Studies from the Morphological Laboratory, Cambridge University

4)
te
4
i

STUDIES

FROM THE

MORPHOLOGICAL LABORATORY

IN THE

UNIVERSITY OF CAMBRIDGE.

EDITED BY

ADAM SEDGWICK, M.A., F.RB.S.

FELLOW AND LECTURER OF TRINITY COLLEGE, CAMBRIDGE.

Vol. Ill.

. London :

C. J. CLAY AND SONS,
CAMBRIDGE UNIVERSITY PRESS WAREHOUSE,
AVE MARIA LANE,

1888

CONTENTS.

PAGE

W. Bateson. On the ee ee of the iene Plates
I.—XII. Parts I. and II. : 3 : é 1
W. Batsson. On the Ancestry of the Chordata 5 Of

Watter Heaprg. The Development of the Mole OEE meh
Stages Eto J. Plates XIII, XIV, XV. . ; 105

Sipnny F. Harmer. On the Life History of Pedicellina. Plates
XVI, XVII. é : : : : ‘ : ; : . 147

ArtHuR E. SHipLEy. On some points in the development of
Petromyzon fluviatilis. Plates XVIII, XIX, XX, XXI. . ~ Uwe

The above are reprinted from the Quarterly Journal of Microscopical Science.

On the

Morphology of the Enteropneusta.'

PART I.
By

William Bateson, M.A.,
Fellow of St. John’s College, Cambridge.

With Plates I—VI.

In the ‘ Quart. Journ. Micr. Sci.’ for April, 1884, I described
the early stages in the development of a species of Balano-
glossus, common on the shores of the Chesapeake Bay. I am
led from various reasons, which I hope to detail when an oppor-
tunity occurs for discussing the classification of the Enterop-
neusta, to regard this species as identical with that described
by Alex. Agassiz, and called by him B. Kowalevskii.

I was again enabled, during the summer of 1884, by the
great kindness of my friend Dr. W. K. Brooks, to pursue my
investigations into the morphology of the Enteropneusta. My
warmest thanks are due to Dr. Brooks for once more affording
me the hospitalities of the Chesapeake Zoological Laboratory,
which was this year situated at Beaufort, North Carolina.
While there I collected a number of specimens of a large
undescribed species of Balanoglossus resembling B. salmoneus
(Giard). This species, a description of which will I hope shortly
appear, I propose to call B. Brooksii. My attempts at rearing
this species from the egg were unsuccessful, and I was pre-

1 A note on the subject-matter of this paper appeared in the ‘Proc. Roy.
Soc.,’ No. 235, 1885,

1

2 MORPHOLOGY OF

vented by illness from making any additional observations on
its development.

On leaving Beaufort, N. C., I returned to Hampton, Virginia,
in the beginning of September, in the hope of finding the later
stages in the development of B. Kowalevskii, and was
fortunate enough to procure a complete series of larve from
Stage H to individuals possessing ten gill-slits, in which con-
dition the generative organs are first present. It is intended in
this paper to give a general account of these stages, together
with the histology of the animal, until two pairs of gill-slits are
developed (fig. 1). From this point the further histological
differentiation of the various organs will be described under
separate headings.

With this account of the organogeny of B. Kowalevskii
will be given, as far as possible, a comparative description of the
same parts in all the species which I have hitherto examined.

External Changes.—The larva of B. Kowalevskii was
described (loc. cit.) as having the egg in an elliptical form,
divided by two transverse constrictions into three segments.
The surface of the body was ciliated, a special tuft of long
cilia being developed at the anterior end, while the posterior
region was surrounded by a transverse band of long cilia.

From further observations it seems probable that this period
(Stage D) assigned as the time of hatching is too early; for
embryos kept in aquaria do aot break the membranous shell
before Stage G is reached. Probably, therefore, the larve
found swimming in Stage D had escaped owing to an artificial
rupture of the shell during the process by which they were
found; an account of which is given in an appendix.

The formation of the mouth as a ventral pore in the an-
terior groove was described (No. 8, fig. 41). It opens directly
into the archenteron, which was previously a closed sac, from
which five mesoblastic pouches had been given off, forming
the anterior, middle, and posterior body cavities respectively.

The external relations of the parts then become changed
until finally the larva has the shape shown in No. 3, figs. 16
and 17, In this stage the proboscis is conical and the middle

THE ENTEROPNEUSTA. 3

segment much shortened. In the anterior dorso-lateral region
of the third segment are the openings of the first pair of gill-
slits, which are simple circular pores leading into the
archenteron.

The larva is still opaque, and pale yellowish-brown in colour.
In this condition it remains for about ten days, at the end of
which time the second pair of gill-slits is formed. The body
has become partly transparent, especially in the region of the
proboscis, through the walls of which muscle-fibres are visible.
Other external changes which occur at this period are the loss
of the anterior tuft of cilia and the gradual disappearance of
the posterior ciliated ring. At the commencement of this
period the larve are to be found in Stage G at a depth
of about six to eight inches in the sand, but towards
its close they work their way into the higher strata of mud,
and do not again go down again until the adult condition is
reached.

As the cilia disappear a peculiar organ is formed as a small
papilla, bearing long cilia and mucous glands, situated at the
central part of the posterior surface (fig. 1, sk.). This organ
serves as a sucker, by which the animal can attach itself to
foreign bodies sufficiently firmly to prevent itself being washed
off by a stream of water from a pipette. The anterior surface
of the proboscis is also slightly suctorial, and by thus fixing
itself posteriorly and extending the proboscis it is able to creep
slowly about, somewhat in the manner ofaleech. The appear-
ance of this organ bears some resemblance to the terminal
sucker described by Graaf as occurring in certain Rhabdo-
celes. It subsequently attains a considerable size, and is
traversed by several wrinkles (figs. 38 and 4, sk.). This organ
afterwards entirely disappears, but as to its mode of disappear-
ance I have no certain observations. It would appear to occur
very suddenly at the stage when the animal possesses seven to
eight gill-slits. I have found animals with eight gill-slits
which possess this sucker, and also animals of apparently the
same age without it; hence it may be inferred that it undergoes
a rapid atrophy at this point.

4 MORPHOLOGY OF

Similar suckers! occur as larval organs in Tunicata, Ganoids,
and Amphibia.

With regard to the meaning of the sucker, considering the
time of life at which it appears, it is probably not ancestral in
origin; as the animal is already, from Stage G onwards, a
distinct Balanoglossus in all its characters. It is more rea-
sonable to conjecture that it is of purely developmental
importance, and indeed its use to a larva of this kind is suffi-
ciently obvious; for the creatures inhabit shallow pools on the
sand-flats, being just buried in the mud, from which position
they would be in danger of being washed away by the incom-
ing tide and so be dried up by the great heat of the sun at low
tide. On attaining a larger size the body can be, and always
is, coiled round a spindle of sand and thus is kept in position,
hence the sucker is no more required. In connection with
this sucker I observed that nearly all the animals found in
these pools were such as are provided with similar means of
fixing themselves, which power is probably essential to life in
such a habitat. An account of the histology of this sucker will
be given subsequently.

During the period which elapses between the appearance of
the first and second pair of gill-slits the body gradually
acquires, as was mentioned above, a considerable degree of
transparency. Owing to this fact several points of internal
structure may be observed. ‘This is especially marked in the
case of the alimentary canal, which can now be clearly per-
ceived to consist of three regions—an auterior branchial tract,
a middle digestive portion, and an intestinal section posteriorly.
The digestive section may be at once recognised by the bright
yellow-brown colour of the secretion which it contains. This
fluid is evacuated after a time when the animal is irritated, but
no experiments were made to determine its physiological pro-
perties.

The partial transparency of the body wall permits also an
indistinct view of the curious supporting rod of hypoblast

1 Balfour was of opinion that in these forms they might be an ancestral
feature (Balfour, ‘Comp. Emb.,’ vol. ii, “ Tunicata”).

THE ENTEROPNEUSTA. 5

which projects into the proboscis cavity. For reasons given
later in this paper I propose to speak of this structure as the
notochord (Nch.). The general appearance of the animal at
the time when the second gill-slit appears, is shown in fig. 1.
The animal is drawn as seen from the actual left side and
displays the increased flexion of the body on its ventral
surface.

The second gill-slit is shown as a small circular pore.

It will also be observed that that part of the body which lay
between the two grooves constituting the middle segment of
the body has now assumed an altered shape. At Stage H is
found little more than a circular ridge on the body separating
the proboscis from the trunk, while in fig. 1 (two gill-slits) it
forms a kind of phlange enveloping the base of the proboscis.
This change in shape is due to the operation of several causes,
which are of great importance in interpreting the processes by
which the final form of the adult is reached.

In the first place, after the formation of the mouth as a pore
on the ventral surface, the constriction by which the proboscis
is segmented off becomes deeper and deeper, until at last it is
only attached by the exceedingly slender stalk shown in figs. 2
and 8. As a consequence of this process coupled with a
forward growth of the ventral lip of the collar, the mouth
comes to be directed anteriorly instead of ventrally (cp. figs.
7,46 and 57). By this process the anterior phlange of the
collar acquires the relation shown in fig. 1, et seq. In
addition to these changes a most important structure is first
formed at this time, namely, the cavity, which from the rela-
tions which it afterwards possesses I shall speak of as the
atrial cavity.

In the later conditions of Stage H the body is perceptibly
wider in the region of the body immediately anterior to the
gill-slits than it is behind them. This increase in width, which
is still very slightly marked, is due to a circular thickening
which passes all round the animal, being most developed at the
sides. By the time of the appearance of the second pair of
gill-slits this thickening has considerably increased, and in the

6 MORPHOLOGY OF

contracted condition of the body is a very marked structure
(fig. 2, op.). As development proceeds, this thickening in-
creases, and at the same time grows backwards in the lateral
regions until (five gill-slits) it has covered half the first gill-
slit. The degree of contraction of the body of course alters its
relations, but in the extended condition in adult specimens its
posterior margin is about on a level with the fourth gill-slit,
As then this structure is a process of the body wall which
forms an opercular fold over the foremost gills, it appears
reasonable to institute a comparison between it and the atrial
walls of Amphioxus, &c. It will therefore be alluded to as the
operculum, and the cavity between it and the body wall as
the atrial cavity.

As previously described, the first gill-slit on its appearance
is a simple circular pore. In this condition it remains for
some days, but gradually, on about the tenth day, its form
changes owing to the growth of a process from the dorsal
margin of the pore, which renders the aperture somewhat
kidney-shaped (fig. 1). Whilst this process is continuing the
second gill-slit appears as another circular pore, similar to the
original aperture of the first gill-slit. I was unable to discover
any priority in the appearance of the gill-slit of either side in
particular, but incline to believing that they appear syn-
chronously on the two sides of the body, as previously
mentioned.

After the appearance of the first gill-slit a definite series of
changes is undergone, and these form a definite period in the
history of the development of the animal, which may be con-
sidered as closing with the formation of the second pair of gill-
slits.

At this point it may be well to recapitulate these changes.
The embryonic life terminates with hatching, and the free
larva with one pair of gill-slits escapes, moving about in the
mud by means of cilia.

Before the second pair of gill-slits appear, it has undergone
the following changes :

THE ENTEROPNEUSTA. 7

(1) Disappearance of ciliated band and apical tuft.

(2) Body walls have become semi-transparent.

(3) Animal has changed its habitat, creeping into the upper
layer of mud.

(4) A fixing organ has formed as a ventral posterior sucker.

(5) Owing to growth of collar-fold, and constriction of pro-
boscis-stalk, the mouth comes to be directed forwards.

(6) Differentiation has occurred in the hypoblast cells,
marking out the alimentary canal into three regions, viz.
branchial, digestive, and intestinal.

(7) Notochord is distinctly visible.

(8) Anal perforation present.

(9) Second pair of gill-slits arise.

(10) Opercular fold forms, as a circular thickening.

From this point onwards the principal changes in external
features consist chiefly in an increase in size, and in continual
progress towards transparency. This latter is so marked a
feature that when three pairs of gill-slits are formed the walls
of the body behind the collar, which were originally opaque,
become perfectly clear and glassy, so that the internal struc-
tures are quite distinguishable. The walls of the proboscis
and of the collar never entirely lose their primitive opacity.
An attempt is made in figs. 1, 2, and 3 to show this gradual
transition. It is presumably due to the consumption of the
food particles which in the earlier condition were distributed
almost uniformly among the cells of the body. This great
transparency is not usual among forms that do not lead a free-
swimming existence, and is possibly, so to speak, accidental,
and correlated to the rapid growth which now occurs. Nume-
rous glandular patches and spots are to be seen in the skin,
They are developed chiefly on the proboscis, collar, and sucker,
those on the latter being refractive and differing somewhat in
appearance from the rest.

The increase in size seems to be rapid, but as to the length
of time that is taken in passing through subsequent stages I
have no record, as the specimens were caught from time to
time, and not reared in aquaria.

8 MORPHOLOGY OF

As the body grows, the number of gill-slits increases. They
are always added in pairs behind the last formed. The newest
has a circular orifice, while the growth of the “ valves ”’ (fig. 4)
from the dorsal margins of the anterior ones continues to
modify their shape. From being circular they then become,
first, kidney-shaped, then horseshoe-shaped, and next, by a
diminution in width from before backwards, together with a
great elongation dorso-ventrally, their openings are made
U-shaped. When this condition is attained, the ‘ valve”
continues to grow downwards, its free end lying inside the
pharyngeal cavity, as will be described when the histology of
the gills is treated of. Frequently, in contracted specimens,
these valves are washed outwards through the gill-slit, and
hang freely out in the water. This condition often occurs
during life. The gill apertures are from the first strongly
ciliated. The cilia move in a constant direction, driving a
current dorsalwards on the anterior line of the U, then down
the anterior margin of the “valve” and up the posterior, and
finally ventralwards on the hinder edge of the gill-slit. The
currents have the same course before the formation of the
‘¢valve,” viz. on looking at a circular gill-slit of the left side,
if the animal’s head is directed to the observer’s left, the course
will be round the aperture in the direction of the hands of a
watch. By this current the water which passes in at the
mouth is carried out of the pharynx; probably, therefore, the
motion of the cilia is in a sort of spiral converging outwardly,
and not circular as it appears to be on looking down upon a
gill-slit.

From the fact that the number of gill-slits varies with the
length of the animal, together with the constant presence in
the posterior branchial region of a regularly arranged series of
gills in all stages from a complete U-shaped opening to a
terminal one which is always circular, I am led to believe that
these structures increase in number throughout the greater
part, if not the whole, of the life of the animal. The greatest
number of slits which I have observed was fifty-seven pairs.
Figures illustrating the development of the branchial

THE ENTEROPNEUSTA. 9

skeleton, &c., will be given when the histology of the gills
is described.

Together with the increase in number of the gills the differ-
entiation between the digestive and intestinal region becomes
more prominent, the bright yellow-brown of the former show-
ing through the transparent body wall, being a most striking
feature in the appearance of the animal. When first per-
ceptible, the digestive tract is a simple tube, separated by a
slight constriction from the intestine. As growth proceeds
this constriction becomes more marked, and when the second
gill-slit is fully formed the separation between the two is
sharply defined (fig. 2). Subsequently a fold arises in the
digestive region which gives it the appearance of being made
up of two saccules (two gill-slits). This condition becomes
more and more marked, and then a third saccule appears
posteriorly (4—5 g.s.). When, however, the animal is seen
from the dorsal side the alimentary canal is seen to have a
wavy contour, the two saccules being thus parts of a slightly
bent tube. Moreover, in longitudinal sections the divisions
between the saccules are parts of a spiral fold which traverses
the whole digestive region. When five gill-slits are formed
there are three saccules, and in animals with ten gill-slits they
are five in number. After this the body walls become much
more opaque, attaining the condition which they present
throughout adult life. It is consequently not possible to follow
the internal development after this stage by means of surface
views.

The walls of the intestinal region are also thrown into folds,
but their arrangement would appear to be irregular. The cells
in this tract bear long flagelliform cilia which appear to drive
a current through the anus.

The anus is first found at about the time of the formation of
the second gill-slit. As previously mentioned it is almost if
not quite in the position in which the blastopore closed, being
posterior, median, and dorsal. When the tail is formed the
anus is immediately dorsal to it, in fact, its ventral margin is
formed by the dorsal side of the tail. In ordinary conditions

10 MORPHOLOGY OF

the anus remains open widely, but when the animal is irritated
it contracts its body and draws in the intestine, closing the
anus, over which there project two flaps of the body wall.
These flaps are very thin and transparent presenting an appear-
ance as of a posterior vesicle (fig. 3a).

In some specimens a curious separation between the meso-
blast and epiblast occurs in this place, which may be due to
re-agents or may have some significance ; this structure will be
treated of together with the other formations in the third
body cavity.

In the transparent animal pulsatory contractions can be seen
in a vesicle lying on the dorsal side of the notochord, but
owing to the imperfect transparency of the body walls in this
region nothing more could be definitely affirmed as to the course
of the blood. As will be seen in considering the internal
structure, a large trunk appears (2—3 g. s.) in. the dorsal
mesentery; pulsations could also be observed in this structure,
which seemed to pass from behind forwards, but occasionally
an appearance was produced as of a reversal in the direction.
But in consideration of the fact that the ventral wall of this
vessel is by the nature of the case adherent to the splanchno-
pleura, while the dorsal wall was fixed in the somatopleura,
no very certain importance can be attached to any observa-
tions of pulsation in the dorsal vessel, since any peristalsis in
the gut might produce this appearance. The same applies to
the ventral vessel, which is said to be contractile in B.
minutus (Spengel). On the whole, however, the balance of
evidence was distinctly in favour of postero-anteror pulsa-
tions in the dorsal vessel. No corpuscles were discovered in the
blood which is colourless. In preserved specimens it appears
as a homogeneous coagulum, which in specimens preserved in
Perenyi’s fluid shows a slight tendency to granulation.

Large amceboid-looking cells are visible, floating about in
the body cavities, especially in the second.

After about seven to eight gill-slits are formed the general look
of the animal changes, mainly owing to the fact that the walls
of the body become more and more opaque. This seems to be

THE ENTEROPNEUSTA. IN

due to thickening of the ectoderm and the appearance of
numerous mucous glands on the skin. The whole skin is
uniformly ciliated from the time when the blastopore closes
throughout life.

At about seven to eight gill-slits the suctorial tail disappears,
probably atrophying rapidly, but as to this process nothing more
can be predicated. It is to be found in some larve with eight
gill-slits, while others in the same stage are without it. The
anus is then terminal, circular, and permanently open.

At ten gill-slits the ovaries are first perceptible, but as yet
are not marked enough to appear in a surface view. In older
animals they form large yellowish-grey projections from the
sides of the body. Their minute structure, together with that
of the testis, will be given later.

The body of the adult is very highly coloured, the proboscis
being of a yellowish-white tint. The collar is a brilliant red
orange (especially in males), with a white line round the edge
of the operculum, while the rest of the body is of an orange
yellow, shading to pale green yellow in the intestinal region,
which is semi-transparent throughout life. The distinction
between the colour of the males and females is very well
marked in B. Kowalevskii, the genital regions being grey in
females and yellow in males. The sexes are of different colour
in all the Enteropneusta, most prominently so in B. sal-
moneus (Giard), in which the males are chrome yellow and
the females salmon coloured.

It may be well, before passing to the internal development,
to mention the peculiar odour which the creatures possess.
This odour is very penetrating and persistent, resembling that
of chloride of lime with a fecal admixture. All the species of
Enteropneusta which I have examined alive possess more or
less offensive odours. This peculiar property is most developed
in B. Brooksii (new species), in which the smell is very
distinct after the animals have been months in spirit, which
has been often changed. The smell of this species is strongly
suggestive of iodoform. It is so powerful as to be a con-
siderable drawback to investigating the species.

12 MORPHOLOGY OF

Another feature which ought not to be overlooked in a
general account of this species (B. Kowalevskii) isits extreme
vitality. Ina bucket of unaérated water in which all other animals
had died some days before, in a hot climate, these creatures were
able to carry on their existence, and parts of the body may be seen
moving about by means of the ciliated skin to which a com-
pletely macerated skeleton of the branchiz is attached. Lobes
of the testis, torn off, will likewise swim about for days. To
what extent the body is capable of regeneration I cannot say.
Specimens were found in which there was at all events an
appearance suggesting that the proboscis had grown again.
Spengel has alluded to regeneration of tissues as occurring in
B. minutus, and I have little doubt that it is also common in
B. Kowalevskii.

With regard to the specific name of this form, it appears
that the figure and description given by Agassiz of B. Kowa-
levskii identify it with the form which is the subject of this
paper. The mode of development ascribed by him to the
species is of course entirely different. Seeing, however, that
he was unable to show the connection between the animals
found by him in the beach and the Tornaria which he reared,
it does not seem by any means certain that these Tornaria
were the larve of B. Kowalevskii. On the whole, it is
at least possible that they were the young of some other
species, e.g. B. Brooksii, which occur; at least as far
north as the Chesapeake, and probably higher still on the
coast.

From a general survey of the group Enteropneusta, which I
hope subsequently to attempt, I think it will appear likely that
B. Kowalevskii stands, in many respects, in a group differing
in several features from the other members, which agree with
one another in these points, e.g. short proboscis, complicated
branchial skeleton, operculum small, liver saccules present, eggs
minute, &c. It is to the latter division i am inclined to
believe that Tornaria alone belongs.

In studying the anatomy of Balanoglossus by means of
sections difficulty arises owing to the variable amount of con-

THE ENTEROPNEUSTA. 13

traction which the body may undergo in preservation. The
trouble chiefly occurs in the case of the proboscis-stalk and
folds of the collar. By comparing figs. 3 and 4, which were
drawn from living specimens in the extended state, with fig. 5,
which is taken from a preserved specimen, these relations will
be understood. In the contracted animal the gill-slits are
always more or less obscure, owing to the great shortening
which takes place in the branchial region, together with the
protrusion of the valves (fig. 5, vlv.).

In older animals this contraction is not nearly so great,
probably owing to the increased firmness of the branchial
skeleton.

Internal Structure.

Stages Fand G—Skin.—The ectoderm is composed of long
fusiform cells arranged two to three deep over the body, except
in the dorsal side of the collar groove, where they are colum-
nar and one layer thick. Also in the posterior dorsal region
the skin is thinner than that of the rest of the body. The
whole surface is ciliated. Beyond the fact that the cells are
more compressed, and closely arranged, the structure is
similar to that of the previous stage.

Nervous System.—The solid cord which began to separate
from the skin in the middle dorsal line at Stage F continues to
sink inwards. No lumen is as yet present init. Throughout
its length it still remains in contact with the skin, fusing with
it at both ends (figs. 10 and 20—24).

Towards the end of Stage G a differentiation begins between
the upper and lower parts of this cord, the upper being formed
of cells, while the lower part consists of lightly stained sub-
stance, which in older animals is distinctly made up of fibres.
Its fibrous nature cannot in this condition be certainly affirmed,
probably owing to defect in preservation. The formation of a
nervous network over the whole body, which afterwards
occurs, is not yet begun.

Hy poblast.—The mouth is ventrally directed (fig. 7), and

14 MORPHOLOGY OF

the anterior wall of the gut (Stage H to F) runs at right angles
to the long axis of the body. As will afterwards appear this
feature is of importance. In longitudinal section, a commenc-
ing differentiation between the branchial and digestive region
is perceptible. . The cells of the former are columnar, while
those of the latter have irregular amceboid processes which
give the inner wall of the gut an irregular contour. The anus
is not yet formed.

In the front end of the third segment of the larva (Stage F),
anterior to the ring of cilia, the sides of the gut give rise to a
pair of dorso-lateral evaginations ; these pouches are the first
indications of the gills. No change has occurred in the skin
covering them. Subsequently they come in contact with the
skin, the walls fuse and then a perforation is formed through
the fused portion, apparently occurring by a process of degene-
ration of the tissue. Fig. 29 is from a section taken through
the side of one of these evaginations. The subsequent appear-
ances are shown in figs. 42 and 43, which are from an older
larva.

Notochord.—Im the later stages of F and G in the
anterior dorsal wall of the gut, arises a most remarkable struc-
ture. For reasons which will appear when its later develop-
ment and fate is considered, I propose to compare this organ
with the notochord of the Chordata, and by this name it will
be subsequently spoken of.

In Stage E it was stated that the anterior wall of the hypo-
blast came vertically to join the skin at the mouth (figs. 7 and
10). As, however, development proceeds, the dorsal wall of
the pharynx becomes partly constricted from the remainder
(figs. 20 and 22). As this process of separation of the dorsal
wall proceeds, the part so separated grows forwards so that it
comes to project slightly in front of the anterior end of the gut
(fig. 30). By this means a hypoblastic tube is formed dorsal
to the gut, with a lumen which opens into the archenteric
cavity (figs. 21, 22, and 30).

Mesoblast.—The lining of the anterior body cavity in
Stages E and F is composed of rounded cells arranged in con-

THE ENTEROPNEUSTA. 15

tact with the ectoderm. These cells are in some parts only
one layer thick (anterior and posterior walls) (fig. 8), while in
others (ventrally) they proliferate rapidly (fig. 9), forming loose
masses of cells, the outer elements of which are elongated,
with rounded heads from which the spherical cells which form
the inner portion are budded. This proliferation continues
until the proboscis cavity is partially filled up. But in the
later phases of Stage F elements are formed other than the
rounded cells above mentioned, in the shape of fibres (fig. 13)
which appear to arise in a curious way, as is shown in figs. 9,
10, 12, &c. The elongated cells gradually become pyriform,
the round ends being for the most part central, while the fine
ends are drawn out into peripheral fibres. The round heads
then appear to separate from the fibres, so that in examining
the mesoblastic structure lining this cavity in late larve of
Stage G, the elements are arranged in the following order:
centrally, a small empty cavity surrounded by a ring of sphe-
rical granular cells; next a layer of pear-shaped cells con-
tinued into peripheral fibres, and externally a layer, composed
almost, and later in life entirely, of radial fibres. Some of
these are inserted into the lower layer of the skin, and are
probably a peculiar form of connective tissue. In the heads of
the pyriform cells brightly refractive granules may be seen;
whether these are food or waste products cannot be affirmed.
Such then is the lining of the proboscis-cavity. In the
anterior third it is evenly distributed over the inner surface,
- but at the back of the posterior third, where the proboscis
tapers abruptly to its stalk, the layer of mesoblastic tissue is
much thinner dorsally than laterally and ventrally (compare
fig. 13 which is through this region with figs. 12 and 11 which
are anterior to it). On passing further backwards, the cavity
in Stage Gis divided into two by a great proliferation of
mesoblast from the dorsal surface, which grows downwards
until it meets the ventral mesoblast. The position occupied
by this structure at first coincides with the point at which the
anterior mesoblastic pouch closed off from the archenteron.
In that stage the anterior body cavity was a simple sac con-

16 MORPHOLOGY OF

tinued backwards into two lateral horns. The mass of cells
which now arise at the original point of separation, therefore,
constitutes a continuation of the division between these two
horns into the anterior simple cavity (fig. 14). As will
shortly be seen, this division of the back of the anterior cavity
into two is correlated to the forward growth of the notochord.

In this septum, which is composed of roundish, hexagonal
cells containing many granules (fig. 16), appears the first
rudiment of a peculiar organ which subsequently is a con-
spicuous and characteristic structure in the proboscis of all
the Enteropneusta, viz. the proboscis-gland (“ heart” of
Spengel!). This gland has at first the relations shown in
fig. 10, gi. It consists (Stage F) of a triangular mass of loose
tissue containing nuclei in which but few cell-outlines can be
seen, and would appear to be formed by a sort of degeneration
of the mesoblast of the septum. As yet it contains no cavity.
Immediately behind and ventral to it is the anterior end of
the notochord (fig. 17).

Fig. 18 is taken through the posterior apices of the meso-
blastic horns behind the gland.

Middle Body Cavities.—These are (Stage E) a pair of
simple cavities divided by a dorsal and a ventral mesentery,
completely closed from both the anterior and posterior cavities,
as they remain throughout life. They are lined by round or
crescentic mesoblast cells. As the proboscis-stalk is con-
stricted, the anterior parts of these cavities are compressed
into two forwardly directed horns. The horn of the left side is
shown in fig. 18. From an early period it projects in front of
that of the right side.

In Stage F a histological differentiation occurs in the
splanchnopleure at the place of union between that part of the
hypoblast which is destined to form the notochord and the
lower section of the pharynx. This appearance is shown in
fig. 25, z). It consists in the prolongation of the ends of the

1 For reasons which will subsequently appear this term is somewhat in-
appropriate. :

THE ENTEROPNEUSTA. 17

mesoblast cells which are in contact with the hypoblast into
curious tails, which are refractive.

I was at first led to suppose that these tails were mus-
cular, and this undoubtedly is the fate of many of the
mesoblastic elements of this region, but no direct evidence of
the contractile nature of these pear-shaped cells was attained
beyond the general suggestion of their shape; on the other
hand, when the notochordal sheath is developed in this region,
an appearance is presented which suggests the possibility of
their having taken part in its formation. Their histological
characters are, however, strikingly similar to those of the cells
which occur at the sides of the notochord in the same region
in Amphioxus at the time when it has eleven pairs of meso-
blastic pouches (Hatschek, No. 4, figs. 124, 126, &c.). These
cells are stated by Hatschek to be muscle-fibres; by analogy
it seems, therefore, likely that this may be the real nature of
the same cells in Balanoglossus.

As development proceeds, proliferations of mesoblast are
formed in the ventral region of the middle body cavities. From
each side in the posterior region of these cavities a tubular
portion is separated off from the rest (fig. 28).

The fate of these parts is not quite certain. It seems, how-
ever, likely that they unite with two forward growths from the
posterior body cavities to form the tissue space in which the
dorsal blood-vessel is ultimately enclosed throughout the collar
region. This tissue space I propose to call the perihemal
cavity (fig. 60, &c., Ph. c.).

The posterior body cavities are simple cavities, similar
to the middle pair. The dorsal and ventral mesenteries
persist throughout life. As yet they contain no special
differentiations.

Period between the Formation of the First and
Second Pair of Gill-Slits.

Skin and Nervous System.—The histology of the skin
in the proboscis and collar regions has not undergone material
2

18 MORPHOLOGY OF

alteration since the last stage. The epiblastic cells are some-
what smaller and more closely packed. In the posterior
regions, however, the ectoderm is thinner than in the
younger animals, owing to the rapid growth which occurs at
this time in the trunk region (fig. 45). Unicellular mucous
glands occur at rare intervals in it. But in the lower layer of
the skin at the posterior surface of the proboscis is formed the
beginning of that network of nerve-fibres which is such a pro-
minent feature in these regions of the body in later life.
Though a network of this kind eventually is formed on the
inner surface of the skin all over the body to a greater or less
extent, it is as yet only to be seen in the base of the pro-
boscis. The exact process by which this layer is deposited is
not certain, but it would appear that cells of the inner layer
elongate and form multipolar cells with long, thread-like, anas-
tomosing tails (fig. 32).

At this stage nuclei are still visible in this fibrous layer,
though in later stages they have almost entirely disappeared
from it (fig. 54, &c.). From this point in development on-
wards these fibres constitute a perfectly defined layer of tissue.
Projecting into it may be seen some of the fibres which were
described as being formed from the mesoblastic lining of the
proboscis cavity. These fibres are presumably supporting
structures. Occasionally an appearance is presented as of an
anastomosis occurring between them and the tails of the nerve-
fibres. Whether this is really the case or not, it can scarcely
be doubted that the muscle-fibres, which are now forming in
the proboscis cavity, receive their innervation from the fibrous
layer of the skin, to which many of them are attached, and
thus such an anastomosis is not @ priori improbable. On the
other hand, the structure of the mesoblastic fibres is rather in-
dicative of a supporting than of a contractile function. But,
as will afterwards appear, there are eventually present in the
mesoblastic elements of these animals cells which present
almost every shade of variety between undoubted contractile
fibres and obvious connective tissue, so that it is by no means
easy to determine the nature of these fibres with precision.

THE ENTEROPNEUSTA. 19

On the whole I am inclined to regard them as supporting
structures.

The separation of the dorsal nervous system is much
more marked during this period than it was before it. It is
now completely separate from the point of junction of the pro-
boscis with the trunk to almost the level of the first pair of
gill-slits. At its anterior end (fig. 86) may be seen the be-
ginning of the process by which the anterior lumen is formed.
This is effected by a forward growth of the collar, together
with a continual sinking and horizontal invagination of the
nerve-cord.

This lumen, thus formed, never extends for more than a
short distance into the cord, which, however, in its middle and
posterior regions in older animals, contains remarkable spaces
lined by columnar cells, more or less separated from each other
by strands of tissue, which will be described, together with the
later development and histology of the nervous system.

The nerve-cord, as always, joins with the skin at both ends,
but from its posterior point of junction the rudiment of its
dorsal continuation in the skin may already be seen in section
as a small area of fibrous tissue in the base of the skin in the
middle dorsal line (fig. 42). A similar strand (fig. 42) may
also be seen on the ventral side, beginning a little in front of
the first gill-slits. The two cords are still quite unconnected.

