the surface shows a perfect
symmetry. The point here to be emphasised is that a whole
section of the body (antimere) may be introduced or omitted,
involving the introduction or loss of all the organs belonging
to such a division.

More remarkable still are the triangular tapeworms de-
scribed by Leuckart and others. A strongly marked bilateral
animal repeats occasionally one of its halves, so that we may
paradoxically speak of the worm as composed of three halves.
Instead of a somewhat flattened bilateral animal, there results

! Bateson points out two cases which are to be distinguished. There may

be a division into two of one antimere, or there may bea redistribution of
the whole material into six parts. The text refers to the latter.

4.70 T. H. MORGAN.

a radial (triradiate) form, having three sets of longitudinal
organs where normally there are only two. Two other facts
in connection with these variations ought to be emphasised.
The scolices of the tapeworms arise on the large bladderworm
by invagination of the surface wall. The relation of the inva-
gination to the surface of the sphere is a radial rather than a
bilateral relation, and it is interesting to find this occasionally
expressed in the triangular scolices. In the second place,
Leuckart records finding on the same bladder (Cysticercus)
both radial and bilateral scolices. This seems to show that the
radial type need not have come in from egg variation, but is
the result of conditions acting at the time of the formation of
the scolices.

Now both of these cases, the sea-urchin and the triangular
tapeworms, show that a complete section of the body may be
repeated and intercalated symmetrically amongst the other
parts. The new portion has appeared at once and fully
equipped, duplicating the structures of the body that lie in a
similar axial position.

These and many other similar cases show us, I think, very
positively that the variations appearing in a radial animal
must have come simultaneously and all together into the anti-
meres.

Moreover I think no one will doubt that the relation
existing between the repeated organs in a radiate animal is at
bottom the same relation existing between the right and left
sides of the body of a bilateral animal.

Mivart (31) and Brooks (9) have emphasised the further fact
that the relation between the right and left sides of the body
is the same relation that exists between the serially repeated
parts of a metameric animal.

If this line of argument be admitted, it puts the problem of
metamerism into a large category of well-established facts.
That the final explanations of these facts is closely bound up
with the solution of some of the most fundamental problems
of biology is self-evident.

To hope, therefore, to solve the problem of metamerism in

A STUDY OF METAMERISM. A471

the simple ways already tried by phylogenists and embryo-
logists is, it seems to me, a vain hope. But a more vigorous
study of the facts of metameric repetition from the standpoint
reached above might lead us in the right direction. A
study of a larger number of facts than we possess at present
will at least tell us whether or not metamerism is to be re-
ferred to this category. If this should prove true, we shall
know where to search for the key to the problem. Even if
there is no immediate hope of reaching a definite conclusion,
yet we shall have gained much if we can find in what direction
the solution lies.
BIBLIOGRAPHY.
1, AnpREws, E. A.—* Report upon the Annelida Polycheta of Beaufort,
North Carolina,” ‘ Proceed. U. S. Nat. Museum,’ vol. xiv, 1891.
2. Batrour.—‘ Comparative Embryology,’ 1879.
8. Batrson.—“ The Ancestry of the Chordata,” ‘Quart. Journ. Micr.
Sci.,’? xxvi.
4. Batrson.—‘ Materials for the Study of Variation,’ 1894.
5. Brpparp.—‘ Proceed. Zool. Soc.,’ 1886.
6. Bennam, W. B.—‘ Annals and Magazine of Nat. Hist.,’ ser. 6, vil,
1891.
7. Bere, R. S.—‘ Untersuchungen iiber den Bau und die Entw. d. Ge-
schlechtsorgane der Regenwirmer,”’ ‘ Zeit. f. wiss. Zool.,’ xliv, 1886.
8. Brooxs.—“ Salpa in its Relation to the Evolution of Life,’ ‘Stud.
Biol. Lab. Johns Hopkins Univ.,’ May, 1893.
9. Brooxs.—‘ Lucifer,” ‘Phil. Trans.,’ 1882 (reprinted); ‘ Morpho-
logical Monographs,’ 1884.
10. Bucuanay, F.— Peculiarities in the Segmentation of certain Poly-
chetes,” ‘Quart. Journ. Mier. Sci.,’ xxxiv, 1893.
11. CatpwreLi.— Blastopore, Mesoderm, and Metameric Segmentation,”
‘Quart. Journ. Micr. Sci.,’? xxv, 1885.
12. Ciavs.— Zur morphologischen und phylogenetischen Beurtheilung des
Bandwurmskérpers,” ‘ Arb. aus d. Zool. Inst. Wien,’ viii, 1889.
18. Cori, C. J.—‘* Ueber Anomalien der Segmentirung bei Anneliden,”
‘ Zeit. f. wissen. Zoologie,’ liv, 3, 1892.
14. Frecpp, ADELENE M.—* Observations on Tenacity of Life and Regene-
ration of Excised Parts in Lumbricus terrestris,” ‘ Proceed.
Acad. Nat. Science,’ 1885.
VOL. 37, PART 4,—NEW SER. II

4,72 T. H. MORGAN.

15.
16.

Li:

18.

19.

20.

21.
22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34,

35.
36.
37.

Harcxet.—‘ Generelle Morphologie,’ 1886.

Harscuex.— Studien tiber Entwicklungsgeschichte der Anneliden,”
‘Arb. a. d. Zool. Inst. Wien,’ Bd. i, 1878.

Hatscnex.— Ueber Entwicklungsgeschichte von Echiurus,” ‘ Arb. a.
d. Zool. Inst. Wien,’ iii, 1880.

HatscueK.—‘ Lehrbuch d. Zoologie,’ 1888.
His, W.—‘ Uisere Korperform,’ 1874.

Haacxe.—* Zur Morphologie der Seeigelschale,” ‘ Zool. Anzeiger,’ viii,
1885.

Hertwie, O. and R.—‘ Die Celomtheorie,’ 1881.

Husrecut.—* On the Ancestral Form of the Chordata,” ‘ Quart. Journ.
Micr. Sci.,’ xxiii, 1883.

Husrecut.—“ The Relation of the Nemertea to the Vertebrata,”
‘Quart. Journ. Micr. Sci.,’ xxvii, 1887.

KenneL, J. v.—‘ Ueber Theilung und Knospung d. Thiere,’ Dorpat,
1887.

Lane, A.—‘ Der Bau von Gunda segmentata,” ‘Mitt. a. d. Zool.
Station Neapel,’ iii, 1881.

Lane.—‘ Ueber d. Hinfluss d. festsitzenden Lebensweise auf. die
Thiere,’ Jena, 1888.

LANKESTER.— On the use of the term ‘ Homology’ in Modern Zoology,”
‘Ann. and Mag. of Nat. Hist.,’ 1870.

LankEsTER.—“The Vertebration of the Tail of Appendicularie,” ‘Quart.
Journ. Micr. Sci.,’ xxii, 1882.

Meyer, Ep.—“ Die Abstammung der Anneliden,” ‘ Biolog. Central-
blatt,’ x, 1890.

MIcHAELSEN.—“ Oligochaeten des Naturhistorischen Museums im Ham-
burg,” iv, ‘ Jahrb. d. Hamburg wiss. Anstalt,’ viii, 1891.

Mivart.—‘ On the Genesis of Species,’ 1871.

Morean.—“ The Development of Balanoglossus,” ‘ Journ. Morph.,’ ix,
1894.

Morean, T. H.—‘“ Spiral Modification of Metamerism, *Journal of
Morphology,’ vii, 2, 1892.

Sepewick.—“On the Origin of Metameric Segmentation,” ‘ Quart.
Journ. Mier. Sci.,’ xxiv, 1884.

Spencer, H.—‘ Principles of Biology,’ vol. ii, appendix B, 1867.
SprencEL.— Die Enteropneusen,” ‘Fauna and Flora,’ Neapel, 1893.

Wi.son.— The Mesenterial Filaments of the Alcyonaria,” ‘ Mittheil, a.
d. Zool. Station Neapel,’ v, 1884,

A STUDY OF METAMERISM. 473

88. Witson, E. B.—‘‘ Embryology of the Earthworm,” ‘Journ. Morph.,’
iii, 1889.

39. Wittey, A.— Studies on the Protochordata:” I, ‘ Quart. Journ. Mier.
Sci.,’ xxxiv, 1893.

40. Woopwaxrp, M. F.—* Description of an Abnormal Earthworm possess-
ing Seven Pairs of Ovaries,” ‘ Proceed. Zool. Soc.,’ 1892.

DESCRIPTION OF PLATES 40—43,

Illustrating Mr. T. H. Morgan’s paper, “‘ A Study of
Metamerism.”

PLATE 40.

Type forms of modification of metameres of Annelids. See pages 395 to 403.

Fie. I, a,B, c.—Compound metamere (split metamere), as seen from dorsal
(A), ventral (B) side, aud reconstructed in c, as seen from above.

Fic. II, a B c.—Modification of last.

Fic. III, a 8 c.—Inserted half-metamere.

Fic. IV, A B c.—Double compound metamere.

Fie. V, a B c.—Failure of lines to meet dorsally.

Fic. VI, a, B, c.—Spiral of two metameres.

Fie. VII, a, B, c.—Spiral, with one more half-metamere on one side than on
other. In all five half-metameres.

Fie. VIII, a, B, c.—Spiral, with same number of half-metameres on each
side. In all six half-metameres.

Fie. IX, a, B, c.—Spiral of type VIL, but longer.
Fic. X, A, B, c.—Spiral of type VIII, but longer.
Fic. XI, a, B.—Spiral of type VIII, but still longer.

Fic. XII, a, 8, c.—Spiral formed by combination of half-compound meta-
meres.

Fie. XIII.—Spiral, in part a double-spiral, formed by introduction of
half-compound metameres.

Fic. XIV.—Spiral, in part a double-spiral, formed by the introduction of
a half-double-compound metamere.

474,

T. H. MORGAN.

PLATE 41.

All figures are from the anterior end of L. (Allolobophora) fetidus.

Fic.
Fie.
Fic.

Fie.
above.

Fie.
Fic.
Fic.
Fic.
Fic.
Fic.
Fic.
Fic.
Fic.
Fic.
Fic.
Fic.
Fie.
Fig.
Fic.
Fic.
Fig.
Fic.
Fic.
Fic.
Fic.
Fic.
meres.
Fie.
Fia.
Fic.
Fic.
Fie.

See text pages 407 to 418.
1, 4 B.—Fourth metamere, incompletely developed on right side.
2, A B.—Second metamere, ditto.
3, A B.—Fifth ditto, left side.
4.—Failure of septal line between 9th and 10th metameres to meet

5.—Ditto between 5th and 6th ditto.
6.—Ditto 4th and 5th ditto.

7.—Ditto 3rd and 4th ditto.

8.—Ditto 13th and 14th ditto.

9.—Ditto 12th and 13th ditto.

10.—Ditto 11th and 12th ditto.

11, 4 B.—Ditto 1st and 2nd below.
12.—Ditto 13th and 14th below.
13.—Ditto 5th and 6th at side.
14.—Spiral of category vi11, involving 7th, 8th, and 9th metameres.
15.—Ditto, involving 12th, 13th, and 14th.
16.—Ditto, 11th, 12th, and 13th.

17.—Ditto, 12th, 13th, and 14th.

18.—Ditto category x, involving 6th, 7th, 8th, and 9th.

19.—Ditto, 11th, 12th, 13th, and 14th.

20, A, B, c.—Ditto category vil, involving 2nd, 3rd, and 4th.

91, a, B, c.—Ditto, 9th, 10th, and 11th.

22, A, B, c.—Ditto category 1x, involving 6th, 7th, 8th, and 9th ditto.
23, A B, C D.—T wo modifications—categories 11 and v.

24, 4 B c.—Short spiral; origin doubtful.

25, AB C.—Spiral virr and compound metamere 1 combined.
26.—Spiral of category vii, very long, involving 11th—24th meta-

27, A B c.—Ditto, long, involving 10th—16th.

28.—Ditto, 13th—24th.