The Proboscis Pore.—tThe first appearance of this struc-
ture is a thickening on the inner surface of the epiblast in the
proboscis stalk, which soon becomes hollow while still attached
to the skin (fig. 34, p. pr.). This epiblastic sac is from the
first asymmetrical, being on the dorso-lateral aspect of the
left side. From the first, the cells of which it is formed are
columnar, and it has no communication as yet (two gill-slits)
with the exterior or with the body cavity.

Hypoblastic Structures.
Branchial Region and Notochord.—The process by
which the mouth comes to be forwardly directed has already
been described. In larve with one to two gill-slits it has already

20 MORPHOLOGY OF

begun (fig. 45). The walls of the branchial region are
distinctly differentiated from those of the rest of the gut,
consisting of long, solid-looking cells arranged in layers of one
to two deep. The lumen of the gut is very small anteriorly
and has an irregular outline in preserved specimens (figs. 36
and 37), but in the middle of the collar region as at present
marked out, its lumen is continuous dorsally with that of the
notochord (fig. 39).

This notochord, which is now a very prominent feature in
sections of the anterior end of the body, arose in the first
instance, as already stated, by a forward growth of the anterior
dorsal wall of the pharynx, which thus shuts off a short diver-
ticulum of hypoblast (fig. 30). The part of the pharynx with
which the walls of this diverticulum are continuous, then
separates itself from the rest by longitudinal constriction, which
at first causes the lumen of the gut to take an 8-shaped figure,
the separation of the dorsal part of the 8 becoming finally
complete from before backwards. This process gives rise to the
appearance seen in fig. 87. The part thus constricted off
becomes then entirely separated except at its posterior end,
where throughout life its lumen opens into the pharynx.

In its anterior region the lumen of the notochord is always
suppressed at this stage, owing to the compression of the ventral
against the dorsal wall. Moreover, in larve of this, as of all
subsequent stages, the lumen is altogether obliterated in part
of its course. This obliteration does not appear to occur pro-
gressively from before backwards, but more or less irregularly,
so that, as in fig. 38, the lumen may have already disappeared
while still present in a region anterior to this (fig. 36). As,
however, in older animals the lumen is always continued far
into the notochord of the proboscis cavity (namely, to a point
anterior to that where it is already obliterated in two-gill
larvee), it is almost certain that the subsequent increase in the
length of the notochord is due to a growth from behind for-
wards, and that all the notochord which is as yet formed
(two gill-slits) is pushed bodily forwards by a proliferation,
probably occurring at the point of union with the gut. The

THE ENTEROPNEUSTA. 21

alternatives, that the growth occurs at the apex or at any point
intermediate between the two ends is unlikely, from the fact
that almost immediately after two gill-slits the tissue of which
it is composed becomes vacuolated and irregular, undergoing
the “degeneration” characteristic of notochordal substance,
presenting therefore by no means the appearance of a growing
tissue.

The length and proportions of the notochord at this stage are
indicated in fig. 45, which is, however, not a truly median
section. Its anterior end already projects far into the anterior
body cavity, pushing in the mesoblastic lining. Fig. 45 is from
a specimen slightly older than that from which figs. 37, &c.,
are taken, and the commencing degeneration of the notochord
tisssue is already begun.

To recapitulate: the growth of the notochord is due to:

1. A forward growth of the dorsal anterior portion of the
archenteron (fig. 30). This is supplemented by—

2. A longitudinal constriction of the dorsal region of the
pharynx, which gradually travels backwards (cp. figs. 21 and 22
with figs. 88 and 39), separating a hollow hypoblastic tube
which remains open to the gut behind.

3. A forward growth from the point of junction with the gut.

In connection with the notochord must be mentioned the
skeletal rods, which now just appear, though it cannot be
positively affirmed that they are of hypoblastic origin. When
first visible, they are two short rods of a deeply-stained, struc-
tureless substance, which lie in the angles between the noto-
chord and the dorsal wall of the pharynx. As first seen in this
position they appear to be formed externally to the hypoblast
cells, against the ends of which they lie. Their posterior ends
are enclosed in the hypoblast (fig. 39), and it is difficult to
understand how this can have been brought about if they were
secreted by the mesoblast cells. This would involve an outward
growth of the hypoblast to inclose them, of which there is no
appearance. As will afterwards be seen the view that they are
of hypoblastic origin is supported by the fact that they con-
tinue growing with the growth of the animal, and that their

22 MORPHOLOGY OF

thickness is, so to speak, inversely proportional to that of the
cellular tissue of the notochord, which becomes thinnest in the
region where they attain their maximum size. This, therefore,
suggests that they are formed at the expense of the notochord.
An analogy moreover at once occurs of the secretion of such
a substance from notochordal tissue in the case of Amphioxus,
in which discs are deposited of very similar histological
character to these masses in Balanoglossus. These two rods
posteriorly bend downwards and then slightly forwards lying
in the hypoblast. They are therefore each cut twice in sections
through this part of the body.

Behind the end of the notochord are the gill-slits, which are
still only circular pores leading tu the exterior (fig. 43). The
lumen of the gut in the branchial region is (in contracted
specimens) much suppressed, and its ventral side, however,
always is grooved, and in this groove the cilia which line the
branchial region are well developed.

When the animal contracts, the branchial region of the gut
posteriorly projects over the front of the digestive region on
the dorsal side (fig. 44).

The digestive region is quite distinct from this point
in development onwards. Its cells have the appearance shown
in fig. 45, being large cells with amoeboid processes containing
large granules. From the back of the digestive region the
intestinal region is now marked out. Its walls are thinner,
and the cells composing it are long and ciliated at their inner
ends. The anus is a large aperture, permanently open when
the animal is extended, situated dorsal to the tail. There is
no epiblastic proctodeum.

Mesoblastic Structures.

Anterior Body Cavity. The mesoblast of this tract may
now be divided into two parts—(1) a peripheral portion
which lines the body walls of the proboscis, and (2) a central
portion which is pushed in by the forward growth of the
notochord ; between these two portions there is a body cavity
which retains a clear central space throughout life.

THE ENTEROPNEUSTA. 23

(1) The peripheral part consists of cells and fibres. The cells
are mostly pyriform similar to those of one-gill larve. They
contain few granules and are much reduced in number. The
fibres are more numerous. They are of several kinds: thick
strands that are obviously muscular, running as yet only in a
longitudinal direction, having their ends inserted in the skin:
and also thinner fibres, which are arranged circularly, as in
the periphery of the middle third; longitudinally, only
occurring asa sort of sheath separating the circular fibres from
the general connective tissue, and occasionally in the central
portions; radially, as over the whole periphery of the pro-
boscis cavity, and finally, very fine fibres forming a loose
network connecting all the mesoblastic elements together. As
may be seen in figs. 31, &c., there is as yet no suggestion of
the peculiar concentric arrangement of the longitudinal fibres,
which gives a section of the adult proboscis its curious appear-
ance. The radial fibres are mostly inserted into the skin
(fig. 82), their possible connection with the epiblast has been
already discussed.

There are still present a few of the large spherical cells
which formed the inner layer of mesoblast in one-gill-slit
larve, but they contain fewer granules, and are only found in
the posterior parts of the cavity.

(2) The central portion in the one-gill-slit larva was made
up of a mass of spherical and polygonal cells surrounding a
central part filled with loose tissue (fig. 16). As the noto-
chord grows forward, this structure, which, from its subsequent
fate may now be termed the proboscis gland, is prolonged with
it. As this forward growth occurs, a basement membrane
forms between the mesoblast outside the proboscis gland and
the loose tissue contained init. Fig. 47 shows a section ante-
rior to the part where this basement membrane is formed.
The mesoblastic cells are here large and granular, having
irregular shapes. The proboscis gland now is hollowed out,
so that it contains a space which extends from the proboscis
stalk to the front end of the notochord. This space contains a few
mesoblastic cells, which are as yet not obviously differentiated.

24, MORPHOLOGY OF

The position of this space is shown in fig. 45. It will be seen
in the sequel that the mass of the secreting tissue of the gland
is formed from the cells covering this space, and forming the
sides of this central portion of the mesoblast. The space itself
contains eventually but few of these secreting cells, and will
be spoken of as the sac of the proboscis gland (fig. 52, gl. s.).

The heart is as yet not represented.

Middle Body Cavities.—The tissue lining these cavities
does not exhibit more than a general progress of differentiation.
Owing to the increased narrowing of the proboscis stalk the
two anterior horns of the cavities are more distinct. In the
dorsal and ventral mesenteries basement membranes occur.
The cells lining these cavities are generally pyramidal or
crescentic, some of them being radially directed and fusiform.

Posterior Body Cavities.—The tissue of the posterior
mesoblastic pouches has undergone the same proliferation and
progressive differentiation as that of the middle cavities. These
processes are not, however, quite so far advanced. The two
horns (peri-hemal cavities) which began to grow forwards
above the gut in the one-gill larva are now much more de-
veloped. They now extend into the back of the collar region,
in front of the first gill-slit. They are filled more or less with
loose mesoblastic tissue containing a few fibres. As before
stated, in the mesentery between them is formed the dorsal
blood-vessel. This structure appears as a split in the
mesentery, and is as yet quite empty in preserved animals.
This split extends already through the whole course of the
perihzmal cavities. On the ventral side also a split is formed
in the lower mesentery of the posterior body cavity to form
the ventral blood-vessel. There is as yet no connection
between the dorsal and ventral vessels.

From this point the details of the subsequent development
and anatomy of the parts will be given in the section dealing
with the separate organs ; but, before doing so, it may be well
to describe briefly the general course of the later history of the
internal structures.

Notochord.—As will be seen, this structure increases

THE ENTEROPNEUSTA. 25

greatly in size, and assumes a vacuolated appearance closely
resembling that of the notochord of young Lampreys and
Elasmobranchs.

The skeletal rods attain a considerable size. Their anterior
ends unite, forming a single bar, while their posterior ends
diverge, partially enclosing the gut. This whole structure
forms the support of the proboscis.

From its development, position, relations to surrounding
parts, histology and function it appears to me to be com-
parable with the notochord of the Chordata, and this name is
strictly appropriate to it. Even if the suggestions which will
be made hereafter as to its phylogenetic significance be not
accepted, this rejection would in no way militate against the
fact that this structure is to all intents and purposes a noto-
chord, which can only be designated as a longitudinal dorsal
supporting rod, derived from the hypoblast.

The nervous system afterwards attains a great develop-
ment (fig.60). The dorsal cord in the collar sinks further and
further from the skin, being (in B. Kowalevskii) connected
to it by a mesentery. The lumen is in this form less developed
than in B. minutus, &c. The ventral cord is the next to
appear, and almost simultaneously with it arises the deposit of
nervous tissue in the skin at the base of the proboscis. This
deposit afterwards attains a great extent, forming athick band
round the proboscis stalk. It may be noticed that this nerve-
ring has practically the same relation to the proboscis that the
ring of ganglia in Nemertines presents, the proboscis of Balano-
glossus being, however, permanently protruded, and the nerve-
ring still in the skin. Both these nerve-rings agree in being
traversed in Nemertines by two and in Enteropneusta by one
pore communicating from the exterior to sacs which were
originally archenteric diverticula.

Body Cavities.—As before mentioned, the left horn of the
anterior body cavity comes to open by the proboscis pore to
the exterior. This opening is median and dorsal in other
species, but in B. Kowalevskii it is on the left side throughout
life. In all species it perforates the nerve-ring of the stalk.

26 MORPHOLOGY OF

Nearly all the proboscis cavity is eventually filled up with
loose tissue. This is composed of a number of concentric rings
of longitudinal fibres and connective tissue. These rings
attain the maximum number of eight. In fig. 51 the general
appearance of an older proboscis cavity is shown. The con-
centric arrangement is, however, not yet attained in the stage
there figured.

A free space is always present between these rings of tissue
and the central structures in preserved specimens.

The proboscis gland becomes a large mass of tissue com-
posed of anastomosing blood-vessels covered with conical cells
fixed on the vessels by their apices. Many of these cells
contain remarkable yellow granules, which are also to be
found outside the cells, sometimes presenting a conglomerate
arrangement. They would seem to be formed in the cells and
thrown out. They are also to be found in the sac of the
proboscis gland. This sac is blind posteriorly, but anteriorly
the loose tissue which it contains passes into unbroken con-
nection with the remarkable cellular layers covering the blood-
vessels. Hence the sac is in communication with the central
body cavity through the tissue spaces of the gland. The
function of this gland is quite unknown. Spengel suggests
that it is an “internal gill.” It does not seem probable to me
that an animal with some sixty pairs of true branchial clefts
would also possess another large and complicated organ of
entirely different structure also for respiratory purposes. The
presence of the brown granules suggests that it may be excre-
tory. If this were so, the excreta might be expected to pass
out by the proboscis pore which opens into the cavity in which
the gland lies. No direct evidence was obtained as to the
normal direction of the flow through this pore. Spengel and
other observers state with regard to B. minutus that water
is taken into the body cavity at the proboscis pore, but my
own observations do not confirm this statement. On the
contrary, particles of Indian ink or carmine held in suspension
in the water in which the animals have lived for days, cannot
be found to enter the proboscis cavity, while similar particles,

THE ENTEROPNEUSTA. 27

if placed in the tissue spaces of the proboscis, are certainly
expelled by the pore. This evidence does not of course
demonstrate, beyond doubt, that inwardly directed currents
never enter the pore but only gives a presumption against
them.

Spengel’s statement of the absence of the pore described by
Kowalevsky and Agassiz, at the apex of the proboscis, is true
for all the species which I have examined.

The Heart.—As mentioned by previous observers, a large
vesicle may be seen pulsating above the water-vessel in Tor-
naria; such a pulsation may be observed in the dorsal side of
the base of the proboscis in B. Kowalevskii at the stage of
two to three gill-slits. Spengel states that this contractile sac is
the upper of the two cavities lying above the notochord (figs. 51,
58, &c.); this he calls the “heart.” In consideration of the
fact (which he also admits) that it contains no blood, gives off
no vessels, and has no muscular walls, this name seems open
to misconstruction. As this so-called ‘‘ heart” is merely a
space filled with loose tissue which is part of the proboscis
gland I have preferred to call it the sac of the proboscis gland,
and to reserve the name “heart” for the sac which lies
between this space and the notochord (figs. 50 and 51, At.).
That this is the actual heart can I think hardly be doubted.
It arises at about three gill-slits as a single horizontal split in
mesoblast between the notochord and the sac of the proboscis
gland. It acquires muscular walls and is always nearly full of
a coagulum similar to that which is found in the remaining
blood-vessels of the body, which can all be traced into connec-
tion with it.

These peripheral vessels are (1) a longitudinal dorsal one,
running from the heart to the tail in the dorsal mesentery
from the back of the collar, and in the collar as a blood-space
surrounded by the perihzmal cavities (fig. 60); (2) a ventral
longitudinal vessel running from the back of the collar to the
tail in the ventral mesentery. These two are connected by
blood sinuses in the skin and in the wall of the gut. I have
not seen the definite circular vessel which other observers state

28 MORPHOLOGY OF

surrounds the gut anteriorly. The principal skin sinuses are a
pair of large ones which extend on each side of the dorso-
lateral regions of the proboscis (fig. 51).

The Collar Pores.—On the outer wall of each atrial cavity
appears a thickening at about eight gill-slits. This thicken-
ing acquires a perforation which leads from the collar body
cavity to the atrial cavity. These perforations acquire a
curious folded lumen and become ciliated constituting the
collar pores. Their opening into the atrial cavity is con-
tinuous with that of the first gill-slit. From analogy it may
be expected that these pores are of an excretory character.
With regard, however, to the direction of the flow through
them, the evidence is as unreliable as that as to the currents
in the proboscis pore. Spengel states that water is taken into
the body cavity at these points, while I was unable to find that
coloured particles ever entered it. Similarly, however, such
particles placed artificially in the collar body cavity were
washed out at these points.

The Middle and Posterior Body Cavities.—These
cavities become in adult life more or less filled with connective
tissue, &c. The cavity of the middle pair becomes practically
obliterated owing to the great development of loose tissue in
it. But in the posterior cavity this proliferation is never so
great. The middle pair of body cavities is far more choked
up in B. Kowalevskii than in B. minutus, but in B.
Brooksii the amount of connective tissue is even greater.
This fact is interesting in the present state of views as to the
morphological meaning of ‘‘ coelom,” as presenting an example
of a body cavity arising in a most typical ‘“‘ mesodermic ”
manner, assuming ontogenetically precisely such an appear-
ance as is presented by the ‘‘mesenchyme” of Platyhel-
minths, &e.

A statement as to the origin of the generative organs
is reserved for the present, as some doubt exists as to the layer
from which they are derived. The general appearance is, how-
ever, suggestive that they are of epiblastic origin.

Before beginning a detailed account of the later development

THE ENTEROPNEUSTA. 29

it may be desirable to discuss briefly the new light which these
facts throw upon the affinities of the Enteropneusta.

In 1881 Metschnikoff published a detailed comparison of
Balanoglossus with the Echinoderms, comparing Tornaria with
Bipinnaria, showing that the resemblance is close, and con-
cluding with the suggestion that Balanoglossus should be in-
cluded among the Echinodermata in a separate division,
“ Bilateralia.” The branchial structures he compared to the
openings from the body cavities of Echinoderms. This view,
as thus expressed, receives no support from further observa-
tions, and would now appear to be untenable.

As mentioned above, all the Enteropneusta possess a sup-
porting structure which is comparable with the notochord in
every way, except in extent and in the persistence of its con-
nection with the alimentary canal. Its resemblance to that of
Amphioxus is especially striking, for in Amphioxus the noto-
chord projects a long way in front of the mouth. It moreover
possesses gill-slits which are not only without parallel, except
among the Chordata, but also in structure, position, and de-
velopment, agree exactly with those of Amphioxus, in which
the slits acquire the same U-shaped form.

The agreement in the position of the blood-vessels and
skeleton of the gill bars is also very close. The fact of their
gradual increase in number from before backwards throughout
life is another common feature.

The position and mode of origin of the central nervous
system is also similar in both forms; the invagination of the
dorsal cord in Balanoglossus being, however, only partial, while
that of Amphioxus is complete.

The mesoblastic pouches suggest the same resemblance,
differing only from those of Amphioxus in number, being one
median and four lateral, while those of Amphioxus are one
median and twenty-eight lateral. As I have already pointed
out, the fate of this anterior pouch is in the two animals closely
similar. In both it is divided into two as the notochord grows
forward. In Amphioxus the division is complete, while in
Balanoglossus it is partial. In both, the backwardly-projecting

30 MORPHOLOGY OF

horn upon the left side becomes lined by ciliated columnar
cells, and opens to the exterior. Moreover, in both animals
this opening has a definite relation to the nervous system. In
Amphioxus it becoms the “olfactory” pit (Hatschek), while
in Balanoglossus it is surrounded by a mass of nervous tissue.
Finally, the collar folds, especially of B. Kowalevskii, would
appear to be comparable with the commencing atrial folds of
Amphioxus, for the most anterior gill-slits open into the cavity
which is thus enclosed.

The pair of ciliated funnels opening from the collar body
cavities to the atrium has been compared above to the excre-
tory tube mentioned by Hatschek in a similar position in
Amphioxus.

A pair of tubes has been described by Lankester in
Amphioxus opening into the back of the atrial cavity, com-
municating with the dorsal body cavities. It may be remarked
that if the collar fold of B. Kowalevskii were prolonged
backwards, as the atrial folds are in Amphioxus, the two collar
funnels would then be carried backwards, and have relation
similar to that of these tubes, which, as suggested by Lankester,
may be excretory.

To recapitulate: striking resemblances to the Chordata, and
especially to the Cephalochord type, are to be found in the
following structures:

(1) The notochord.

(2) The gills and branchial skeleton and blood supply.

(3) The central nervous system.

(4) The origin of the mesoblast.

(5) The peculiar fate and remarkable asymmetry of the
anterior pouch.

(6) The atria.

(7) The excretory funnels.

In each of these cases, excepting that of the branchial struc-
tures and the excretory funnels, the condition is that which
would be produced by a partial or arrested development of the
corresponding structure in Amphioxus.

The above considerations appear to justify us in including

THE ENTEROPNEUSTA. 31

the Enteropneusta among the Chordata. I would, therefore,
tentatively suggest the following table :
Chordata.—Hemichordata (Enteropneusta).
Urochorda (Ascidians).
Cephalochorda (Amphioxus).
Vertebrata.

It is not now proposed to enter into a more detailed dis-
cussion of the morphology of the group, or of the light which
an acceptance of this suggestion throws on the origin of the
Chordata. A fuller examination of these points is reserved for
a subsequent occasion.

It may nevertheless be advisable to point out that since,
according to Spengel, the tissue of the “ water vessel” of Tor-
naria forms the lining of the proboscis cavity of B. minutus,
this “water vessel” is therefore the same structure as the
anterior body cavity in the form just described. If, then, the
“water vessel”? of Tornaria is comparable to the “ water
vessel” of Bipinnaria, which has a similarly asymmetrical de-
velopment upon the left hand side of the body, which view has
been held by all previous observers, it would therefore appear
to follow that the water vessel of Bipinnaria is primd facie
comparable with the asymmetrical anterior body cavity of
Amphioxus.

Later Development and Comparative Account of
the Organs in the various Species.

The species which I have examined are the following :
B. Kowalevskii (Alex. Agassiz.) (Coast of North America).
B. Brooksii, n. sp. (ditto).
B. minutus (Kowalevsky.) (Bay of Naples).
B. salmoneus (Giard.) (Iles de Glenans, off South Coast
of Britany).
B. Robinii (Giard.) (ditto).
B. Kowalevskii differs from all the others in having—
(1) a relatively long proboscis ;
(2) no hepatic sacculations ;

32 MORPHOLOGY OF

(3) a simple branchial skeleton, not connected by longi-
tudinal bars, as in the other forms ;

(4) very short collar funnels, the external opening of
which is directed transversely instead of posteriorly,
as in the others, in consequence of

(5) the greater extent of the backwardly-directed atrial
fold.

As far as can be determined from Agassiz’? account, his
species agrees in these points with the one which is the subject
of this paper.

The Notochord and Axial Skeletal Rods.

B. Kowalevskii.—The general course of development of
the notochord has been already described. Fig. 47 shows the
histological characters of the cells at two gill-slits, when they
are still large full-looking cells with large nuclei.

Fig. 48 is from a section of the proboscis stalk in the region
of the pore of a larva with three gill-slits. It exhibits the
commencing degeneration of the notochordal tissue and the
increase in size of the structureless deposit which constitutes
the skeletal rods.

In this region the skeletal rods unite to form a single
median rod, which is continuous with the notochordal sheath.

In figs. 49—53 the appearance of the notochord at four gill-
slits is illustrated. The degeneration is now far advanced.
Nuclei are rare in the notochord, and the cells are vacuolated,
as shown by the fact that the nuclei occur in the nodes of the
cell outlines. The protoplasm of the cells merely forms a kind
of network containing a few nuclei. The remainder of the
space is probably occupied by some homogeneous non-proto-
plasmic substance, such as may be supposed to fill up the
notochordal tissues of other forms.

Figs. 49—52 are from sections taken in front of the lumen
(cp. fig. 57). In fig. 53 the lumen is reached. It will be ob-
served that the lumen at this point still ends asa fine tube. In
later life a great thickening of the notochord takes place at this
point, and the lumen then acquires a downward extension

THE ENTEROPNEUSTA. 33

(figs. 57 and 56). Immediately behind this downward exten-
sion lies the anterior end of the united skeletal rods, which
here attains its greatest thickness, almost filling the sheath of
the notochord, the tissue of which is here almost suppressed.

In old specimens the shape of the anterior parts of the
notochord becomes rather irregular in section.

In that part of its course which lies behind the proboscis the
notochord in the adult is more or less elliptical in section,
containing a large and somewhat irregular lumen. Its tissue
is here greatly reduced, and this reduction appears to progress
regularly as the animal grows older. In fig. 60 the appearance
of the notochord in such an old adult is shown. Degeneration
has progressed far, leaving the notochord as a space surrounded
by vacuolated cells enclosed in asheath. With this sheath are
connected the skeletal rods, which attain a great size. Cen-
trally, on the dorsal side, between the notochord and the gut,
lies the principal rod; this is formed by the uniting of the two
rods (figs. 37, 38, &c.), whose development has already been
described. This fused portion is now diamond-shaped in sec-
tion ; its lower angle causes a dorsal ridge to project into the
mouth cavity. Laterally are placed two long rods, which
are continued into the central rod and notochordal sheath
anteriorly. To these lateral rods are attached large bunches
of longitudinal muscles, by which, doubtless, the notochord
may be pulled backwards, and the proboscis retracted so as to
shut the mouth (fig. 60).

Posteriorly the median rod divides into two, and the opening
from the notochordal lumen into the gut lies in the angle
formed by the separation of these two diverging rods
(fig. 57).

A considerable deposit of “ structureless” substance takes

_ place, filling up the spaces in the proboscis stalk, and forming
a partial sheath around the perihemal cavity. Whether this
substance is chitinous or of some other material I am unable
to say. The ensheathing parts of it have exactly the histolo-
gical appearance presented by the “structureless” substance,
which in Amphioxus is continued from the notochordal sheath,

3

34. MORPHOLOGY OF

&c., between the myotomes. The rods, however, are seen in
adult specimens to have concentric markings in section, sug-
gesting that they are formed by a deposit around a central
core. In that part of the rod which lies just in front of the
point of divergence of the two horns there are two such cores.
The cores stain more deeply than the rest of the rods.

In B. minutus the position and general appearance of the
notochord is similar to that in B. Kowalevskii, with the
exception that the lateral rods are not developed to the same
extent. The histology, however, is somewhat different. In
figs. 61—63 the appearance of the tissue is shown. Spengel
has stated that he is unable to find, in B. minutus, any
structure resembling the notochord of the Chordata, figuring
that organ as though composed of columnar cells. My own
observations give no support to this statement. No specimen
of notochord of any of the species which have been examined
by me present the appearance indicated by Spengel. On the
coutrary, specimens preserved severally in picric acid, corro-
sive sublimate and acetic acid, Perenyi’s fluid and osmic acid,
all equally show this body as made up of vacuolated tissue
strongly suggestive of the notochord of Chordata. This is
especially the case with regard to that of B. minutus. Fig.
61 is taken from a section of B. minutus in front of the lu-
men. The sheath is here very slightly developed. It will be
observed that the nuclei are fewer in number than in B. Kowa-
levskii, and that they are gathered round the upper centre. The
section shown in fig. 62 is from the same in the region of the
proboscis stalk. It shows the diamond-shaped section of the
skeletal rod and the concentric markings upon it. The noto-
chord itself is here elliptical in section, and the strands of
protoplasm running across it are well seen.

Fig. 63 shows its appearance in front of the point at which
it opens. Dorsally its wall is very thick and solid, while
ventrally it is comparatively thin. Behind this part the
skeletal rod divides into two divaricating horns as in B.
Kowalevskii.

In the other species the structure of the notochord is

THE ENTEROPNEUSTA. 385

essentially the same as in B. minutus. In the anterior
region of the notochord of B. Brooksii there are hardly any
nuclei at all. I hope to publish figures illustrative of the later
development and anatomy of the other parts at no distant date.

Mertnops oF INVESTIGATION.

From the characters of the unfertilized egg of B. Kowa-
levskii it was extremely improbable that the earliest stages
of development could be passed anywhere else but in the mud
which the parents inhabit. Though the examination of
Agassiz had failed in the attempt to find any but adult
Balanoglossus in this situation it seemed worth repeating.
Accordingly a large quantity of mud inhabited by Balanoglossus
was placed in a glass vessel of water and worked up, avoiding
rotatory currents, until the whole was in suspension. A number
of Balanoglossus which had previously been minced very finely
were then thrown in and the whole was then left to settle for a
few minutes. I then siphoned off the water and lighter parti-
cles in suspension, which consisted chiefly of vegetable débris,
stopping the siphon when the layer of chopped Balanoglossus
was reached, which could easily be seen by the bright orange
colour of the fragments. This portion was drawn off and
examined separately. It was found to contain great numbers
of larve and embryos of Balanoglossus, minute Nemertines,
free Nematodes, &c. In this manner all the animals
living in several hundredweight of mud may, in an hour or
two, be collected into about a pint of water and sorted with a
simple microscope. This was generally performed by rotating
a little of the water in a shallow saucer with a slight peripheral
groove. The larve then all lie in the groove which may be
passed under the lens by rotation. After a little practice it
becomes unnecessary to discolour the water with fragments of
the animal required, as of course the right layer can easily be
detected by the size and character of the particles composing
it. It appeared to be worth mentioning this mode of obtaining
mud larve as its application on a large scale does not seem to
be generally employed. It should be remarked that small

36 MORPHOLOGY OF

Nemertines, &c., are usually found still suspended long after
most of the mud is precipitated. The water should then of
course be poured off and left to settle separately. The degree
to which any particular animals required may thus be easily
separated off from the rest is very great. The disadvantage of
such a method is that a certain number of larve are sure to be
broken in the stirring necessary to effect suspension. I have
little doubt that the larvee which were observed free-swimming
earlier than stage F were thus liberated from the egg-shell.

Preservation.—All my specimens of B. minutus were
obtained from Naples, being very kindly prepared for me by
Mr, Weldon with picric acid.

Most of the larve of B. Kowalevskii were placed for
less than a minute in corrosive sublimate sat. sol. two parts,
mixed with one part glacial acetic acid, washed with water and
successively passed through 30 per cent., 50 per cent., 70 per
cent., and 90 per cent. spirit. On the whole the results
given by this reagent were the best. The softer parts, however,
are best preserved in those specimens which were treated with
Perenyi’s fluid one hour, then 90 per cent. spirit for twelve hours,
the 90 per cent. spirit being then changed. In the case of
adults preserved with Perenyi’s fluid, the fluid was changed
once or twice. ;

Osmic acid did not give good results, but probably this was
due to bad manipulation. The sections were cut in continuous
series with Caldwell’s automatic microtome.

THE ENTEROPNEUSTA. 37

On the

Morphology of the Enteropneusta.
PART II.

By

William Bateson, M.A.,
Fellow of St. John’s College, Cambridge.

With Plates VII—XII.

Tur following. paper is descriptive of the figures which
illastrate my concluding account of the morphology of the
Enteropneusta.

As an abstract of some of these facts was given with the first
part a certain amount of repetition has become unavoidable.

Since the publication of Part I, I have been able to make
some further observations on the histology of the fresh tissues
of the Brittany species (B. salmoneus and Robinii). For
this opportunity I am indebted to the kindness of the di-
rectors of the Zoological Laboratory stationed at Concar-
neau, Finistére. And especially my thanks are due to Dr.
Chabry for affording me these facilities.

The Skin and Nervous System.

The skin of all the species is entirely ciliated.
In the fresh condition I have chiefly studied it in B.
4,

38 MORPHOLOGY OF

Robinii, and it will be better first to describe its features in
this form. Its structure is best seen by killing the tissue in a
mixture of one part of 1 per cent. osmic acid and one of sea-
water, then washing with sea-water, and staining with picrocar-
mine. This tissue on being teased out in glycerine shows the
structure figured in figs. 76 and 77. The cells are very long,
and most, if not all of them, extend the whole length of the
skin (cf. fig. 75). The heads of these cells in the natural
living state are closely in contact with each other, but on
pressing out the tissue both in living and also in preserved
specimens these heads may be stretched away from each other,
but each remains attached to its neighbours’ by more or less
regular anastomoses. It thus is brought about that the
surface of the skin is made up of a sort of honeycomb of
tissue, each of the nodes being the outer end of an ectoderm
cell. The cells are very difficult to separate finely, but the
skin may easily be broken up into small rectangular pieces.
On separation each cell is very thin; its outer end is slightly
pyramidal, and is continued into a thin fibre which gives off
anastomoses with adjacent cells and dilates at intervals. In
one of these dilatations, generally the last, the nucleus is
placed. Below this point the cell is continued into a very fine
filament which may be traced for some distance. Many of
these filaments terminate in small round knobs, which are
possibly due to reagents.

In sections of hardened specimens these filaments may be
followed into the layer of nerve-fibre, which is always more or
less developed at the base of the ectoderm cells over the whole
body. These cells compose the larger part of the skin of the
proboscis and collar. Amongst them are distributed cells
which probably secrete mucus, &c. ‘These cells are of several
kinds. First,in the skin of the proboscis are large goblet cells
whose nucleus alone stains (fig. 75, mu’). Next, in the skin of
the back of the collar and of nearly all the rest of the body
excepting those parts in which concentrations of nervous
tissue are found, almost the whole tissue is made up of large
cells full of some substance probably lubricating also, which

THE ENTEROPNEUSTA. 39

does not stain. These cells are sufficiently represented in figs.
72 and 72a, which are, however, from B. minutus. In parts
of the skin which are of this kind the long cells of the ecto-
derm are comparatively few in number, and thus the skin has
a spongy consistency which is very characteristic. This is
true of the skin behind the collar in B. minutus, B. sal-
moneus, and B. Robinii. There is a general similarity
between the skins of all these forms, and probably their struc-
ture is the same asin B. Robinii. This statement, however,
only rests on the evidence of sections, as no teased preparations
were made of B. minutus. In the skin of the collar and
proboscis especially a small number of nuclei may be seen in
the higher layers of the skin. Whether these belong to young
cells of the tailed series or of the secreting type was not deter-
mined. Another set of small, generally bitid secreting cells,
are found in the proboscis skin; the contents of these cells are
granular.