29, a B c.—Spiral category vu, involving 14th, 15th, and 16th.
30, A B c.—Ditto category vitl, involving 15th, 16th, and 17th.
31, a B.—Ditto, 15th, 16th, 17th, and 18th.

A STUDY OF METAMERISM. A75

Fic. 32, a B.—Ditto, 15th, 16th, and 17th.

Fic. 33, A B.— With compound metamere 10—22.

Fic. 34, a B.—Ditto 7—2. Also category v.

Fig. 35, a B c.—Ditto 9—.8,. Also category II.

Fic. 36, A B c.—Spiral vil, involving 11th, 12th, and 13th metameres.

Fig. 37, A B c.—Compound metamere 12—32.

Fic. 38, A B c.—Category 1.

Fic. 39, A B c.—Compound metamere 15—35.

Fic. 40, A B.—Spiral vir, involving 8th, 9th, and 10th. Also another
modification.

Fie. 41, A B c.—Four compound metameres I and spiral vit.

Fic. 42, a B c.—Compound metamere 3—3, followed by complicated spiral.

Fic. 43, a B c.—Two spirals vit, and compound metamere II.

Fic. 44, 4 B.—Compound metamere 6—. Spiral viz. Compound meta-
mere 13—13.

Fie. 45, AB c.—Intercalated half-metamere 15. tu. Spiral involving
15th—19th metameres.

Fic. 46, a B c.—A much modified anterior end. See construction c.

PLATE 42.
All figures of L. foetidus except Fig. 74 (L. terrestris).

Fie. 47.,—All of the abnormalities drawn from one very abnormal worm.
The numbers inserted between the spirals, compound metameres, &c., give
the number of normal rings that have been omitted.

Fics. 48—50.—Vasa deferentia opening on different metameres. The
figures as seen from below.

Fies. 51, 52.—Vasa deferentia doubled on one side. Seen from below.

Fires. 53—55.—Both vasa deferentia on 10th, 12th, and 14th metameres
respectively. Projections from above.

Fies. 56—58.—Vasa deferentia on alternate segments. Diagram as seen
from above, so that right and left of Figs. 53—60 are reversed as compared
with Figs. 48—52 (below).

Fries. 59, 60.—Vas deferens doubled on one side. Projections from above.

Fics. 61—64.—Dissections of compound metameres to show arrangement
of septa, &c.

Fies. 65—70.—Dissection of spirals to show condition of septa.

Fig. 71.—Failure of surface line to meet above. Septa were normal.

476 T. H. MORGAN.

Fic. 72, A B.—Disagreement between surface line (a) and septa (B).

Fie. 73.—Dissection of a compound metamere. Septa do not agree with
surface lines.

Fie. 74, 4 B.—Vas deferens doubled on right side. L. terrestris.

Fies. 75—78.—Abnormal arrangement of metameres in young worms at
time of emergence from coccoon. Fig. 75, category 1; Fig. 76, category vir;
Fig. 77, category vii; Fig. 78, category x (Pl. 40).

PLATE 43.

Fics. 79—81.—Regenerating posterior ends of the body of L. fetidus,
See text for description.

Figs. 82—85.—Abnormal arrangement of the metameres in Amphinome.

Fies. 86 —88.—Abnormal arrangements of the rings of leeches (Macrob-
dilla decora).

Fic. 89, A 8 c.—Compound metamere 5—5 in abdomen of locust.

Fics. 90—95.—Abnormal arrangements of the rings of the lobster,
Homarus americanus.

Fic. 96.—Nine segments of the arm of a brittle-star (Opliiuridea) from
Jamaica, as seen from below. Every fourth segment is a pigmented colour-ring.

Fic. 97.—Nine segments of another individual, showing the middle pig-
mented ring broken,

Fic. 98.—To show a further separation of colour-rings.

ON THE C@LOM, GENITAL DUCTS, AND NEPHRIDIA. 477

On the Celom, Genital Ducts, and Nephridia.
By

Edwin 8S. Goodrich, F.L.S.,
Assistant to the Linacre Professor of Comparative Anatomy, Oxford.

With Plates 44 and 45.

Tue chief object of this paper is to call attention to a
theory of the homology of the coelom which has been gradually
gaining ground abroad, but has not, I venture to think,
received in this country the notice which it deserves. The
theory I refer to is, that the cavity which we know as the
coelom in the higher Coelomata is represented by that of the
genital follicles in the lower types of that grade. Ever since
Hatschek wrote the often-quoted words: “ Die secundire
Leibeshohle verhalt sich wie die Héhle der Geschlechtsdriise
der niedrigeren Formen,” and pointed out that true meso-
blastic metamerism is really due to the repetition of the gonads
(48), so many favourable new facts have been brought to light,
that from a suggestion the statement has become a well-
established theory. In a most interesting and suggestive
paper on the ancestry of the Annelids, Dr. E. Meyer has set
forth the theory in some detail (81). After showing that the
ccelomic cavities in the Polychetes are quite comparable in
development and function with the genital follicles of the
Planarians, he further maintains that the theory throws a
flood of light on various otherwise obscure questions, such as
the bilateral character of the coelom, its invariable connection
with the genital cells, the absence of a truly unpaired prosto-

478 EDWIN S. GOODRICH.

mial celom, &c. He then treats of the nephridia and genital
ducts, and it is with this part of the question that I wish more
especially to deal in this paper. Meyer holds that the ne-
phridium of the Platyhelminths is represented by the so-called
head-kidney in the first segment, and by the tube of the
nephridium in the trunk segments of the Annelids, while the
genital duct of the Platyhelminths is represented by the wide
funnel of the trunk nephridium which develops, independently
of the tube from each genital or ccelomic follicle, and becomes
grafted on to it afterwards. Unfortunately, the author restricts
his remarks almost entirely to the Polychztes and those forms
which appear directly to lead up to them. Although the
theory has been at all events partially adopted by other writers
(Lang, 71; Korshelt and Heider, 67), no one, as far as I am
aware, has pushed it to its logical conclusion, and applied it to
all the groups of Celomata. This is what I shall attempt to
do in this paper. First of all, however, there is one thing to
be noticed with regard to Meyer’s general statement about
the nephridial funnel, namely, that since the publication of
his own researches on the Polycheta, and of those of Vejdovsky
and others on the Oligocheta, there can be no doubt that the
nephridial funnel in the latter forms part of the true nephri-
dium (is, in fact, derived from the end-cell), and is not a
grafted genital duct, as is the case in some at least of the
Polycheta.

An unprejudiced review of the well-established and recently
ascertained facts concerning the development of the excretory
organs and genital ducts of the Colomata must, I think,
inevitably lead us to the conclusion that we have been confus-
ing two organs of totally different origin under the one name
nephridium—the one organ the true nephridium, the other
the morphological representative of the genital duct, which
may be called the peritoneal funnel, to avoid confusion.
Further, that while on the one hand in certain groups such
as the Planaria, Nemertina, Hirudinea, Cheetopoda, Rotifera,
Entoprocta, besides the genital ducts or peritoneal funnels, we
find true nephridia in the adult; on the other hand, in such

ON THE C@LOM, GENITAL DUCTS, AND NEPHRIDIA. 479

groups as the Mollusca, Arthropoda, Ectoprocta, Echinoderma,
and Vertebrata, there are in the adult no certain traces of true
nephridia. In these latter groups, as we shall see, the peri-
toneal funnels (primitive genital ducts) take on the excretory
functions of the nephridia which they supersede.

In the following brief review of the various classes of Ceelo-
mata, I shall endeavour to show that the two kinds of organs
can always be distinguished ; that the first, the nephridium, is
primitively excretory in function, is developed centripetally
as it were, and quite independently of the ccelom (indeed, is
probably derived from the epiblast), possesses a lumen which
is developed as the hollowing out of the nephridial cells, and
is generally of an intracellular character, is closed within, and
may secondarily acquire an internal opening either into a blood
space or into the ccelom (true nephridial funnel as opposed to
the peritoneal funnel) ; and that the second kind of organ, the
peritoneal funnel, is primitively the outlet for the genital pro-
ducts, is invariably developed centrifugally as an outgrowth
from the coelomic epithelium or wall of the genital follicle, is
therefore of undoubtedly mesoblastic origin, and possesses a
lumen arising as an extension of the ccelom itself.

In the series of diagrams illustrating this paper, based on
the most recent and accurate researches, it has been my con-
stant endeavour to interpret the author’s results correctly, and
not to distort the facts in favour of the theory here advocated.

PLANARIANS.

The nephridia of the Planarians, as is well known, are
formed of a main duct, which branches out into fine tubules
ending blindly internally in flame-cells (fig. 1); they do not
develop beyond this “ pronephridial ’’ condition—protonephri-
dium of Hatschek (55). The arrangement of this single pair
of nephridia is extremely variable; the two organs may join
and open by a median external pore near the mouth or behind,
or they may open by a number of pores at the sides. Gunda
segmentata, a most interesting form described by Professor
Lang (69), possesses longitudinal main trunks into which

480 EDWIN S. GOODRICH.

open the fine branches ending in flame cells, and from which
pass segmental ducts to the exterior, corresponding to the
segmentally arranged gonads. It is by the breaking up of
such a system into separate organs that Lang would derive the
nephridia of the higher Ceelomata.

Unfortunately, we know little about the origin of the ne-
phridia in this group. Lang has described, in Discocelis,
paired ingrowths of the epiblast, which he believes give rise to
the nephridia (70). This observation strongly supports his
theory as to the phylogenetic derivation of the nephridia from
epidermal glands ; and, indeed, it seems pretty certain that an
incipient excretory organ to be eflicient must have been
derived from, or at all events situated close to, the surface
layer in order to get rid of its excretory products.

It is in the Planarians, a group undoubtedly primitive! in
some respects, that we should expect to discover the celom in
its first stages of development, and, in fact, we do seem to be able
to trace it from its first appearance. In some Accela (Graff,
42,44), and other simple forms, the gonads consist merely of
the genital cells lying freely in the parenchyma. In others,
these cells become surrounded by an epithelium formed by the
adjacent cells; the epithelial sacs, one on either side of the body,
may then become hollow, while the wall grows out to form two
tubes, the genital ducts (peritoneal funnels). Another important
stage is presented by these organs in Gunda segmentata
(Lang,69). Here the genital follicles are repeated segmentally,
the first pair being ovarian, the rest testicular sacs. If these

1 One of the most useful lessons of modern research has been to teach us
with what great care the word “primitive” should be applied to any group of
existing animals. A few years ago naturalists readily derived one group of
living animals directly from another, apparently more primitive; but their
genealogical trees are now becoming reduced to bushes, in which the branches
spring from a common base. Nevertheless, it is true that certain groups
may retain, either in their general organisation or in some particular details,
characteristics of the ancestors from which they have diverged. The Plana-
rians, with their complicated nephridial and genital apparatus, their deeply-
sunk nervous system, yet generally archaic plan of structure, are a striking
case in point.

ON THE C@LOM, GENITAL DUCTS, AND NEPHRIDIA. 481

follicles were larger, Gunda segmentata could be calledatruly
segmented animal.! The inner ends of the genital ducts are
formed as outgrowths from the genital follicles : ‘‘ Der Oviduct
ist bei Gunda segmentata, wie bei Planariatorva anfangs ein
solider Zellenstrang. Zweifellosensteht er durch Wucherung aus
dem soliden ovarium selbst, ahnlich wie die Samenleiter Aus-
wiichse der Hoden sind” (observations confirmed in his later
work, 70).2 The oviduct becomes hollow and ciliated, and grows
backwards to the genital pore; the vasa efferentia fuse to form
the main sperm-duct.

Tosum up,then. In the Planarians the excretory organs area
pair of pronephridia, probably derived from the epiblast; the
gonads arise from a mass of cells in the mesoblast, which may
become hollowed out into a genital follicle (coelomic sac) from
the wall of which arise the genital cells. The follicle grows out
to form the genital duct (peritoneal funnel), which joins an
epiblastic invagination at the genital pore (fig. 1).°

1 £. Meyer believes (81) that the ancestor of the Annelids possessed a pair
of long genital follicles, and that metamerism was brought about by their being
broken up at intervals, chiefly to facilitate its serpentine motion; each portion
would then have acquired its own duct to the exterior. It seems to me more
probable that the metameric arrangement of the genital follicles is more
directly due to that “tendency” to repetition by a sort of budding, which is
seen in the case of the gonads, the penes (Anonymus), and even the pharynx
(Phagocata; Woodworth, 112) amongst the Planarians, and again, perhaps
amongst the Mollusca (for a full discussion see Bateson, 3).