There is one other point of importance in treating of the
skins of these forms, viz. the constant presence in teased pre-
parations of large spindle-shaped cells (fig. 77, c). As the
result of many observations it appeared nearly certain that
these had really been broken off from the ends of the long
ectoderm cells. Unless care was taken in the preparation this
frequently happened, many of the ectoderm cells being broken
and therefore without nuclei, and hence the probability that
this was the origin of the spindle-shaped cells. Since these
fusiform cells are generally most abundant at that level of the
skin at which the nuclei of the long cells are placed, the
appearance is suggested that they form a second layer of ecto-
derm cells ; but for the reasons above stated it seems likely that
this is erroneous, and that there is no such definite second layer.

The resemblance between this skin and that of some
Nemertines, e.g. Monopora vivipara (Salensky, ‘ Arch de
Biologie,’ 1884), is very close. In this animal the same
spongy appearance is produced, and it is possible that the
deeper layer of ectoderm may be capable of the same expla-
nation,

40 MORPHOLOGY OF

The skin of B. Kowalevskii differs in some ways from
that of B. minutus, &c., especially that of the trunk, in which
the large goblet cells are comparatively rare. In all parts of
the skin round, unicellular glands are more or less frequent, but
their contents stain more or less deeply with hematoxylin, &c.
These cells often fall out, leaving empty spaces. In the collar
of B. Kowalevskii the skin is very thick and is full of very
long cells (figs. 80 and 81) containing granular contents, which
stain very deeply.

Fig. 79 shows a section of part of the proboscis skin in
which the layer of nerve-fibre is very thick. In the upper
part of this kind of skin there is a definite row of long nuclei
which with some reagents assume a dice-box shape, probably
due to preservation. To what extent these cells reach the
whole depth of the skin cannot be affirmed, but many of them
can be traced into fibres which run into the layer of nerve-
fibres.

Nervous Concentrations in the Skin.—As has been
already mentioned, in all the parts of the skin a greater or
less quantity of unstained substance may be found in the base
of the skin. The substance contains no nuclei (excepting a
few in the nerve-sheath of the base of the proboscis), and may
oe seen, especially in fresh osmic acid preparations, to consist
of fine fibres. Into it run the tails of ectoderm cells. In the
next place fibres may frequently be seen running out of it
througl. the basement membrane, and losing themselves
amongst the mesoblastic tissues. The question as to the
nature of these fibres is one of great interest. They may
either be mesoblastic fibres penetrating into the ectoderm as
supporting structures, or they may be epiblastic fibres leaving
the skin, in which latter case they are in all probability
nervous.

Somewhat similar fibres have been described by Ludwig in
the similar tissues of Asterias, and he is of opinion that they
are connective tissue. The possibility, however, that these
fibres in Balanoglossus are nervous is supported—firstly, by
the fact that they always taper inwards and not out-

THE ENTEROPNEUSTA. Al

wards ; secondly, that as a matter of fact, in B. Robinii at
all events, the ectoderm cells may themselves be traced into
tails of this kind; thirdly, the general absence of nuclei in
the “‘ punktsubstanz,” for if these fibres are supporting cells,
nuclei might be expected to be found in their course ; fourthly,
there is an a priori difficulty as to the nerve supply to the
muscles in these animals, for, though the body of some of the
species is very thick, no definite nerve-cords are to be
found crossing the body cavities, with the exception of the
“dorsal roots’? mentioned hereafter. How, then, are the
muscles innervated? It seems, then, at least possible that
the nerve supply is derived directly from the skin, in which
case the fibres leaving the “ punktsubstanz ” naturally suggest
themselves as the transmitting agents. Finally, the view that
these fibres are ectodermic is rendered likely from the fact
that their origin may occasionally be traced from a very high
level in the skin, though the appearance which is sometimes
produced in sections as of their actual continuity with the
undoubted ectoderm cells may not be quite reliable. In a
few instances these fibres appear to anastomose with meso-
blastic elements, though this cannot be quite definitely
affirmed. On the whole, the balance of evidence seems in
favour of the view that they are ectodermic. If this be
correct the skin of Balanoglossus is to be regarded as a
collection of sensory cells ending in long fibres, which may
either be connected to the central nervous system, pro-
bably by the longitudinal fibres of the “ punktsubstanz,”
or may pass directly through this as motor fibres into the
muscles.

The next point relates to the question as to the intervention of
some third cell in their course functioning as a ganglion cell. In
B. Robinii, in which the examination of this subject is most
complete, as stated above, the occurrence of such cells could
not be shown; but this is, of course, by no means conclusive
in face of the antecedent probability of their occurrence. The
*‘ punktsubstanz,” then, would mainly consist of afferent fibres
passing to the central nervous system, and the motor fibres

42, MORPHOLOGY OF

probably pass directly through it. As will be shown in the
next paragraph its distribution agrees with this view.

In the account of the general development the central
nervous system was shown to have arisen chiefly by a solid
delamination from the skin, added to which its anterior, and
to some degree its posterior, ends are being continually
invaginated as growth continues, so that each end is tubular.
This tubular form results not so much from the longitudinal
closure of a tube as from a forward and backward growth of
skin at the extremities of the delaminated cord. Soon after
delamination histological differentiation occurred between the
upper cellular and lower fibrous parts of the cord. While
this was proceeding (2,9. s.) fibrous tissue was deposited to
form the ventral cord at the point of this structure, which
was most anterior (viz. the back of the collar). While this is
proceeding the deposition of similar tissue in the region of the
dorsal cord commences at the posterior attached end of the
central nervous system. Next, the deposition of fibrous tissue
extends itself forwards on to the proboscis, being first laid down
in the dorsal middle line of the proboscis stalk (v. figs. 34 and
35, pkt.). On the appearance of the atrial fold the ventral
and dorsal cords become united by a fibrous ring in the inner
angle of the fold. This ring, therefore, may be supposed to
bring up the fibres from the ventral cord to the central
nervous system, which it enters at its posterior end, together
with the dorsal cord (v. diagram, fig. 65).

The greatest concentration following upon these occurs in
the skin of the base of the proboscis. In the larva with four
gill-slits (fig. 99, P. rg.) it is already well marked. Concen-
trations are formed in the line of the gill-slits (figs. 72a and
104), and slight fibrous anastomosing tracts run irregularly,
following the line of the wrinkles from both the dorsal and :
ventral cords. These wrinkles taper towards both the cords
and are permanent, being, in fact, limiting lines between
patches of glandular cells.

Now, all these tracts of fibres are thickened as they approach
the central nervous system, and dwindle peripherally. If this

THE ENTEROPNEUSTA. 43

diminution were due to the continual separating of efferent
fibres from the cords it would reasonably be anticipated that it
would be greatest in the case of those parts of the body which
lie behind the collar (i. e. behind the central nervous system) ;
for these cords have almost the whele body to supply, but,
on the contrary, it is the nervous sheath of the proboscis
which presents the greatest concentration, and this continually
thickens on approaching the collar, though the proboscis is
conical and its base is towards the collar. This may be
taken to show that this sheath of nerve-fibre is afferent, and
is continually increasing in thickness owing to the incoming of
sensory fibres from the ectoderm cells lying above it. Its
sudden increase on the proboscis stalk is due to the sudden
tapering of the base. This feature is particularly well seen in
B. minutus. On any other hypothesis it would seem un-
likely that this great deposition of nerve-fibre should occur in
a region which is generally covered up by the anterior folds of
the collar.

The Central Nervous System.—The changes occurring
in this structure in B. Kowalevskii after its separation
consist in an increase in size and in histological differentiation.
As the result of these changes its anterior end comes to have
the structure shown in fig. 60. Among the cells lining the
anterior end of the lumen are always some few gland-cells.
The cellular part of this cord is continuous, of course, with
the cellular part of the skin, and the fibrous part or white
matter, as we may call it, with the fibrous layer of the skin.
Behind the lumen it has the appearance shown in fig. 78.
The white matter does not enclose the upper part of the cord.
Above it are a number of pyriform cells, probably ganglionic,
whose tails project into the white matter. Central to these
the cells are more or less irregularly grouped into strands
enclosing spaces. The histology of this central part of the
cord is very difficult, and I have not been able to determine
how these spaces are filled. In B. minutus (v. fig. 67) they
are so definite as to make it certain that they are not due to
reagents. :

44, MORPHOLOGY OF

Among this loose tissue of the centre of the cord are remark-
able stellate groups of cells (fig. 78, sted.) whose heads are thus
placed radiating from a small lumen, which is_ generally
sharply defined on three sides and usually irregularly bounded
at some part of its margin. The nature of these stellate
groups did not appear. They are commonest in the sides of
the “ grey” tracts, viz. at the points where the white matter
is bent up (v. fig. 74, 6). It is possible that the spaces thus
enclosed may in some indirect manner communicate with the
neural tube.

The histology of the cord is nearly the same in all the
species. In B. salmoneus and B. Brooksii, however, there
is always a quantity of yellowish granules embedded in the
central substance (on the analogy of Nemertines this substance
may function like hemoglobin). The shape of the cord in
section varies in the different forms and in different parts of
its course (v. fig. 74).

From the lower surface of the white matter of all species
many fibres may be seen leaving the cord and losing them-
selves among the subjacent muscular tissues. In B. Kowa-
levskii alone no connection exists between the dorsal side of
the cord and the skin. In B. minutus this is accomplished
by three cords of skin substance. Their outsides are covered
with a fibrous sheath (Spengel), and this is in connection with
the fibrous layer of the skin. As Spengel has stated, these
cords contain a more or less distinct lumen. I have not been
able to trace these out upon the skin, though they occasionally
appear to lead to the cavities enclosed by the radiating cells.
These cords I propose to term the dorsal roots. They occur
in B. minutus, Robinii, salmoneus, and Brooksii.
Their homology will be discussed when the other morpho-
logical questions arising out of these facts are treated of.

The histology of the rest of the nervous system has been
sufficiently described.

The relations of the parts are explained ie figs. 60, 64, 65,
67, 73, &e.

There are no special sense organs.

THE ENTEROPNEUSTA. 45

As the “dorsal roots” do not occur in B. Kowalevskii
their development has not, unfortunately, been observed.

The Hypoblastic Structures.

The notochord has been described already, as also the
mode by which the mouth comes to be anteriorly directed.

The cavity into which the mouth leads is lined by very thick
walls (figs. 90, 67, &c.), composed of long cells supported by
some intracellular substance, probably the same as that of the
notochord. In B. Kowaievskii it leads continuously into
the branchial chamber, but in the other species, in which the
branchial chamber is separated by longitudinal ridges (fig. 91),
from the lower cavity of the branchial region (which thus has
the well-known figure-of-8 shaped cavity). The anterior end
of the branchial cavity comes to be almost enclosed in the
pharyngeal cavity. As the result of this on either side the
branchial cavity projects as two blind horns, which are en-
closed in the pharyngeal cavity.

The structure of the gill-slits has been sufliciently described
by Kowalevsky, Agassiz, and Spengel.

To these accounts there is little to add. The figures 84 and
85 illustrate the mode by which their final structure is
attained. It is practically impossible to follow their structure
by means of transverse sections, but longitudinal sections and
surface-views make them easily intelligible. Each gill-slit of
B. Kowalevskii is U-shaped and surrounded by a skeletal
secreted structure, as shown in fig. 85. In my last paper I
stated that, though the origin of these structures was uncertain,
the balance of evidence favoured the view that they were hypo-
blastic. Since the above was written I have been led to
regard them as more probably mesoblastic, owing to some of
the appearances since observed. It should be noticed that the
body cavity is continued into the valves always, but never into
the bars separating adjacent gill-slits in which the bordering
bars are in contact. ‘This is due to obliteration of the cavity
by the skeletal bars. This feature is very useful in distin-
guishing these parts in sections.

46 MORPHOLOGY OF

The atrial cavity must be described in this connec-
tion. As stated in the general account, its origin is due to
the backward growth of the collar-fold to form an operculum.
In B. Kowalevskii (v. fig. 88) it is more marked than in
B. minutus, but in B. salmoneus the collar-fold does not
reach as far as the first gill-slit, which consequently opens
directly to the exterior (fig. 107). In B. Kowalevskii it
covers about three gill-slits. (In fig. 88 only one gill-slit is
thus shown; this is owing to the slight obliquity of the
section. )

The relation of the opercular fold in B. minutus is shown
in figs. 73 and 104.

The dorsal wall of the branchial chamber is thickened in
the middle line to form a ridge (figs. 89 and 92). This ridge
contains a groove in its posterior part. It is no doubt a sup-
porting structure, and may conceivably be homologous with
part of the backward extension of the notochord in other |
Chordata.

The digestive tract follows upon the branchial region.
The branchial chamber ends in a short blind sac above it,
and it is in this sac that the new gills are added after three
pairs are formed (v. fig. 44). The walls of the digestive tract
in B. Kowalevskii are thrown into an irregular spiral fold
(v. figs. 82 and 108), which is not continued into the intes-
tinal region as a definite feature.

The cells of the digestive region are arranged (fig. 82) in a
single layer for the most part. They contain large granules
and bear a few long cilia. In the walls of the gut in this
region are numerous blood-vessels. The lumen of the gut in
this region varies greatly in size, probably with the digestive
processes (cf. Salensky, loc. cit.), the liver being in
B. Kowalevskii occasionally obliterated.

In B. Kowalevskii there is no distinct sacculation to
form the liver, but in B. minutus the dorso-lateral walls of
the digestive region are pushed out to form the characteristic
liver outgrowths. These structures are not regularly paired.
Their walls are full of secondary foldings (v. fig. 93). The

THE ENTEROPNEUSTA. A

cells lining these folds are similar to those of the digestive
tract, containing large granules and fluid-looking vacuoles.

The skin covering these liver-saccules is very thin, and in
B. salmoneus it may often be seen fused with the
hy poblast, forming openings which place the cavity
of the liver diverticula into actual connection with
the exterior. The histological appearances are such as to
leave no doubt that an actual fusion occurs. When the
extreme softness of the tissue is remembered, it seems likely
that these perforations may, in the first instance, be due to
wounds which have healed so as to form fistulae. [In a single
case of B. minutus a fistula of this kind was found forming °
a perforation from the intestine to the body cavity. In this
animal the fusion between hypoblast and mesoblast was quite
complete. ]

The liver of B. salmoneus is dark green in colour, and this
colour is due to minute round granules or drops in the hypo-
blast. In B. Robinii the tint is generally dark brown.

The histology of the intestine, which is usually more or
less diamond-shape, two of the angles being dorsal and ventral,
isin no way remarkable. From the first the wallis formed
of asingle layer of cells, ciliated, and smaller than those of
the digestive region (v. fig. 83). The anus opens imme-
diately above the tail until this structure disappears, and then
it opens widely in a terminal position (v. figs. 83 and 6).

The Tail and Anal Lappets.

The tail is present in the period between one and eight pairs

of gill-slits. Its skin is full of unicellular glands. The third

_ pair of body cavities are prolonged into it, and the mesentery

between them remains. The anal lappets (fig. 38, a) also dis-
appear with the tail.

48 MORPHOLOGY OF

Mesoblastic Structures.

Muscles.—The muscle-fibres of the proboscis are not
gathered into bundles. They consist of circular, radial, and
longitudinal fibres. The circular fibres are few in number, and
chiefly occur in the external parts of the middle third of the
proboscis.

The radial fibres are very few in B. Kowalevskii, but in
B. salmoneus and B. Robinii they are common, and have
a very characteristic appearance (v. fig. 94, a). Their peri-
pheral ends are very long and fine, occasionally branching.
Their central ends taper suddenly from a thick part containing
a nucleus to a very fine fibre. These fibres are always plain
fibres. Probably the peripheral ends are inserted into the
skin, and the central end into the meshes of connective tissue
which permeate the body cavity (v. fig. 79).

The longitudinal fibres of B. Kowalevskii are arranged in
concentric rings, and united to each other by a peculiar con-
nective tissue, which contains stellate cells with large nuclei.
These concentric rings seem to be more numerous in old than
in young animals, reaching the observed maximum of eight.
This concentric arrangement is not a distinct feature until
adult life is nearly reached. These fibres appear in section to
have the same structure as those shown in fig. 94, 6, which is
taken from B. Robinii. The muscles of B. Kowalevskii
were unfortunately not examined in the fresh state.

In B. minutus the longitudinal muscles do not form such
definite concentric rings as in B. Kowalevskii, but all the
mesoblastic tissues filling the proboscis cavity are broken upin
preserved specimens into radial segments. This is not the case
in living B. Robinii, and hence is probably due to reagents
in B. minutus; as, however, I have never had an opportunity
of seeing the latter in the fresh state this cannot be affirmed.

In passing inwards from the outside to the centre of the
proboscis the structures are thus arranged :

THE ENTEROPNEUSTA. 49

1. Ectoderm.—Ciliated tailed cells.
Glandular cells.
Nerve-fibres as a layer.
Basement membrane.

2. Narrow tissue space crossed by ingoing fibres from ecto-
derm, and by supporting fibres in all directions, together with
a very few circular fibres (v. fig. 51).

3. Tract densely filled with radial and longitudinal muscles
(in B. Kowalevskii concentrically disposed. in rings) and
connective tissue.

4. The tissue space into which the central organs project.

5. The central organs: ©

(a) Proboscis gland with its sac.
(0) Heart.
(c) Notochord.

The muscles of the collar body cavity in B. Kowalevskii
are not gathered into bundles or definitely arranged, excepting
those which are attached to the lateral rods of the axial skeleton
(fig. 60). These large muscles are inserted into the back of
the collar. The whole cavity between the pharynx and the
skin, being originally second pair of mesoblastic pouches, be-
comes obliterated, being filled with muscles and connective tissue.

In B. salmoneus, B. Robinii,and B. Brooksii this also
occurs, but in B. salmoneus (fig. 106) the longitudinal
muscles are grouped into bundles. These bundles form two
series, the one on the somatic and the other on the splanchnic
side, and in the narrower parts of the cavity the groups of the
two series dovetail into each other (fig. 106), being each
gathered around a connective tissue septum projecting into the
cavity.

These fibres in B. Robinii occasionally, after osmic acid,
show a slight striping (fig. 94, c).

In B. minutus the longitudinal muscles of the collar lie in
a layer immediately under the skin and under the pharyngeal
wall. The cavity is crossed by many radial fibres, upon which
some cells are placed, but is not so much filled up as in the
other species,

50 MORPHOLOGY OF

The muscles of the third body cavity are not markedly
different from those of the collar. In B. Kowalevskii alone
a large muscular band runs along each side of the ventral
nerve-cord, forming a projection from the body (v. fig. 108).

The perihemal cavities are similarly almost filled with
tissue, and always contain more or less longitudinal muscle-
fibre. These are gathered into two bundles, and are inserted
into the notochord sheath in the proboscis stalk. They are
most developed in B. minutus, &c. (w. figs. 67 and 68).

The Mesenteries.—The dorsal mesentery persists through-
out life in B. Kowalevskii and B.salmoneus. In the
other species it disappears inthecollar region. The
ventral mesentery persists in the trunk in all species, but
is always obliterated in the collar.

In B. minutus the body cavity of the trunk in the hepatic
region is again divided in consequence of an attachment
between the lateral angles of the diamond-shaped intestine to
the body wall (v. fig. 93). In this position two large lateral
vessels run.

As Spengel has stated, strands of connective tissue run in
B. minutus from the body wall between the follicles of the
ovaries, forming a sort of radial septa. These septa are pro-
bably not of morphological importance, beyond indicating the
“ accidental” way in which such septa may arise (cf. Poly-
gordius, &c.).

All the body cavities are full of corpusculated fluid, as Spengel
has observed. These corpuscles, when living, are full of bright
granules and vacuoles, and exhibit amceboid movements.

The Proboscis Gland.—In B. Kowalevskii (fig. 47,
gis.), at about the age of two gill-slits, a space appears in the
proliferation of mesoblast lying dorsal to the anterior end of
the notochord, when the latter is pushed forwards into the
anterior body cavity. This space is the first rudiment of the
sac of the proboscis gland. Soon after its appearance it be-
comes enclosed in a membrane, which is added first at the
posterior part of the sac (cp. figs. 45, 31, and 47). Its cavity
is therefore a tissue space arising in the wall of the body cavity,

THE ENTEROPNBEUSTA. 51

and it is in communication with the body cavity by means of
the interstices between the cells bounding its anterior end.

Its further development is involved with that of the heart
which had better be now described. The heart arises in ani-
mals with three pairs of gill-slits, as a horizontal split in the
tissue between the notochord and the sac of the proboscis
gland. Its walls are very thin (v. fig. 52). From the first it
appears to contain blood, which is apparently non corpusculated,
and can be coagulated by reagents. Whether the heart is
originally in connection with the dorsal vessel or not could
not be determined. Its walls soon become slightly muscular
(v. figs. 67 and 97), and the pulsations, which can be dimly
discerned through the skin in the living state, are doubtless
occurring in this vesicle.

After the formation of the heart a plexus of vessels in con-
nection with it is formed among the mesoblastic cells covering
the tip of the notochord (fig. 50). As this occurs the cells
standing on the capillaries assume a pyriform shape, the sharp
ends being fixed to the vessels and the wide ends free. These
wide ends acquire a very transparent appearance, as though
filled with fluid (fig.49). These bunches of capillaries eventu-
ally acquire a great development and communicate with two
larger blood-vessels (fig. 58, 6.v.), and with a sinus in the
periphery of the gland.

The sac of the proboscis gland anteriorly becomes filled up
with a quantity of loose tissue, in which some granules of a
yellowish colour are embedded.

In B. minutus these yellow granules are of much com-
moner occurrence (v. fig. 98). The capillaries of the gland are
more regularly arranged.

In B. salmoneus the capillaries are still more regular,
running parallel to each other to the periphery of the gland,
where they are united in a plexus of larger vessels (ep., figs.
95—97). The outer cells of the gland are modified to form a
peculiar tissue (fig. 97). They are large cells, which stain
deeply and have a nucleus usually on their outline. The cells
standing on the capillaries contain some yellow granules, and

52 MORPHOLOGY OF

larger granules or even masses of them are to be found in the
spaces surrounding them.

The gland of the living B. salmoneus is light green in
colour.

The nature of these glands is entirely obscure. These
yellow granules occur amongst nearly all the mesoblastic
tissues. In B. Robinii (collar) they may be found in the
fresh state, presenting the appearance shown in fig. 100. They
are never crystalline.

An attempt was made to investigate the chemical nature of
these bodies, but with only negative results. They may, per-
haps, be excretory, and it is possible that they are more or less
removed by the proboscis pore and collar funnels respectively.
This does not explain their presence in large masses in the
trunk body cavity (v. fig. 93, a), from which no pore has been
observed to open. Occasionally granules of this character
occur in the ectodermic structures, suggesting that they are a
product of the activity of all the tissues.

The proboscis pore was shown to arise at two gill-slits
as a small vesicle in the skin of the proboscis stalk upon the
left side (v. fig. 34) ; at three gill-slits it acquires an opening to
the exterior, and at four gill-slits its tissue fuses with the lining
of the left posterior horn of the anterior body cavity (v. fig. 99),
placing this cavity in communication with the exterior.

In B. Kowalevskii this pore is permanently on the left
side of the body; in B. minutus, &c., it is median.

The collar funnels arise as thickenings in the outer wall
of the arterial cavity opposite the opening of the first gill-slit
(v. fig. 101). These thickenings soon become perforated
(8, g.s.). At their origin they are simple conical funnels, but
they soon acquire a crescentic lumen owing to a thickened
inward folding of their outer wall. This is not conspicuous in
B. Kowalevskii (cp. figs. 88 and 104). Their histology is
sufficiently indicated in the figures.

As previously mentioned, the blood-vessels consist of
(1) a dorsal vessel leading from the heart to the tail; (2) a
ventral vessel running from the back of the collar to the tail ;

THE ENTEROPNEUSTA. 58

(3) in B. minutus a pair of large lateral vessels (v. fig. 93)
in the digestive region. These are connected by plexuses in
the skin and under the epithelium of the gut. In the operculum
this capillary system of the skin forms a more or less definite
circular vessel. In parts of their course these vessels are always
more or less filled with a fibrous-looking substance, apparently
cellular, which lines the walls (fig. 71). The generative
organs lie in blood-sinuses derived from the subcutaneous
plexus.

I stated (Part I) that the branchial blood-supply resembled
that of Amphioxus. From further observation I have come
to the conclusion that this is a mistake, and that the
vessels supplying the gills are all derived from the dorsal
vessel, as Spengel has stated, being, in fact, merely the skin
capillaries of the dorso-lateral regions. The main vascular
trunks are all formed from the mesoblast of the first cavity
and of the third pair of cavities. The capillaries under the
skin and round the gut are formed in sitt in the mesoblastic
walls in which they occur.

The Generative Organs.

The Ovaries.—The animals are all dicecious. The origin
of the ovaries is not certain, but there is very strong evidence
that they are epiblastic. At all events, from almost their
earliest appearance, they are connected with the skin in the
dorso-lateral regions (ov. fig. 110). It is almost impossible to
believe that an attachment of this kind is secondary, and I
have never seen an ovarian follicle entirely separate in the
body cavity.

Soon after its appearance it consists of a mass of loose round
cells. A cavity next appears in its interior, as though due to
a disintegration, and after the appearance of this cavity the
cells bounding it develope into ova (figs. 111 and 112).

The egg-shell appears soon as a close-fitting membrane. The
germinal spot is enclosed in a remarkably tough membrane in
all the species examined. Though the ovaries are connected

5

54 MORPHOLOGY OF

with the skin by ducts the ova are dehisced by the breaking
away of whole follicles, which then disintegrate. In the
branchial region of B. minutus there is a general corre-
spondence between these ducts and the gill-slits, as Spengel
has observed.

The testes are lobed masses placed in the same situation as
the ovaries. The outer zone of each testicular follicle is made
up of spherical cells (figs. 108 and 109, a), which contain several
(? eight) deeply-stained dots. These cells are young spermato-
blasts, and the dots, which increase in size in the spermato-
blasts of the inner zone, are the heads of spermatozoa which
are finally set free into the central cavity. Here they are
arranged in curious strings, which wave above parallel to each
other in preserved specimens (fig. 108). The testes, when
mature, break up in B. Kowalevskii as masses, but in B.
Robinii they exude from the skin as a yellow slime.

Mucus.—All the species secrete vast quantities of mucus
when irritated. That of B. Robinii sets to form a mass of
tough consistency, which collecting grains of sand forms a sort
of tube. In this the animal can move slightly. The body of
this species is very flat in the generative region, and is
naturally folded up dorsalwards within the tube. The mucus
of this form, which comes out after prolonged irritation, turns
to a reddish-violet colour on exposure to the air, which is
very characteristic.

In B. Brooksii, Robinii, and salmoneus the sides of
the body are produced dorsalwards into flaps which nearly
meet in the branchial region, and thus cover the gill-slits and
dorsal nervous system.

THE ENTEROPNEUSTA. 55

List or Papers RErerReD To.

. Acassiz, ALEX.—“ Hist. of Balanoglossus and Tornaria,” ‘Mem. Amer.

Acad.,’ vol. ix.

2. Batrour, F. M.—‘ Monograph on Elasmobranch Fishes,’ 1878.

3. Bareson, W.—“ Early Stages in Dev. of Balanoglossus,” ‘Quart. Journ.

11.

12.

Mier. Sci.,’ April, 1884. (‘Studies from the Morphological Labora-
tory,’ vol. ii, part i.)

. Hatscnex, B.—“ Stud. tb. Entw. d. Amphioxus,” Claus’s ‘ Arbeiten,

Wien, 1881.

. Hatscuex, B.—“ Mitth. tib. Amphioxus,” ‘ Zool. Anz.,’ Sept. 29th, 1884.
. Kowauevsxy, A.—‘ Anatomie des Balanoglossus,” ‘Mem, Acad. Imp,

Sci.,’ St. Petersburg, 1866.

. Lanxester, E, Ray.—‘ New Points in Structure of Amphioxus,”

‘Quart. Journ. Mier. Sci.,’ xv, p. 257.

. Mrrscunixorr, H.—‘‘ Ueb. d. Metam. einiger Seethiere,” ‘Z. f. W. Z.,’

20, 1870.

. Metscunixorr, E.—‘ Ueb. d. Syst. Stell. v. Balanoglossus,”’ ‘Zool-

Anz.,’ 1881.

. Scott, W. B.—‘ Beitrage z. Entw. d. Petromyzonten,’’ ‘ Morph. Jahrb.,’

Bad. vii.

SPenGEL, J. W.— Bau. u. Entw. v. Balanoglossus,” ‘Tag. d. Naturf.
Ver.,’ Miinchen, 1877.

Spence, J. W.—‘Z. Anat. d. Balanoglossus,” ‘ Mitth. a. d. Zool. Sta.
z. Neapel.,’ Bd. v, Hft. 3 and 4.

56 MORPHOLOGY OF

EXPLANATION OF PLATES I—VI,

Illustrating Part I of Mr. Bateson’s Memoir on the “ Mor-
phology of the Enteropneusta.”

Complete List of Reference Letters.

a. Anus. ce. Archenteron. 4/. Alimentary canal. a. 7. Anal lappets.
bc.', 7, *, The anterior, middle, and posterior body cavities respectively. The
letters 7 and 7 affixed to these letterings denote that the parts are of the left
or right side. 47. Branchial chamber. C. Apical tuft of cilia. ©. WV. §. Cen-
tral nervous system. c.g. Groove between the collar and the trunk. il.
Transverse band of cilia. Circe. Circular muscle-fibres. C/. Collar. Cl'. The
posterior fold of the collar, which eventually forms the operculum. CV. s/. Skin
of collar. D. d.v. Dorsal blood-vessel. D. mes. Dorsal mesentery. D. x. s.
Dorsal nervous system. dig. Digestive tract of alimentary canal. #2. Ecto-
derm, ex. g. Refractive granules in mesoblastic cells. 4 Mesoblastic fibres.
g. s. Gill-slit. gy. s. 7. Branchial supporting rod. gl. Proboscis gland. gi. s.
Sac of proboscis gland. gz. Ganglion (?) cells. H. Hypoblast. A¢. Heart.
Int. Intestine. J. 7. Lateral rods of the skeleton. 7. Lumen of notochord.
M'. M". M'". Mesoblast derived from the anterior, middle, and posterior pouches
respectively. Jo. Mouth. msc. Muscle-fibres. mz. Mucous gland. x. s.
Nervous system. x. cz/. Neural canal. We. pr. Pore by which the notochord
lumen opens into the pharynx. ch. Notochord. WV. pr. Neural pore. Op.
Operculum. ph¢. Fibrous substance of the nervous system. P. pr. Proboscis
pore. P. h. b. c. Perihemal body cavity. P. sé. Skin of proboscis. Sep.
Septum between the horns of the anterior body cavity. Sh. Sheath of
notochord. Sr. Sucker. S%. Skin. §.7. Supporting rod. Sp. Tissue-space
in the proboscis cavity. 7¢s. Testis. V.vs. Ventral vessel. V. x. s. Ven-
tral nervous system. Viv. Valve of gill-slit. «. Pyriform cells of splanch-
nopleure of middle body cavity.

With the exception of Figs. 5 and 6, the outlines were all drawn with
Zeiss’s camera lucida. I have to thank Mr. Edwin Wilson of the Lithographic
Department of the Cambridge Scientific Instrument Company for drawing for
me two beautiful figures (Figs. 5 and 6) of the whole animal from preserved
specimens, and also for Figs. 1, 2,3, and 4, which he has prepared for me
from my own sketches, the original outlines of which were traced from living
specimens.

Fre. 1.—Whole animal seen from the side, immediately after the appearance
of the second pair of gill-slits. (Obj. A, oc. 2.)

THE ENTEROPNDEUSTA. 57

Fic. 2.—Similar view of an older animal, with two gill-slits. (Obj. A, oc. 2.)

Fig. 3.—Similar view of a three-gill-slit larva. (Obj. A, oc. 2.)

Fic. 3a.—Posterior end of the same in a retracted state, seen in profile
(from a preserved specimen). (Obj. A, oc. 2.)

Fic. 4.—Side view of larva with five pairs of gill-slits. The fold of the
operculum covering part of the first gill-slit is semi-transparent. (Obj. A, oc. 2.)

Fic. 5.—Side view of preserved specimen with nine pairs of gill-slits
Owing to the contraction of the body and the protrusion of some of tke
valves, few only of these are visible. (Obj. AA, oc. 2.)

Fic. 6.—The adult animal (¢). (x 2 diameters.)

Fic. 7.—Longitudinal vertical section (not quite median) of a larva in
Stage F. (Obj. C, oc. 2.)

Fie. §.—Cells of mesoblast of anterior pouch, from a transverse section of
a larva shortly after this pouch is closed off from the hypoblast. (Obj. F,
oc. 2.)

Fie. 9.—The same tissue from the posterior third of the proboscis of a
rather older larva, showing the proliferation and commencing differentiation of
the mesoblastic elements. (Obj. D, oc. 2.)

Fic. 10.—Section similar to Fig. 7, from a rather older larva to show
anterior, dorsal structures (StageG—H). (Obj. D, oc. 2.)

Figs. 11—20 represent transverse sections of a larva in Stage G—H, with
the exception of Figs. 12 and 16, which were drawn under Obj. F, oc. 2. All
these figures were drawn under Obj. CC, oc. 2, In most of these figures the
ectoderm is only indicated ona short are of the circle of the whole section.
They are numbered from before backwards.

Fie, 11.—Transverse section of proboscis cavity. Mesoblastic elements
distributed nearly uniformly all round the interior.

Fic. 12.—Portion of these mesoblastic elements more highly magnified, to
exhibit the differentiations.

Fie. 13.—Section taken behind Fig. 11. The mesoblastic layer is thinner
dorsally than elsewhere.

Fie. 14.—In this section the septum separating the two horns of the cavity
is reached.

Fie. 15.—Still further back the first rudiment of the proboscis gland is
reached (cp. Fig. 29).

Fic. 16.—The rudiment of the gland lying in the septum more highly
magnified.

Fic. 17.—Section across the proboscis stalk. The anterior end of the
notochord is reached. The back of the gland is nearly passed. The meso-
blastic horn of the left side is apparently divided into two parts, this appear-
ance is due to shrinking.

Fig. 18.—The lumen of the notochord is reached. The extreme ends of the
two anterior mesoblastic horns and the posterior apex of the left horn from

58 MORPHOLOGY OF

the first cavity are all cut in this section. (The irregular folds of skin
are due to the contractions of the body; it will be understood by comparing
the figures of the whole animal that only the central part within the folds is
the real stalk of the proboscis.)

Fic. 19.—The section crosses the end of the archenteron, lying in front of
the mouth. (The whole of this part of the archenteron becomes eventually
pushed forward to form the notochord.)

Fic. 20.—The mouth is here traversed, as also the anterior end of the
nervous system. (A space, due probably to shrinking, is visible in the dorsal
mesentery.)

Figs. 21—29 are transverse sections of a larva, slightly older than the fore-
going. Fig. 25 was drawn under Zeiss’s Immersion 2 and oc. 2 ; the others were
drawn under Obj. D and oc. 2. They are numbered from before backwards.

Fie. 21.—Section taken just behind the mouth, The lumen of the noto-
chord is here shut off from the archenteron.

Fic. 22.—The notochord still open to the archenteron.

Fic. 23.—The nervous system is attached to the skin.

Fic, 24.—The nervous system is already nearly separated from the skin.

Fic. 25.—Part of the foregoing enlarged, to show the peculiar pyriform
cells of the splanchnopleure (c).

Fie. 26.—Nervous system still in the skin.

Fie. 27.—The nerve-cord is separated from the skin.

Fic, 28.—Two parts of the middle body cavities may be here seen separat-
ing from the rest, probably forming part of the perihemal cavities.

Fie. 29.—The anterior ends of the third pair of body cavities are here cut
as a solid mass of mesoblast on each side.

Fie. 30.—A longitudinal vertical, nearly median, section of a larva, in the
same stage as that shown in Figs. 21—29. The differentiation of the walls of
the digestive tract may be here seen.

Figs. 31—44 are from transverse sections of a larva which has just
acquired the second pair of gill-slits. They are numbered from before
backwards.

Fre. 31—Transverse section of the proboscis cavity. The loose tissue in
the sac of the gland is shown. The membrane which is deposited round it is
visible (cp. Fig. 47). (Obj. D, oc. 2.)

Fie. 32.—Small portion of skin and mesoblast on a larger scale, to show
structure of the nervous layer. (Obj. F, oc. 2.)

Fie. 33.—Region behind that shown in Fig. 31. (Obj. D, oc. 2.)

Fic. 34.—Through the proboscis stalk. The anterior horns of the middle
body cavities are here cut.

Fic. 35.—Through the proboscis stalk and the anterior phlange of the
collar which forms the lower lip.

THE ENTEROPNEUSTA. 59

Fie. 36.—Through the anterior end of the nervous system (in the region
of the mouth), showing the manner in which the lumen arises in the
nerve-cord.

Fic. 37.—Through the anterior part of the collar, showing the nearly com-
plete separation of the notochord with the hypoblast.

Fie. 38.—Section taken behind the previous one. The nervous system is
here separated from the skin.

Fic. 39. Through the junction of the lumen of the notochord with that of
the gut. The skeletal rods, which here bend downwards, backwards, and
then slightly forwards, are therefore cut twice on each side.

Fig. 40 is taken rather in front of the branchial sacs; it shows the two
anterior horns of the posterior body cavities, which form the perihemal
cavities.

Fig. 41.—Section through the gill-sacs in front of the clefts, The nervous
system here is fused to the skin dorsally.

Fic. 42.—Through the extreme posterior end of tne second body cavities,
and the anterior end of the ventral blood-vessel and nervous cord.

Fic. 43.—(Section not quite transverse.) Through the left gill-slit.

Fic. 44.—Section taken through the extreme posterior end of the branchial
region of the gut, showing how this overlaps the digestive region. The dila-
tation in the sides of the branchial region here shown are parts of the second
pair of gill-slits.

Fic. 45.—A longitudinal vertical section of the whole animal. (Two gill-
slits.) (The section is not truly vertical, as it cuts the gill-slit.) It exhibits
the relation of the notochord and other parts. (Obj. CC, oc. 2.)

Fie. 46.—By an oversight no figure bearing this number appears in the
plates. The figure referred to as such in the text at pages 85, 98, and 103,
is Fig. 45.

Fic. 47.—A transverse section of the extreme tip of the notochord, &c.,
in the larva from which Figs. 31—45 were taken. It shows the histology of
the notochord and the arrangement upon it of the mesoblastic tissues, which
are pushed in by it. (Obj. D, oc. 2.)

Fic. 48.—Transverse section across the proboscis pore of a larva with three
gill-slits. The two skeletal rods are here fused and have attained a considerable
size. (Obj. F, oc. 2.)

Figs. 49—53 represent transverse sections taken from B. Kowalevskii,
at the stage of four pairs of gill-slits. (When the remaining structures are
dealt with a fuller explanation will be given of these and of the subsequent
figures ; on the present occasion they are only introduced to explain the account
of the notochord given in the text.) Figs. 49—53 are numbered from before
backwards.

60 MORPHOLOGY OF ©

Fies. 49 and 50 illustrate the histology of the anterior end of the notochord,
(Obj. F, oc. 2.) These sections are in front of the heart.

Fic. 51 shows a section of the whole proboscis, to illustrate the relations of
the parts. The body cavity may be seen to be nearly filled up with muscle-
fibres and connective tissue, with the exception of a small central space, in
which lies the notochord bearing the central mesoblastic structures, viz. the
heart and proboscis gland with its sac. (Obj. C, oc. 2.)

Fic. 52 exhibits the central structures of Fig. 51. (Obj. F, oc. 2.)

Fic. 58.—Here the lumen of the notochord is reached. At this point the
sac of the proboscis gland is attached dorsally to the skin of the proboscis.
(Obj. F, oc. 2.)

Fig. 54.—Transverse section of proboscis stalk of an older animal, taken
behind the epiblast sac of the proboscis pore, the back of which appears in
section. The skeletal rod has still in this region two central “cores.” (Obj.
D, oc. 2.)

Fic. 55.—A transverse section of the notochord of an adult B. Kowa-
levskii, taken in front of its lumen (cp. Fig. 57). (Obj. CC, oc. 2.)

Fie. 56.—Transverse section of the same as foregoing, through the dilated
lumen in front of the skeletal rod.

Fic. 57.—Longitudinal vertical median section of the back of the proboscis
and the front of the collar of B. Kowalevskii, to show the relation of the
notochord. The other parts arediagrammatic. (Obj. A, oc. 2.)

Fic. 58.—Transverse section of the junction of the proboscis stalk with
the collar. (Semi-diagrammatic.) Exhibits relations of the notochord and its
sheath to the skeletal rods, nervous system, &c. (Obj. A, oe. 2.)

Fie. 59.—A diagram of a transverse section taken through the anterior
region of the collar, to show the relation of the parts figured in Fig. 60.
(Obj. A, oc. 2.)

Fic, 60.—The dorsal structures of the anterior part of the collar shown in
Fig. 59, viz. the nervous system, notochord, perihemal cavities, &c. (The gut
is torn away from the skeletal rod, presumably by shrinking.) The lateral rods
are here connected to the notochordal sheath. (Obj. D, oc. 2.)

Figs. 61—63 show transverse sections of parts of B. minutus.

Fic. 61.—The notochord in transverse section in front of its lumen. (Obj.
CC, oc. 2.)

Fic. 62.—The same in the region of the proboscis stalk.

Fic. 63.—Through the notochord, &., in the front of the collar, showing
the thickened dorsal half of the organ. (Obj. CC, oc. 2.)

THE ENTEROPNEUSTA. 61

EXPLANATION OF PLATES VII—XII.

Fies. 64—112,

Illustrating Part IL of Mr. Bateson’s Paper ou “ The Mor-
phology of the Enteropneusta.”

Complete List of Reference Letters.

a. Anus. ai. Alimentary canal. a¢. Atrial cavity. dc. 1, 3, 3. The anterior,
middle, and posterior body cavities respectively. 4g. Rods bordering the gill-
slits. 7. cls. Border cells of proboscis gland (B. salmoneus). 4. v. Blood-
vessel. C. N.S. Central nervous system (?.e. the cord of the collar region).
Cap. Capillaries of proboscis gland. Cvzre. Circular muscle-fibre. CU. f. Collar
funnel. C.7g. Ring of nervous tissue round the collar. D. d. v. Dorsal
blood-vessel. D. mes. Dorsal mesentery. J. z.s. Dorsal nervous cord.
D.r. Cords connecting central nervous system with the skin. D. rdg. Dorsal
ridge of hypoblast in branchial region. dig. Digestive region of alimentary
canal. #. Ectoderm. . Fold in wall of collar funnels. 4. s. Gill-slit.
g. 8-1, 7, First and second gill-slits respectively. yg. sc. Lining of gill-sac.
g. sr. Supporting rods of gills. gy. vs. Germinal vesicle. g. sp. Germinal spot.
gl. Proboscis gland. gi. s. Sac of proboscis gland. gz/. Granules in central
nervous system of B. salmoneus. gv. Granules, probably excretory.
At, Heart. ié. Intestine. 7. 4. v. Lateral blood-vessel. 7. rdg. Lateral
ridges separating the branchial chamber from the lower cavity of the gut in
the branchial region. /. msc. Longitudinal muscle-fibres. Zv. Liver. mm. spz.
Spermatoblast cells. Mo. Mouth. msc. Muscle-fibres. mw. Mucous glands of
skin. mu’. Goblet cells of skin. mz’. Long glands of collar skin of B
Kowalevskii. 1. cl. Neural canal. MWch. Notochord. J. pr. Neural
pore. 2. sh. Nervous sheath of proboscis. 0. Opening of collar pores.
Op. Operculum. ov.? Ingrowth of skin, probably an ovary. ov. Ovarian
follicle. ph. Pharyngeal region of gut with thick walls. yer. Perforation
into liver saccule. p&t. Fibrous substance of the nervous system. P. rg.
Ring of nervous tissue round proboscis. ph. c. Perihemal body cavity.
Sci. Liver saccule. Sf Surface of skin with anastomoses of ectoderm cells.
Sk. Skin. Shr. Sucker. S. 7. Supporting rod of notochord. Sp. viv. Spiral
fold in wall of gut in the digestive region. S¢. Stripes occasionally seen in
preserved muscle-fibres. Sve/. Stellate masses of cells in central nervous
system. 1. pr. Tube of proboscis pore. ¢s. Testis. V. 6d. Ventral band of
longitudinal muscle of B. Kowalevskii. V.6.v. Ventral blood-vessel,

62 MORPHOLOGY OF

V. g. Nervous concentration in the line of the gill-slits. 7. msc. Ventral
muscles. Vlv. Valve of gill-slit. V.7z.s. Ventral nervous cord.

Fie. 64.—Diagrammatic longitudinal vertical section of B. minutus, to
show the arrangement of the nervous system. [The openings of the
gill-slits are indicated, though of course not visible in a section of this kind. ]

Fic. 65.—Diagram of nervous system of B. Kowalevskii as seen
from the dorsal surface. The ventral cord and the ring round the pharynx
are indicated in broken lines. The sheath of nervous tissue covering the
proboscis is indicated by shading, as though the tissues were transparent.
The gill-slits are shown on one side only.

Fies. 66—73 illustrate the structure of the skin and nervous tissues of B.
minutus.

Fig. 66. Nearly median longitudinal vertical section of the middle third
of the central nervous system, showing origin of two of the cords
connecting central nervous system with the skin. Their union with
the skin is not here shown. (v. Fig. 68.) Obj. A, long tube, oc. 2.

Fig. 67. Longitudinal vertical section through the side of the central
nervous system, showing the relation of the neural and proboscis pores
to each other, &c. The wall of the heart is cut in this section.
As the section is taken through the side of the central nervous system
its continuation into the dorsal nerve-cord is not visible. Obj. A,
oc. 2.

Fig. 68. Transverse section of the central nervous system at end of
neural tube. Obj. A, long tube, oc. 2.

Fig. 69. Transverse section of the central nervous system behind the
neural tube, showing attachment of dorsal cord to the skin. Obj. A,
oc. 2.

Fig. 70. Longitudinal section of the anterior end of the ventral nerve-
cord. Obj. D, oc. 2.

Fig. 71. Transverse section of ventral nerve-cord. Obj. A, oe. 2.

Hig. 72. Longitudinal section of skin in lateral region. Obj. A, oc. 2.

Fig. 72 4. Transverse section of skin in the space between the gill-slits.
Obj. D, oc. 2.

Fig. 73. Longitudinal horizontal section through the back of the collar,
showing the relations of the peripharyngeal nerve-ring.

Fie. 74 (a, 4, c).—Three sections taken through the anterior, middle, and
’ posterior thirds respectively of the central nervous system of B. salmoneus.

Fic. 75.—Section of a wrinkle of the skin of the middle third of the pro-
boscis of B. salmoneus. Obj. D, oc. 2.

Fie. 76.—Teased out osmic acid preparation of the skin of the collar of
B. Robinii. The cells remain attached to each other by their heads. The

THE ENTEROPNEUSTA. 63

network, sf, is formed superficially by the anastomosing heads, each of
the nodes being the head of a cell. Obj. D, oe. 2.

Fic. 77 (a and 4).—Cells of preparation similar to Fig. 76, more separated.
(c) Spindle-shaped cells from lower layer of skin, probably broken off from
cells resembling a and 4. Obj. F, oc. 2.

Fic. 78.—Transverse section through middle third of the central nervous
system of B. Kowalevskii. Obj. D, oc. 2.

Fre. 79.—Longitudinal section of skin of posterior third of proboscis of
B. Kowalevskii. Obj. D, oc. 2.

Fic. 80.—Horizontal section through the skin of the collar of B.
Kowalevskii. Obj. D, oe. 2.

Fic. 81.—Vertical section of the above. Obj. D, oc. 2.

Fic. 82.—Section taken tangentially to the flexure of the body of young
B. Kowalevskii (8, g. s.), showing the spinal folding in the digestive region
of the gut. Obj. B, oc. 2.

Fie. 83.—Longitudinal section of the tail of young B. Kowalevskii
(4, g. 8.). Obj. D, oc. 2.

Fig. 84.—Longitudinal section through the wall of the posterior region of
the branchial sac, showing the relations of the valves and skeletons of the
gills (B. Kowalevskii, 10,9. s.). Obj. A, oc. 2.

Fie. 85.—Diagrams of successive stages in the development of the gill-
slits of B. Kowalevskil.

Fig. 86.—Macerated preparation of the gill-skeleton of B. Kowalevskii.
Obj. A, oc. 2.

Fie. 87.—Longitudinal section of adjacent valve and gill-bar of B.
Kowalevskii. Obj. D, oc. 2.

Fie. 88.—Longitudinal horizontal section through atrial cavity of B.
Kowalevskii in the plane of the opening into it of the collar funnel and
first gill-slit. Obj. B, oc. 2.

Fie. 89.—Transverse section through the back of the branchial sac of
B. Kowalevskii (10, g. s.). Obj. A, oe. 2.

Fie. 90.—Vertical section of pharyngeal wall of B. minutus. Obj. D,
oc. 2.

Fic. 91.—Vertical section of one of the lateral ridges, separating the
branchial sac from the lower part of the branchial region. Obj. A, oc. 2.

Fic. 92.—(a) Transverse section of the dorsal ridge of the branchial region
of B. Kowalevskii. (4) The same of B. minutus. Obj. A, oc. 2.

Fic. 93.—Transverse section through the junction of a liver saccule with
the gut, through the back of the adjacent saccule (B. minutus). Obj. A,
oc. 2.

64 MORPHOLOGY OF

Fic. 93 a.—Longitudinal section of some liver saccules of B. salmoneus
One of these is perforated at its end. Obj. A, oc. 2.

Fie. 94.—Muscle-fibres of B. Robinii (osmic acid preparations). Obj. F,
oc. 2. (a) Three isolated radial muscle-fibres from the proboscis cavity.
(4) Two adjacent fibres belonging to the longitudinal system of the collar.
(c) Two fibres from the same region as (6), which on treatment with osmic
acid show an appearance of striping.

Fie. 95.—Section of the central part of the proboscis gland of B. sal-
moneus, anterior to the notochord. Obj. D, oe. 2.

Fre. 96.—Outer part of the proboscis gland of B. salmoneus, anterior to
the notochord, to show the arrangement of the border cells (07. cls.).
Obj. C C; oc. 2.

Fie. 97.—A radial segmeut of the proboscis gland of B. salmoneus in
the region of the notochord. (For the relations of this tissue wide “ Later
Stages,” &c., Figs. 51 and 52.) Obj. D, oc. 2.

Fie. 98.—Group of cells from the interior of the proboscis sac of B.
minutus. Obj. D, oc. 2.

Fic. 99.—Longitudinal vertical section through the left side of the pro-
boscis stalk of B. Kowalevskii (4, g. s.), to show the internal opening of
the tube of the proboscis pore. Obj. D, oc. 2.

Fie. 100.—Concretions from the living mesoblastic tissues of the 2nd body
cavity of B. Robinii. Obj. D, oe. 2.

Fic. 101.—Transverse section through the collar funnels and first gill-shit
of B. Kowalevskii (10, g. s.). Obj. A, oc. 2.

Fig. 102.—Transverse section through the collar funnels and upper end of
the atrial. cavity of B. Kowalevskii (young adult). Obj. B, oc. 2.

Fig. 103.—Transverse section of B. minutus, passing through the in-
ternal opening of one of the collar funnels. (Hypoblastic structures indi-
cated roughly.) Obj. A, oc. 2.

Fie. 104.—Longitudinal horizontal section of collar funnel of B. minutus
at the level of the opening of the first gill-slit. Obj. D, oc. 2.

Fic. 105.—Transverse section through middle of collar funnel of B.
salmoneus. Obj. B, oc. 2.

Fic. 106.—Transverse section behind Fig. 105. Obj. A, oc. 2.

Fig. 107.—Transverse section through external opening of the collar funnel
of B. salmoneus. Obj. A, oc. 2.

Fie. 108.—Half diagrammatic transverse section of generative region of
male B. Kowalevskii.

Fie. 109.—(a) Spermatoblast cells, forming the outer zone of the testicular
follicle. (6) Spermatoblast cells, forming the inner zone of the testicular
follicle. (c) Spermatozoa in the interior of the follicle. Obj. F, oc. 2.

THE ENTEROPNEUSTA. 65

Fie. 110.—Transverse section of young ovary of B. Kowalevskii (young
adult). Obj. D, oc. 2.

Fie. 111.—Group of ovarian follicles of B. Kowalevskii, older than the
above. Obj. D, oc. 2.

Fie. 112.—Ripe ovarian follicle of B. Kowalevskii. Obj. A, oc. 2.

PLATE XII.
Diagrams.—Skin coloured light blue; nervous system, dark blue;
hypoblast, light red; blood-vessels, dark red; mesoblast, green.
Fic. 1.—Blastosphere.
Fic. 2.—Gastrula.

Fic. 3.—Longitudinal vertical section through gastrula; blastopore being
nearly closed.

Fic. 4.—Ditto ; blastopore closed.

Fic. 5.—Ditto, later stage; mesoblast forming.

Fie. 6.—Longitudinal horizontal section through a somewhat later stage.

Fie. 7.—Longitudinal horizontal section through larva in Stage G.

Fic. 8.—Transverse section of collar of foregoing in plane of line d d.

Fic, 9.—Longitudinal vertical section of Stage H.

Fie. 10.—Transverse section of collar of foregoing in plane of line d d.

Fig. 11.—Longitudinal horizontal section of adult in plane of heart.. (This
plane would not really take in the perihemal cavities, but their relations are
thus made clear.)

Fie. 12.—Transverse section of foregoing in plane of line d d.

Fic. 13.—Longitudinal horizontal section of junction of collar and trunk in
a larva of about 4 gill-slits.

Fic. 14.—Suimilar section to foregoing of adult, showing formation of oper-
culum, atrial cavity, and collar funnels.

Studies M.L. Vol. IIL. P11.

LS

F. Huth, hth Edin™

Studies M.L Vol. Ill. Pl. I.

Fig. 10. Fug. IL

rae 05
Sp PEG, ‘a

ee ant . is e
oe AA ie v NC i
a Vi! 151 a
— AN yee
Mo.
F Huth, lith® Edin™

Studies M.L.Vol. III. PI. II.

ei

P. Huth Lith" Edin®

he SO eae i _ =
"?
q
*

a
i »ge

—

—_

il Oe

-
_
—~

——
a

x
aS
xs)

es

Janta dia” ‘
as? pp

Studies M.L.Vol. III. Pl. IV.

z
c

i

morons

, ra Vf \
Mit}

FP Huth, Lith? Edin™

a 5
Bee s
re is

aD)
Bs tie

~

ae ee

Ty 7

Lae

t
Co

GA, ih
(2 Cit LOD ¢

Utgth
di (itts }8

UF

\ \
AY)
. LA

oN

i!

Li

ZA

W

INSTRUMENT COMPANY.

THE CAMBRIDGE SCIENTIFIC

Stoo aia al | 4 |

LE, AMS Dil Md Le, ii.

Nis

“st
‘o
aod
i
5
o
2
2
s
rs)
>

Studies M.L. Vol. III, PL-VII.

ee
: * . cS IOC econ ie

eee f me " F% FRR aS ; sa okay! -
0-0-0 00-004) 0-009 HO NG 9-9 QOH" Te

} at
Dns. :
Dns
Vins.
Por
LTY-
i (Da 6 ermtias ace
ee RT ep eee He >
aC tel ee ae \

FB Huth, Lith? Edin®

W Bateson dei

Studies M. L Vol. IL, Pl. VII.

eee

)

Ba.

Oger

9
Ore

x

e

was.

Dby,

Fég. 79,

*9
a

¥ Huth, lnth® Edin*

=
ro
5
°
2
2
=
is}

Studies M.L. Vol. III, PL. IX.

7

eg

@
)

My

Bek

! |
ea nae | '"
vee iH x Hise EAL einen Pala h “i AN ee

0,904 € reagoteet fo £900 toffee fon

eee, Lites ioaneeer ee

a%e feskesionnntate ye

TW TON To

Laat

i i}
A \\ AN t l i j
Aa SA NA A Hl

9 96 8 01,
i) 0 ca, 08 a8 ane hae %s,0,00

8

“Anermte

080 syonetatiat Aadgt® a0 Pe Wtean date?

Lig. de.

F Huth, lath? Edir=

W.Bateson del

Studies M.L. Vol.Ill, Pl X.

ich Cap.

F Huth, Lith® Edin

Fig, 102.

fa,

he i

x he OS Se ED >» 3 aS

OIA rH ok Baa TTT

’ Pres he AT beaks / ype ¥
er

/

s

,

W. Bateson del

Studies M.L.Vol. III, Pl XI.

Bee

od
v4
z
=
==
=
=

.
Re bens
a

ee ates UP sies yy ;
a se aie mlaeneeien is en tL 4! \
h\

F Huth, Lith® Edin®

Ne en
cat

Studies M.L.Vol.IIL, Pl. X11.

a a ae
a7"

Pgh 5p PGS. 4, ge.

W Bateson del F Huth Lath” Bdin®

THE ANCESTRY OF THE CHORDATA. 67

The Ancestry of the Chordata.
By

William Bateson, M.A.,
Fellow of St. John’s College, Cambridge.

Tur ANCESTRY OF THE CHORDATA.

Preface.—In view of the facts relating to the structure of:
the Enteropneusta which form the subject of the accom-
panying paper and of those which have preceded it, it seemed
necessary to attempt some analysis of their import and bearing
upon morphological problems, and especially upon the vexed
question of the ancestry of the Chordata.

But at the outset it was impossible to attempt such an
analysis without first clearing the way by a discussion of the
morphologic meaning of Segmentation. Since the Enterop-
neusta are essentially “ unsegmented ”’ animals and the Verte-
brata are “ segmented,” this preliminary discussion was neces-
sary. Moreover, having shown reason for not accepting the
view that the vertebrate segmentation was of such a kind as to
necessitate the existence of a series of segmented ancestors to
account for it, it became also necessary to treat the whole
question of the origin of segmentations of this class upon a
wider basis. This must be the apology for the introduction
into this paper of some matter and speculation not otherwise
immediately relevant to the subject.

The decision that it would be profitable to analyse the
bearing of the new fact in the light of modern methods of
morphological criticism, does not in any way prejudge the
question as to the possible or even probable error in these
methods.

Of late the attempt to arrange genealogical trees involving

68 WILLIAM BATESON.

hypothetical groups has come to be the subject of some ridi-
cule, perhaps deserved. But since this is what modern mor-
phological criticism in great measure aims at doing, it cannot
be altogether profitless to follow this method to its logical
conclusions.

That the results of such criticism must be highly specu-
lative, and often liable to grave error, is evident.

Part ].—TuHe SEGMENTATION or AMPHIOXUS AND THE VERTE-
BRATA, COMPARED WITH THAT OF THE ANNELIDS.

From the time when the theory of descent in some form or
other became generally accepted amongst zoologists, the ques-
tion of the pedigree of the Vertebrates has been the subject of
much speculation and controversy. The amount of attention
which has been bestowed on this question has perhaps been
greater than is warranted by the actual importance of the
problem considered as a contribution to general biology ; but
when it is borne in mind that the question is that of the history
of the human race, the fascination which has been found in it
is not surprising.

Beyond, however, this more sentimental side, there is
another source of special interest to be found within the
terms of the problem itself; namely, that which is afforded by
the obscurity of the solution; for when the relation of any
one group to the rest of the animal kingdom is sought, in most
cases there are some cardinal features of anatomy common to
it and to some other group, which appear to point to some
affinity between them. For example, the structure of the
Tracheata at once suggests Crustacean affinities, while there is
a strong apparent resemblance between the whole Arthropoda
and the Annelids. Even a group so isolated as the Mollusca
has points of obvious harmony with other groups as soon as
the characters of the Trochosphere are known, and similarly
with most other groups. LHach and all of these “obvious ”
resemblances may be illusory, but still they furnish something

THE ANCESTRY OF THE CHORDATA. 69

which, temporarily, is satisfying, and at least provides a point
of departure for criticism. But in the case of the Chordata
there are none of these common features. The three characters
which unite them, the notochord, the gill-slits, and the rela-
tions of the nervous system, are limiting and exclusive, and
without parallel in any forms outside the Chordate group.
So strongly has this fact been felt by many of those morpho-
logists who have already dealt with the pedigree of the group,
that they have practically abandoned the attempt to find
homologies for these features among the Invertebrates ; for it
is impossible to take seriously such suggestions as, for example,
that the notochord may be compared to, generally, the sacs of
the Capitellidz, the “ siphons ” of any of various Invertebrates
the “ giant-fibres” of Earthworms, or the crystalline style of
Anodon, Each of these structures has been in turn suggested,
together with many others, as offering something with which
to compare the notochord. In the same way Semper argues
that the vetebrate gill-slits have an obvious similarity to
certain pores which he has found in the heads of certain
Oligocheta (Nais), while other authors see a striking resem-
blance between them and the Chetopod segmental organ, and
so on.

In seeking, then, for the proximate ancestors of Chordata,
the Chordate features have been disregarded, and another
character of the vertebrate animal has been selected as
offering a more probable basis of operations. The character
which has in this way been chosen as the point of departure is
that of metameric segmentation. By thus setting aside
the questions arising out of the notochord, &c., and speculating
upon the segmentation of the body, the conclusion is soon
reached that some Annelid was the immediate ancestor sought.

This view has found its chief exponents in Dohrn and
Semper, and has been generally supported by Haeckel and by
most of the popular exponents of evolution.

It would be unprofitable to recapitulate here the numerous
morphological difficulties as to the primitive mouth, &c., which
arise if this theory be received. Many objections of this kind

6

70 WILLIAM BATESON.

have been raised and have been variously replied to, and in
this condition the matter rests. By those who support it, it is
assumed that the common feature of segmentation is so binding
and unique a property as to suffice to link together groups
whose morphology is otherwise widely different.

Tn the following pages it is proposed to examine the pro-
priety of employing the character of metameric segmentation
as one of first importance in forming a phylogeny of this kind.
And before referring to the evidence derived from the fact that
the three characteristic features of Chordata are found in
Tunicata and Enteropneusta, which are unsegmented forms, it
will be best first to discuss the meaning of the phenomenon—
“segmentation’’—for if resolved into its elements it will be
found to be by no means a peculiar feature of a few groups,
but rather the full expression of a tendency which is almost
universally present.

The term ‘‘metameric segmentation” has been used to
describe several anatomical features, which reach their highest
development in the Annelids, the Arthropods, and the Verte-
brata. If an attempt be made to reduce this expression to its
simplest terms it appears to mean, in the first place, that cer-
tain organs of the body are serially repeated from before back-
wards, and in the second place that, in the case of the
Vertebrates and Annelids at all events, the body cavity is at
some period of life divided into a series of compartments, each
of which is closed off from its neighbours. But when a more
precise account of this phenomenon is required, and when it
becomes necessary to particularise as to which of the various
organs of the body is thus repeated, difficulty at once arises
from the fact that this repetition is irregular, and even within
narrow limits may vary considerably. In the case of many of
the errant Polychets all the mesoblastic organs, together with
certain apparently serially homologous parts of the nervous and
digestive systems may recur for a seemingly indefinite number
of times in one individual, or even the whole animal may be
repeated in a chain, thus giving the highest expression to
the phenomenon, On the other hard, as in Lumbricus,

THE ANCESTRY OF THE CHORDATA. 71

&e., one or more of the mesoblastic organs may not be
repeated; while in both Oligochets and Polychets there is
a marked tendency to a division of labour between and
specialisation of structure of individual segments or even
regions of segments in various parts of the body. It thus
appears that even among Annelids alone the fact of segmenta-
tion is not a circumscribed idea, but may include several phe-
nomena which clearly differ from each other in degree, and
possibly are also unlike inkind. For while in the case of Nais,
&c., this repetition is complete, and is thus used as an obvious
and simple mode of reproduction, yet in other worms it appears
only to be concerned in increasing the length of one individual
without adding to the number. Now, if these two conditions
are merely various expressions of the same phenomenon the
question at once arises as to which is its more primitive mam-
festation. Was segmentation originally a repetition of all
the organs for purposes of reproduction, which process
has become subsequently commuted into mere increase in
bulk, or is this complete repetition to be regarded as the
final term in a series of which the first was increase in
bulk? Segmentation, as we know it, may clearly be viewed
from either of these two standpoints. With regard to the
Annelids, many authors have held that the former is the correct
one; the question whether this is so or not cannot be dis-
cussed here, but in the case of the Chordata examination will
show that their segmentation is of the latter class, and is the
result of a summation of repetitions ; and, being so, it is by no
means a unique condition, which can unite forms otherwise
unlike, as Chordata and Annelids, but is rather a result of the
common tendency to repeat parts already present, which ten-
dency occurs more or less in almost all animals. But before
communicating the features of Chordate anatomy, which point
to this as the mode of origin of the segmentation of the class,
it will be best to establish the fact that repetitions of this sort
are common, and to examine the comparative evidence as to
the manner in which they occur. It will then be seen that
segmentation on the plan found in the Vertebrates are really

72 WILLIAM BATESON.

extremely common, and appear to arise suddenly and in forms
nearly allied to those in which they are not found.

Firstly, among the ciliated Platyhelminths a striking case is
offered by Gunda segmentata, in which, as described by
Lang, the diverticula of the gut, the testes, the yolk-glands,
the tubules of the excretory organs, the transverse commis-
sures, and the nerve-cord, are all regularly and synchronously
repeated. Now, this case stands alone merely in the com-
pleteness of the repetition. All through the Turbellaria are
to be found many instances of animals with great numbers of
gut diverticula, with testes and yolk-glands scattered all over
the body, with branched excretory systems, with anastomosing
nervous networks, &c. Not only this, but instances are
common in which some of these structures are repeated regu-
larly, and others irregularly or not at all, as, for example,
Polycelis pallida (Quatrefages), in which the ovaries are
scattered and the testes are not, while the reversed condition
is more frequent. It becomes probable that the repetitions of
these organs did not phylogenetically occur simultaneously,
but that repetition occurred at various times in each set of
organs.