2 Jijima’s account of the development of the sperm-ducts differs somewhat
from that of Lang, but he derives both from the mesoderm (60).

3 The spaces contained in the connective tissue or parenchyma have been
sometimes compared with the ceelom; these spaces seem rather to represent
the vascular system of the higher Coclomata. I need not treat here of the
homology of the vascular system, which is probably of quite separate origin
from the celom. Professor Ray Lankester’s view, that the blood-system is
simply a liquefaction, as it were, of the mesoblast, seems to me to agree per-
fectly with the facts. Moreover, the theory held by many authors that it is
directly derived from the blastoccel appears to be quite untenable. As Pro-
fessor Lankester has pointed out to me, if this were the case we should
expect to find the blood-spaces best developed amongst the Diploblastica ;
now it is just in the (adult) Colenterates that it is entirely absent. Indeed
we may say that the blood-space, or hemoccel, does not appear in phylogeny

482 EDWIN S. GOODRICH.

About the Cestodes and Trematodes it need only be said that
they are built (so far as concerns the question discussed in this
paper) upon essentially the same plan as the Planarians;?
while as to the Nematodes, of the development of which we
know too little, it seems probable that here also the genital
follicles represent the coelom, while the body cavity is a blood-
space corresponding in its relations to the parenchyma of the

Planarians.
RoviFrEeRaA.

The Rotifers, recently described in great detail by Platte (87,
88) and Zelinka (115), agree in the general structure of the
nephridia, genital follicles, and genital ducts, so closely with the
Platyhelminths that they may be dismissed with a very few
words. The nephridia are a pair of branching tubes ending
internally in flame cells, and opening behind into the cloaca.’
The colom is represented by a pair of genital follicles,
one of which only is generally developed; the wall of the
follicle is produced backwards to form the genital duct, or
peritoneal funnel opening into the cloaca.

The development of the nephridia has not yet been thoroughly
worked out. Zelinka has traced them to a group of cells of
doubtful origin. He adds: ‘das Exkretionssystem konnte ich in
so fern mit Sicherheit auf das Ektoderm zurickfihren, als es
bestimmt nicht auf das Entoderm bezogen werden kann ”’ (115),
until the mesoblast has been formed, and, generally speaking, that the greater
the development of mesoblast the more definite is the vascular system. I
should rather consider the connection of the blood-spaces with the blastoccel
as arising for purely mechanical reasons, so to speak, and in no way of phylo-
genetic significance. The temporary continuity of the blastoccel with the
blood-space during the ontogeny of certain of the higher forms would seem to
be connected with that method of development by the folding of germ-layers
or sheets of tissue, which no one would look upon as primitive. During this
process, the cavities which will form the vascular system later on, are ine-
vitably continuous with the space left between the germ-layers.

1 Fraipont maintains that the flame end-cells of the nephridia of the Cestodes
and Trematodes communicate internally by means of a small lateral aperture

39).
: 3 As in the case of some Platyhelminths, the nephridia have been stated to
open internally (Eckstein, 28).

ON THE C@LOM, GENITAL DUCTS, AND NEPHRIDIA. 488

ENTOPROCTA.

This group also, from our present point of view, differs little
in its organisation from the Planarians.

A pair of nephridia—short tubes, generally with an intracellu-
lar lumen and ending in a flame-cell or a group of cells with
cilia,—open by amedian pore in front of the genital pore (fig. 4)
(Hatschek, 47; Harmer, 46a; Ehlers, 29; Davenport, 26).
Possibly in some forms they open internally (Joliet, 61),andin
Urnatella they may be connected with a system of branching
tubules ending in flame-cells (Davonport, 26). The origin of
the nephridia is doubtful; Hatschek traced them in the embryo
to cells which he believed to be mesoblastic (47).

Embedded in parenchymatous tissue are a pair of genital
follicles (fig. 4). From each a typical peritoneal funnel leads to
a median pore (Ehlers, 29).

Mo.uvsca.

We now have to deal with a group of animals in which, as I
shall endeavour to show, the embryo is provided with a pair of
true nephridia; later, when the two ceelomic sacs belonging to
the unsegmented trunk have acquired a considerable size,
the peritoneal funnels formed from their walls take on the
excretory function, whilst the nephridia degenerate !,

True nephridia have been described in all the groups of
Mollusca except the Isopleura and the Cephalopoda, and have
recently been the subject of a special study from Dr. R. von
Erlanger (34, 35). They consist of short tubules formed, as a
rule, of one or of a small number of cells, pierced by a canal
which communicates with the exterior by a pore behind the
velum (fig. 20). The inner end of the nephridium is provided
with a flame-like bunch of cilia, or with a flagellum. An
internal opening into the blood-space or head-cavity has been
observed in some forms, such as Lymneeus, Helix (de Meuron,
79; Jourdain, 62; Fol, 38; Sarasin, 92; Wolfson, 111; v.
Erlanger, 35, &c.), Teredo (Hatschek, 50), and perhaps in

1 This view evidently obviates the difficulty some have felt as to the presence
of two pairs of so-called nephridia in an animal composed of one segment.

484 EDWIN S. GOODRICH.

Cyclas (Ziegler, 116). In other cases, such as Paludina, and
Bythinia (v. Erlanger, 32, 33), the nephridium appears not to
open internally, but to remain in the pronephridial stage.

The first stages in the development of the nephridium of the
Mollusca have, unfortunately, not yet been satisfactorily
worked out. Rabl (88a) derives the nephridium in Planorbis
from a large cell which also gives rise to the mesoblast.
Wolfson traced that of Lymnzus, which, he says, is derived
from a large in-wandering epiblastic velar cell on either side
(111). Hatschek (50), v. Erlanger (32, 33), and others trace
it to cells which they consider to be of mesodermal character,
but the exact origin of which is not clear.! In some cases a
late epidermal invagination is said to take place at the nephri-
diopore which forms the peripheral end of the duct (v. Erlanger,
32, 34).

We must now examine the development of the excretory
organs of the adult Mollusca, which appear to be nothing but
peritoneal funnels. An accurate description of the develop-
ment of Paludina has recently been given by R. von Erlanger,
and although the ontogeny is much modified owing to the
asymmetry of the adult, I shall begin with it as it is the only
detailed account we have. The genital follicles or ceelomic
sacs (pericardium) arise as a cavity on either side, a hollowing
out of the mesoblast (fig. 20). These cavities enlarge and fuse
below the gut. On either side a thickening takes place on the
ventral wall of the coelom, which here grows out in the form of
a typical peritoneal funnel (fig. 21). The right peritoneal
funnel then enlarges and fuses with an epidermal invagination,
which forms the outer end of the excretory organ. ‘‘ Was die
Niere anbelangt,” says von Erlanger, “so bin ich der Ansicht,
dass der secernirende Abschnitt derselben aus dem Mesoderm

1 IT think we may safely say that there is nothing which precludes the pos-
sibility of the nephridia being primitively derived from the epiblast, as they
appear to be in the Platyhelminths, Rotifers, and Cheetopods; although the fore-
casts may sink in at a very early stage and thus become included in that “ An-
lagecomplex” which we call mesoblast. In Helix, de Meuron derives them
from the epiblast, but he is not positive about the internal extremity (79).

ON THE C@LOM, GENITAL DUCTS, AND NEPHRIDIA. 485

stammt und dass diejenigen Beobachter, welche ihr aus dem
Ektoderm enstehen lassen, entweder nur den ausfiihrenden
Theil der Niere beriicksichtigt haben oder, was noch haufiger
geschicht, die beiden Abschnitte nicht in ihren Zusammenhaug
erkannten: Der ausfithrende Theil wird namlich von Allen,
mit Ausnahme von Rabl, aus einem Theil der Mantelhdhle
abgeleitet ” (82). On the left side, to which the genital
function is restricted, the gonad develops from the wall of the
celom ; then, together with the rudimentary left peritoneal
funnel, it becomes constricted off from the main division of
the colom (the pericardium), forming a small genital sac.
From the wall of this sac the genital duct grows out, and
joins an epidermal invagination, like the peritoneal funnel of
the right side. Doubtless the genital duct is really the left
peritoneal funnel, as v. Erlanger suggests : ‘Ich konnte....
feststellen, dass die Anlage der Genitaldriise in der urspring-
liche linken Halfte des Perikards ensteht, und zwar ungefahr
da, wo sich die rudimentire linke Niere zuriickgebildet hat.
Eben so ensteht auch die Anlage des Ausfiihrganges an der
Stelle, wo der rudimentire Ausfiihrgang der linken Niere sich
befand, und scheint einfach aus diesem hervorzugehen.”” These
observations have been fully confirmed in the case of Bythinia
(33), and the conclusions, as we shall see shortly, are supported
by the comparative anatomy of the Mollusca in general
(fig. 22).

Ziegler has described the development of Cyclas cornea,
which in some respects is less modified, the adult being
symmetrical (116). Here also the ccelom (pericardium) arises
as a right and left follicle. The two cavities enlarge, surround
and fuse below the gut. On either side the organ of Bojanus
develops as a peritoneal funnel, which meets an epidermal
invagination (fig. 22).

In the Mollusca, then, we find at an early stage a pair of
true nephridia (Urniere, head-kidneys), possibly of epiblastic
origin. The ccelom develops as two cavities in the mesoblast,
genital follicles, from the walls of which grow out two
peritoneal funnels, the organs of excretion and carriers of the

486 EDWIN S. GOODRIOH.

genital cells of the adult. It can hardly be doubted that
primitively the renal organs (organs of Bojanus, nephridia of
authors) functioned as genital ducts. Such is, indeed, still the
case in Chetoderma, the Neomenians, and the Zygobranchia.
Also in the most primitive Lamellibranchs, such as Nucula
and Solenomya, the peritoneal funnels still retain their
original function. As Pelseneer has shown (86), gradually a
separate genital duct has been split off, which in Anodon,
Cardium, &c., opens independently. Intermediate stages are
found in such forms as Pecten, and Spondylus, where the
genital cells are shed into the kidney itself; and in Arca,
Ostrea, &c., where the kidney and genital duct open into a
common cloaca. Likewise in the Chitons a separation has
taken place of the genital region of the ccelom from the renal ;
the gonad then acquires special ducts, which may not be
homologous with peritoneal funnels. In the Cephalopoda,
although the coelom remains continuous, special apertures serve
for the escape of the genital cells ; whether these should be con-
sidered as a second pair of peritoneal funnels is also doubtful.!
(The aberrant form Rhodope veranii would appear to be
amongst the Mollusca, if it be really of that class, the only one
which retains true nephridia in the adult. It is provided with
a pair of branching tubes, opening by a common pore, and
ending internally in flame-cells. The coelomic or genital fol-
licles, ovarian and testicular, are small and numerous. Their
ducts join to a common hermaphrodite duct, which again di-
vides into two openings by male and female pores [48, 12].)

DInopui.vs.

For our purpose it will be convenient to treat this genus
separately, and not with the Archiannelida which will be in-
cluded with the Polycheta. It is of great interest, since, in
some species at least, the nephridia are metamerically repeated
whilst the colom is represented by a single pair of genital
follicles.

1 Such a case may be compared to that of certain Nemertines, where nu-

merous genital follicles and as many genital ducts are present in one segment,
and to the similar arrangement in the Vertebrates.

ON THE C@LOM, GENITAL DUCTS, AND NEPHRIDIA. 487

In Dinophilus apatris, Korschelt has described flame-cells
which he believed to be connected with a longitudinal canal
opening posteriorly (66). E. Meyer, however, has figured in
D. gyrociliatus five pairs of typical closed nephridia, or
pronephridia, ending in flame-cells internally, and disposed
metamerically according to the outer signs of segmentation (80).
Harmer (46) describes five pairs of very similar nephridia in
the female D. teniatus. In the male there are four pairs of
nephridia, perhaps with internal openings.!

The testes are a large right and left sac, which fuse across the
median line ; as Harmer himself says, it is ‘‘ possible that in the
connective-tissue lacune of the body of Dinophilus we have the
representative of the so-called ‘primary body-cavity,’ whilst in
the fully developed male the ‘secondary body-cavity’ is repre-
sented by the cavity of the testis, with which the funnels of the
vesiculze seminales are connected.”’ It might be added that these
funnels, which form the inner openings of the sperm-ducts, have
all the appearance of true peritoneal funnels, comparable to the
genital ducts of the Platyhelminths, and the other groups
already spoken of.

The paired ovarian cavities appear to be provided with only
very degenerate ducts, reduced to mere pores in D. vorticoides
and D. apatris (compare the Archiannelida and the female
ducts of certain Oligochzta such as the Enchytreeids).

The structure of Dinophilus might be explained in one of
two ways. Hither it has acquired a number of nephridia,
whilst retaining the primitive single pair of genital follicles ;
or it is a degenerate form which has lost its metamerically
repeated genital follicles, whilst retaining a number of separate
nephridia. Our present knowledge does not enable us to con-
clude for certain which of these explanations is the correct one.”

Histriodrilus Benedeni (Histriobdella homari) may

1 The fifth pair is possibly, according to Harmer, represented by the distal
portion of the genital ducts leading to the penis (compare with certain Poly-
cheeta where the nephridium fuses with the peritoneal funnel).

2 In a paper which has just appeared Schimkewitsch describes a ladder-
like nervous system, and traces of segmentation in the developing mesoblast

(‘ Zeit. f. w. Zool.,’ Bd. lix, 1895).
VOL. 37, PART 4,—NEW SEk. KK

488 EDWIN 8S. GOODRICH.

be closely related to Dinophilus. It possesses five pairs of
pronephridia (four in the female), and a distinct coelomic or
genital cavity also opening by one pair only of peritoneal
funnels to the exterior (Foettinger, 37). Here it seems to be
pretty certain that there was originally a segmented cclom ;
there is still a ventral chain of ganglia.

NEMERTINA.

The nephridia of the Nemertines consist essentially of a
longitudinal canal on either side of the anterior region of the ali-
mentary canal, opening to the exterior by one or by several pores
situated laterally at more or less regular intervals (von Kennel,
63; Oudemans, 85; Hubrecht, 59; Burger, 15,17). Internally
they have been stated to open into the blood-vascular system ;
the latest researches, however, do not support this view (15),
and Birger has shown that in several species the nephridial
canals give off fine branches ending in bunches of flame-cells
(17) (figs. 2, 3, and 24; in these diagrams the nephridia are
represented in the same region as the gonads).

Although the development of the nephridia has not been
followed out in detail, Hubrecht (58) and Birger (18) have
traced their origin from direct invaginations of the epiblast.

In this group of elongated worms the genital follicles are
numerous, and generally arranged in pairs, alternating with
the intestinal ceca. Each follicle communicates with the ex-
terior by a duct or peritoneal funnel, formed as an outgrowth
from its wall at a comparatively late period (figs. 2, 3, and 24).
It is hardly necessary to emphasise the striking similarity
between this metameric arrangement of follicles with their cor-
responding ducts and the almost identical metameric coelomic
follicles and genital ducts of the Chzetopods.!

OLIGOCH ATA.
Thanks to the numerous researches of Profs. Hatschek,
1 R. 8. Bergh, in 1885, in a paper which I have not seen, pointed out the
resemblance between the genital follicles of the Nemertines and the eelomic

cavities of the Chetopods; but he considered the ducts of the follicles to be
homologous with the nephridia of the latter group (6).

ON THE C@LOM, GENITAL DUCTS, AND NEPHRIDIA. 489

Bergh, Wilson, Vejdovsky, and others, we have now a very
detailed history of the development of the nephridia in the
Oligochetes.

The observations of Hatschek, Wilson, and Bergh do not coin-
cide in many particulars; but all these discrepancies have been
so admirably reconciled by Vejdovsky in his careful work on
Rhynchel]mis and other forms (101), that we can, I think, be now
quite confident that we have a satisfactory account of the
embryology of these organs ; more especially since these results
have been in many respects amply confirmed by the researches
of Whitman, Bergh, and Biirger on the Hirudinea.

The development of the nephridium in Rhynchelmis has
been most carefully described by Vejdovsky (101). In the
first stage he figures it consists of a large cell (trichterzelle or
funnel-cell) within or on the posterior surface of the septum.
This “ funnel-cell”’ divides and gives off a string of small cells
behind, from which is developed the canal of the nephridium. A
large vacuole appears between the two cells formed from the
funnel-cell itself, in which a flame-like flagellum is developed.
Vacuoles now arise in the posterior string of cells, fuse together,
and form the lumen of the canal which communicates with the
end-chamber containing the “ flame” (figs. 6 and 25). Finally
this chamber opens into the ccelom, and the posterior loop joins
the skin; a communication is established with the exterior by
means of a secondary invagination of the epidermis—the end-
vesicle. Quite similar is the development of the nephridium in
Stylaster and Tubifex (Vejdovsky, 100).

Bergh (9 and 10) traced back the nephridia in Criodrilus and
Lumbricus to a large cell, the funnel-cell, lying close to the
epiblast and between each successive pair of solid mesoblastic
somites. When these become hollowed out, the funnel-cell
buds off posteriorly a chain of cells, the future canal of the
nephridium ; vacuoles appear in these cells, as in Rhynchelmis,
to form the lumen. Meanwhile the funnel-cell itself, which
has retained its large size, pushes through the mesoblast to
reach the ceelomic cavity in the segment in front; here it divides,
acquires cilia, and becomes the funnel of the adult nephridium,

490 EDWIN &. GOODRICH.

The posterior canal grows to the surface, where it opens through
the epidermis ; in some cases there is here an invagination of
the epidermis to form the end-vesicle (Vejdovsky).

As to the origin of the funnel-cell—the forecast of the whole
true nephridium: it arises from the primitive cell row, or
nephric cord, formed by the repeated division of one of the
teloblasts on either side. In the earlier stages this teloblast
and the nephric cord to which it gives rise lie on the surface of
the embryo; thus the funnel-cells are epiblastic in origin.
From the nephric row one cell enlarges and enters into con-
nection with each successive segment, as described above (fig. 5).
To judge from the figures of Bergh, Wilson (108 and 109), and
Vejdovsky, in some forms, such as Dendrobena and Lumbricus,
the funnel-cells give off the chain of posterior cells, whilst
separating from the nephric row, thus remaining for some time
in connection withit. In other cases, such as Criodrilus, the
funnel-cells appear to separate first.!

In the embryo of most Oligochetes the nephridia of the
first segment are developed precociously to perform the
excretory functions at an early stage (fig. 25). Vejdovsky
(100 and 101) has described these organs in Rhynchelmis,
Chetogaster, AZolosoma, Nais, Allolobophora, Lumbricus,
Dendrobzena, &c., and Bergh described those of Criodrilus (9).
They consist of fine canals with an intracellular lumen, or some-
times of wider tubes; they'are often ciliated, and occasionally end
internally in a flame-cell; they appear to be always blind
within. Externally they open either by a median dorsal pore
or by two lateral pores on the first segment. In fact they
closely resemble the closed pronephridia of the Platyhelminths,
Entoprocta, and other groups we have already examined, or the
pronephridial stage of the trunk nephridia.

The origin of the cells which form the (larval) nephridia of
the first segment has not been traced; but since they arise (in
some cases at least) before the division of the promesoblast cell,

1 Bergh denies the derivation of the funnel-cell from the nephric row.
Wilson rightly traced the development of the main body or canal of the nephri-

dium from the nephric row, but failed to discover the origin of the funnel-
cell itself.

ON THE C@LOM, GENITAL DUCTS, AND NEPHRIDIA. 491

Vejdovsky considers it probable that they are derived from the
epiblast, a conclusion which agrees with the known develop-
ment of the posterior nephridia.

The mesoblast in the Oligochetes is formed, as in all
Annelids, as two germ bands, which become broken up into
separate somites. The hollowing out of these gives rise to the
coelomic follicles, which increase in size, surround the gut (a
stage resembling what we find in the Nemertines), and fuse
below it (figs. 6 and 25). The transverse septa, between
adjacent follicles, become pierced, allowing a communication
from one to the other (Kowalevsky, 68; Hatschek, 48; Wilson,
109; Vejdovsky, 101). From the wall of certain of these
follicles the gonads are developed, whilst others remain sterile.
The number and position of the fertile follicles varies consider-
ably according to the family of the worm in question, and even
amongst different individuals of the same species (Woodward
observed an earthworm with seven pairs of ovaries; 118, 114).

The genital ducts (peritoneal funnels) develop as a thicken-
ing of the celomic epithelium in the fertile segments, which
grows outwards towards the epidermis, with which it fuses
(figs.6,7,and 25). Vejdovsky, who has followed thedevelopment
of these organs in several forms, such as Stylaria, Cheetogaster,
the Enchytreids, and Tubificids, says: “Die Anlage des
Samenleiters wiederholt sich nach dem oben Dargestellten in
iibereinstimmender weise bei allen bisher beoachtete Familien.
Die zuerst zum Vorschein kommende Anlage des Samentrichters
besteht aus einer Zellvermehrung des Peritoneums an den
Dissepimenten der betreffenden Segemente” (100). His
observations have been confirmed by Bergh (8) and Lehmann
(74) in the Lumbricids.t In the case of the male ducts, these
organs may be further complicated by the fusion of two con-

1 Beddard (4) tried to show that the genital ducts were derived from the
nephridia in Acanthodrilus. The few facts he brings forward from the very
scanty material at his disposal do not, I think, prove his case. ‘The theory
of Claparéde that the genital ducts of the Oligochetes were the modified
nephridia of the genital segments was founded on an erroneous notion,
since thoroughly disproved by Vejdovsky’s observations.

492 EDWIN 8S. GOODRICH.

secutive peritoneal ‘funnels, and by the invagination of the
epidermis at the genital pore to form an atrium and penis.

We may now sum up the main characteristics of the nephridia
and genital ducts in this group, which has been treated at
length owing to its great importance (figs. 5, 6, 7, and 25).

The nephridia of the Oligochetes are probably of epiblastic
origin. They develop from large cells (‘‘funnel-cells” ),
arranged metamerically outside and between each pair of
somites. They pass through a more or less disguised prone-
phridial stage (comparable to that permanently retained in
flatworms, &c.); in the first (most forms), and sometimes in
the trunk segments (Chetogaster) they never develop beyond
that stage. In the other segments the nephridia grow towards,
and open into, the ccelom by means of a funnel formed from the
original ‘‘funnel-cell.”!

The genital ducts, on the other ai are peritoneal funnels
of undoubted mesoblastic origin, which grow outwards from the
metameric genital follicles to open to the exterior. They thus
have no connection with the nephridia, and differ from them
entirely in their development.

HIRUDINEA.

So closely do the ceelom, genital ducts, and nephridia of the
Leeches agree in their development with those of the Oligo-
cheetes, that their history need only be rapidly sketched.

Birger (16 and 19) has carefully traced, in several forms,
the origin of the whole nephridium proper (funnel and canal)
from a large “ funnel-cell,” which comes to lie in the hinder
wall of each ceelomie follicle. Just as in the previous group of
worms, this cell buds off a row of cells behind which constitute
the canal; the “funnel-cell”’ then divides up into a ring of
small cells, which form the funnel of the adult nephridium
(figs. 8and 9). This organ remains closed in some forms, such
as Hirudo, but opens into the ccelom in others, such as Nephelis.

1 The branched so-called plectonephric condition of the nephridia in certain

earthworms has recently been shown to arise by the secondary subdivision of
originally paired nephridia (Vejdovsky, 102; A. G. Bourne, 18).