Again, among Nemertines in some species the saccules of
the gut, the generative organs, and the circular blood-vessels
are all repeated together and with great regularity, so as to
produce a segmented whole. In other species these repetitions
are not all formed or are more or less irregular, thus pointing
to the fact that these repetitions have been acquired within
the limits of the group. The development (v. especially
Salensky, ‘Arch. de Biologie,’ 1884) precludes at once the
possibility of the ancestral form of Nemertines having been
‘‘seomented ;” hence they, together with the Planarians, offer
a type of a high degree of repetition being acquired within the
limits of a group. Nor do these forms alone exhibit this
feature as one peculiar to themselves, for there are few groups
in which it is not found. Even among Mollusca, which are,
perhaps, the most typically unsegmented of all forms, the
Chitons may be instanced as examples showing that such com-

THE ANCESTRY OF THE CHORDATA. 73

plicated organs as shells may be repeated within the limits of
a small group. Moreover, in some Chitons bunches of cal-
careous sete recur along the sides symmetrically to the
scutes, producing an appearance not far removed from that of
Arthropoda.

Another case is to be found among the Nudibranchs, in which
the liver diverticula, which are peculiar to and characteristic of
the group, not only recur in an obviously segmental manner,
but may be arranged in several ways among the Molide,
being in some (as Alolis papillosus, Molis pulcher, &c.)
arranged in more or less regularly paired oblique rows, while
in others (as Dendronotus) the liver ceca stand in paired,
arborescent tufts, which are as definitely symmetrical in their
repetition as any system of organs of a Vertebrate. In cases
ot this kind the regularity of these repetitions is obviously
secondary, and all the other anatomical features show no trace
of segmentation, which constitutes the great interest of cases
of this kind from the point of view of the present argument.

The cases which have been so far mentioned have all been
selected from bilateral animals, with a definite long axis in
the direction of which they move. But the belief that repe-
titions of this sort are of constant occurrence as a factor in
effecting modifications of general form, derives most remark-
able support from the facts of the anatomy of radiate animals,
especially of the Echinodermata. From embryonic evidence it
may be regarded as almost certain that these animals are
descended from a bilateral ancestor, and that their present
form has been since acquired. Whenever this change took
place it came to pass in some entirely unknown manner that
the various organs came to be repeated round a central
axis. However this may have been brought about, the fact
remains that the number of such repetitions did not become a
fixed and definite feature common to all the divisions of the
group. For while the number five appears to be the limit of
the repetition in the Echinoidea, Ophiuridea, and Crinoidea,
among the Asteroidea the arms of different genera have not
the same number, nor do they necessarily occur in multiples

74 WILLIAM BATESON.

of any number. For example, while in the divisions Asteride
and Asterinide the prevailing number is again five, among the
Solasteride we find that the arms of Solaster may be
thirteen or nine (as in 8. endeca), in Heliaster from twenty-
nine to forty. Not only is this true of living forms, but in
the case of the fossil Cystidea the plates were irregularly
arranged and the perforations of the feet scattered, and in the
Blastoidea the basal plates were three, though bearing
five radials and interradials. All these facts point to a
history of the occurrence of repetitions among the various
parts around a central axis. And perhaps more remarkable
still is the extreme variability to be seen among individual
members of living species. .

For example, though Asterias rubens ordinarily possesses
five arms specimens possessing six or seven arms are very
common, while individuals with only four are not rare (the
latter may possibly, however, arise from mutilation). In like
manner specimens of Brisinga coronata are said to have
from nine to twelve arms. Thus, in these cases the arms, with
all the organs which they contain, may be spasmodically
repeated as a mere individual variation.

All these animals move on the oral surface, and though, of
course, the body may be regarded as arranged bilaterally
round a longitudinal axis, yet in the locomotion of the animal
this fact is not conspicuous (?) But in the Holothurians in
which a long axis does again assume importance, though
repetitions of this magnitude do not occur, yet there is a
tendency for certain organs to arrange themselves in a series
of longitudinal repetitions closely imitating segmentation.
In this connection the Elasipoda (Holma Théel, ‘ Challenger
Monographs’), which crawl about on the “ trivial” surface in
the direction of the long axis are of great interest. The body
of these animals is long and flat, and its margins are produced
into long processes, resembling parapodia, which are regularly
arranged in pairs down the sides. The regularity of this
arrangement is so great that some of the species figured by
Théel might easily be thought at first glance to be segmented

THE ANCESTRY OF THE CHORDATA. 75

worms!. Thus, in animals whose long axis has been sup-
pressed, it appears that repetition may arise of most of the
organs of the body radially arranged; next, that not only the
specific but also the individual number of these variations is
liable to great variations, pointing to the fact that the power
to repeat in this way is one which may be easily called into
action producing great differences of form.

It may also ,be observed in this connection that similar
casual repetitions are frequent in the case of the Gonozooids
of Hydromeduszaz, in which animals also they are radially
arranged. As in the case of the Echinodermata this is shown
by the great diversity in the specific and individual number of
those organs which are radially repeated. The latter may be
seen, for example, in Clavatella prolifera. The Medusa of
this animal creeps about on its tentacles, which are long and
stiff, and which carry short suctorial processes on their oral
faces which support the animal, giving it the appearance of an
Ophiurid. The number of these tentacles and of the radial
canals varies with age, from six to eight (Hincks). In the
specimens which have come under my own observation in the
undetached buds the number of these arms was five, while
those of the free Medusa was generally six. The number of
the organs in Cladonema radiatum, another creeping form,
is also very variable, the number of oral lobes being five or
seven, and that of the tentacles and canals eight or ten
(Hincks).

The facts of Echinoderm and Ceelenterate anatomy above
quoted, suffice to illustrate the statement that in animals
whose organs are already radially repeated, variations con-
sisting in the repetition of one or more of the peripheral
organs is of common occurrence, and may affect large numbers

1 Jn relation to this acquisition of the appearance of longitudinal repetition
or segmentation by a radiate animal, an example of the inverse phenomenon
may be given. Among the Operculate Cirripedes, though in the Balanide the
arraugement of the six plates composing the “cone” are so placed as plainly
to indicate the original long axis, yet in the Coronulide this feature becomes
obliterated, and the plates are disposed in a radially symmetrical manner.

76 WILLIAM BATESON.

of organs as in the case of the arms of Asteroidea, and
may be of specific occurrence as in Asterias rubens and
Brisinga coronata, or even ontogenetic as in Clavatella, &c.

All the instances of repetition of organs which have been
so far selected, whether in the case of animals with a marked
long axis or in the radiate forms, have been examples of the
recurrence of parts or organs in some more or less definite
relation to the axis of symmetry of the animals. These have
been chosen especially as more markedly illustrating the
possibility that the segmentation of some forms at all events
may have been derived from the continual recurrence of this
phenomenon until it became more or less regular and trans-
missible to the offspring as the definite course of development.
But it must be remembered that repetitions of this kind are
of an extreme type. The recurrence of whole sets of organs,
as in the case of the arms of Asterias or the gastric pouches
and generative organs of the Nemertines, must be regarded as
the higher manifestations of this phenomenon, and conse-
quently of more or less occasional occurrence. Since, how-
ever, it isin these cases that the nearest approach has been
made to metameric segmentation as we now see it, they have
necessarily been selected as of the first importance. But if
repetitions of this magnitude are of rare occurrence, repetitions
of smaller parts or organs are extremely common, if not uni-
versal. There is hardly one of the larger or more organised
types in which whole tracts of the body are not composed of
almost precisely similar and “serially homologous” parts,
which are of very variable number. The scales and fin-rays
of fishes, the tufts of hair and markings on many-caterpillars,
the teeth of Vertebrata, the joints of the Arthropod appen-
dages, or of the stems of a Crinoid, the ossifications in the
ambulacra of the Echinodermata, and many others, suggest
themselves at once.

Especially noticeable are the casual repetition of large com-
plex structures, such as the mammary glands and of exoskeletal
organs, as the horns and dermal scutes of Vertebrates. The

THE ANCESTRY OF THE CHORDATA. 77

number of these is liable to great variations, not even being
constant in the species. For example, certain deer and also
certain sheep have specifically more horns than two; and in
the case of Iceland sheep the horns may be three, four, or five
(Youatt, ‘The Sheep’). By the nature of the case none of
these repetitions can be atavistic; and it is interesting to notice
how, just as it was shown that irregular repetitions of parts
about the axes of symmetry of the body often take up regular
secondary relations to them, recurring either in segmental
pairs or in radial symmetry, so these minor repetitions take up
regular relations (secondary in some cases, probably primitive
in others) to the axes of the limb or part of the body in which
they occur. Thus the ossifications in the Crinoid stem or
the Starfish arm are so regularly related to the axis of the
part that in the latter case they have suggested to Haeckel his
extraordinary view of the phylogeny of the group, appearing to
him precisely similar to the segmentation of a Chetopod. The
case of the scales of fishes and the hairs and markings of cater-
pillars should perhaps have been more properly quoted in the
former connection, as being an instance of irregular repetitions
which have become definitely related to the symmetry, as in
the case of the Sturgeon, and among caterpillars the Tussocks
and the Spherigide. One very curious instance may be
quoted of a series of repetitions which, though essentially
arranged with reference to the axis of a limb, have yet a defi-
nite relation to the long axis of the body. This instance is
that of the Vertebrate tail, which has often been adduced by
opponents of the Annelid theory of Vertebrate descent. Now,
the structures which repeat themselves in the Vertebrate tail
with great variability of number, namely, the vertebre with
their neural and hemal arches, the segmental vessels and
nerves, &c., are precisely those structures upon whose repeti-
tion in the trunk the view of the primitive character of tue
segmentation of the Vertebrata mainly depends.

In the foregoing pages the attempt has been made to show
that greater or less repetition of various structures is oue of
the chief factors in the composition of animal forms, that these

78 WILLIAM BATESON.

repetitions may be of greater or less extent, affecting single or
many organs, and may be at first irregular, and finally culmi-
nate in regularity, and that even this regularity may afterwards
vary so as to become a symmetry of a different order. It is
further contended that between repetitions in these varying
degrees it is impossible to draw any hard and fast distinction,
for nothing more can be affirmed as yet about them than that
they are repetitions. The reason for their appearance is as yet
unknown, and the laws that control and modify them are
utterly obscure. But in view of what has been adduced it is
surely not too much to say that enough of their mode of
working can be seen to enable us to realise that they are at
least powerful enough to have produced anatomical features of
high importance, and further that the metameric segmentation
of the Vertebrata is distinctly of the kind which could be
brought about by their operation. That in this case they have
attained a degree of completeness far exceeding that which
they elsewhere present must be admitted ; but there is no evi-
dence to show that this result differs in kind from that which
occurs on a smaller and more restricted scale in almost all
animals. Whether the repetitions which occur in the Annelids
and Arthropoda are also the products of this force in a still
higher degree cannot yet be certainly stated.

General Conclusions as to the Mode of Occurrence
of Repetitions of Organs.

In the present state of biological knowledge no guess can be
hazarded as to the cause of the facts above quoted. The solu-
tion of the problem must be sought in a fuller knowledge of
the laws of growth and variation, of which we are still igno-
rant. As yet only one or two features in these repetitions may
be mentioned as possibly of importance, though even these can
only be selected in the most tentative manner,

In this connection the first noticeabie fact is that the struc-
tures repeated in the Triploblastica are very generally of
mesoblastic origin, and that when other structures have
become involved this would appear often to be a secondary

THE ANCESTRY OF THE CHORDATA. 79

occurrence. To such an extent is this true that in a recent
contribution to this subject (Caldwell, ‘Quart. Journ. Mic.
Sci.,’ 1885), a suggestion has been made which proposes to
give a simple physical explanation of all the phenomena of
segmentation. Caldwell suggests that owing to the early
acquisition of the long axis of the body and the consequent
elongation of the blastopore, the mesoblast has become, so to
speak, left behind in blocks, in consequence of the more rapid
growth of the epiblast. That this extremely simple theory
will not account for all cases of repetition is shown, firstly, by
the fact that though the repeated structures are generally me-
soblastic, yet they are not always so; secondly, that the meso-
blast does not thus originally segment as a whole, but rather
that separate organs repeat themselves separately, as has been
already urged, especially in the case of the Turbellaria; and
finally, these repetitions are by no means universally embryonic
or even larval features, but their whole history rather points to
their having very generally originated in the adult condition,
and to the view that they have come to be thus earlier in
development, the opposite of which is assumed by such a
hypothesis as Caldwell’s.

This belief that these repetitions have had their origin in
variations which occurred in the first instance late in life is
founded upon several considerations. Firstly, the cases in
which the generative organs are repeated are very numerous ;
in fact, both organs or the testis, at all events, are repeated in
nearly all the cases in which much repetition is found (in most
Dendroceles, Chetopods, Nemertines, Balanoglossus, Am-
phioxus), even if few other systems are repeated. In the
case of these organs it is most likely that the repetition first
arose in adult life, and, in fact, in most of them it does still
so arise; that is to say, the masses of cells which are to form
generative organs are not specially broken up at an early age.
And in the second place, the original late origin of repetitions
is likely from the fact that most of them still so arise; it is
only in exceptional cases as that of the mesoblastic pouches of
Vertebrata, Phoronis, Enteropneusta, and the horns of the

80 WILLIAM BATESON.

water- vessel of Echinodermata, that some of the repetitions are
presented early in the development.

Besides the probability that most repetitions occur in the
first instance in adults, or, at least, in mature individuals, it
may also be noted as a general feature of them that they are at
first very similar to, if not identical with, each other. For on
their first appearance in an individual they do not generally
arise phylogenetically in the condition which may be supposed
to have been that in which the original organs of the same
series first arose, but rather from the first they are found as
fully differentiated copies of the other members of the series,
and not as rudiments. For example, the horns and teeth of
mammals, whose number varies greatly, are, in those forms
which possess additional ones, not repeated as tubercles or as
plates, but rather as fully developed horns, teeth, &. Though
this is not universally true it is yet sufficiently well marked a
feature to be of great importance in estimating the probability
of the recurrence of such a complicated organ as a vertebra
with its correlated parts within narrow limits of race. But no
less noticeable is the tendency towards a subsequent differen-
tiation and division of function among members of a series of
similar parts as soon as the series is formed or any new
member is added to it. This is of course to be seen in the
ease of the tentacles of Hydromedusz, the division of the
ambulacra of Echinoderms into bivium and trivium cul-
qinating in the bilateral symmetry of Holothurians, differen-
tiation between vertebree, &c.

Beyond this little can be predicated of the mode of occur-
rence of repetition of parts. Nothing is attained by analysis
of the known facts which can be felt to be in any way a basis
from which to interpret them. This much alone is clear, that
the meaning of cases of complex repetition will not be found
in the search for an ancestral form, which, itself presenting
this same character, may be twisted into a representation of
its supposed descendant. Such forms there may be, but in
finding them the real problem is not even resolved a single
stage ; for from whence was their repetition derived? The

THE ANCESTRY OF THE CHORDATA. 81

answer to this question can only come in a fuller understand-
ing of the laws of growth and of variation which are as yet
merely terms.

Preliminary Remarks onthe Repetition of Organs
of the Chordata.

In the foregoing pages it has been attempted to show (1)
that repetition of organs and sets of organs is of common
- occurrence among animals, and (2) that however far back a
segmented ancestor of a segmented descendant may possibly
be found, yet ultimately the form has still to be sought for in
which these repetitions had their origin. Hence it follows
that in no case must it be held 4 priori impossible that an
unsegmented form showing no degeneration should be related
to a segmented stock. But when inquiry is made in the
special case of the Chordata as to the condition of the repe-
titions found among them, it will be seen that so far are they
from suggesting that their immediate ancestor of the group
must have been segmented, that they even preclude this view.
As will be shown, there is a history of the actual steps by
which several of the organs (the nervous system, the axial
skeleton, and the mesoblast) acquired their repetitions within
the group, and certain other structures (the notochord, &c.)
persist in an unsegmented form. So that instead of regarding
a fully segmented form as their possible ancestor it is neces-
sary to search for a form in which these particular sets of
structures at least are not repeated.

For in the first place, taken generally, the development of a
Vertebrate consists in the gradual appearance of repetitions,
first of one organ and then of another, until at last a climax is
reached. The mesoblast divides into blocks, paired peripheral
nerves grow out, and segmented tubules arise in connection
with the excretory ducts, but the mesoblastic plates were at first
unbroken, the medullary plate continues without transverse
divisions, though its peripheral organs may be repeated, and
the excretory ducts are single tubes with single openings. That
many of these structures roughly correspond with each other

82 WILLIAM BATESON.

is no doubt true, but these correspondences are only partial,
and, as will be shown in the sections on the nervous system
and vertebral column, a history is preserved to us of the steps
by which some, at least, of these repetitions have been attained
and of stages in which these correspondences were still more
irregular.

The attempt to find the ancestor of the Chordata resolves
itself first into the question as to whether the Chordate features,
viz. notochord, gill-slits, and nervous system of a particular
type were first associated in a form which possessed repetitions
in a high degree or not. Now, since the notochord is always
unsegmented, it is 4 priori likely that it arose in an unseg-
mented form; for, having in view the early period of develop-
ment at which it arises and the situation which it occupies in
the body, and the fact that it is found in the dorsal wall of the
gut, the sacculation of which is one of the commonest features
in segmented forms, it could hardly have thus arisen without
participation in such segmention. On the hypothesis of
Annelid descent the facts of the morphology of the notochord
are inexplicable; for, seeing that no homologue of the noto-
chord exists among Anvelids, on the theory that Vertebrates
are their descendants, the notochord must have arisen sub-
sequently to that segmentation, to account for which the
Annelid ancestor is postulated. If this were so the notochord,
by every rule of phylogenetic interpretation, might be expected
to arise late in development, and to exhibit marked segmenta-
tion, instead of which it is almost the earliest organ formed,
and is absolutely unsegmented.

Similarly from the first, the medullary plate is distinctly a
single structure, and without suggestion of transverse division.
Not until the peripheral nerves arise is any serial repetition to
be found in it, and were it not for theoretical considerations it
would not have been supposed that the nervous system of a
two-day Chick was a segmented structure. Further, in Am-
phioxus and the Marsipobranchs the serial repetition, even of
the peripheral nerves, is not regular and opposite, the further
meaning of which facts will be discussed later,

THE ANCESTRY OF THE CHORDATA. 83

Lastly, the gill-slits are by their nature repeated structures ;
but, seeing that nothing resembling them occurs outside the
group,’ their origin and, 4 fortiori, their repetition has been
acquired within it.

It becomes then probable, from preliminary examination of
the morphology of the three typically Chordate features, that
their first origin was not in a segmented form. There is also
one other structure which certainly points in the direction of
an unsegmented animal as the immediate ancestor of the Ver-
tebrate. This structure is the liver. Now, the liver is essen-
tially a unique structure in the body which is not repeated.
On the Annelid theory of Vertebrate descent it would have to
be supposed that the liver either arose as an enlargement of
one of the segmental saccules of the gut, or by the coalescence
of several. The evidence attainable on this point is distinctly
against either of these possibilities; for the liver of all the
Vertebrates, and especially of Amphioxus, is markedly and
obviously a single structure, not formed by the coalescence of
several, while its asymmetrical position and general appearance
favour the view that it is a structure newly formed within
the limits of the group, rather than a relic of a paired
sacculation.

Having then disposed of the a priori objections to regard-
ing an unsegmented form as a primitive member of the group,
the attempt will be made to show that the Enteropneusta
occupy this position. After this we will proceed to consider
the light which this admission will give on the history of
the steps by which the organs of the other Chordata ac-
quired their present arrangement, and finally to determine the
relation which the various forms included under this head bear
to one another.

The Enteropneusta as Members of the Chordata.
The general features of the anatomy of the Enteropneusta
place them in a very isolated position. They are extremely

1 Vor Semper’s suggestion that the ccelomic pores on the heads of some
Oligochts are of the same nature caunot be seriously considered,

84. WILLIAM BATESON.

like one another, but apparently very unlike any other group
of animals. Before Tornaria was known to be a stage in their
development they were assumed to be worms of some kind, but
after Metschnikoff had succeeded in proving Tornaria to be
the larva of a Balanoglossus this was felt as an impossible view
of its affinities. Up to this time Tornaria had been regarded
by Joh. Miller, who first described it (‘ Berl. Akad.,’ 1849,
1850), and by others who examined it as a varied form of
Bipinnaria, which, indeed, it very closely resembles, differing
only in the presence of eye-spots, and of a peri-anal ring of
cilia; both of which structures are liable to great variation.
When, then, Metschnikoff discovered its real destiny, it
appeared at first sight necessary to suppose the Enteropneusta
closely connected with the Echinodermata, and accordingly
Metschnikoff (‘ Zool. Anz.,’ 1880) proposed to include them
in a division Bilateralia under the Echinodermata, the re-
mainder of the group forming a parallel division, Radiata.
But this generalisation with regard to the group was made
solely on the characters of the larva, and almost without
reference to the structure of the adult, which, indeed, was
little known. So certain, however, did the conclusion seem,
that Metschnikoff was led to suppose that the gill-slits of
Balanoglossus were mere amplifications of the water-vascular
system of Echinoderms, which could hardly have been sug-
gested had it not been felt that no other solution was possible.
Since this time the anatomy of the adult has become more fully
known, and another mode of development has been shown to
occur, and from neither of these additional sets of facts can
any confirmation of the Echinoderm theory be derived. Hence
we must conclude that the characters of Tornaria are not to
be looked to solely in attempting a solution of the problem.

In the development of Balanoglossus Kowalevskii the
following important features occur: (1) the origin of the cen-
tral nervous system is by longitudinal delamination from the
skin in the dorsal middle line; (2) at the anterior end of the
body a portion of hypoblast is constricted off on the dorsal
side to form a supporting structure, i.e. a notochord ; (3) the

THE ANCESTRY OF THE CHORDATA. 85

gill-slits are formed as regular fusions and perforations of the
body wall and gut from before backwards. Hence the three
features which alone distinguish Chordata from other animals
are present, and associated from an early period in develop-
ment. Added to this the minor features of Chordate anatomy
are also represented by (1) the origin of the mesoblast; (2)
the remarkable asymmetry of the anterior parts; (3) the
opercular fold; (4) the excretory funnels opening into the
atrial cavity thus formed. From all these facts we may form
a preliminary conclusion that the Enteropneusta bear some
relation to the Chordata. We will now discuss what relation
this is, and before doing so we must determine what relative
importance is to be attributed to the two modes of develop-
ment known to occur, the one largely embryonic the other
pelagic.

In our present state of ignorance as to the mode of develop-
ment of Tornaria and of the details of its later stages, it is
difficult to compare these two modes, but the question as to
which is to be regarded as primitive is probably a part of the
larger question as to the comparative likelihood of the pre-
servation of ancestral features in the free or in the protected
developments. This question cannot be fully gone into here.
No general answer has as yet been giveu to it, and since the
balance of probability is very nearly divided between these two
possibilities we may be right in assuming either of them to be
correct. For the purposes of the following argument it will be
assumed that, on the whole, development within an egg-shell,
as involving a less complicated struggle with environmental
forces, is less subject to variation than that in the open sea, and
consequently is more likely to preserve ancestral features.
Besides this, in the special case before us, the adult structure
is practically conclusive against Echinoderm affinities, to
which the pelagic development would point if regarded as
primitive.

Assuming, then, that the development of B. Kowalevskii
is more primitive than that involving a Tornaria stage, the
following features are of great importance :

7

86 WILLIAM BATESON.

(1) The animal is ciliated and inhabits muddy sand.

(2) The preoral lobe is enormously developed.

(3) The notochord arises at the anterior end of the hypo-
blast and grows forwards,

(4) The origin of the central nervous system consists in the
delamination of a solid cord of epiblast in the dorsal middle
line of the middle third ; this, by invagination of its two ends,
s afterwards extended as a tube in both directions.

Other collections of nerve-fibre are afterwards deposited in
various parts of the body, and finally a general network of
nerve-fibre occurs at the base of all the skin of the body,
especially in the line of the gill-slits.

(5) The mouth originally faces ventralwards, but comes
afterwards to open forwards, being not a sucking but a dig-
ging mouth.

(6) The gill-slits for along time are only one pair, but
subsequently are repeated in pairs, increasing in number with
increase in the size of the body.

(7) The mesoblast arises as one unpaired pouch, followed by
two pairs of pouches.

(8) The blood-system is entirely peculiar, consisting of an
anterior heart and a dorsal and ventral vessel, and in B.
minutus of two lateral vessels in the intestinal region. The
two former are united by a plexus of trunks, which are placed
under the skin and below the walls of the gut.

(9) The generative organs are repeated through a large part
of the body ; in the branchial region more or less following the
repetition of the gill-slits.

(10) Of the excretory system little can be affirmed. The
cells of the mesoblast appear to have a power of forming
concretions, probably excretory, in their substance, and
then throwing them into the body cavity. Here they
form small aggregations. A large gland (containing a
plexus of vessels), apparently performing their function,
exists in the proboscis cavity attached to the end of the
notochord.

From the proboscis cavity opens an asymmetrical ciliated

THE ANCESTRY OF THE CHORDATA. 87

pore, placed on the left side of the body, whichin B. Kupfferi
is stated to be paired.

From the middle body cavities open a pair of pores into the
atrial cavity, which is partly enclosed by

(11) A rudimentary operculum.

Having these facts in view, and having set aside the pre-
liminary objection that no high degree of segmentation is pre-
sent in Balanoglossus, we may consider their bearing on
theories as to the ancestry of the Chordata.

Previous Suggestions as tothe Ancestry of the
Chordata.

Setting aside the possibility of Annelids having been geneti-
cally connected with the Chordata, the most notable alternative
suggestion is that of Balfour, that the Nemertines might be
thus regarded. This view has been supported and extended by
Hubrecht. It has thus been thought that the Chordate nervous
system might have arisen by the longitudinal coalescence of
two such cords as are present in Nemertines. But even the
facts of other Chordate developments almost preclude the
view that their nervous system is a double structure; the
medullary plate of Amphioxus is distinctly single, and it is only
in the medullary folds of higher and more complex forms that
even an appearance of a double structure is produced, while no
really double origin occurs. This being so, the mode of origin
in Balanoglossus is practically conclusive against the theory of
double origin. It is possible, and even likely, that Nemertines
bear some distant relation to Chordata, as will be further dis-
cussed subsequently, but if this is so it can no longer be sup-
posed that their nervous system is other than a special develop-
‘ment within the group.

In most speculations as to the origin of Vertebrata, it is
assumed that all the lower forms of Chordata are degenerate.
The supporters of the Annelid theory especially are compelled
to resort to this view severally in the case of the Ascidians
Amphioxus, and the Marsipobranchs. These, with the excep-
tion of the Enteropneusta, are the only forms which could have

88 WILLIAM BATESON.

been used to throw light on the origin of the group, and they
had to be expressly excluded because the suggestion as to the
origin of the group had been made without regard to them.
In the case of Amphioxus and the Marsipobranchs this theory
of degeneracy will not bear examination.

It rests solely ia the one case on the fact that Amphioxus
has no developed sense organs and lives buried in the sand, and
in the other on the semi-parasitic habit of life of the group.
This degeneration is postulated to explain the lower degree of
segmentation presented by these forms; and the fact remains
that of all animals the worms which live most underground are.
the most segmented types which are known. Hence it cannot
be assumed without ontogenetic evidence that degeneration in
this direction has occurred. This ontogenetic evidence is en-
tirely absent. Degeneration in this sense means a phylo-
genetic change of plan; and this change of plan should then
leave a mark on the ontogeny, as occurs in Echiurus, &c.; but no
event in the development of Amphioxus or of Lampreys points
to any such change of plan. ‘The development of these forms
is a steady progress up to the point which the creatures finally
reach, and in a case of this kind it is gratuitous to postulate
degeneration in order to support a preconceived view of the
morphology of the group. (Even in the Ascidians, though a
well-marked change of this kind does occur, yet it is not a
deviation from a segmented to a less segmented form ; for with
the doubtful exception of Appendicularia, Ascidian tadpoles
are quite without trace of segmentation.)

Again, no such evidence of a change of phylogenetic plan is
found in the case of the Enteropneusta. Highly modified, no
doubt, the adult animals are, but not degenerate. For these
reasons the presumption of universal degeneracy on the part
of all the lower Chordata will be dismissed, and an attempt
made to systematize the facts as they are found.

THE ANCESTRY OF THE CHORDATA. 89

The Habits of Life and Form of the Body of the Primitive
Chordata.

Habits of Life.—The presence of gill-slits in all the
Chordata may be taken as positive evidence that they arose in
an aquatic habitat. Moreover, such a structure as the noto-
chord cannot be conceived as having arisen in a fixed form.
Hence they probably led a more or less free existence. This
being so, they may either have been pelagic creatures, as the
larve of Amphioxus, or may have crept in mud as the larve of
B. Kowalevskii. Between these two possibilities there is
little or no determining evidence. The only feature which
seems likely to affect the question is the question as to the
original point in the body at which the notochord first segre-
gated itself from the gut. Unfortunately the evidence upon
this point is divided. For if we suppose that the condition in
Balanoglossus is primitive, and that notochord began as a rod
in the dorsal wall of the anterior end of the hypoblast, then
this origin would more or less point to a burrowing habit, the
notochord functioning as a support for the head in this opera-
tion; but if the separation of the notochord in the middle of
the body, as in Amphioxus, be held to be primitive, then this
would point to a pelagic habit, the notochord serving as a
fulcrum, from which the movements of the animal in swimming
might be maintained. The absence of fins on the young Bala-
noglossus and on the young Amphioxus, though pelagic,
appears to point slightly in favour of a burrowing habit, though
no reliance can be placed on such slight negative features.

Primitive Mouth.—There is one more point that does
point in favour of a pelagic habit, namely, the fact that the
anteriorly-directed digging mouth of both Balanoglossus and
of Amphioxus is of secondary origin, being formed by a modi-
fication of a more primitive ventrally-directed mouth.

Balfour, having the mouth of Lampreys and Tadpoles in
view, held that the original Vertebrate mouth was suctorial.
This the ventrally-directed mouth might have been; but this

90 WILLIAM BATESON.

fact does not interfere with the obvious possibility of a digging
mouth having again intervened, from which such a mouth as
that of the Lampreys could easily be derived.

Taking into consideration, then, the fact that in the most
primitive forms the mouth is anteriorly directed, and that in
the Lampreys it is also anteriorly directed, though of different
function, we may tentatively suppose that though the mouth of
the possibly original pelagic form was directed ventralwards, and
was possibly suctorial, yet probably the mouth of the Marsipo-
branchs is derived from a digging ancestor, in which the mouth
of the hypothetical pelagic form had come to be anteriorly
directed in correlation with an acquired burrowing habit. In
any case the facts of the Enteropneusta entirely confirm Bal-
four’s view, that the Vertebrate jaws have been developed com-
paratively long afterwards.

The Skin.—That the skin was originally ciliated there can
be little doubt; also it is probable that at first plexuses of
nerve-fibre were formed at the base of the ectoderm cells, such
as may be seen in many if not in all animals with ciliated skins
of this type.

The Nervous System.—The next question relates to the
position and mode of the first formation of a differentiated ner-
vous system. The evidence of Enteropneusta, Ascidians, and
Amphioxus is united in showing that this first occurred in the
dorsal middle line, and not by the coalescence of two lateral
cords. The structure of the nervous system of Balanoglossus
further shows us a stage in the process by which this nervous
cord separated from the skin. By many authors it is supposed
that this was accomplished in the first Chordata by an invagi-
nation, but the evidence of Balanoglossus is decidedly for the
view that a process of delamination preceded this ; and, indeed,
this being the simple process, might naturally have been ex-
pected to have occurred first. In Balanoglossus we seein
the trunk the cord still in the skin, in the collarthe
cord delaminated, and at the ends of this cord the
process of invagination commencing and leading to
the presence of a lumen. More than this, the mode of

THE ANCESTRY OF THE CHORDATA. 91

origin of the peripheral nerves is also seen ; for those portions
of nervous tissue which remain in the skin consist of fibres and
a few cells. Into the nervous tissue thus composed run the
tails of ectoderm cells, and out of them, on their inner sides,
run many fibres into the subjacent mesoblastic tissues. Now,
the fibres entering this nerve-substance on its outer side are
plainly sensory, or at all events afferent, and the fibres
passing from it on its inner side are presumably motor, or at
least efferent, seeing that they innervate the mesoblast.

It is clear, then, that on the separation from the skin of a
cord thus composed the relations of the efferent fibres will not
be changed, as they still remain in contact with the mesoblast.
But, on the other hand, if this nerve-cord be entirely separated
from the skin the supply of outer or afferent fibres is cut off
from it, unless cords of epiblast remain to connect it with the
skin. Applying this reasoning to the particular case of the
separation of the dorsal cord, we see that the afferent
fibres are entering it on its dorsal side, and that the efferent
fibres are leaving it on its ventral side. If, then, the cord
sinks in from the skin, the efferent fibres coming out on the
ventral side to supply the muscles can still do so without
being gathered into cords, remaining irregular as they
do in Balanoglossus, but without dorsal cords connecting
the main cord with the skin afferent impulses could only enter
at the two ends which remain connected with the skin; hence
I submit that it is probable that the three median cords in
Balanoglossus minutus, &c., are to be regarded as the
homologues of the dorsal roots of other Chordata. It is
at once evident, from the physical exigencies of the case,
that if the nervous system arose in this way the dorsal roots
were from the first sensory, and that they did not arise
as differentiations of roots of mixed function, as has
often been supposed. If this is true, then, as the cord phy-
logenetically comes away from the skin from before back-
wards the number of these dorsal cords will increase, until
finally the cord lies connected all along the body with the skin
by a series of median dorsal cords placed at intervals,

92 WILLIAM BATESON.

Now, returning to what is found in Balanoglossus, it is to be
noted that, first, the cord separates from the skin as a solid
rod connected at the two ends to the skin, and upon this con-
dition invagination supervenes at the two ends, forming a
neural tube in these regions. Let us follow the effect which
an extension of this system of invagination along the cord will
have upon the origin of the dorsal roots; for it is nearly certain
that invagination in this case is secondary to delamination ; the
condition in Amphioxus, in which the medullary plate folds up
after being enclosed, offering a stage of transition between the
condition found in Balanoglossus and that of an Elasmobranch,
for example. Since the invagination of a plate of tissue differs
from the separation of a cord in the fact that it is not the
central line, but the two edges of the plate, which remain last
in connection with the skin, it follows that, as the process of
invagination phylogenetically arrives at the point of attachment
of any one of these median dorsal roots, it must take up its
new attachment at one of these two edges. It is thus not
possible, supposing these views correct, that the dorsal roots
could in the first instance have been paired, except on the
hypothesis that as the process of invagination phylogeneti-
cally reached its point of attachment each dorsal root split into
two; which is almost impossible, and which the condition of
Amphioxus shows not to have occurred. The other alternatives
would be (1) that all the dorsal roots should remain attached
on one side to the cord; (2) that they should be attached
irregularly to one side or the other; and lastly (3) that
they should have been attached alternately to either side.
From the nature of the case they could not be opposite.
Now, the fact of their alternate arrangement in Amphioxus
is almost a proof that the latter alternative was the one
which occurred. (It may be observed that, as a physio-
logical convenience, they probably supplied the two sides
of the body alternately while yet attached in the middle
line.) Thus the opposite origin of the dorsal roots is
almost certainly secondary to an alternate arrangement.
The fact that it is the foremost pairs which are opposite in

THE ANCESTRY OF THE CHORDATA. 93

Amphioxus seems to indicate that the process by which they
became so occurred first anteriorly.

Let us now follow the history of the ventral roots as pre-
served to us. In Amphioxus the large nerves or dorsal roots
supply the skin and certain sense organs placed among the
muscular tissue (Rohon) ; but into each myotome, opposite
each dorsal root, runs a bunch of loose nerve-fibres from the
cord. This was stated by Rohon, but denied by Balfour.
Improved methods of section cutting leave no doubt, however,
that Rohon’s observation was correct, and, indeed, these fibres
may be easily seen. The presence of these bunches of fibres
clearly gives us another step in the formation of the “seg-
mented” nervous system. For in the simplest case, that of
Balanoglossus, the muscles are not gathered into bunches, and
the nerve-fibres likewise are irregular. In Amphioxus the
muscles are already gathered into bundles, and the motor
nerves follow them in this arrangement, but remain distinct
from the dorsal roots. This therefore is a stage towards the
gathering of the efferent fibres into a “ventral root;” in Bdel-
lostoma this is already done, and though the dorsal roots are
already approximately, though not quite opposite each other,
yet the ventral roots are not at the same level with them.
Besides this, in Lampreys, the anterior and posterior roots are
still not united into a common cord, though in Myxine they
are thus arranged (Schneider and others).

In this the nervous systems of Balanoglossus, Amphioxus,
Lampreys, and Myxine form a graduated series leading up to
the condition found in higher Vertebrates, showing the evolu-
tion of the nervous system of Vertebrata from a solid cord in
the skin to its condition as a closed tube whose walls give off
a series of ‘‘segmental” nerves arising by roots of different
functions.

[It will be seen that if this view be accepted it becomes very
doubtful whether efforts to analyse the segmentation of the
head can lead to any result, seeing that it almost follows that
the head was differentiated as such before any complex meta-
merisation was present ; and, indeed, were it not for theoretical

94 WILLIAM BATESON.

considerations, it could hardly have been supposed that the
head of a three-day chick, for example, was a highly segmented
structure, seeing that the regular segmentation of the body
conspicuously stops at its junction with the trunk. No doubt
the cranial nerves may, by arbitrary divisions and combinations,
be shaped into an arrangement which more or less simulates
that which is supposed by some to have been present in the
rest of the body, but little is gained by this exercise beyond
the production of a false symmetry. |

The Axial Skeleton.—The notochord of the Enterop-
neusta is so partially developed that it is not difficult to con-
ceive that its presence in the middle third of the body may
indicate a stage in its phylogenetic appearance. If while in
this condition it was used as a fulcrum in swimming it seems
further conceivable that if this organ grew backwards the con-
dition of the Ascidian Tadpole’s tail would be produced,
though no stress can be laid on this view. As will be shown later
on, it is likely for other reasons that the Ascidians separated
themselves from the other Chordata before Amphioxus, or
even the Enteropneusta.

By extending the separation of the notochord the condition
of Amphioxus is reached. And next, the axial column of the
Marsipobranchs shows us the notochord enclosed in a meso-
blastic sheath as yet unsegmented. This process is fore-
shadowed by the presence of rings round the neural canal,
placed between the nerves whose segmentation they follow.
Finally, in the other Vertebrata the column itself is segmented,
so that this is another instance of the appearance of a typical
segmentation in a system of a Vertebrate whose origin within
the limits of the group is unmistakeably traceable.

The Myotomes.—Intermediate conditions between the
condition of the muscles of Balanoglossus and of Amphioxus
are as yet unknown. I submit, however, that it is not im-
possible to conceive the formation of myotomes by a simple
mechanical process of gathering the muscular fibres into
bundles. Their origin as archenteric pouches may then be
supposed to have originated from the fact that the ancestral

THE ANCESTRY OF THE CHORDATA. 95

mesoblast already arose thus, and when new bundles of
muscles formed in the adult began to arise in the larva they
arose in the same manner as the primitive mesoblast. That
provision is made for the production of more mesoblast than
that of the original fourteen pairs of pouches is shown by the
presence of mesoblastic pole-cells in Amphioxus (Hatschek).
In any case the existence of Balanoglossus proves that the
notochord, gill-slits, and Chordate nervous system were present
together before the myotomes were formed.

The Gill-slits.—It is unfortunate that the facts of the
Enteropneusta seem to throw no new light on the original
meaning of gill-slits. That they do not do so tends, however,
to show that probably gill-slits were from the first developed
as such, and not as modifications of any previously-existing
organ, as has been sometimes held.

The folded skeletons of the gill-slits of Balanoglossus are re-
markable in their resemblance to those of Amphioxus. Until the
development of these latter is fully known no further com-
parison can be instituted. It is clear from their origin in
Balanoglossus that no “‘myotomes” are obliterated between
them (as has been suggested by some, with the hope of in-
creasing the symmetry of the body), for plainly their repeti-
tion preceded that of the myotomes.

The Excretory System.

Upon the origin of the excretory system of Vertebrata
nothing can be affirmed from a study of Balanoglossus. The
excretory systems of Vertebrata cannot be easily derivable
from anything found in either Balanoglossus, Ascidians, or
Amphioxus. The absence of any regular excretory system in
these three forms may, perhaps, be correlated with the extraor-
dinary development of their respiratory systems, which may
possibly assist in this function. The one fact which is de-
rivable from the morphology of Balanoglossus, Ascidians, and
Amphioxus, is that it is nearly certain that the excretory
system of other Chordata has been developed within the group,

The Pituitary Body and Proboscis Pore.-—Though

96 WILLIAM BATESON.

no insistance is placed on the following suggestion, the plausi-
bility of it is such that it cannot be omitted. On a previous
occasion I have called attention to the fact that the pore which
in Amphioxus leads into the left anterior body cavity is ob-
viously homologous with the proboscis pore of Balanoglossus,
which leads from the left horn of the anterior body cavity. In
some species of Balanoglossus the opening of this pore is placed
medianly, though opening into the left horn. Now, supposing
the preoral lobe to atrophy, as in an Ascidian, so that the
neural pore came to open into the buccal cavity, as occurs in
these forms, it is clear that any pore placed dorsally between
the neural pore and the mouth will then be directed ventrally,
and open into the pharynx below the end of the nervous
system. This is precisely the position occupied by the ciliated
pit of an Ascidian, which leads into the gland described by
Julin (‘ Arch. de Biol.,? 59). Hence with this pore and gland
of an Ascidian the proboscis pore and gland of Balanoglossus
may be compared. Next, supposing the end of the nervous
system to dilate and form a brain which bends up by a cranial
flexure it follows that on the atrophy of the proboscis (or rather
before the proboscis was formed, this being peculiar to En-
teropneusta) this pore will lie in the dorsal wall of the stomo-
dzeum, i. e. in the position of the pituitary body. More than
this, any gland attached, as is the proboscis gland, to the end
of the notochord, will, when this is flexed by the cranial
flexure, be bent backwards with it to the place where its end
comes to lie, i.e: above the pituitary involution. In this way
the double structure of the pituitary body becomes intelligible.
If these views are correct the pituitary body and its pore is to
be regarded as the rudiment of a primitive excretory organ,
which originally opened dorsally.

I have elsewhere shown the prima facie resemblance of the
anterior body cavity with its pore in Amphioxus to that of
Balanoglossus, which in the Tornaria development is formed
from the water-vessel (Spengel). This water-vessel is precisely
similar to that of Echinoderms, being otherwise without parallel
among animals.

THE ANCESTRY OF THE CHORDATA. 97

The Affinities of the Chordata.

Having thus examined the history of those organs which the
morphology of Balanoglossus enables us to trace, let us
consider the relations of Chordata (1) to other groups, (2) to
each other.

Of the Echinodermata.—Unlikely though it may seem,
if any reliance can be placed on the characters of pelagic
larve, we must assume some affinity between Echinodermata
and Chordata, for Tornaria is not very like, but practically
identical with, Bipinnaria. The case is like that of Mollusca,
which may be supposed to be allied to Annelids, as is indicated
by the trochosphere larva.

Of the Nemertines.—So much has been said by previous
writers as to the Chordate affinities of Nemertines that the
subject cannot be omitted. The suggested homology of the
nervous system has already been dismissed. Hubrecht has
further suggested (1) that the notochord is homologous with
the proboscis sheath of Nemertines, (2) that the cephalic
pits are gill-slits, (3) that the proboscis is the pituitary
body.

With regard to (1), what can be adduced from a study of
Enteropneusta seems rather to be opposed to this view. If
this were true, the notochord must have arisen in some such
body as that of a Rhabdocel, into the wall of the endoderm
of which a przoral lobe could be invaginated, rather than as a
hard thickening which is constricted off to form a lumen.
Into the free end of such a structure it is impossible to con-
ceive the invagination of a proboscis, which is what Hubrecht’s
suggestion seems to require. All that can be said is that the
notochord of Baianoglossus suggests that it arose as a support-
ing structure and not as a modification of something else.

But supposing the larva in Stage G to represent a phyloge-
netic phase, several points of Nemertine anatomy can be
derived from it. At this stage it has one pair of gill-slits, a
short nerve-cord, one median anterior mesoblastic pouch, and
two pairs of posterior pouches. Now, on the hypothesis of

98 WILLIAM BATESON.

Hubrecht that the esophageal pouches of Nemertine were the
homologies of gill-slits, and supposing the proboscis invagi-
nated and around its base a quantity of nerve-tissue deposited
as in Balanoglossus, the proboscis would then have the same
relation to the nerve-ring as that found in Nemertines. Hu-
brecht’s view of the pituitary body falls if the alternative here
given is accepted. Though the points of anatomical resem-
blance are not striking, yet when taken with the ciliated skin,
the ventral mouth and position of the generative organs they
form a basis for comparison.

If these resemblances were found to be real the nervous
system of the Nemertines would have to be supposed to have
arisen within the limits of the group. As both animals
possess a nerve-plexus in the skin this does not seem impos-
sible. Also the excretory system lately described by Oude-
mans (‘ Quart. Jour. Mic. Sci.,’ 1885), would have thus arisen
as a specialization of parts of the body cavity; since in
Balanoglossus this function appears to be generally distributed
over the body cavity, this also might be conceived.

Of the Tunicata.—Next, since all the Chordata at some
period of their development agree with the larva in Stage H,
in possessing a dorsal nerve-cord more or less invaginated, one
or more pairs of gill-slits and a notochord, let us pass on to
Stage H, in which the notochord is forming at the anterior end
of the gut. From such an animal as this the Ascidians may
have been descended. For, as has been suggested by van
Beneden and Julin (‘ Archives de Biologie,’ 1885) it may be,
that all the Ascidians have but a single pair of gill-slits; for
that Appendicularia has only one pair is known ; while in some
genera the atrial cavity arises as an increase in the size of the
pair of ciliated chambers by which the gill-slits open; and
this increase may take place in the hypoblastic half of the
chambers, or in the epiblastic; by the fusion of these two
chambers the atrial chamber of these genera is formed. Van
Beneden and Julin then suggest that the atrial pore is the
actual opening of the two fused gill-slits, and that the rows of
slits placing the pharynx in communication with the atrial]

THE ANCESTRY OF THE CHORDATA. 99

chamber are to be regarded as secondary perforations.
Whether this ingenious theory be adopted or not, the fact
remains that Appendicularia is almost certainly a very primi-
tive Tunicate, and also that the arrangement of the pharyngeal
perforations of other Ascidians makes it unlikely that they are
homologous with the gill-slits of higher forms.

The increase in size of the tail, which would speedily
follow the first use of the backward directed notochord as a
swimming organ is not difficult to understand. In connection
with the increase of the tail the curvature of the gut would
also be intelligible. From atrophy of the preoral lobe in cor-
relation with the future sessile habit, coupled with increase of
the lower lip to bear the suckers, the relations of the neural
pore to the mouth would result. The gland of the preoral
lobe would then, as before described, be placed below the
nerve-ganglion and open into the pharynx.

It has been remarked by Seeliger (‘ Jen. Zeit.,’ 1885) that
the body of the Ascidian tadpole appears to consist of one head
and two trunk segments. It may be observed that though the
reasons for this belief are not very obvious, this view, if correct,
would coincide with the possibility of its descent from such a
larva as Balanoglossus, Stage G, which also possesses one head
and two trunk segments.

However the various points that have been raised in the
preceding paragraph may be decided, it has seemed necessary
to point out what conclusion with regard to the structure of
Ascidians may be drawn from the development of Balano-
glossus. That these are so meagre is to be regretted ; the only
tangible point appears to be the confirmatory evidence that it
offers to the view that the atrial folds of Tunicata are not
homologous with those of Amphioxus.

In this way only can the absence of mesoblastic repetitions
in Tunicates be accounted for. Their development gives no
support to the view that their ancestors possessed repetitions
of this kind.

Of the Enteropneusta.—That the Enteropneusta might
possibly have had an ancestor in an animal possessing the

100 WILLIAM BATESON.

structure of Stage H is of course shown by their ontogeny.
They are derived from it chiefly by increase in size of the
preoral lobe, change in direction of the mouth, growth of a
rudimentary operculum, serial repetition of the gill-slits, and
appearance of the generative organs also as a serial repetition.
That any animal possessing a large preoral lobe should
acquire a thick sheath of nervous tissue (especially when con-
sisting of fibres for the most part) is easily understood. As
shown in the foregoing pages, this mass of tissue is probably
mainly composed of afferent fibres connecting the proboscis
with the dorsal cord. As soon as the ventral nerve-cord arose
as a concentration of nerve-tissue, this would naturally be
followed by another circular concentration in the nervous
sheath connecting the ventral cord with the central, invagi-
nated, nervous system, also as an afferent mechanism.

In all probability the enormous increase in size of the larger
species was a comparatively recently acquired feature, as also
the peculiar odours which they emit; to this latter power it is
possibly not too much to attribute the preservation of such a
group.

Of the Cephalochorda.—The relations of the Cephalo-
chorda is the next subject for consideration.

The young Balanoglossus agrees with Amphioxus, especi-
ally in the following anatomical features :—

(1) The digging mouth.

(2) The repetition and folding of the gill-slits.

(3) The repetition of the generative organs.

(4) The peculiar fate and remarkable asymmetry of the
anterior mesoblastic pouch and proboscis pore.

(5) The presence of atrial folds.

(6) The absence of (a) any developed sense organs; (b) any
excretory glands differentiated as such.

(7) In the presence of excretory tubes opening into the
atrial cavity.

On the other hand it differs from it n—

(1) The relative size of the przoral lobe.

(2) The degree of its mesoblastic repetition,

THE ANCESTRY OF THE CHORDATA. 101

(3) The degree of the invagination of its nervous systen
and the extent of the neural tube.

(4) The extent and degree of isolation of its notochord.

(5) The extent of the atrial folds.

(6) The absence in B. Kowalevskii of any definite liver
sacculi, and the presence in B. minutus, &c., of liver saccules
differing from those of Amphioxus.

The points of resemblance taken together are so consider-
able as to suggest that they were possessed by a common
ancestor of the Hemichordata and Cephalochorda. On the
other hand, the points of difference are nearly all differences of
degree, and (1), (2), (3), (4), (6) are points in which the
Vertebrata agree with Amphioxus. In the case of (5), how-
ever, the Vertebrata more nearly agree with Balanoglossus.

Of the Vertebrata.—The common ancestor, then, of the
Cephalochorda and the Vertebrata may be presumed to have
possessed the features of mesoblastic repetition, invaginated
nerve-cord, and consequent extension of the neural tube, raised,
so to speak, to the degree in which they are found in both those
divisions. Also it may be believed that the preoral lobe had
somewhat diminished and that the atrial folds were still small.
The origin of such a liver as that of Amphioxus, as a speciali-
sation of part of the wall of the digestive region of a young
B. Kowalevskii is easy to imagine, for the histology of these
two tissues is still almost identical. [The presence of peculiar
liver saccules in B, minutus, &c., presents no difficulties, as
their absence in the more primitive B, Kowalevskii shows
that they have arisen within the limits of the group.] Animals
possessing those features would answer nearly to the Proto-
chordata of Balfour, though the structures now attributed to
it are somewhat different.

The Protochordata thus constituted would then differ from
the Enteropneusta in the possession of a serially-repeated me-
soblast, in addition to serially-repeated gill-slits, and possibly
generative organs; also in the complete separation of the
nervous system and notochord. The serial repetition of the
gill-slits, the small operculum, &c., they must be presumed to

8

102 WILLIAM BATESON.

have acquired from the ancestor common to them and the
Enteropneusta.

In this way the connection of the Protovertebrata of Balfour
with the other division becomes explicable on the new facts
derived from the Enteropneusta.

The peculiar fact that so many of the features of the
Enteropneusta differ from those of the Cephalochorda in
degree of expression only is very remarkable, and suggest
that their further evolution towards the Protochordate type
proceeded by correlated variations affecting the several
systems,

From the Protovertebrata thus constituted, which in all
probability possessed an unsegmented mesoblastic sheath for
the notochord and a brain, the Cyclostomata may be easily
derived without the necessity of any hypothesis of great
degeneration, which cannot be well supported.

Balfour has fully discussed the question of the origin of his
hypothetical group of Protognathostomata, and upon the
question of their immediate origin no new light can be thrown.

The above suggestions entail many difficulties. The chief of
these is that they involve the hypothesis that the rudiment
ofthe notochord of the Archichordata developed itself
as a separate structure, once in the case of the Ascidians, and
again in the case of the Protochordata. In the first case,
owing to the atrophy of the przoral lobe and use of the tail in
swimming, it came to lie in that organ, and in the second case
extended through the whole length of the body. Also does this
suggestion of the origin of the Tunicates involve the proposition
that the rudiment of the dorsal nerve-cord extended itself
twice along the body, once in the case of the Ascidians, and
again in the case of the Protochordata. If this occurred there
is no difficulty in supposing it to have been twice invaginated,
this bemg a more less common feature among nervous
systems.

Another difficulty which affects all these suggestions arises
from the epiblastic origin of the generative organs of Enterop-
neusta, in which they resemble the Echinoderms.

THE ANCESTRY OF THE CHORDATA. 103

Though it is likely that many of the suggestions here made
may be shown hereafter to be wrong, still it has seemed well,
on the whole, to analyse the facts as they stood, and to endea-
vour to reconstruct the past stages, whose existence is indi-
cated by the lacune in the sequence of these facts, avoiding
as far as possible a reliance upon phylogenetic changes of
whose occurrence we have no evidence.

The foregoing views are, perhaps, more clearly expressed in
the following table, which is not meant so much as a genealo-
gical tree as to serve as an exhibition of the logical relation
of the various forms, showing their points of divergence.

Protognathostomata

(of Balfour).
Cyclostomata.
Cephalochorda.
Enteropneusta.
Tunicata

PN ee oo

Form with one gill-slit.

=~

a
ws
a
how
¥

—

¥ ~~,

ala che Wile arid

ae ae

The Development of the Mole (Talpa Europea).
STAGES E to J.

By

Walter Heape, M.A.,
Resident Superintendent of the Plymouth Laboratory of the Marine Biological
Association of the United Kingdom.

With Plates XIII, XIV, and XV.

Dvurine the preparation of the following paper I have been
conscious that a considerable proportion of the matter included
is of little special interest ; at the same time it has appeared
to me that the course of the development of certain organs in
the Mole deserves to be recorded, and in order to do so satis-
factorily I have been compelled to mention much which is not
different from embryological phenomena already observed in
other Vertebrates.

I have further been led to hope that a somewhat complete
account of the development of one of the Insectivora will not
be without value.

To facilitate reference I have described the development of
the embryo in stages, which, in continuance with the stages of
growth described in a former paper (No. 8), will be called
Stages H, F, G, H,and sy. A summary of the various sections of
this paper will be found on p. 132.

ExtEeRNAL FEATURES.

Stage E—The youngest embryo which I have figured (fig. 1)
lies flat upon the surface of the blastodermic vesicle. The

embryo is ‘76 mm. long, and is narrow in the centre and wider
9

106 WALTER HEAPE.

at eachend. A shallow medullary groove runs down the centre
of the long axis of the embryo, which in its turn is narrow in
the centre and wider at either end. On each side the medul-
lary groove in the central narrow portion of the embryo, three
protovertebre may be seen already formed.

The hinder end of the embryo is thickened owing to the
growth of the mesoblast of the primitive streak, while anteriorly
it is flattened out to form the cephalic plate. The shaded
portion surrounding the embryo (a.p.) is the extent of the
area pellucida at this age. ;

Fig. 2 represents a slightly older embryo of the same stage
of growth (1:82 mm. long). The medullary folds have here
begun to form, they are raised somewhat, and in the centre of
the embryo are already approximated. At the anterior end the
floor of the medullary groove, on either side, is swollen, and on
the outer and anterior edge of the two masses so formed a deep
narrow groove indicates the commencement of the formation
of the optic organs and will be referred to as the “ optic
grooves.”

This early appearance of the organ of sight is, so far as I am
aware, peculiar, and is worthy of notice; even at this age the
grooves are directed outwards and downwards, and have their
origin from the most anterior portion of the medullary groove.
The curved condition of this embryo is due to careless manipu-
lation whilst it was in a fresh and soft state.

Stage F. — Fig. 3 represents an embryo of this stage of
growth ; it is 1:96 mm. long. The medullary folds have met,
although they have not yet coalesced, in the middle of the
embryo, and have extended thence forwards.

The anterior end of the medullary canal is, however, still
widely open, and the two thickenings of the floor and sides of
this portion are shown. The optic grooves are also indicated
in the same manner as they were in the previous figure.

It will be observed that the sides of the medullary canal at
the anterior end have grown forwards in advance of the floor.
At the hind end the medullary canal is widely open, forming

THE DEVELOPMENT OF THE MOLE. 107

the sinus rhomboidalis. On either side of the embryo, just
behind the widely open anterior end of the medullary canal, a
ridge extends backwards and onwards over the blastodermic
vesicle; these ridges are the first traces of the two tubes which
will eventually form the heart (compare fig. 5, At.).

Figs. 4 and 5 are two drawings of an embryo of about
the same age as that last described (Stage r). The length of
the latter is, however, greater than that of the former embryo,
being 2°12 mm., while the medullary groove is not so far ad-
vanced in development. My object in drawing fig. 4 is not
only to show these points but to represent the amnion, which
is as yet developed only at the hind end of the embryo,
and has already grown nearly half way over the back of the
embryo.

Fig. 5 is a transparent view of the same embryo, and indi-
cates the position of the first five protovertebre, and of the
commencing tubes (At., At.) which eventually will form the heart.
The blind lateral prolongations of the medullary groove at the
cephalic end are the optic grooves. In this figure also the
floor of the sinus rhomboidalis at its posterior end is seen to
contain a much thickened, forwardly projective knob, which, as
will be shown in sections, is the anterior end of the primitive
streak. The medullary folds may therefore be described as
extending posteriorly behind the front end of the primitive
streak.

Stage G.—Stage a is represented by the embryo drawn in
fig. 6. The hinder portion of the medullary canal is much
the same as before; anteriorly, however, development has pro-
gressed, and the edges of the medyllary folds have come
together and partially fused at the anterior end of the em-
bryo. At the extreme end, however, a pore is left, owing
to the more rapid growth of the sides than of the floor of the
canal as pointed out above. At this stage, therefore, the
neural canal is still open to the exterior, both anteriorly and
posteriorly.

The optic grooves are now closed, and have given rise to the

108 WALTER HBAPE.

optic vesicles; these are shown as two bud-like vesicles pro-
jecting outwards and backwards, and slightly downwards from
the front end of the neural tube; behind them the swelling of
the fore-brain is discernible, while still further backwards and
at the edge of the body of the embryo the two tubes of the
heart are indicated.

The folding off of the embryo from the yolk-sac has at this
stage made some progress, and, indeed, the whole of the head
of the embryo as far back as the line so. pl. now lies projected
freely above the blastodermic vesicle.

Stages H and J.—These stages are depicted in figs. 7 and 9,
the embryo represented in the former figure being 2:2 mm.
long, that in the latter figure 3°06 mm. long. The more com-
plete closure of the medullary canal and the constriction of its
anterior region into fore-, mid-, and hind-brains is to be noticed.
The optic vesicles are still seen in fig. 9; in fig. 7 they are
barely noticeable, owing to the curved position of the embryo
when drawn. ‘

The increase of the protovertebre and the gradual reduction
of the sinus rhomboidalis is also seen, while the thickened
anterior end of the primitive streak is now enclosed within
the posterior walls of the medullary canal, and projects up-
wards as a rounded knob at its hinder end.

The direction of the increase of the protovertebre is a
difficult matter to determine, but a careful examination and
measurement of figs. 5, 7, and 9 leads me to believe that in all
probability the increase is almost altogether posteriorwards
during those stages. The embryo (fig. 7) of Stage u has,
however, apparently one protovertebre more anteriorly than
the embryo of Stage F (fig. 5), and the embryo (fig. 9) of
Stage J one more than that of Stage u (fig. 7). The embryos
of Stages = and F are more difficult to compare (figs. 1 and
5), but I think it is highly probable the increased number in
the latter is due to a backward growth.

The amnion at Stage u completely covers the embryo (fig.
7), an anterior limb having grown over the head as the

THE DEVELOPMENT OF THE MOLE. 109

posterior limb grew over the tail at an earlier period (Stage F,
fig. 4).

The anterior fold of the amnion (vide p. 128) is the so-called
pro-amnion of Beneden and Julin (No. 2). It must be noted
that up to the close of Stage J no signs of a folding off of the
tail end of the embryo can be observed, and, indeed it is not
until considerably later that this process takes place.

The first junction of the two tubes to form the heart takes
place during Stage u, and is shown in fig. 8; while the side
view of the head of the embryo drawn in fig. 10 (Stage 3)
shows the relation of the heart to the visceral arches, and the
arrangement of the latter.

There are at this stage five visceral arches. Faint grooves
indicating the partial formation of two and even three visceral
arches may be discerned during Stage u, but it is not until
Stage 5 is reached that they can be satisfactorily outlined.

For the comparison of these figures with figures of other
mammalian embryos I would refer to papers Nos. 3, 4, 5, 6, 7,
9, and 10.

Tue EFiIsiast.

Soon after the epiblast is first definitely produced it is in the
form of a plate of columnar cells of uniform thickness over the
whole embryonic area, and passing abruptly at the edge into
the flattened epiblast cells which cover the remainder of the
embryonic vesicle. This stage is figured in a former paper,
No. 8, fig. 30.

During the primitive streak stage of growth and the early
formation of the medullary groove, the lateral epiblast becomes
reduced in thickness and at the edge of the area the cells
gradually assume a flattened condition and blend without
a break with those of the vesicle (l.c. figs. 32—36, and
43—46).

The appearance of protovertebre and the deepening of the
medullary groove is attended by a further modification of the
epiblast of the embryo.

During Stages E to c the median portion becomes thickened

110 WALTER HHEAPE.

and forms the medullary plate (fig. 15) while the lateral por-
tions become gradually still more reduced in thickness, until in
those portions of the embryo where the medullary groove has
attained its greatest depth prior to its conversion into a canal
the lateral epiblast is formed for the most part of a single row
of somewhat cubical cells, continuous, without modification,
either over the vesicle or across the embryo as the inner fold
of the amnion (fig. 17).

Where the lateral epiblast joins the wall of the medullary
groove there is now an abrupt transition from the columnar
cells lining the latter to the cubical cells of the former.

Subsequently, Stages u.J.,in that portion of the embryo
where the neural canal is formed, the closure of the medullary
groove causes the approximation of the lateral portions of the
epiblast, which fuse, and thus form a continuous layer across
the embryo. The cubical epiblast cells at the same time become
much flattened on the dorsal surface of the embryo (figs. 26, 29,
and 47), while (1) in the trunk, the cells of that portion of epi-
blast which overlies the somatic mesoblast remain cubical
(figs. 26, 27, 29, and 47); and (2) in the anterior region, the
cells of certain portions which either give rise to sensory
structures (figs. 25 and 46), or which surround externally the
visceral arches (figs. 23 and 46) assume again a columnar
form.

In that region of the trunk where the medullary canal is
still open the lateral epiblast cells remain as before, cubical.

The Medullary Groove.—At the commencement of Stage z
a deep medullary groove exists about the middle of the embryo ;
anteriorly and posteriorly it is shallower however, finally ter-
minating in the latter direction upon reaching the anterior end
of the primitive streak, while in the former direction all trace
of the groove is lost some considerable distance behind the front
end of the embryo.

Beyond the anterior end of the medullary groove the epiblast
is thickened to form the “ cephalic plate.”

Fig. 1 is a transparent view of an embryo with three proto-
vertebrz, and shows the relations above mentioned ; 1 have

THE DEVELOPMENT OF THE MOLE. 111

also figured three transverse sections, which indicate the struc-
ture and form of (1) the cephalic plate (fig. 12) ; (2) the groove
in its anterior portion (fig. 14); and (3) the groove in its
posterior portion (fig. 15).

In fig. 12 the thick cephalic plate is shown, becoming folded
off from the yolk-sac; fig. 14 is taken from the region in front
of the protovertebre, and depicts the wide and shallow groove,
the wall at the bottom of which is considerably thinner than
at the edge of the groove ; and in fig. 15, taken from the region
of the second protovertebra the medullary groove is V-shaped,
and the columnar cells of which it is formed pass abruptly into
the lateral epiblast cells, thus indicating the extent of the
“medullary plate.”

At the hind end the wide and shallow medullary groove
forms the so-called “sinus rhomboidalis.” A section through
this region of an embryo during Stage Fr is shown in fig. 18.

At the close of Stage & the groove has considerably increased
in length, and during Stage F it reaches to the anterior end
of the embryo (figs. 3 and 16). The latter figure is a trans-
verse section through the anterior end, and shows—

(1) The median medullary groove.

(2) The commencement of the curvature upwards of the
lateral portions of the cephalic plate and the formation of the
two “optic grooves” (op. gr.), seen in surface view in fig. 4,
which give rise when the neural canal is closed, to the optic
vesicles.

The Medullary Canal.—The medullary plate is now sharply
marked off from the lateral epiblast from a considerable dis-
tance in front of the first protovertebra backwards to the pos-
terior end of the embryo, and the groove itself commences to
close in the region of the protovertebre.

The closure is effected by the approximation of the peri-
pheral edges of the medullary plate, a sharp angle being thus
formed at the junction of the lateral epiblast with the edge of
the plate (fig. 17).

The closure commences at a late period of Stage e@ in the
region of the first provertebra, extending thence forwards and

112 WALTER HEAPE.

backwards ; it proceeds very rapidly, being at the end of this
stage, although open at its immediate anterior end (fig. 6),
closed from there posteriorly until the fourth protovertebra is
reached, after which point it gradually widens out into the
sinus rhomboidalis (figs. 28 to 33).