ON THE C@LOM, GENITAL DUCTS, AND NEPHRIDIA. 493

Meanwhile the lumen of the canal in the posterior chain of
cells becomes hollowed out. The peripheral end of the canal
fuses with an invagination of the epidermis, the vesicle, by
means of which it opens to the exterior. The first origin of
the “ funnel-cells,’’ from which the nephridia are formed, has
not been traced in detail; it seems quite probable that they are
derived from the nephric rows described by Whitman (107) in
Clepsine. (They are possibly the large cells mistaken by
Whitman [106] for the forecasts of the testes, as suggested by
Bergh.)

As in the Oligocheta and Polycheta, so in the Hirudinea
the nephridia of the anterior segments develop precociously in
the larva. Bergh (7) has traced their origin as outgrowths from
the epiblastic cell-rows. In Aulastoma there are four pairs,
which never develop beyond the pronephridial stage, i.e. do not
open internally. Bergh was also unable to find external
openings (compare the nephridia of Capitella, 30 and 81).

The ccelom develops in a normal manner as a hollowing out
of the paired metameric blocks of mesoblast. Most of these
ceelomic or genital follicles are fertile: an anterior pair develop
the ovaries on the peritoneal wall; several posterior pairs
develop the testes. That part of the ccelom which surrounds
the gonads generally becomes partially separated off as a peri-
gonadial ceelom (Bourne, 12a; Birger, 16). That the genital
ducts are peritoneal funnels is shown by their development,
although it seems to be somewhat modified. The oviducts
arise from the ccelomic epithelium surrounding the ovary, and
fuse with the two ends of a forked, but median invagination of
the epidermis (fig. 9) (Biirger, 19). The vasa efferentia are
similarly formed from the testes (fig. 28); they grow outwards
and forwards, fusing with those in front. The most anterior
join the ends of a median forked invagination of the epidermis
(Nussbaum, 84; Biirger, 19). The complete genital ducts thus
closely resemble those of some Planarians (Gunda), and differ
essentially from those of the earthworm only in the number of
peritoneal fnnnels which contribute to their formation (compare
also the Vertebrates).

494, EDWIN S. GOODRICH.

ARCHIANNELIDA AND PoLycH#Ta.

The first origin of the nephridia in these worms is not so
well known as in the case of the Oligocheta. It is, however,
to be remarked that in the only case where the forecast of the
nephridium appears to have been traced from the beginning,
it has been found to arise from the epiblast (the head-kidney
in Nereis; Wilson, 110).!| This would agree with what we
have seen occurs in most, if not all, of the groups we have
already examined.

E. Meyer (80) has given us a most excellent description of
the development of the nephridia in Polymnia and Psygmo-
branchus. They arise on either side from large cells, situated
close to the epiblast between each pair of mesoblastic somites,
with which they have no connection at this early stage. These
large cells, which seem to me obviously homologous with the
“ funnel-cells” of the Oligocheta and Hirudinea, divide, form-
ing a short chain of cells within which an intracellular lumen
becomes hollowed out (figs. 10, 11, and 26). The mesoblastic
somites become hollowed out to form the genital or celomic
follicles, from the posterior wall of which the ccelomic epithe-
lium becomes pushed out, forming a typical ciliated peritoneal
funnel, which fuses with the internal blind end of the nephri-
dium (figs. 11 and 26). This specialised portion of the coelomic
epithelium forms the wide-mouthed funnel of the adult nephri-
dium (fig. 12). The lumen of the nephridial duct becomes
intercellular by the multiplication of the cells which constitute
its wall, and breaks through at its point of junction with the
funnel on the one hand (opens, in fact, here into the ccelom),
and establishes a communication with the exterior through the
epidermis on the other. Such is the history of the wide-
mouthed segmental organs of compound origin of the trunk.
The nephridia of the first segment (or sometimes of several

1 Hatschek (48, 51, 54), Salensky (91), and von Drasche (27) all consider
the cells from which the nephridia are developed to be of mesoblastic origin,
but the evidence on this point is not convincing. Possibly, however, as sug-

gested for the Mollusca, they have secondarily come to be derived from the
mesoblast,

ON THE C@LOM, GENITAL DUCTS, AND NEPHRIDIA. 495

anterior segments) develop in the same way as the posterior,
but never pass beyond the pronephridial stage (head-kidneys) ;
they end blindly internally with a typical flame-cell (fig. 26).
Meyer figures as many as five pairs of such pronephridia in
Nereis cultrifera (80)'. Although the head-kidneys of the
Mollusca undoubtedly occasionally open internally, v. Drasche
and Hatschek seem to have been mistaken in describing an
internal opening in these organs in the Cheetopods (see Frai-
pont, 40; Meyer, 80).

Fraipont appears to attribute a very similar history to the
wide-mouthed trunk ‘“‘nephridia”’ of Polygordius as Meyer has
described for the trunk “nephridia” of the Tubicolous Poly-
chetes, though his statements are less precise: “ Le méso-
blaste est representé de plus au niveau des muscles obliques
par une masse de cellules assez confuse. C’est dans ce groupe
de cellules situées au dessus des muscles obliques contre les
champs musculaires longitudinaux que ce différencient les
entonnoirs des organes ségmentaires et plus tard encore les
organes sexuels, C’est un simple épaississement du péritoine ”
(40).

We see, then, that in the Polycheta nephridia are developed
from large cells, which may be compared to the “ funnel cells,”
giving rise to the nephridia in the Oligocheta. Whilst,
however, in the latter the pronephridium acquires an opening
into the ccelomic follicle independently of the peritoneal funnel,
which acts as a genital duct ; in the former, the Polycheta, the
pronephridium may acquire an opening into the ccelomic
follicle in the region where the peritoneal funnel is formed,
fuse with it, and become an organ of double function—excretory
and genital. In many cases division of labour leads to the
restriction of the genital function to one set of ccelomic follicles
and their funnels, and of the excretory function to another set

1 There can now, I think, be no doubt that the head-kidneys are simply the
precociously developed nephridia of the first segment; they do not open into
the ccelom for the very good reason that at this stage there is, as a rule, no
ccclom for them to open into. They preserve the same relations as the Platy-
helminth nephridia (Hatschek, 48, 51, 54; Meyer, 80).

496 EDWIN 8S. GOODRICH.

of celomic follicles and their funnels. This may lead to a
corresponding differentiation of structure ; in the first set the
peritoneal funnel becomes the most important part, in the
second the nephridial portion (such specialisation has been
well described in many forms by Eisig, Meyer [80], Trautzsch
[98], Cunningham [25], Marion and Bobratzky [78], &c.).

The fusion between the nephridium and peritoneal funnel
does not occur in all Polychetes ; fortunately, we appear still
to have all the intermediate stages between this condition and
that of the Oligochetes.

Meyer (80) has shown that in Nereis the nephridia of the
first five segments have the typical pronephridial structure
with a flame end-cell, and that in the posterior segments
this end-cell (judging from his figures) opens into the
ceelom (true nephrostome ; compare Rhynchelmis).' I have
also described in the Lycoridea (41) a ciliated region of the
ccelomic epithelium which I believed to be the peritoneal:
funnel. It is, however, to Hisig that we owe the description of
what appear to be intermediate stages. In Dasybranchus and
Tremomastus we have conditions in which the peritoneal funnels
(Genitalschlauche) are separate from the nephridia and open
independently (fig. 14), and in which the two organs are con-
nected but still open separately (fig. 13). Perhaps in certain
segments of these forms and of Capitella we have the more usual
Polychete arrangement, in which the peritoneal funnel no
longer acquires an independent opening (81).

ARTHROPODA.

It will be best to begin our review of this group with a brief
recapitulation of the development of Peripatus, which has been
so excellently described by Mr. Sedgwick (94), and Dr. von
Kennel (64). Soon after the metameric somites have been
hollowed out to form the celomic follicles, the upper half of
each coelomic cavity becomes nipped off from the lower half.
From the wall of each of these lower ccelomic sacs a peritoneal

1 Such would appear to be the condition in Protodrilus, where the nephri-
dial funnels figured by Hatschek (52) are small, and provided with a
flagellum.

ON THE C(@LOM, GENITAL DUCTS, AND NEPHRIDIA. 497

funnel is formed as an outgrowth, which fuses with the epidermis
(figs. 15 and 16). V. Kennel maintains that in Peripatus
Edwardsii the mesoblastic funnel is met by an epiblastic
invagination, and the question arises as to whether this
invagination represents a true nephridium, in which case the
segmental organ of Peripatus would be a compound organ
similar to that of Psygmobranchus (see above), or whether it
is merely a secondary invagination of the epidermis, such as
occurs more or less pronounced in almost every case where a
tube opens on to its surface (vesicle of the nephridia in the
Hirudinea, peripheral end of the genital ducts of the Oligo-
cheta, &c.). I am inclined to take the latter view, and con-
sider the segmental organs of Peripatus as purely peritoneal
funnels which have assumed the excretory functions. Whilst
these organs have developed in this way, the dorsal or genital
halves of the somites in the posterior segments have become
fused, forming two genital tubes communicating posteriorly
with the undivided ccelomic follicles of the last segment. The
peritoneal funnels of this segment retain their primitive func-
tion, and develop into the genital ducts (fig. 17). The peri-
pheral ends and median portion of the ducts are probably
derived from the epidermis.

The history of the genital and excretory organs of the other
groups of Arthropoda can easily be brought into agreement
with the development of these parts in Peripatus. However,
whereas in the latter all the celomic follicles give off peritoneal
funnels, in the Crustacea, Arachnida, Myriapoda, and Hexa-
poda the peritoneal funnels are only fully developed in a very
few segments. The shell glands and green glands of the
Crustacea have been shown to bear the relations of peritoneal
funnels (Grobben, 45; Weldon, 104; Marchall, 77; Allen, 1)
and develop from the ceelomic follicles. ‘ Bei Daphnia,” says
Lebedinsky (78), “entwickelt sich . . . die Schalendriise als die
Ausstiilpung der Somatopleura, welche sich zur Max.’ richtet
und hier sich mit dem Ectoderm vereinigt.” The same author
describes the development of the coxal gland of Phalangium as
a typical peritoneal funnel derived from a ccelomic follicle (73),

498 EDWIN 8. GOODRICH.

which description agrees with that of Laurie (72) of the origin
of the coxal gland in Scorpio, and of Kingsley (65) in Limulus.

The genital cells are derived from the walls of the ccelomic
follicles! of many segments (Heymons, 57; Wheeler, 105).
The dorsal portion of these follicles generally fuse to form con-
tinuous tubes, or a median genital sac (as in the Crustacea).
The ducts are the peritoneal funnels of one segment. The
particular segment selected, so to speak, for this purpose varies
much in position in different groups, and also according to sex.
Wheeler, in his admirable account of the development of the
Orthoptera (105), shows that the ccelomic follicles of all the
abdominal segments at an early stage begin to develop peri-
toneal funnels, but that those of one segment only reach the
exterior and form the genital ducts (fig. 30).

As far as we can see, therefore, there are no certain traces of
true nephridia in the Arthropoda. The segmental organs, the
green glands, the shell glands, and the genital ducts are all
developed as peritoneal funnels.”

SIPUNCULUS.

The development of the excretory organ of Sipunculus
nudus has been described by Hatschek (53). The nephridium
arises from a large cell (? “ funnel-cell” ) which comes to hein
the wall of the ccelomic follicle. This cell divides, forming a
chain of cells in which a lumen is developed. The outer end
joins and opens on to the epidermis; the inner end grows
towards and opens into the celom. The ciliated funnel is
formed from the cclomic epithelium. From this it would
appear that the excretory organ of Sipunculus is of a double
origin, formed by the junction of the nephridium with the
peritoneal funnel, as in most Polychetes. It functions as the
carrier of both genital and excretory products.

1 Sedgwick holds that they are, in Peripatus capensis, derived from the
hypoblast. If this be the case, it must be considered as due to some secondary
modification.

2 The possibility of the segmental trachez of the Arthropods being derived
from the true nephridia should not be lost sight of. ‘The trachee arise com-
paratively late, as a rule, as invaginations of the epidermis, and it seems not

ON THE C@LOM, GENITAL DUCTS, AND NEPHRIDIA. 499

PHORONIS.

In the larva of Phoronis, Caldwell describes a pair of nephri-
dia, slender ciliated canals blind internally (21). Their origin
is still doubtful. The ccolom is developed as two pairs of
follicles, of which the larger “ posterior” pair is alone fertile.
The excretory organs of the adult consist in Phoronis
psammophila and Ph. Kowalevsk1ii of a pair of peritoneal
funnels leading to the exterior from the ‘ posterior” ccelomic
follicles (Cori, 23). According to Caldwell (22), they are
developed in connection with the nephridia of the larva,
and would appear to be of a compound nature like those of
Polychetes. In Phoronis australis both the anterior
and posterior pairs of celomic follicles are provided with
their peritoneal funnels, which open by a common duct
(Benham, 5). These funnels in Phoronis serve both as renal
and genital ducts.

Ectorrocra.

No true nephridia appear to have been found in these
Polyzoa. On the other hand there are two peritoneal funnels,
an excellent account of which has been given by Cori in
Cristatella (24). They differ from those of Phoronis only in
that they open by a common median pore.

BRacHIOPODA.

Here the ceelomic follicles, formed as paired archenteric
pouches, are provided with a pair of wide-mouthed peritoneal
funnels opening to the exterior. They function as the carriers
both of excretory and of genital products, and possibly the
distal end of the organ represents the true nephridium (also in
the Ectoprocta), which has fused with the peritoneal funnel
(Morse, 83; Blochmann, 11).

impossible that they may be formed not from the nephridia, but from those
late epidermal invaginations which, as we have seen, so generally occur in
connection with the external opening of the peritoneal funnels. In the Ar-
thropods above mentioned these funnels disappear in those segments which
possess trachez.

500 EDWIN S. GOODRICH.

SAGITTA.

Archenteric diverticula give rise to the three pairs of coelomic
follicles present in the adult Sagitta. The genital cells are
precociously developed, and come to lie in the two posterior
pairs of follicles (Hertwig, 56). Whether the genital ducts—
which in the male segment, at all events, open into the ceelom
by ciliated funnels—are peritoneal funnels, or are partly formed
from true nephridia, cannot be decided with our present in-
complete knowledge of their development.

EcuHINODERMA.!

Only a very brief reference can be made to this highly modi-
fied group. In the larva we find a right and left coelomic
follicle, the enterocceels (derived from unpaired or paired
archenteric diverticula), which may give rise to a second pair
of cceelomic follicles by constriction. The anterior follicles
then develop peritoneal ciliated funnels (fig. 18), which
open to the exterior, fusing with paired epiblastic invaginations.
As arule only the left peritoneal funnel becomes developed
(Field, 36; Bury, 20). The genital cells are developed from the
wall of the posterior coelomic follicles (MacBride, 75, 76); but
how far the genital ducts can be likened to peritoneal funnels
is quite uncertain.

There appears to be no trace of true nephridia.

VERTEBRATA.!

As is well known, the coelom in the Vertebrates arises by the
hollowing out of a series of metameric blocks of mesoblast.
In the lower forms (Balanoglossus and Amphioxus) several of
the anterior celomic follicles are formed directly as pouches
from the wall of the archenteron. From these follicles, pro-
duced by either method, peritoneal funnels are developed

1 Tf the treatment of these last two groups (Echinoderma and Vertebrata)
seems to be somewhat too brief and dogmatic, it is that space will not allow
me to discuss the subject in extenso. Moreover, we are here treading on
such uncertain ground that I do not feel competent to treat of the structure
of these animals in full detail, and offer these remarks merely as a suggestion.

ON THE C@LOM, GENITAL DUCTS, AND NEPHRIDIA. 501

which communicate with the exterior. The gonads develop
from the wall of the follicles; in the lower forms a large
number are fertile, but in the higher forms the genital products
become restricted to a very limited region.

No true nephridia have been discovered in this group ; for,
until the development of the interesting segmental tubules
described by Weiss (103) and Boveri (14) in Amphioxus is known,
it is not possible to decide for certain on their nature.

Hemichorda or Enteropneusta.—lIn the anterior or pro-
boscis region is a coelomic cavity which appears to represent a
fused pair of follicles (fig. 23), as is evidenced by the fact that
it communicates with the exterior by paired peritoneal funnels
(ciliated proboscis pores) in Balanoglossus Kupfferi and
B, canadensis, and occasionally in Ptychodera minuta and
B. Kowalevskii (Spengel, 96 and 97), and by its development
as a bilobed sac in B. Kowalevskii (Bateson, 2).!_ The collar
region contains a second pair of ccelomic follicles, also provided
each with a peritoneal funnel or collar pore (Bateson, 2; Morgan,
82; Spengel, 97). Behind these is a third pair of follicles
which do not develop funnels, but become of large size and
encroach on neighbouring segments, sending back blind poste-
rior prolongations. The posterior region contains a series of
fertile ccelomic follicles, the gonads (fig. 23), each provided with
a peritoneal funnel leading to the exterior.” Although the
metamerism of this region is not very definitely pronounced,
possibly having become obscured through degeneration, the
genital follicles bear a remarkable resemblance, both in their
structural relations and in their arrangement, to those of the
Nemertines, as has already been noticed by Schimkewitsch (93
aud 98a). As in the Nemertines, so also in the Enteropneusta,
several genital follicles may occur in one segment.

1 Compare the development of the corresponding anterior pair of cavities in
Amphioxus (Hatschek, 49).

? If this explanation be correct, the gill-slits of Balanoglossus are meta-
meric and intersegmental as in other Vertebrates. If not correct, we have an
almost incredible state of things in which the genital cells are shed into a
cavity which is not the ccelom.

502 EDWIN S. GOODRICH.

Craniata.—Owing mainly to the recent researches of van
Wijhe (99), Riickert (89 and 90), Semon (95), and others, it
is now generally concluded that the Craniates originally pos-
sessed a metameric series of pronephric tubules, or peritoneal
funnels, opening independently to the exterior. In existing
forms the funnels, which grow out from the ccelomic follicles,
sometimes reach the epiblast as solid rods (fig. 19) ; they then
fuse with each other to a common duct which opens into the
cloaca (fig. 31), constituting what is known as the pronephros
and its pronephric duct. In his excellent paper on the de-
velopment of the renal system of Ichthyophis, speaking with
regard to E. Meyer’s theory, Semon says: ‘‘ Die segmentalen
Ausfiihrgange der segmentalen Genitalfollikel oder Ursegmente
iibernehmen neben ihrer urspriinglichen auch noch exkre-
torische Funktion, sie werden zu Vornierkanalchen ”’ (95).

To conclude our general survey it may be said: that the
ceelom can be traced from its smallest beginning as a cavity or
cavities in which are developed the gonad-cells: it grows
gradually in size and importance until it becomes the body
cavity in which the viscera rest ; that the genital ducts (with
a few possible exceptions due to secondary modifications) are
homologous throughout the Celomata; that the nephridia,
which have often been confused with these ducts, can always,
when they occur, be distinguished from them ; and finally, that
the coelom may secondarily acquire a renal function, in conse-
quence of which the peritoneal funnels supersede the nephridia
proper as excretory ducts.

List or WORKS REFERRED TO.

1, Auten, EK. J.—‘‘Nephridia and Body-cavity of some Decapod Crustacea,”
‘Quart. Journ. Mier. Sci.,? N. 8., vol. xxxiv, 1892-93.

2. Bateson, W.—‘‘ The Early Stages in the Development of Balano-
glossus Kowalevski,” ‘Quart. Journ. Mier. Sci.,’ N. S., vol. xxiv,
1884. ‘The Later Stages,” &c., id., vol. xxv, suppl. 1885, and vol.
xxvi, 1886.

_ 8. Bareson, W.—“ The Ancestry of the Chordata,” ‘Quart. Journ, Mier.

Sci.,’ vol. xxvi, 1886.

ON THE C@LOM, GENITAL DUCTS, AND NEPHRIDIA. 5038

. Bepparp, F. E.—“ On Certain Points in the Development of Acantho-

drilus multiporus,” ‘Quart. Journ. Mier. Sci.,’ N.S., vol. xxxiii,
1892.

. Benuam, W. B. B.—‘‘ The Anatomy of Phoronis australis,” ‘Quart.

Journ. Micr. Sci.,’ N. S., vol. xxx, 1889-90.

. Bereu, R. §.— Die Exkretionsorgane der Wirmer Kosmos,” Bad. ii,

1885.

. Bereu, R. $.—‘* Die metamorphose von Aulastoma gulo,” ‘ Arb. Zool.

Inst. Wirzburg,’ Bd. vii, 1885.

. Bereu, R. 8.—“ Untersuchung tiber den Bau und die Entwicklung der

Geschlechtsorgane der Regenwiirmer,” ‘ Zeit. f. Wiss. Zool.,’ Bd. xliv,
1886.

. Brereu, R. 8.—* Zur Bildungsgeschichte der Exkretionsorgane bei Crio-

drilus,” ‘ Arb. Zool. Inst. Wirzburg,’ Bd. viii, 1888.

. Bereu, R. $.— Neue Beitrage zur Embryologie der Anneliden,” ‘ Zeit.

f. wiss. Zool.,’? Bd. 1, 1890.

. Brocumany, F.—‘ Untersuchungen iiber den Bau der Brachiopoden,’

Jena, 1892.

. Boumie, L.—“ Zur feineren Anatomie von Rhodope Veranii (Kdlliker),”

‘Zeit. f. wiss. Zool.,’ Bd. lvi, 1893.

12a. Bournz, A. G.—“ Contributions to the Anatomy of the Hirudinea,”

13.

, 14.

15.

16.

Li:

18.

19.

20.

21.

‘Quart. Journ. Mier. Sci.,’ vol. xxiv, 1884.

Bourne, A. G.—‘ On Certain Points in the Development and Anatomy
of some Karthworms,” ‘Quart. Journ. Mier. Sci.,’ N.S., vol. xxxvi,
1894.

Boveri, Th.—“ Die Nierenkanalchen des Amphioxus,” ‘Zool. Jahrb.,’
Bd. v, 1892.

Bircer, O.— Untersuchungen iiber die Anatomie und Histologie der
Nemertinen,” ‘ Zeit. f. wiss. Zool.,’ Bd. 1, 1890.

Bireer, O.—“ Beitrage zur Entwicklungsgeschichte der Hirudineen,”
‘Zool. Jahrb.,’ Bd. iv, 1891.

Bircer, O.— Die Enden des exkretorischen Apparates bei den Ne-
mertinen,” ‘ Zeit. f. wiss. Zool.,’ Bd. liii, 1892.

Bircer, O.—“ Studien z. e. Revision der Entwicklungsgeschichte der
Nemertinen,” ‘ Ber. Nat. Gess. Freiburg,’ Bd. viii, 1894. ;

Burezr, O.—“ Neue Beitrage zur Mntwicklungsgeschichte der Hiru-
dineen,”’ ‘ Zeit. f. wiss. Zool.,’ Bd. lviii, 1894.

Bory, H.— Studies in the Embryology of the Echinoderms,” ‘ Quart.
Journ. Mier. Sci.,’ N.8., vol. xxix, 1888-89.

CaLpwELL, W. H.— Preliminary Note on the Structure, Development,
and Affinities of Phoronis,”’ ‘ Proc. Roy. Soc.,’ vol. xxxiv, 1883.

vou. 37, PART 4,—NEW SER, i

504 EDWIN §. GOODRICH.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.
36.

37.

38.

39.

40.

41.

42.
43.

CaLpweLt, W. H.— Blastopore, Mesoderm, and Metameric Segmenta-
tion,” ‘Quart. Journ. Micr. Sci.,’ N. 8., vol. xxv, 1885.

Cort, C. J.—‘‘ Untersuchungen tiber die Anatomie und Histologie der
Gattung Phoronis,” ‘ Zeit. f. wiss. Zool.,’ Bd. li, 1891.

Cort, C. J.—“ Die nephridien der Cristatella,” ‘ Zeit. f. wiss. Zool.,’ Bd.
lv, 1893.