At the close of Stage u a narrow slit-like pore is all that
remains open at the anterior end (fig. 20), while posteriorly it
is closed as far back as the eighth protovertebra; and at the
end of Stage s the whole groove is converted into a canal until
the last, the fourteenth, protovertebra is reached.

The sinus rhomboidalis is now narrow and shallow (figs. 48
and 50). The swelling in the floor at the hind end of the
sinus rhomboidalis is caused by the mesoblast of the front end
of the primitive streak (figs. 33 and 35 ; 48 and 50).

When the canal is first formed, its lumen—except in the
anterior region which is described below—is a narrow slit and
its walls are thicker at the sides than they are dorsally and
ventrally (fig. 28) ; soon afterwards, however, during Stage u,
the middle portion of the lateral walls thickens still more and
projects into the narrow lumen of the canal, thus converting it
into an hour-glass form (fig. 29).

The cells of the cord are much elongated, and their nuclei, in
general, oval (fig. 43).

I may in this place mention there appears to me to be great
likelihood of the migration of mesoblast cells into the walls of
the medullary canal during Stages u and ys. Sections of an
embryo belonging to the former stage present strong evidence
of this process (fig. 43). Two masses of mesoblast cells are to
be seen in very close connection with the lateral walls of the
canal in the region of the neck, and from these masses I feel
inclined to believe certain cells grow into the tissue of the
nervous system.

As I will show below, these masses of cells are in connection
with two blood-vessels, which are in process of formation, and
it would appear highly probable that these ingrowing meso-
blast cells give rise to the blood-vessels of the spinal cord.

The Brain.— When the medullary groove first closes in

THE DEVELOPMENT OF THE MOLE. 113)

(Stage G) it is wider in front of the first protovertebra than it
is in the latter and posterior regions, and faint indications of a
division of the brain into portions may be discerned in section,
and to some extent also in the surface view of this stage; the
hind-brain, with its somewhat thinner roof, is of considerable
length and blends into an anterior portion in which the roof is
thicker. Stage H shows some little advance upon this; the
cranial flexure has begun (fig. 34) and the cavity of the brain
has increased in size, the roof of the hind-brain also is thinner
and wider than before (fig. 23).

At the close of Stage s three divisions of the brain are indi-
cated (fig. 49). There is a well-marked cranial flexure, and at
what is now the anterior end of the animal the mid-brain is
situated. The cavity of the mid-brain is partially separated
from that of the fore-brain by a constriction of the walls at the
junction of the two, but the structure of the wall is very
similar in both portions. The hind- and mid-brains pass into
one another without any such constriction, but the thin roof of
the former distinguishes it from the latter. The lower wall of
the hind-brain at the posterior end is now much folded. The
lower wall of the fore-brain is curved downwards, forming a
short and wide diverticulum which marks the first appearance
of the infundibulum. The apex of the infundibulum comes
into close connection with the anterior end of the alimentary
tract and with the notochord overlying it (fig. 49).

The Optic Vesicles.—The optic grooves seen in the head in
surface view in figs. 2 and 5 are the rudiments of the optic
vesicles ; they are shown in section in fig. 16. Later (Stage
H), when the medullary groove forms a closed canal in the
head region, these grooves become wide lateral diverticula pro-
jecting from the anterior portion of the brain, and constitute
the optic vesicles (fig. 20). They are situated dorsally on
each side the middle line, and are projected outward and
somewhat downwards and backwards.

Such a condition is clearly shown in surface view in fig. 9.
Sections of this stage show a very similar condition as regards
the development of the vesicles; they merely extend slightly

114 WALTER HEAPE.

further outwards, but do not at this stage fuse with the
external epiblast (fig. 21).

The wall of the optic vesicles is similar in structure to the
wall of the remainder of the fore-brain.

It is interesting to note that for a considerable period after
Stage s the optic vesicles show but very slight advancement on
the condition then attained; their growth appears now to be
retarded in as marked a degree as it was advanced in the early
stages. The early appearance of the optic grooves will probably
be recognised as a mammalian distinction when the embryo-
logy of more species of Mammalia has been worked, but the
sudden checking of the development in the Mole we may
expect is due to the specialisation of this species. Any modi-
fication of an important sensory organ would doubtless rapidly
affect the development of the organ, but such an extended
modification as is apparent here says much for the primitive
nature of the habits of the animal.

The Ear.—The first indication of the ear arises during Stage
H as a thickening of the external epithelium on each side the
hind-brain (fig. 25). The thickening extends along a great
portion of the hinder half of the hind-brain, and during Stage
J increases in thickness and becomes grooved along the greater
part of its length (fig. 46).

The Cranial and Spinal Nerves.—I do not propose to describe
the development of the cranial and spinal nerves in this paper.
I hope to make a separate communication upon this portion
of the development at some future time.

Tue Hyposmast.

The hypoblast in the earliest condition of Stage x is similar
to what it was in Stage p (described in my former paper, No.
8), and is composed of a single layer of flattened cells extending
on all sides over the embryonic area (figs. 13, 14, and 15).

The cells in the median line give rise to the notochord, and
the changes they undergo will be described in detail in another
section of this paper.

The formation of the deep medullary groove in Stage p and

THE DEVELOPMENT OF THE MOLE. 115

the thickening of the vertebral portions of the mesoblast
causes the hypoblast cells underlying those structures to be
stretched out as it were and flattened (No. 8, fig. 45).

In Stages & and F this condition may still be seen where the
groove is deepest in front of the protovertebre (fig. 13) ;
anteriorly the groove becomes shallower and the hypoblast
cells more rounded in consequence (fig. 14), while posteriorly
the formation of protovertebre forces the lateral hypoblast
downwards, and the axial hypoblast cells are again thickened
(figs. 15 and 17).

In the region of the sinus rhomboidalis the medullary groove
again projects considerably below the level of the peripheral
body wall, and, forcing the hypoblast cells downwards also,
flattens them.

This condition in the anterior region and posteriorly below
the sinus rhomboidalis is, however, soon modified; the thick-
ening of the peripheral mesoblast and the gradual depression
of the body wall brings the lateral portions of the hypoblast
on a level with the axial portion throughout the length of the
embryo, and at the close of Stage 3 the cells of the whole
layer, wherever it is not converted into the alimentary canal,
become rounded.

The Alimentary Canal.—The first trace of the alimentary
canal appears during Stage p (vide No. 8, fig. 46) at the
anterior end of the embryo asa short tubular diverticulum.
In the paper referred to I described this tube as the noto-
chord, an error which I have corrected here and in more detail
on p. 118 of the present paper in the section devoted to that
organ.

This structure is indicated in figs. 1] and 12, Stage zr. The
diverticulum has but a small lumen, and is situated close
against the cephalic plate; the cells of which it is formed are
columnar.

Stages ¢ and u witness further changes ; the fore-gut is now
considerably longer (fig. 34). It is rounded anteriorly (fig.
22), but farther backwards is widened out laterally (fig. 19)
and becomes flattened and crescent shaped, the lateral horns

116 WALTER HBAPE.

of the crescent being projected upwards and somewhat closely
approximated to the lateral epiblast of the embryo (figs. 19,
23, and 24).

The epithelium of the dorsal border of the sac is thinner
than that of the ventral border, the difference being more
apparent in the hinder portion than in the front portion of
the sac. The points of the lateral horns are lined with
cylindrical cells.

There is no distinct evidence at this stage (H) of outgrowths
of the fore-gut in the position of the future visceral arches,
but slight indications of the invagination of the epiblast may
be seen corresponding to the grooves mentioned in the descrip-
tion of the surface view of an embryo of Stage u.

On the ventral surface at the anterior end of the fore-gut in
Stages a and u (figs. 19 and 22) two slight invaginations of
the epiblast may be seen one on either side of the middle line,
and a few sections further backwards the epiblast and hypo-
blast are closely applied in the middle line, and there is a deep
median groove in the epiblast (fig. 23).

At the close of Stage 5 there is a still further change in these
relations. The lateral outgrowths of the fore-gut are now
directed towards invaginations of the epiblast which corre-
spond to the grooves mentioned in the description of a surface
view of an embryo of this stage (Stage 3). The outgrowths
are directed outwards and downwards from the lateral portions
of the lumen of the canal (fig. 46). The hypoblast and epiblast
have met and are partially fused in the case of the anterior
diverticula, although there is as yet no perforation consti-
tuting a definite cleft, but in the more posterior diverticula
the hypoblast does not meet the epiblastic involution.

Now also the fore-gut is a little longer, and the fusion of
epiblast and hypoblast on the ventral surface near the front end
is closer, although the perforation to form the mouth has not
yet taken place (fig. 49).

This invagination of the epiblast is clearly seen in an embryo
of this stage to be in the form, anteriorly, of two shallow
grooves which converge posteriorly, these forming a deep

THE DEVELOPMENT OF THE MOLE. 117

median invagination (figs. 44, 45). These grooves are formed
along the anterior border of the first visceral arch. The epi-
blast and hypoblast are in close contact along the whole of the
V-shaped groove, but become actually fused posteriorly at the
apex, where the mouth will eventually be formed (compare
figs. 44, 45, and 49).

It will be seen by the foregoing description that the mouth
is formed somewhat behind the anterior end of the fore-gut at
the apex of a V-shaped groove on the ventral surface of the head,
the diverging limbs of which groove are directed forwards.
The section of the gut which is placed anteriorly to the
mouth is identical with the blind tube first formed by the
folding-off of the embryo from the yolk-sac, and this anterior
diverticulum exists for some time after the ventral enlargement
of the gut towards the external groove.

These facts appear to indicate that a more primitive mouth,
the terminal position of which is indicated by the primary
anterior diverticulum of the fore-gut, has been replaced by a
secondary formation, the paired origin of which is rendered
possible by the two converging grooves in the epiblast of the
ventral surface.

If these observations are correct, they must be considered to
some extent confirmatory of Dr. Dohrn’s theory of the paired
origin of the existing mouth of the Vertebrata, but I would
suggest that such evidence cannot be used as argument for the
paired formation of the primitive Vertebrata mouth, the
terminal position of such being exceedingly probable.

Asin the earlier stage, the cells forming the dorsal wall of
the fore-gut are throughout thinner than those lining the
remainder of the cavity, and in the posterior section of its
length are much flattened ; on the other hand the cells of the
ventral wall, the lateral horns, and the outgrowths to form the
visceral clefts, are cubical or even columuar in form.

The Notochord.—The notochord, as I have before described
(No. 8, figs. 37—48), is a hypoblastic structure and is primi-
tively in connection with the hypoblast and the lateral plates
of mesoblast of the embryo. During Stage p it becomes first

118 WALTER HBAPE.

separated from the lateral mesoblast, then reduced in thickness,
and finally converted (1) in the anterior region into an are
formed of a single row of columnar cells; (2) towards the
central deepest portion of the medullary groove into a single
row of considerably flattened cells; while (3) in the hinder
region it remains thickened and forms posteriorly the
anterior wall of. the neurenteric canal, thus joining the
epiblast.

During this stage (Stage p), the notochord is, throughout
its whole length, never actually isolated from the hypoblast,
but remains a portion of that layer, although an obviously
specialised portion; it is in fact the remnant of the primitive
hypoblast (I. c.).

In this same paper (l.c. fig. 46) I described as a portion of
the notochord a short tube formed of columnar cells lying
below the medullary plate at the anterior end of the embryo.
I must here correct that error. This tube does not represent
the notochord solely, but constitutes the anterior end of the
alimentary tract (figs 11 and 12), and, as I shall show below,
the cells only of the dorsal portion of this tube give rise
eventually to the anterior end of the notochord.

During Stages = and F the relations of the notochord
remain very much the same as they were during Stage p (figs.
14,15, and 17) ; it is noticeable, however, at the close of Stage Fr,
that in the trunk of the embryo, where the medullary groove
is deep, the axial hypoblast has increased in thickness
(fig. 17).

The deepening of the medullary groove towards the anterior
region which occurs during Stage e@ causes the notochord cells
situated there to be reduced in the same manner as they were
reduced in the central region during Stage p. Similarly the
axial hypoblast is reduced in bulk in the posterior region of
the embryo, while in the central region, where the protovertebre
are forming, there is a further increase in the size of the
notochord.

At this stage of growth (Stage ce) the notochord exhibits a
tendency to become separated from the hypoblast layer in the

THE DEVELOPMENT OF THE MOLE. 119

same manner, although not with precisely the same result, as
when the neurenteric canal was formed in Stage p.

The process in the latter stage involved the ingrowth of the
lateral portions of hypoblast and the conversion of the axial
portion, containing the neurenteric canal, first into an arc and
then into a complete tube. Now the lateral hypoblast grows
inwards below the axial portion of primitive hypoblast and
unites to form a continuous layer, merely causing the isolation
of the axial portion as either a solid rod or band of cells which
lies freely between the hypoblast and the medullary canal. It
is, however, true that a lumen may appear in some of the por-
tions of the notochord which are rod like, although its conver-
sion thus into a tube is, so far as I can determine, a secondary
matter, and is not connected with the method of isolation.

The isolation of the notochord first occurs in the region of
the first protovertebra during Stage c, and extends during
Stages H and s anteriorly and posteriorly. The separation
does not, however, appear to be a continuous process, and the
shape of the isolated notochord is very various. To demon-
strate these facts I have figured several sections of an embryo
with nine protovertebre (Stage u, figs. 24 and 36 to 42).

In this embryo, in front of the first protovertebra, the noto-
chord is isolated for some distance as a rod or thickened band
(figs. 24 and 37), in which a lumen may occasionally be seen
(fig. 836: compare also figs. 23 and 25, which are drawings of
sections through another embryo of this stage).

In the region of the first protovertebra, it is in the form of
a flattened band consisting of a single row of cells (fig. 38),
and this condition persists, except here and there, where the
notochord is not completely isolated (fig. 39), until the fourth
protovertebra is reached ; here it increases in size. From this
point it is more frequently attached to the hypoblast (fig. 40),
and posterior to the seventh protovertebra is not isolated at all.
Immediately behind the seventh protovertebra it is in the form
of an are (fig. 41), which further backwards flattens out, and
the mass, increasing in size, joins the front end of the primitive
streak (fig. 42).

120 WALTER HEAPE.

Such is the condition of the notochord during Stage nH. At
the close of Stage 3, however. the whole of the notochord,
except at the immediate anterior end, backwards to the ninth
protovertebra, is isolated as a rod of varying size and shape
(figs. 46 and 47). Behind the ninth protovertebra it becomes
band shaped and continues in this form, still distinct from the
hypoblast, for some distance behind the last (fourteenth) proto-
vertebra. It then again assumes the form of a rod, although
of much larger size than in the anterior region, in the centre
of which a lumen may here and there be seen, and joins the
anterior end of the primitive streak becoming thus connected
there with the epiblast, hypoblast, and lateral mesoblast
(fig. 50).

The phenomena I have here described, viz.: (1) the presence
of a mass of primitive undifferentiated hypoblast in the
median line (Stage p) ; (2) its reduction to a thin, even single
layer of cells (Stages p, &, and F), and (3) the conversion of
those cells into the notochord (Stages eg, oH, and 3); these
phenomena, in my opinion, indicate without doubt that this
organ is of hypoblastic and not of mesoblastic origin.

Further, during the isolation of the notochord, (a) the
appearance, vague though it be, of an are of notochordal cells ;
(6) the fact that the isolation of the solid rod or band com-
mences at the two sides and gradually extends across the
median line (figs. 39—4.2) ; and (c) the occasional appearance
of a lumen in this rod,—these appearances indicate that it is
formed in the same manner as the notochord of Amphioxus,
that is to say by the ingrowth of the lateral hypoblast and the
constriction of the axial mass of primitive hypoblast cells.

I have already discussed the views of other observers upon
this subject (No. 8) and need not again refer to them.

Figs. 39 and 40 are especially interesting in regard to the
isolation of the notochord. In both these drawings the
process of isolation is shown taking place; in both the noto-
chordal tissue is in the form of a pair of knobs connected by a
median more slender portion; and in both cases when the
notochord is actually isolated it will be isolated as a band of

THE DEVELOPMENT OF THE MOLE. 121

greater or less substantiality. It will be noticed the knobs
are more or less free from the underlying flattened hypoblast
cells, while in the median line there are no flattened cells, thus
showing the process of the growth of the lateral hypoblast
below the axial primitive hypoblast.

The relation of the notochord at the front end of the embryo
requires special notice ; it will be best understood by a reference
to figures of longitudinal sections through embryos of Stage
E (fig. 11), Stage um (fig. 34), and Stage s (fig. 49). In
fig. 11 the notochord is not separated from the roof of the
fore-gut; in fig. 34 it remains attached to the anterior wall
of the fore-gut, although isolated posteriorly ; but in fig. 49 the
notochord, although joined at its anterior extremity to the
hypoblast, is separated from it throughout its extent posteriorly.

The hooked anterior end of the notochord, so characteristic
of this organ, is seen to be due, in the Mole, to the fact that it
is derived from the anterior wall of the alimentary tract after
the cranial flexure has commenced.

At the close of Stage 3, therefore, the notochord is continu-
ous with the epiblast at the front end of the embryo, by means
of the front wall of the fore-gut, which is fused with the
epiblast at the point where the mouth will eventually be
formed ; and posteriorly, at the anterior end of the primitive
streak, where epiblast, hypoblast, and mesoblast are ali joined
together (compare figs. 49 and 50).

The close relation of the fore-brain to the notochord, a rela-
tion brought about not so much by the cranial flexure as by
the ventral enlargement of the brain at this point, will be
referred to in another communication, which I hope shortly to
make, upon the pituitary body of the Mole.

There is one other point of interest in the growth of the
notochord in the Mole, and that is its size compared with the
nervous system. The relative size of the notochord compared
with the nervous system is less in the higher than it is in the
lower Vertebrate embryos. In the Mole the notochord is rela-
tively smaller than it is in any other Vertebrate embryo I am
acquainted with, and it appears to me the reduction in size is

10

122 WALTER HEAPE.

due to the comparatively early development of the nervous
system. During the early part of Stage p there is a consider-
able mass of primitive hypoblast along the axial line of the
embryo, but the rapidly forming medullary groove pressing on
to this mass before it has become formed into a rod capable of
resisting much pressure, causes it to bulge inwards and thus
flattens out its cells, administering an effective check to the
development of the organ. Such a check occurs during Stages
p and g. Subsequently the thickening of the lateral meso-
blast plates and the consequent depression of the lateral hypo-
blast, removes the strain from the axial cells and admits of the
isolation of the slender rod or band which exists for the
greater portion of the length of an embryo Mole at the close
of Stage s.

Tur Mesosiast.

At the close of Stage p the mesoblast in front of the
primitive streak is in the form of two lateral plates which are
connected together across the middle line by means of a mass
of undifferentiated hypoblast, except during a short space
where they are separated by the deep portion of the medullary
groove.

At the periphery these mesoblastic plates are split into two
layers, an upper somatic and a lower splanchnic layer, along
the whole of their extent posterior to the cephalic plate. The
split is entirely peripheral, however, and does not extend into
the embryonic area.

The Mesoblastic Somites and the Body Cavity.—During
Stage & the splitting of the mesoblast extends further forwards,
and also inwards towards the medullary groove. I have never
been able completely to satisfy myself that this splitting ever
extends to the innermost portion of the mesoblastic plates ;
but, as I have before explained, the small size of the embryo
and the dense compact nature of the middle layer renders it
exceedingly difficult accurately to determine such a point.

The nearest approach to a continuous split of the mesoblast
from the axial portion to the periphery which I have seen is

THE DEVELOPMENT OF THE MOLE. 123

represented in fig. 13; and here it will be seen, although there
is no positive division into somatic and splanchnic layers, yet
such a division is indicated in the section by the arrangement
of the nuclei of the cells on each side a line, which is repre-
sented by a narrow band of a lighter shade than the surrounding
tissue.

In sections of three other embryos which I have examined,
about the centre of the medullary groove there is similarly an
indication of the splitting of the mesoblast from the periphery
to the axial portion, the cells being arranged in two parallel
rows along the inner edges of the two layers of mesoblast,
although no cavity is actually formed. Thus, although it
cannot be said that a split actually occurs through the whole
plate of lateral mesoblast in the Mole, yet there is without
doubt a tendency to such splitting in embryos of Stage z about
the centre of their body.

In the same stage of growth (Stage £) is to be observed :

(1) The separation of the axial and peripheral portions of the
mesoblastic plates, these two portions being connected by a
narrow neck of cells, the intermediate cell mass; and (2)
the formation of protovertebre by means of clefts in the
axial mesoblast at right angles to the long axis of the embryo,
which divide this portion into cubical masses. The indication
of the splitting of the mesoblast at the same time becomes
more definite, and results in a cavity (fig. 15) within both the
protovertebre and the peripheral mesoblast, a cavity which
does not, however, extend through the intermediate cell mass
(compare also fig. 17, Stage F).

The cells of the protovertebre are radially arranged round a
narrow elongated cavity, and form in a transverse section through
the middle of a somite a triangular mass, the apex of which is
situated at the base of the medullary groove.

The cells of the peripheral mesoblast in the region of the
protovertebre are columnar on their inner side and border, a
narrow slit extending to the periphery. At the edge of the
area the cells become flattened, and form a thin somatic and
thicker splanchnic layer, extending over the yolk-sac.

124 WALTER HBEAPE.

An examination of consecutive sections reveals, in front of
the protovertebrz, the axial and peripheral mesoblast in the
form of a continuous solid plate, with no cavity in the axial
portion; while in the peripheral portion the cavity gradually
recedes outwards (fig. 14) until it no longer exists within the
limits of the embryonic area.

Behind the protovertebre the cavity in both axial and peri-
pheral mesoblast becomes at once and simultaneously oblite-
rated, and two thick solid lateral plates of mesoblast extend
backwards, and join the mesoblast of the primitive streak.

As the medullary groove closes in, the protovertebre become
more cubical and compact (compare sections of Stage u, show-
ing these points (figs. 31, 30, and 28) ), and the narrow slit re-
duced to a small central pore, which about this time becomes
very generally partially filled up by a core of cells derived from
the lower and inner portion of the protovertebra.

The protovertebre then (Stage 4) commence, first in the
anterior region, and gradually assuming in subsequent stages the
same relation posteriorly, to divide into two portions, an outer
and dorsal arched portion composed of columnar cells, and a
lower and inner portion formed of irregularly rounded cells (fig.
29; compare also fig. 52 of Stage 3), the former giving rise
mainly to the muscle-plates, the latter to the bodies of the
vertebrz and the connective tissue surrounding them. It will
be shown subsequently, however, that the inner portion also
participates in the formation of the muscle-plate.

A very marked cavity exists between the two portions on
the outer side of the somite (fig. 29), and the vertebral portion
of the mesoblast is continued ventrally below the neural canal
towards the notochord.

The cavity is derived from the small cavity present in the
earlier stage (Stage B) ; and it is worthy of notice it is not first
obliterated and then again formed, as has been stated by some
observers to be the case in the Chick, nor does it entirely dis-
appear, as has been supposed to be its fate in Mammalia
(vide No. 1, p. 553).

Anterior to the protovertebre scattered mesoblast cells

THE DEVELOPMENT OF THE MOLE. 125

exist below the neural canal, closely approximated to the slight
rod-like notochord (figs. 24 and 26), while in the region of the
protovertebre the mesoblast is more compact, and does not
extend so far beneath the medullary canal (fig. 28).

In Stage s the anterior protovertebre exhibit still further
changes :—(1) The vertebral portion of the somite has increased
very considerably in depth ; (2) the cavity has almost entirely
disappeared, remaining only as a mere slit (fig. 47) within; (3)
the muscle-plate, which is now formed of two rows of columnar
cells. The second row lies inside the first, close beside and
parallel to it. It is formed from the cells of the vertebral por-
tion of the somite, which have hitherto occupied this position.
The two rows are continuous with each other at their dorsal
and ventral ends, and the cavity before spoken of lies between
them, reduced to a narrow slit.

Posterior to the three anterior protovertebre the muscle-
plate consists of only a single layer of columnar cells, as was
the case in the earlier stage (H).

The muscle-plates are therefore first formed anteriorly.

When examining this stage my attention was drawn to the
histological characters of the cells of the outer layer of the
muscle-plate in the anterior protovertebre.

These cells were observed with an ordinary Zeiss D lens
to be continued outwards into more or less fine processes, and
upon examining sections with a high power (Powell and Lea-
land’s ;4,th oil immersion and Reichert’s ;4,th oil immersion)
it was found that these fine processes were branched or simple
prolongations of the mesoderm cells, which on the one hand
joined with the ectoderm cells, and on the other formed a
fine network immediately below the ectoderm.

These processes are voluntary muscular fibres, which are thus
early developed from the outer portion of the muscle-plate.
This structure is fairly satisfactorily represented in fig. 51,

The fact being observed that these mesoderm cells actually
joined the ectoderm cells led me to make a renewed exami-
nation of my sections of earlier stages, and I found that from
the time the hypoblastic mesoblast was formed in Stage c

126 WALTER HEAPE.

(No. 8) it was always possible to trace processes from meso-
derm cells into the overlying epiblast cells.

The elongated mesoblast cells shown in fig. 51 are more
extended in Stage s than they hitherto have been.

In Stage u (figs. 27—29) a tendency to elongate may be
observed in these cells and so also in Stages & and F (fig. 17),
but it is not until Stage s is reached they can actually be
described as muscular processes. Further, this condition in
Stage gs only exists in a marked degree in the first few anterior
protovertebre ; further backwards these processes gradually
decrease in length.

With regard to the inner layer of the muscle-plate, certain
of the cells already show a differentiation into elongated
muscular fibres, but they are not all of them as yet so meta-
morphosed.

The protovertebre remain at the close of Stage 3g still
separated from one another throughout their depth, and
between each, short blood-vessels run, which are dorso-lateral
branches from the dorsal aorta (fig. 52).

The mesoblast at the front end of the embryo now extends
between the notochord and the floor of the neural canal (figs.
44, 45, and 49), the embryo having increased dorso-ventrally.

I find no trace of the body cavity in the head. As was
stated above, the splitting of the mesoblast never extends to
the axial portion of this part of the mesoblast, and no cavity,
as far as I have been able to see, makes its appearance
secondarily.

Pericardial Cavity.—The separation of the pericardial cavity
from the remainder of the body cavity has only commenced
during Stage y, and at the close of that stage the mesenteries
in which the ductus Cuvieri run from the body wall to the
sinus venosus, divide the body cavity into two dorsal sections,
one on each side, the pleuro-peritoneal cavities, and one median
ventral section, the pericardial cavity.

These three sections are all continuous at the anterior end
into a single cavity surrounding the heart, which is prolonged
a considerable distance further forwards.

THE DEVELOPMENT OF THE MOLE. 127

Posteriorly the pleuro-peritoneal cavities are each continuous
with the body cavity contained between the diverging folds of
the somatopleure and splanchnopleure.

The Primitive Streak.— During Stages © and F the relations
of the primitive streak are almost exactly similar to those
described for Stage p (No. 8), the only difference being the
extension of the medullary folds backwards round the front
end of the primitive streak (fig. 18). The lumen of the
neurenteric canal disappears, but the point where it originally
existed is shown by the fusion of the epiblast, hypoblast, noto-
chord, and primitive streak mesoblast at the front end of the
latter (Stage 3, fig. 50).

In my paper, No. 8, I endeavoured to prove the mesoblast
of the primitive streak did not extend beyond the point where
the neurenteric canal was situated, and I showed that over the
whole of that part of the embryo situated anterior to the
primitive streak, mesoblast was formed from the hypoblast
(‘‘hypoblastic mesoblast ”’).

Now if this be true, it follows that the mesoblast of the primi-
tive streak takes no part in the formation of the body of the
embryo anterior to the neurenteric canal, and that the growth
of the embryo is caused by a multiplication of cells anterior
to the primitive streak.

The mesoblast of the primitive streak is, however, a con-
siderable and hitherto a constantly increasing mass, and it
extends backwards and outwards beyond the embryonic area.
It thus occupies the position where eventually the allantois is
formed, and it is, in fact, the primitive streak mesoblast which
forms the walls of that organ.

During the stages now under discussion (e—3) the primitive
streak becomes partially—almost eutirely—divided into two
portions, an anterior and a posterior portion. The division is
caused by the formation of two pits—(1) a dorsal pit which
eventually gives rise to the anus, and (2) a ventral pit which
projects upwards and backwards into the primitive streak, and
forms the cavity of the allantois (figs. 35 and 50).

These two pits constrict the blastoderm and partially divide

128 WALTER HBAPE.

the primitive streak into ashort anterior portion which projects
upwards along the floor of the medullary groove at its hind
end, and into a larger posterior portion which forms the wall
of the allantois (figs. 33 and 35, Stage u, and figs. 48 and 50,
Stage 5).

The dorsal pit I have mentioned gives rise to the anus; this
structure is therefore formed in the middle of the primitive
streak in the Mole, in the same position as Weldon (No. 11)
pointed out the anus of Lacertilia is formed.

The Amnion.—The amnion is first formed, as I have before
described (p. 109), at the hind end of the embryo; extending
thence forwards, and being met by the lateral folds of the
amnion which also grow, in the first place, from behind forwards
(figs. 26,28, 35, and50). This portion of the amnion is formed
as in the Chick of a double fold of somatopleure. Immediately
upon the junction of the two lateral folds and the formation
of true and false amnion, the epiblast of the false amnion,
which is shown in fig. 28, unites eventually with the neigh-
bouring uterine tissue, and the thin sheet of somatic mesoblast
alone remains between the uterine wall externally and the
splanchnic mesoblast within.

At the front end of the embryo a different structure is found
to exist. The lateral folds in this region are similar to the
posterior portion of these folds, but the median anterior fold
of the amnion is different inasmuch as it is formed solely of
epiblast and hypoblast (fig. 34). Although the amnion at the
anterior end is not formed until some considerable time after
the mesoblast of the embryo has extended to the front end of
the embryonic area, and although this mesoblast has ex-
tended laterally over the vesicle throughout the whole length
of the embryonic area, it only extends forwards for a very
short distance, and does not grow between that portion of the
epiblast and hypoblast which gives rise to the anterior fold of
the amnion. Consequently, when the head of the embryo
becomes projected anteriorly over the yolk-sac, as it does
first in Stage e (fig. 6), and then bends downwards, forming
for itself a pit on the surface of the yolk-sac, the walls of this

THE DEVELOPMENT OF THE MOLE. 129

pit constitute the anterior fold of the amnion, and are formed
solely of epiblast and hypoblast. This portion of the amnion
does not come in contact with the wall of the uterus.

The relations of these parts have recently been very fully
described by Beneden and Julin in Rabbit and Bat embryos
(No. 2). These authors have named this anterior fold of the
amnion the “ pro-amnion,” and have most ingeniously, and as
it appears to me correctly, compared it with the internal stalk
of the “ trager” of inverted types of mammalian embryos.

I should mention that the mesoblast present in the median
line in the longitudinal section of an embryo of Stage g&
(fig. 11) is concerned in the production of the heart, the
anterior fold of the amnion having its origin in front of
this mesoblast (compare figs. 11 and 34).

The Allantois.—The allantois is, in an embryo of Stage F, a
short wide diverticulum of the hypoblast projecting into the
posterior portion of the primitive streak mesoblast behind the
point where the epiblast and mesoblast curve over to form the
amnion, and therefore also behind the point where the anal
pit is forming.

This diverticulum increases in size during Stages eG, H, and J,
and forms at the latter stage a very considerable vesicular
cavity opening by a narrow neck into the (future hind-gut)
yolk-sac beneath. The hypoblast diverticulum is formed of
rounded cells, and is surrounded by a mass of mesoblast through
which blood-vessels already ramify. The relation of these
parts is shown in figs. 35 and 50.

The Vascular System.—In the earliest embryo I have examined
of Stage z, viz. one with only a single protovertebra, the posi-
tion of the heart is already indicated, and vessels are already
formed in the splanchnic mesoblast of the blastoderm outside
the embryonic area. Blood-corpuscles are, moreover, to be
seen within these vessels even at this early age.

At the close of Stage r, the rudiments of the dorsal aorta
are present, lying some distance on each side the notochord
and extending from a point on a level with the front end of
the heart backwards to the last protovertebra (fig. 17). They

130 WALTER HEAPS.

do not, however, as yet form continuous tubes. From the front
end of the aorta on each side a short vessel is given off which
lies dorsal to the aorta and immediately below the nervous
system ; it does not, however, extend far. Thereis no commu-
nication between the aortz and the heart tubes at this stage.

At the commencement of Stage n the two tubes of the heart
have met at their anterior end, and form a single wide tube for
a short distance (figs. 24 and 25), a single pair of aortic arches
are formed and the dorsal aorte extend backwards as two
separate tubes some distance beyond the last protovertebra ;
just before they terminate they give off two vitelline arteries.

A series of short diverticula project from the aorta dorso-
laterally between the somites, and ventrally, below them at
this stage and during Stage 3 (compare figs. 2731 and 52).

From near the front end of the aortz, a little posterior to
the point where the aortic arch runs into it, two internal carotid
arteries are projected forwards and extend to the under surface
of the optic lobes (fig. 21, z.c. a.) ; while from about the same
point two vessels run backwards joined at intervals with the
aorte (fig. 24, v. a.) on each side of, and closely applied to, the
now closed neural canal. These vessels run back to a point
just in front of the first protovertebra and are doubtless the
vertebral arteries.