Cunnincuam, J. T.—‘Some Points in the Anatomy of Polycheta,”
‘Quart. Journ. Mier. Sci.,’ N. 8., vol. xxviii, 1887.

Davenport, C. B.—“On Urnatella gracilis,” ‘Bull. Mus. Comp.
Anat., Harvard Coll.,’ vol. xxiv, 1893.

DrascuE, R. von.—“ Beitrage zur Entwicklung der Polychaten,” Wien,
1884-5.

Ecxstrin, K.—‘ Die Rotatorien der Umgegend von Giessen,” ‘ Zeit. f.
wiss. Zool.,’ Bd. xxxix, 1883.

Enters, E.—<Zur Kenntniss der Pedicellineen,” ‘Abh. Kgl. Gesell-
schaft d. Wiss. Géttingen,’ Bd. xxxvi, 1890.

E1sic, H.—‘‘ Die Segmentalorgane der Capitelliden,” ‘ Mitth. Zool. Sta.
Neapel,’ Bd. i, 1879.

Este, H.—“ Die Capitelliden,” ‘ Fauna und Flora des Glofes von Neapel,’
xvi, 1887.

Exuancer, R. von.—“ Zur Entwicklung der Paludina vivipara,” 1
und ii Theil, ‘Morph. Jahrb.,’ Bd. xvii, 1891.

Eriancer, R. von.—“ Beitrage zur Entwicklungsgeschichte der Gas-
tropoden,” i Theil, ‘ Mitth. Zool. Sta. Neapel,’ Bd. x, 1892.

Ertancer, R. von.—“ Bemerkungen zur Embryologie der Gastropoden,” i,
‘ Biol. Centralbl.,’ Bd. xiii, 1893.

Er.ancer, R. von.—Ib. ii. ‘Biol. Centralbl.,’ Bd. xiv, 1894.

Fietp, G. W.—‘“The Larva of Asterias vulgaris,” ‘Quart. Journ.
Mier. Sci.,’ N. S., vol. xxxiv, 1892-93.

Forrtincrr, A.—‘‘ Recherches sur l’organisation de Histriobdella
homari,” ‘ Arch. de Biol.,’ t. v, 1884.

For, H.— Développement des Gastropodes pulmonés,” ‘Arch. Zool.
Exper.,’ vol. vili, 1879-80.

F¥rRatpont, J.— Recherches sur l’appareil excréteur des Trématodes et
des Cestoides,” ‘Arch. de Biol.,’ t. i, 1880.

Fratront, J.—“ Le Genre Polygordius,’’ ‘Fauna und Flora des Golfes
von Neapel,’ xxiv, 1887.

Goopricn, B. §.—“ On a New Organ in the Lycoridea,” ‘ Quart. Journ.
Micr. Sci.,’ N. S., vol. xxxiv, 1892-93.

Grarr, L. von.—“ Monographie der Turbellarien,” Leipzig, 1882.

Grarr, L. von.—“ Ueber Rhodope Veranii Koell,” ‘ Morph. Jahrb.,’ Bd.
viii, 1883.

44.

ON THE C&LOM, GENITAL DUCTS, AND NEPHRIDIA. 505

Grarr, L. von.—‘ Die Organisation der Turbellaria acoela,’ Leipzig,
1891.

45. GrosBen, C.—‘‘ Die Antennendriise der Crustaceen,” ‘Arb. Zool. Inst.

46.

Wien,’ Bd. iii, 1880-81.
Harmer, 8S. F.—‘‘ Notes on the Anatomy of Dinophilus,” ‘ Journ. Mar.
Biol. Ass.,’ N.8., vol. i, 1889-90.

46a. Harmer, S. F.—‘ On the Structure and Development of Loxosoma,”

47.

48.

49.

50.

51.

52.
53.

54.

55.
56.

57.

58.

59.

60.

61.

62.

63.

64.

‘Quart. Journ. Micr. Sci.,’? N. 8., vol. xxv, 1885.

HatscHek, B.—“‘ Embryonalentwicklung und Knospung der Pedicel-
lina echinata,” ‘ Zeit. f. wiss. Zool.,’ Bd. xxix, 1877.

Hatscuex, B.—‘‘ Studien tiber Entwicklungsgeschichte der Anneliden,”
‘Arb. Zool. Inst. Wien,’ Bd. i, 1878.

Hatscuex, B.—‘ Studien tiber Entwicklung des Amphioxus,” ‘ Arb.
Zool. Inst. Wien,’ Bd. iv, 188].

Hartscuex, B.—“ Ueber Entwicklungsgeschichte von Teredo,” ib., Bd.
ii, 1881.

Hatscuex, B.—“‘ Ueber Entwicklungsgeschichte von Echiurus,” ib.,
Bd. iii, 1881.

Hatscuex, B.—“Protodrilus Leuckartii,” ib., Bd. iii, 1881.

Hatscuex, B.—‘‘ Ueber Entwicklung von Sipunculus nudus,” ib.,
Bd. v, 1884.

Hatscuek, B.—“ Entwicklung der Trochophora von Eupomatus,”’ ib.,
Bd. vi, 1886.

HatscuEex, B.—‘ Lehrbuch der Zoologie,’ Jena, 1888.

Hertwic, O.—‘ Die Chatognathen,” ‘Jen. Zeit. f. Naturw.,’ Bd. xiv,
1880.

Heymons, R.—“ Die Entwicklung der weiblichen Geschlechtsorgane von
Phyllodromia (Blatta) germanica,” ‘ Zeit. f. wiss. Zool.,’ Bd. liii,
1892.

Husrecut, A. A. W.—‘ Contributions to the Embryology of the Ne-
mertea,” § Quart. Journ. Micr. Sci.,’ N. S., vol. xxvi, 1886.

Husrecut, A. A. W.—“ Report on the Nemertea,” ‘‘“ Challenger”
Reports,’ vol. xix, 1887.

Jisima, J.—‘‘ Untersuchungen iiber den Bau und die Entwicklungs-
geschichte der Siisswasserdendrocélen,” ‘ Zeit. f. wiss. Zool.,’ Bd. Ix,
1884.

Jouiet, L.—‘‘ Organe segmentaire des Bryozoaires entoproctes,”
‘Arch. Zool. Expér.,’ t. viii, 1879-80.

Jourpatn, 8S.—“ Sur les organes segmentaires, etc., des Limaciens,”
*C. R. Acad. Paris,’ t. xeviii, 1884.

KeEnneEL, J. von.—* Beitrage zur Kenntniss der Nemertinen,” ‘Arb,
Zool. Inst. Wirzburg,’ Bd. iv, 1877-78.

KENNEL, J. von.— Entwicklungsgeschichte von Peripatus Ed-

506

65.
66.
67.
68.
69.
70.

ah
72.

73.

74.

75.

76.

77.

78.
79.
80.

81.

82.

83.

EDWIN 8. GOODRICH.

wardsii und P. torquatus,” ‘Arb. Zool. Inst. Wiirzburg,’ Bde. vii
and viii, 1885-86.

Kines ey, J. S.—“ The Embryolog y of Limulus,” ‘Journ. of Morph.,’
vol. viii, 1893.

Korscuett, E.—‘ Ueber Bau und Entwicklung des Dinophilus
apatris,” ‘Zeit. f. wiss Zool.,’ Bd. xxxvii, 1882.

KorscHett, E., und K. Herper.—‘‘ Lehrbuch der vergl. Entwicklungs-
geschichte,” ‘Spec. Theil.,’ Jena, 1893.

Kowatevsky, A.—‘‘ Embryologische Studien an Wiirmern und Arthro-
poden,” ‘ Mém. Acad. St. Pétersb.,’ 7 s., t. xvi, 1871.

Lane, A.—“ Der Bau von Gunda segmentata,” ‘Mitth. Zool. Sta.
Neapel,’ Bd. iui, 1882.

Lane, A.—* Die Polycladen,” ‘ Fauna und Flora des Golfes von Neapel,’
xi, 1884.

Lane, A.—‘ Lehrbuch der Verg]. Anatomie,’ Jena, 1888.

Lauriz, M.—“ The Embryology of a Scorpion,” ‘Quart. Journ. Mier.
Sci.’ N.S., vol. xxxi, 1890.

LeBepinsky, J.—‘ Die Entwicklung der Coxaldriise bei Phalangium,”
‘Zool. Anz.,’ Bd. xv, 1892.

Lrnmann, O.—“ Beitrage zur Frage von der homologie der segmental-
organe und Ausfiihrgange der Geschlechtsproducte bei den Oligo-
chaeten,” ‘Jen. Zeit. f. Naturw.,’ Bd. xxi, 1887.

MacBript, HE. W.—‘‘ The Development of the Genital Organs, &c., in
Amphiura squamata,” ‘Quart. Journ. Micr. Sci.,’ vol. xxxiv,
1892-93.

MacBripz, EK. W.—“ The Organogeny of Asterina gibbosa,” ‘ Proc.
Roy. Soc.,’ vol. liv, 1893.

Marcuau, P.—*‘ Recherches anatomiques et physiologiques sur |’appareil
excréteur des Crustacés Décapodes,” ‘Arch. Zool. Exp.’ (2), vol. x,
1892.

Manion, A. F., et N. Bopretzky.—“ Etude des Annélides du Golfe de
Marseille,” ‘ Ann. Sc. Nat. Zool.’ (6), vol. ii, 1875.

Meuron, P. pE.—‘ Sur les organes rénaux des embryons d’Helix,’
*C. R. Acad. Paris,’ t. xeviii, 1884.

Meyer, E.—‘ Studien iiber den Korperbau der Anneliden,” parts 1 and
2, ‘ Mitth. Zool. Sta. Neapel.,’ Bd. vii and viii, 1887-89.

Meyer, E.— Die Abstammung der Anneliden,” ‘ Biol. Centralblatt,’
Bd. x, 1890. Translation in the ‘American Naturalist,’ vol. xxiv,
1890.

Morean, T. H.—* The Development of Balanoglossus,” ‘Journ. of
Morph.,’ vol. ix, 1894.

Morse, E. 8.—‘‘On the Oviducts and Embryology of Terebratulina,’
‘Am. Journ. Sci. and Arts’ (3), vol. iv, 1872,

ON THE C@HLOM, GENITAL DUCTS, AND NEPHRIDIA. 507

84. Nusspaum, J.—‘Zur Entwicklungsgeschichte der Geschlechtsorgane
der Hirudineen,” ‘ Zool. Anz.,’ vol. viii, 1885.

85. Oupemans, A. C.—‘ The Circulatory and Nephridial Apparatus of the
Nemertea,” ‘ Quart. Journ. Micr. Sci.,’ N. S., vol. xxv, 1885.

86. PetsenerR, P.—‘ Contributions 4 l’étude des Lamellibranches,” ‘ Arch.
Biologie,’ t. xi, 1891.

87. Puatre, L.—Beitrige zur Naturgeschichte der Rotatorien,” ‘Jen.
Zeitschr. f. Naturw.,’ Bd. xix, 1886.

88. Prarre, L.—“ Ueber die Rotatorienfauna des bottnischen Meerbusens,”
‘ Zeit. f. wiss. Zool.,’? Bd. xlix, 1889.

88a. Rast, L.—‘ ‘Ueber die Entwicklung der Tellerschnecke,” ‘Morph.
Jahbrb.,’ Bd. 5, 1879.

89. Rickert, J.—‘‘ Ueber der Enstehung der Exkretionsorgane bei Se-
lachiern,” ‘Arch. f. Anat. u. Phys. (Anat. Abth.),’ 1888.

90. Ruckert, J.—‘ Entwicklung der Exkretionsorgane,” ‘ Hrgeb. Anat. u.
Entwickl.’ (Merkel und Bonnet), Bd. 1, 1891.

91. Satensxy, W.—“ Etudes sur le développement des Annélides,” ‘ Arch.
Biologie,’ vol. iii, 1882; vol. iv, 1883.

92. Sarasin, P. u. F.—‘ Ergebn. Naturw. Forsch. auf Ceylon,’ Bd. i, Wies-
baden, 1888.

93. ScnimKewitscu, W.—‘ Ueber Balanoglossus Mereschk.,” ‘ Zool. Anz.,’
vol. xi, 1888.

98a. ScHimKEWwitscH, W.—‘‘ Ueber die Morphologische Bedeutung der
Organsysteme der Enteropneusten,” ‘ Anat. Anz.,’ Bd. v, 1890.

94. Sepewick, A.—* The Development of the Cape Species of Peripatus,”
‘Quart. Journ. Mier. Sci.,’ N. S., vols. xxv—xxviii, 1885-88.

95. Semon, R.—‘‘Studien iiber den Bauplan des Urogenitalsystems der
Wirbelthiere,” ‘Jen. Zeit. f. Naturw.,’ vol. xxvi, 1891.

96. SpencEL, J. W.—“ Zur Anatomie des Balanoglossus,” ‘ Mitth. Zool. Sta.
Neapel,’ Bd. v, 1884.

97. SpenceL, J. W.—“ Die Enteropneusten,” ‘Fauna u. Flora des Golfes
von Neapel,’ xviii, 1893.

98. TrautzscH, H.— Beitrag zur Kenntniss der Polynoiden von Spitz-
bergen,” ‘Jen. Zeit. f. Naturw.,’ Bd. xxiv, 1889-90.

99. Van Wine, J. W.—“ Ueber die Mesodermsegmente des Rumpfes und
die Entwicklung des Exkretionssystems bei Selachiern,” ‘Arch. f.
mikr. Anat.,’ Bd. xxxiii, 1888.

100. Vrespovsxy, F.—‘System und Morphologie der Oligochaeten,’ Prag.,
1884.

101. Vespovsxy, F.—‘ Entwicklungsgeschichtliche Untersuchungen,’ Prag.,
1888-92.

102. Vespovsky, F.—‘‘ Zur Entwicklungsgeschichte des Nephridial-Appa-
rates von Megascolides australis,” ‘Arch. mikr. Anat.,’ Bd. xl,
1892.

508

103.

104.

105.

106.

107.

108.

109.

110.
111.

112.

118.

114.

115.

116.

EDWIN S. GOODRICH.

Weiss, F. E.—‘ Excretory Tubules in Amphioxus lanceolatus,”
* Quart. Journ. Micr. Sci.,’ N.S., vol. xxxi, 1890.

We pon, W. F. R.— Colom and Nephridia of Palemon serratus,”
‘Journ. Mar. Biol. Ass.,’ vol. i, 1889.

Wueeter, W. M.—‘ A Contribution to Insect Embryology,’ ‘Journ.
of Morph.,’ vol. viii, 1893.

Wuirman, C. O.—“ The Embryology of Clepsine,” ‘Quart. Journ.
Mier. Sci.,’ N. S., vol. xviii, 1878.

Wuirtman, C. O.—“ A Contribution to the History of the Germ-layers
in Clepsine,” ‘Journ. of Morph.,’ vol. i; 1887.

Witson, KE. B.— The Germ-bands of Lumbricus,”’ ‘ Journ. of Morph.,’
vol. i, 1887.

Witson, E. B.—‘ The Embryology of the Earthworm,” ib., vol. iii,
1889.

Witson, HE. B.—* The Cell-lineage of Nereis,” ib., vol. vi, 1892.

Wotrson, W.—“ Die Embryonale Entwicklung des Lymnaeus stag-
nalis,” ‘ Bull. Ac. Imp. Se. St. Petersb.,’ vol. xxvi, 1880.

Woopwortu, W. M.—“ On the Structure of Phagocata gracilis,”
Leidy, ‘ Bull. Mus. Comp. Anat. Harvard,’ vol. xxi, 1891.

Woopwarp, M. F.—“ Description of an Abnormal Earthworm possess-
ing Seven Pairs of Ovaries,” ‘ Proc. Zool. Soc.,’ 1892.

Woopwarp, M. F.—‘“ Further Observations on Variations in the Geni-
talia of British Earthworms,” ib., 1893.

ZELINKA, C.—“ Studien tiber Raderthiere,” ‘Zeit. f. wiss. Zool.,’ Bd.
liii, 1892.

ZIEGLER, H. E.—“ Die Entwicklung von Cyclas cornea,” ‘Zeit. f.
wiss. Zool.,’’ Bd. xli, 1885.

ON THE C@LOM, GENITAL DUCTS, AND NEPHRIDIA. 509

EXPLANATION OF PLATES 44 & 45,

Illustrating Mr. Edwin 8S. Goodrich’s paper on ‘The
Ccelom, Genital Ducts, and Nephridia.”

REFERENCE LETTERS.

c. Coclomic or genital follicle. ep. 7. Epidermal invagination. xeph. True
nephridium. p./. Peritoneal funnel developed from the celomic epithelium.

In all the diagrams the outline of the ccelom, the peritoneal funnels, and the
genital cells, are drawn in red. The epiblast and hypoblast are darkly shaded.
The mesoblast is lightly shaded. The true nephridia are drawn in black.

Fic. 1.—Diagrammatic plan of the cclom, genital ducts, and nephridia in
the Planaria.

Fic. 2.—Similar plan of a young Nemertine.

Fic. 3.—Plan of a mature Nemertine.

Fie. 4.—Plan of the celom, &c., in the Entoprocta.

Fig. 5.—Plan of the ccelom and nephridium in an embryo Oligochete.

Fic. 6.—Later stage of the same, showing the nephridia in the pronephri-
dial condition.

Fic. 7.—Plan of the ccelom, &c., in an adult Oligochete (Lumbricus).

Fic. 8.—Plan of the ccelom and nephridia in an embryo leech.

Fic. 9.—Plan of the ccelom, &c., in the Hirudinea.

Fig. 10.—Plan of the ccelom and nephridia in an embryo Polychete.

Fic. 11.—Later stage of the same, showing the pronephridium about to
fuse with the peritoneal funnel.

Fic. 12.—Plan of the ceelom, &c., in the Polycheta possessing wide-mouthed
“compound nephridia,” the nephridium having fused with the peritoneal
funnel.

Fic. 13.—Plan of the celom, &c., in certain segments of Tremomastus and
Dasybranchus in which the nephridial funnel is continuous with the genital or
peritoneal funnel.

Vic. 14.—Plan of the celom, &c., in certain segments of Dasybranchus
caducus, in which the nephridial funnels are independent of the peritoneal
funnels.

Fic. 15.—Plan of the ccelom and developing peritoneal funnels in an embryo
Peripatus.

Fic. 16.—Plan of the ccelom and peritoneal funnels in a trunk segment of
Peripatus. The ccelomic follicles on either side have become separated into a
ventral and a dorsal portion, forming the genital tube.

510 EDWIN 8S. GOODRICH.

Fic. 17.—Plan of the ccelom and peritoneal funnels in the posterior seg-
ment of Peripatus.

Fic. 18.—Plan of the ccelomic follicles (enterocels) and developing peri-
toneal funnels in an embryo Echinoderm (Asterias).

Ftc. 19.—Plan of a stage in the development of the ccelom and peritoneal
funnels (pronephric tubules) of an Elasmobranch.

Fic. 20.—Plan of the ccelom and nephridia (pronephridial stage) in an
embryo Mollusc.

Fie. 21.—Plan of the celom and developing peritoneal funnels (excretory
organs of the adult) of a Mollusc.

Fic. 22.—Farther stage of the same, after the peritoneal funnel has fused
with the epidermal invagination.

Fie. 23.—Plan of a longitudinal section of Balanoglossus, showing the
ccelomic follicles and peritoneal funnels.

Fic. 24.—Plan of a longitudinal section of a Nemertine, showing the
celomic follicles and peritoneal funnels before the latter have reached the epi-
dermis (the nephridia are represented in the same region as tie genital follicles).

Fic. 25.—Plan of a longitudinal section of the nephridia, ceelom, and de-
veloping peritoneal funnels in a larval Oligochete.

Fic. 26.—Similar plan of a larval Polychete.

Fic. 27.—Plan of a longitudinal section of the ccelom and peritoneal funnels
in an embryo Vertebrate.

Fic. 28.—Plan of a longitudinal section of the ccelom, peritoneal funnels,
and nephridia of a Leech.

Fic. 29.—Plan of a longitudinal section of the celom, &c., of an Oli-
gochete.

Fic. 30.—Plan of a longitudinal section of the ccelom and developing peri-
toneal funnels of an embryo Insect.

Fic. 31.—Plan of ccelom and peritoneal funnels (pronephric tubules) of a

Vertebrate.

EN DEXVLO V Orbsin 37s

NEW SERIES.

Alcyonium, chemistry of mesoglcea of,
by Langdon Brown, 389

Alcyonium digitatum, anatomy
of, by S. J. Hickson, 343

Assheton, Richard, a series of five
embryological memoirs, 113, 165,
173, 191, 223

Aurelia aurita, variation of the
tentaculocysts of, by E. T. Browne,
245

Balanus, mouth parts of, by Groom,
269

Benhamia cecifera, by Benham,
103

Benham on Benhamia cecifera,
103
* on the cerebral convolutions
of the chimpanzee and ourang, 47

Branchiostomide, a revision of the
genera and species of, by Kirkaldy,
303

Browne on the variation of the
tentaculocysts of Aurelia, 245

Brown, Langdon, on the chemical
constitution of the mesoglea of
Alcyonium digitatum, 389

Calman on Julinia, a new genus of
Ascidians, 1

Cellular theory, inadequacy of, by
Sedgwick, 87

Cerebral convolutions of the Chim-
panzee, “ Sally,” by Benham, 47

Chimpanzee, cerebral convolutious of,
by Benham, 47

Coccidia from mice, by Clarke, 277

Ccelom, genital ducts, and nephridia,
by Goodrich, 477

Clarke on Coccidia of mice, 277
», on Sporozoa, 287

Earthworm, a new species, Ben-
hamia cecifera, by Benham,
103

Frog’s embryo, growth in length of,
by Assheton, 223

Genital ducts, nephridia, and ccelom,
by Goodrich, 477

Goodrich on the cclom, genital ducts,
and nephridia, 477

» on Vermiculus pilosus,

253

Groom on the mouth parts of
Balanus, 269

Hermaphroditism in Mollusca, by
Pelseneer, 19

512

Hickson on the anatomy of Alcyo-
nium digitatum, 343

Julinia, a new genus of Ascidians,
by W. T. Calman, 1

Kirkaldy on the genera and species
of the Branchiostomide, 303

Layers in the embryonic rabbit and
frog, by Assheton, 113

MacBride on Sedgwick’s theory of
the embryonic phase, 325

Mammalian embryo, its attachment
to the walls of the uterus, by
Assheton, 113

Mesoglea of Alecyonium, chemistry
of, by Langdon Brown, 389

Metamerism, a study of, by T. H.
Morgan, 395

Mollusca, hermaphroditism in, by
Pelseneer, 19

INDEX.

Morgan on metamerism, 395

Nephridia, ccelom, and genital ducts,
by Goodrich, 477

Pelseneer on hermaphroditism in
Mollusca, 19

Primitive streak of the rabbit, by
Assheton, 191

Rabbit, early stages of, by Assheton,
113

Sedgwick on the inadequacy of the
cellular theory of development, 87

Sedgwick’s theory of the embryonic
phase, by MacBride, 325

Sporozoa, notes on, by Clarke, 287

Vermiculus pilosus, by Good-
rich, 253

PRINTED BY ADLARD AND SON,
BARTHOLOMEW CLOSE, E.C., AND 20, HANOVER SQUARE, W.

Tesh
ae | phan
aie se
aoe by, ah Aa
is) A

ry

Pe! | | TEL |||] Hl!
Wate 26 275 495

Ay ee a an
ee i eee
a TE alee aii eos

mens
reyes
ee
Sy epee prem

a ei ie ch et NA DA EO”

Sets we ae Pte
b Patierthgt Aen alindveliatin-btnok alla aa atm, ah sat
a ho ee

em a oe - ow
anon eee iki atreenplote ene!
re al al ? Siibetent chethcanen alii ana diadaninand