Stage gs shows little alteration ; the heart is still in the form
of a straight tube somewhat longer than in Stage u, but without
curvature or any sign of a division into chambers; there is still
also only one pair of aortic arches, and two separate aorte are
still. present throughout the extent of their course.

A number of small vessels are now given off from the internal
carotid arteries, and the aortz in their anterior portion also
send short branches into the surrounding tissue. The vessels
which I have before described, running backwards on each side
the nervous system, are frequently in communication with the
aorte, and it is these vessels which appear at this stage to
project diverticula into the substance of the walls of the spinal
canal (vide above) (fig. 43).

The vitelline arteries are given off about on a level with the

THE DEVELOPMENT OF THE MOLE. 131

ninth and tenth protovertebre as a series of branches, after
which the aortz immediately become reduced to very minute
proportions.

The venous system, which is barely distinguishable in Stage
H, is very slightly developed in Stage 3. The vitelline veins
run in the converging folds of the splanchnopleure to the
posterior end of the heart on a level with the second and third
protovertebre.

The only veins in the trunk of the embryo are two slightly
developed anterior cardinal veins which are situated on the
outer edge of the anterior protovertebre (fig. 47, a. c. v.).
They communicate with the ductus Cuvieri where the vitelline
veins run into the heart between the second and third proto-
vertebree, and run forwards as far as the first protovertebra.

Traces of a posterior cardinal vein may be seen for some
little distance behind the ductus Cuvieri; but as a vessel it
exists only for a few sections, and is situated at the point
where the somatopleure commences to turn upwards to form
the amnion

Thus it may be observed the arterial system is in a far more
advanced condition than is the venous system in the body of
the embryo.

The Structure of the Heart—In Stage & the heart merely
consists of a small tube in the thickened splanchnic mesoblast
on either side, in front of the protovertebre (fig. 14). Then
(Stage r) the thickened portion is bulged outwards into the
body cavity and splits up into two layers. The outer layer
bounding the body cavity forms the wall of the heart itself, the
inner the flattened epithelial lining of the cavity of the heart.
The space between these two layers increases and in Stage u
(fig. 25) is considerable. In this figure the epithelial layer is
connected with the outer layer of the heart by long proto-
plasmic processes stretching from cell to cell across the space.
In Stage s the wall of the heart has increased in size more in
proportion than has the inner epithelial layer. The latter is
now an elongated bag within the space contained by the outer
wall and connected with the latter by marvellously delicate

132 WALTER HEAPE.

simple or branched cell processes (fig. 47). At the points
where the cavity of the heart is continuous with the vessels
entering into and emanating from the heart, the epithelial
layer is continuous with the wall of these vessels. As I have
stated above, the heart shows no indication of curvature or of
division into chambers.

The Blood-Corpuscles are formed from stellate mesoderm
cells. The nuclei of these cells become darker, the stellate
processes are then withdrawn and a meagre coating of proto-
plasm surrounds the now rounded nucleus. Such conditions
and changes are shown in many of the figures I have drawn ;
notably in fig. 25 in the heart, and in fig. 28 in the vitelline
vessels.

SuMMARY.

External Features. — The early appearance of the optic
grooves (Stage E) which give rise to the optic vesicles; the
existence of five visceral arches in Stage 3; the formation of
the amnion first at the hind end of the embryo; and the
folding off of the head end of the embryo only, are the chief
points to be noted. The enclosure of the front end of the
primitive streak within the medullary fold; the formation of
protovertebre, chiefly from before backwards; the closure of
the medullary groove; the appearance of three divisions of the
brain, and the formation of the heart are also detailed.

The Epiblast.—The epiblast of the embryo (Stages z—e)
becomes formed into a median thickened portion, the medul-
lary plate, and into lateral portions which are formed of
cubical cells and are continuous with the flattened epiblast
cells which cover the vesicle. The closure of the medullary
groove (Stages H and 3) causes the union of the lateral epiblast
which thus forms a continuous layer across the embryo. The
medullary groove commences about the centre of the embryo,
widening out into the sinus rhomboidalis behind and into the
cephalic plate anteriorly. The optic grooves are formed one
on each side of the middle line in the cephalic plate (figs. 4
and 16).

THE DEVELOPMENT OF THE MOLE. 133

The Medullary Canal.—The closure of the medullary groove
commences in the region of the first protovertebra during
Stage g and proceeds anteriorly and posteriorly, and at the
close of Stage 5 a complete canal is formed as far back as the
last (fourteenth) protovertebra. The lateral walls of the canal
thicken, and are converted into an hour-glass form in places.
The migration of mesoblast (nutritive) cells into the walls of
the canal is noted in Stages H and J.

The Brain.—The three divisions of the brain are indicated in
Stage 5, and a well-marked cranial flexure is then present.
The infundibulum is just apparent at this stage in close con-
nection with the front end of the alimentary canal and noto-
chord (fig. 49).

The Optic Vesicles are formed from the optic grooves by the
closure of the medullary canal. These organs first appear
extremely early, but their development is soon checked, doubt-
less in consequence of the habits of the adult animal.

The Ear in Stage 5 is merely indicated as a deep groove in
a thickened mass of mesoblast on either side of the hind-brain.

The Cranial and Spinal Nerves are not described.

The Hypoblast may be divided into axial and peripheral
portions. The peripheral hypoblast, a single layer of flattened
cells, extends on all sides over the embryonic area during
Stage nr. The deepening of the medullary groove stretches
these cells and flattens them still more, but the thickening of
the lateral mesoblast forces the lateral hypoblast down, removes
the strain, and its cells become rounded.

The Notochord is formed of axial hypoblast cells. In Stage
c a mass of axial hypoblast cells are continuous with two
lateral masses of mesoblast—derived from lateral hypoblast—
and with the lateral hypoblast layer also. In Stages p and &
the axial mass becomes isolated from the lateral mesoblast
plates, and gradually decreases in size below the deepening
medullary groove until in that portion where the groove is
deepest, 1. e. near the centre, a single layer of flattened cells is
all that exist.

It does not, however, become reduced to this extent through-

134 WALTER HEAPE.

out its length ; at the posterior end it remains thickened, and
by the ingrowth of the lateral portions the axial cells first
form an arch and then a complete tube, which is the neuren-
teric canal and which communicates dorsally with the exterior
and ventrally with the yolk-sac.

This tube is the homologue of the median dorsal diverticulum
of the alimentary tract in Amphioxus, i. e. the structure which
gives rise to the notochord of that animal, and it is noteworthy
that in the Mole it disappears almost entirely before the noto-
chord is formed.

The single layer of cells to which the greater part of the
axial hypoblast is reduced at the close of Stage p (No. 8) again
increases in bulk during Stages £ to J, and gives rise to the
notochord.

As was the case with the lateral hypoblast, the flattening
of these cells and their increase in bulk appears to be due, first
to the stretching effect of the rapidly deepening medullary
groove, and secondly to the release from that strain caused by
the depression of the lateral portions of the embryo.

The isolation of the notochord first occurs in the region of
the first protovertebra during Stage Gc, and extends anteriorly
and posteriorly during Stages H and J.

The isolation is caused by the ingrowth of the lateral hypo-
blast below the axial cells, and the latter are isolated either
as a solid band or rod, although a lumen may here and
there appear in it afterwards.

At the close of Stage s the notochord is completely separated
from the hypoblast, except at two points, viz. at the anterior
end, where it is connected with the hypoblast and epiblast,
where these two layers fuse to form the mouth, and posteriorly
where it is joined to both epiblast, hypoblast, and mesoblast,
at the front end of the primitive streak (figs. 49 and 50).

The origin of the notochord and the manner of its isolation
appear to be sufficient reason to regard it as entirely homolo-
gous with the notochord of Amphioxus.

For a review of other opinions on this point I would refer to
a discussion in my former paper (No. 8).

THE DEVELOPMENT OF THE MOLE. 135

The hooked anterior end of the notochord is due to its origin
from the front wall of the fore-gut. Its close approximation
to the fore-brain is noted.

The relatively small size of the notochord to the nervous
system in the Mole is pointed out, and it is suggested the early
development of the latter is the cause of the check adminis-
tered to the growth of the former, a check from which it appears
never entirely to recover.

The Alimentary Canal first appears in Stage p as a short
tubular diverticulum, projecting below the cephalic plate nearly
to the anterior end of the embryo.

The tube enlarges and extends backwards during the pro-
gress of the folding off of the embryo during Stages & to j,
and the cranial flexure causes a ventral enlargement, which
is somewhat posterior to the original anterior diverti-
culum.

The mouth and the visceral clefts are not formed at the close
of Stage ys, but the epiblast and hypoblast have fused at the
point where the mouth will eventually be formed, and several
lateral outgrowths from the now widened fore-gut exist ; in the
case of one of these, the anterior one, the hypoblast has reached
the epiblast, and the two layers are partially fused at that
point.

The mouth is formed at the apex of a Y-shaped groove, the
diverging limbs of which are directed forwards; these grooves
are the anterior border of the first visceral arch.

The primary anterior diverticulum would indicate the
existence primitively of a terminal mouth, while the two
grooves, at the junction of which the mouth is formed,
would suggest a paired origin for the existing mouth of the
animal.

The Mesoblastic Somites and Body Cavity.—The lateral plates
of mesoblast are split horizontally into somatic and splanchnic
layers, but the split is not actually carried through both peri-
pheral and axial portions of the plates, being merely indicated
in Stage E in the axial portion. The mesoblast of the head
also is not split, and no cavity is formed there.

136 WALTER HEAPE.

Protovertebre are formed and the axial and _ peripheral
portions of the mesoblast plates are separated from one another
by the intermediate cell mass.

A cavity appears in the protovertebre, Stage 8, which still
exists at the close of Stage g.

The formation of the muscle-plate commences at Stage
from the outer layer of cells of the protovertebra, but during
Stage J, in the three anterior protovertebre, a second row of
cells derived from the inner (vertebral) portion of the somite,
takes part in its formation, the two rows being continuous with
one another at their dorsal and ventral ends.

The muscle-plates are first formed anteriorly. The outer
cells of the muscle-plate in Stage 5 are prolonged into fine pro-
cesses, which are connected with the overlying epiblast cells,
and constitute voluntary muscular fibres (fig. 51). Certain of
the cells of the inner layer are also differentiated into elongated
muscular fibres.

I would further remark the mesoblast and epiblast cells in
front of the primitive streak appear always to be connected
together by processes.

The Pericardial Cavity only commences to form during
Stage s, and is not at the close of that stage entirely separated
from the remainder of the body cavity.

The Primitive Streak has the same relations in Stages = and F
as in Stage p, except that the medullary folds grow backwards
round its front end. The neurenteric canal disappears, but its
original position is indicated by the fusion of the germinal
layers at the front end of the primitive streak.

The mesoblast of the primitive streak does not give rise to
the mesoblast of the body of the embryo in front of the primi-
tive streak, in my opinion, but extends backwards and outwards
and forms the wall of the allantois. The anus is formed in the
middle of the primitive streak.

The Amnion is first formed at the hind end and from thence
extends forwards. This portion of the amnion is formed of a
double fold of somatopleure ; the epiblast of the outer fold
unites with the epithelium of the uterus. The anterior fold

THE DEVELOPMENT OF THE MOLE. 137

of the amnion, however, is formed only of epiblast and hypo-
blast, and has been called by van Beneden and Ch. Julin, who
first described this structure, the ‘‘ pro-amnion.”’

The Allantois commences in Stage F as a short wide diverti-
culum projecting upwards and backwards into the primitive
streak. This diverticulum enlarges during Stages eG to 3; it
is lined with hypoblast cells (figs. 35 and 50).

The Arterial System.—The dorsal aortz commence in Stage F,
and remain double until after Stage 3; they are connected with
the heart by a single pair of aortic arches during Stages H
and jg, and give off vitelline arteries at their posterior end.
Internal carotid arteries and vertebral arteries are formed, and
it is from the latter of these vessels the mesoblast cells are
derived which migrate into the walls of the neural canal.

The Venous System is very slightly developed. Vessels are
to be seen in the splanchnopleure over the yolk-sac at an early
date, but vitelline veins connected with the heart are not seen
until Stage a. Two short anterior cardinal veins are present
in Stage J, and traces of two posterior cardinals, but nothing
more.

The Heart, which is formed of two tubes widely asunder in
Stage s, is composed of a single tube for a short distance in
Stage H, and is somewhat longer, but still straight and without
sign of division into chambers at the close of Stages. The
thickened splanchnic mesoblast which gives rise to the heart,
splits into two layers at an early age. The outer of these
layers forms the outer wall of the heart, the inner the flattened
epithelium of the cavity of the heart. ,

When the heart enlarges, as it does rapidly, a wide space
exists between these two layers, but they are connected together
by exceedingly fine processes of their cells which stretch across
the space.

The Blood-Corpuscles appear to be formed from stellate
mesoblast cells directly.

In conclusion, I may mention that I propose eventually to
follow the further development of the organs of the Mole, one
by one, and in doing so, to pay more attention to the researches

11

138

WALTER HEAPE.

of other investigators than has appeared to me advisable in the
present paper.

=

LITERATURE.

F. M. Batrour.—‘ Comparative Embryology,’ vol. ii, 1881.

. Ep. van BENEDEN AND Cuas. Jutin.—* Recherches sur la formation

des annexes feetales chez les mammiféres,’ ‘Archives de Biologie,’
vol. v, 1884.

. Tu. L. W. Biscnorr.— Entwicklungsgeschichte des Kaninchens-eies,’

1842.

4, Tu, L. W. Biscuorr.— Entwicklungsgeschichte des Hunde-eies,’ 1845.

11

. Tu. L. W. BiscHorr.— Entwicklungsgeschichte des Meerschweinschens,’

1852.

. Tu. L. W. Biscoorr.—‘ Entwicklungsgeschichte des Rehes,’ 1854.
. W. F. Havsmann.—‘ Ueber die Zeugung und Entstehung des wahren

weiblichen eies bei den Saugethieren und Menschens,’ 1840.

. W. Hearr.— Development of the Mole,” ‘Quart. Journ. of Micr,

Sci.,’ 1883.

. v. Henson.—“ Beobachtungen tiber die Befruchtung und Entwicklung

des Kaninchens und Meerschweinschens,” ‘ Zeit. f. Anat. und Ent-
wickelungsgeschichte,’ vol. i, 1876.

A. Korircer.— Hntwicklungsgeschichte des Menschens und der Hoheren
Thiere, 1879.

W. F. R. Wetpoy.—“ Note on the Early Development of Lacerta
muralis,”’ ‘Quart. Journ. Mier. Sci.,’? 1883.

THE DEVELOPMENT OF THE MOLE. 139

DESCRIPTION OF PLATES XIII, XIV, & XV,

Illustrating Mr. Walter Heape’s Paper on “ The Development
of the Mole (Talpa Europea),” Stages E to J.

List of Reference Letters.

a.are. Aortic arch. a.c.v. Anterior cardinal vein. al. c. Alimentary
canal. ail. Allantois. ail. v. Allantoic vessels. am. Amnion. am. fl. False
amnion. az.p. Anal pit. a.p. Area pellucida. aud.ep. Auditory epithelium.
aud. inv. Auditory involution. 6.c. Body cavity. cc. pl. Cephalic plate.
d.a. Dorsal aorta. ep. Hpiblast. f. dr. Fore-brain. 4%. dr. Hind-brain.
ht. Heart. hy. Hypoblast. 7. c. a. Internal carotid artery. 7@. c. m. Inter-
mediate cell mass. m. Mesoblast. m. dr. Mid-brain. m. gr. Medullary
groove. m.pl. Medullary plate. msc. pl. Muscle-plate. x. c. Neural canal.
uch. Notochord. op. gr. Optic grooves. op. v. Optic vesicles. p.c. Peri-
cardial cavity. pro. am. pro-amnion. . st. Primitive streak. yp. v. Proto-
vertebra. S%. 74. Sinus rhomboidalis. so. m. Somatic mesoblast. so. pl.
Somatopleure. sp. m. Splanchnic mesoblast. sp. p/. Splanchnopleure. 2. a.
Vertebral artery. vs. ach. Visceral arch. v¢. a. Vitelline artery. vt. v. Vi-
telline vessels. v¢. vz. Vitelline vein.

Figs. 1—10 were sketched with Zeiss’s a* lens and eye-piece 2 by myself,
and were most carefully shaded by Mr. H. A. Chapman under my supervision.

Figs. 11—35, 44—50, and 52 were sketched with Zeiss’s B lens and eye-
piece 2.

Figs. 36—43 and 51 were sketched with Zeiss’s D lens and eye-piece 2.

Fic. 1, Stage E.—Transparent view of embryo ‘76 mm. long. It has three
protovertebre. The medullary groove is narrow in the middle of the body,
widening out at either end. The anterior end of the primitive streak projects
as a dark knob into the wide sinus rhomboidalis. The flattened cephalic plate
(c. pl.) and the area pellucida (a. p.) are to be observed. .

Fic. 2, Stage H—Surface view of embryo 1°82 mm. long. The cephalic
plate has now two deep lateral grooves, the optic grooves (op.gr.). The
curved condition of the embryo is due to careless manipulation.

Fic. 3, Stage F.—Surface view of embryo 1°96 mm. long. The medullary
groove has commenced to close in the region of the protovertebre (which are
not shown in this drawing), but the edges of the groove have not yet
coalesced. ‘The optic grooves are seen at either side of the cephalic plate.

Fic. 4, Stage F.—Surlace view of embryo 2°12 mm, long. Although some-

140 WALTER HEAPR.

what bigger than the embryo drawn in Fig. 3, the medullary folds have not
advanced so far along the body of the embryo as they have in the latter
embryo. At the anterior end, however, they are slightly more advanced, and
where the folds meet in front a narrow slit is to be seen. The amnion is
shown covering the posterior half of the embryo, and the wide sinus rhom-
boidalis is indicated below it.

Fic. 5, Stage F.—Transparent view of the same embryo, seen from below.
Four fully formed protovertebre are present, anda fifth is indicated behind
the posterior one. The primitive streak projects into the sinus rhomboidalis.
The optic grooves appear as narrow lateral prolongations of the medullary
groove at its anterior end. The heart (4¢.) commences to form at this stage,
and is indicated by a thickening of the blastoderm on either side the embryo
just behind and outside the optic grooves.

Fic. 6, Stage G.—Surface view of embryo 2°33 mm. long. The medullary
groove is closed up to the anterior end, where a small pore remains connecting
the medullary canal with the exterior. The sinus rhomboidalis is still widely
open behind. The head has now been folded off from the yolk-sac as far as
the line so. pl., which shows the point of divergence of the folds of somatopleure.
Faint indications of the divisions of the brain are shown, and the laterally
projecting optic vesicles are very distinct (op. v.).

Fie. 7, Stage H.—Surface view of embryo 2°2 mm. long. Ten proto-
vertebra are present. The closure of the medullary groove has advanced.
The sinus rhomboidalis is narrowed, and the primitive streak forced upwards
as a rounded knob at the posterior end of the latter. The head is more
rounded, and shows partial division into fore-, mid-, and hind-brains. The
amnion has been torn away, and the jagged edge of the somatopleure sur-
rounds the body of the embryo.

Fic. 8, Stage H.—View of the under surface of the head of the same
embryo, showing the heart and the diverging folds of somatopleure and
splanchnopleure.

Fie. 9, Stage J—Surface view of an embryo 3:06 mm. long. The sinus
rhomboidalis is much narrowed, and the medullary groove closed for the
greater portion of its length. The optic vesicles and fore-, mid-, and hind-brains
are well shown. Thirteen protovertebre are present. The primitive streak is
in the same condition as described for Fig. 7, also the amnion has been torn
away as it was in that figure,

Fie. 10.—Lateral view of the head of the same embryo, showing the heart
and five visceral arches.

Fie. 11, Stage E.—Median longitudinal section of the anterior end of an
embryo with three protovertebre. The cephalic plate projects slightly over
the blastoderm in front, the folding-off process having already begun in this
embryo. The commencement of the fore-gut is indicated at al.c. A small
portion of mesoblast exists between the epiblast and hypoblast of the blasto-

THE DEVELOPMEN'T OF THE MOLE. 141]

derm at the front end of the embryo; beyond that point no mesoblast is
present in the middle line.

Fie. 12, Stage E.—Transverse section through the cephalic plate of an
embryo with three protovertebre. At the point where this section is taken
the flat cephalic plate is completely folded off from the yolk-sac. The narrow
fore-gut is shown as a tube (a/. c.) immediately below the cephalic plate. A
few scattered mesoblast cells extend between the two layers of epiblast.

Fie. 18, Stage E.—Transverse section through an embryo 1:97 mm. long,
with only a single protovertebra. The section is taken in front of the proto-
vertebra, and shows the indication of a split of the mesoblast into somatic and
splanchnic layers throughout its whole depth. The medullary groove is wide
and deep. The notochord is formed of flattened cells.

Figs. 14 and 15, Stage H.—Tranverse sections through the same embryo
which is drawn in Fig. 1.
Fig. 14 is taken in front of the protovertebra, where the mesoblast is
split into somatic and splanchnic layers only at the periphery.
Fig. 15 passes through a protovertebra. The body cavity extends in-
wards as far as the intermediate cell mass (¢.c. m.) in the peripheral
mesoblast, and a small cavity is also present within the protovertebra.

Fies. 16, 17, and 18, Stage F.—Transverse sections through embryos with
five protovertebre.

Fig. 16 is a section through the head. The cephalic plate is grooved in
the middle line and at either side where the wide optic grooves are
situated. When the external edges meet in the middle line these optic
grooves will be converted into vesicles communicating by a wide aper-
ture with the central canal. The notochord is not yet separated from
the hypoblast.

Fig. 17 is through the trunk ; the medullary groove is narrower, and the
notochord more defined than in Fig. 15, which is a section through a
similar region of an embryo of Stage HE.

Fig. 18 is a section through the sinus rhomboidalis, and shows the
anterior end of the primitive streak and the amnion.

Fie. 19, Stage G.—Transverse section through the head of an embryo with
eight protovertebre, 2°49 mm. long. The head at this point is completely
folded off, and the medullary groove (still open) will at this point give rise to
the mid-brain. A few sections further forward the optic vesicles are cut,
projecting outwards from the central canal, and it is on account of the
proximity of these structures that the wide space between the external
epiblast and the walls of the medullary canal is present here. This space is
here filled with stellate mesoblast cells. The two grooves in the epiblast on
the under surface on either side the middle line converge posterioriy, and
where they meet the mouth will eventually be formed.

142 WALTER HBAPE.

Fics. 20—33, Stage H.—Transverse sections through three embryos of this
slage.

Fig. 20 is a section through the front of the head; it passes through the
point of origin of the optic vesicles, and shows at the same time the
pore through which the neural canal is open to the exterior at this
stage.

Fig. 21 passes through both the mid- and fore-brains and through the
centre of the optic vesicles, which are here seen to be directed out-
wards, downwards, and backwards.

Fig. 22 passes through the hind-brain and the front end of the fore-gut
(al. c.). The notochord is not yet separated from the axial hypoblast
here. The front edge of the first aortic arch is shown. ‘This vessel is
very wide, and may be seen for many sections.

Fig. 28 is also a section through the hind-brain, but at its posterior end.
The notochord is here isolated from the hypoblast. The two grooves
in the ventral epiblast on either side the middle line, which were seen
in Fig. 22, have met in this figure and form a single deep groove
closely in contact with the ventral wall of the fore-gut, and here the
mouth will be formed. These grooves define the anterior border of the
first visceral arch (¢s. ach.). The alimentary canal in this figure is
very considerably wider than in Fig. 22. In Figs. 20 to 28, the head
of the embryo is folded off from the yolk-sac.

Fig. 24 is taken from a different embryo from what Figs. 20—23, and 25
are taken. The front end of the heart is shown. The section is not
quite transverse, and the first aortic arch is shown on the right side
and not on the left side. The extremely wide fore-gut and the separa-
tion of the heart into two portions, shown here, is also due to this fact.

Fig. 25. The alimentary canal is here open ventrally. In Fig. 24 the
splanchnopleure had formed a complete layer, but the somatopleure had
not met below the gut. In this figure the splanchnopleure as well as
the somatopleure are still divergent. The heart is here in the form of
two tubes, and the two layers of which it is formed may here be seen.
The formation of blood-corpuscles from stellate mesoblast cells also
may be observed. The thickened epiblast on either side the neural
canal is the commencement of the auditory organ.

Fig. 26. A section immediately in front of the first protovertebra. The
vitelline vein is seen in the splanchnopleure, branching out over the
yolk-sac. The fold of the somatopleure to form the amnion is also
indicated (am.).

Fig. 27. A section through the anterior protovertebra.

Fig. 28. A section through the middle of the embryo. The large vitel-
line vessels are shown in the splanchnic layer of mesoblast over the
yolk-sac. The true (am.) and false (am. fis.) amnion are both shown

THE DEVELOPMENT OF THE MOLE. 143

here. The false amnion is formed of flattened somatic mesoblast and
columnar epiblast cells, the latter will eventually fuse with the
uterine epithelium.

Fig. 29 is also a section through the middle of the embryo. The cells of
the protovertebre are here seen to be somewhat elongated on the right
side of the section, while on the left the cavity of the protovertebra is
shown partially filled by a cove of mesoblast cells. This was also
shown in Fig. 28. The hour-glass shape of the neural canal at this
point is also to be observed.

Figs. 30 and 31 are from the hinder portion of the trunk of the embryo.
The neural canal is not closed, the protovertebre are not so com-
pletely isolated from the neighbouring mesoblast, and the notochord,
which is larger than in former sections of this stage, is not at all
isolated from the hypoblast in Fig. 31. In both these sections the two
dorsal aortz, which were present in all the sections from Fig. 23 to
Fig. 29, are here giving off branches to the yolk-sac. The vitelline
arteries (v¢. a.), and posterior to this point, the aorte themselves, no
longer exist.

Fig. 32 is a section behind the former sections, and just in front of the
primitive streak. ‘I'he widely open medullary groove is here called the
sinus rhomboidalis.

Fig. 33 is a section through the primitive streak; the medullary folds are
growing round it and will shortly completely enclose its front end.

Fic. 34, Stage H.—A median longitudinal section through the head of an
embryo, in which the following points are shown:—The cranial flexure; the
fore-, mid-, and hind-brains; the notochord separated from the hypoblast,
except along the front wall of the alimentary canal ; the ventral prolongation
from the primitively straight fore-gut (a/. c.), and the NRE; -amnion formed of
epiblast and hypoblast only (pro. am.).

Fie. 35, Stage H.—Median longitudinal section eee the hind end of an
embryo. ‘The dorsal pit (anal pit) and the ventral pit (allantoic pit) separate
the anterior from the posterior portions of the primitive streak.

Fies. 36—42, Stage H.—Transverse sections through various regions of an
embryo, to show the formation of the notochord.

Fig. 36 in the anterior region shows a lumen within the notochord.

Fig. 37. In the anterior region: notochord is rod like, and is separated
from the hypoblast.

Fig. 38. In the anterior region: notochord is a flattened band-like siruc-
ture ; it is separated from the hypoblast.

Fig. 39. In the middle region: the notochord is not yet separated from
the hypoblast in the middle line, although it is so separated at either
edge. The lateral ingrowth of the hypoblast is shown.

Fig. 40, from the posterior region, shows relations similar to those seen

144. WALTER HEAPE.

in Fig. 39, only the notochordal mass is itself considerably larger the
lateral ingrowth of the hypoblast is here also indicated.

Fig. 41. From still further posteriorwards: the notochord is not yet
isolated from the hypoblast, but formed into an arc.

Fig. 42. From the hind end of the embryo, immediately in front of the
primitive streak : the notochord is a large thickened axial mass, with
no indication of the growth of the hypoblast below it.

Fic. 43, Stage H.—A transverse section through the medullary cord of
an embryo with eleven protovertebre, from the region in front of the first
protovertebra and behind the hind-brain. Between the lateral mesoblast
plate and the cord is a small space, in which several nuclei are seen. The
space is continuous with blood-vessels in process of formation, and the nuclei
show a tendency to pass into the medullary cord. One such nucleus is shown
in the drawing.

Fics. 44 and 45, Stage J.—Transverse sections through the hind-brain of
an embryo with fourteen protovertebre. The alimentary canal is narrow in
front (Fig. 44), and wider posteriorly (Fig. 45). The two grooves in the
epiblast on the under surface in the anterior section converge in a single deeper
groove in the posterior section, where the fusion of the epiblast and hypoblast
takes place, and where the mouth will eventually be formed. The dorso-
ventral elongation of the fore-gut and the notochord is due to the plane in
which the section was cut, caused by the cranial flexure. The presence of
mesoblast cells between the notochord and the floor of the brain is to be
noticed.

Fic. 46, Stage J, is a section, not completely transverse, through an embryo
with fourteen protovertebre, passing through the hind-brain and the auditory
mvolution. The first aortic arch is shown on one side, and a lateral prolonga-
tion of the fore-gut to form the first visceral cleft on the other side.

Fic. 47, Stage J.—Transverse section through an embryo with thirteen
protovertebre in the region of the second protovertebra. Fore-gut crescent
shaped. Anterior cardinal veins and dorsal aorte present. Hmbryo is com-
pletely folded off from the yolk-sac. The heart is enclosed in the peri-
cardium. The thick outer wall and flattened epithelial layer of the heart are
here seen to be connected by fine processes of the cells forming one or other
of these layers.

Fic. 48, Stage J.—A transverse section through the primitive streak of an
embryo with fourteen protovertebre. The thickened lateral mesoblast will be
seen, by comparing this section with that drawn in Fig. 50, to be concerned in
the formation of the allantois. Allantoic vessels (a//. v.) are to be seen in this
section.

Fic. 49, Stage JW—A median longitudinal section through the head of an
embryo with fourteen protovertebre. ‘The division between the fore- and
mid-brains and the folded floor and thin roof of the hind-brain is shown. The

srk rasa

ong?

amy

ae:
~

Studies M.L.Vol.IIL, Pl. XIl.

noch

F Huth, Lith? Edin®

ae
2 ue
Ly

C)

2 é

Studies M.L VoL.IIL, P1.XIV.

F. Huth, Lith’ Edin®

————

|
|
|

Studies M.L. Vol. IIL, Pl. XV.

F Huth, Lith Litht

r

a1

as

on

me va

>
bad

iS hd

ar
~~

THE DEVELOPMENT OF THE MOLE. 145

cranial flexure, ventral prolongation of the anterior end of the fore-gut, and
the hooked anterior end of the notochord is also indicated. The notochord is
seen to be continuous at its anterior end with the hypoblast and epiblast at
the point where the mouth will eventually be formed. The existence of
mesoblast cells between the notochord and the floor of the brain is to be
noticed.

Fie. 50, Stage J—A median longitudinal section through the hind end of
an embryo with fourteen protovertebre. The allantoic cavity has increased
in size, and numerous blood-vessels are seen in its walls. The mesoblast
surrounding the allantoic cavity is derived from the primitive streak meso-
blast. The relation of the epiblast, hypoblast, notochord, and mesoblast at
the front end of the primitive streak is seen to be precisely the same as I
indicated in a diagram (Fig. 50) of my former paper (No. 8). The neurenteric
canal is obliterated.

Fic. 51, Stage J—A transverse section of a portion of the muscle-plate of
an embryo with fourteen protovertebre. The cells of the muscle-plate
(msc. pl.) are extended into long processes which are continuous with the
epiblast cells (ep.) lying above them. These processes are commencing
muscular fibres.

Fie. 52, Stage J—Longitudinal section through an embryo with fourteen
protovertebre. The section is taken through a line about half way between
a perpendicular and horizontal longitudinal line, and bisects the muscle-plates
and aorta of one side of the body. The hypoblast lining the alimentary canal
is cut through slightly to one side of the middle line, and the notochord
therefore is not shown. The arched muscle-plate and muscular processes of
its cells, the aorta and its dorso-lateral prolongations between the proto-
vertebra are shown.

oe

ON THE LIFE-HISTORY OF PEDICKELLINA,. 147

On the Life-History of Pedicellina.
By

Sidney F. Harmer, B.A., B.Sc.,
Fellow of King’s College, Cambridge, and of University College, London.

With Plates XVI and XVII.

Durine the summer of 1885, spent in Rocquaine Bay,
Guernsey, I succeeded in obtaining material for the study of
the metamorphosis of Pedicellina echinata, a form which
occurs in great abundance (in Rocquaine Bay) on Coralline
growing under the shade of other seaweeds in tide-pools.

The larve of Pedicellina invariably refused to fix them-
selves when kept in a small quantity of water, and I therefore
ultimately adopted the following method for procuring the
various stages necessary for the investigation.

Adult colonies were placed, after the removal of all super-
fluous parts of the Coralline on which they were growing, in a
small vessel, the mouth of which was closed by a piece of linen.
The vessel was then left for a day or more in a tide-pool, after
which a careful search (with the aid of a low power) over the
Coralline was generally rewarded by the discovery of several
young Pedicellina, which had resulted from larve hatched in
the tide-pool, and which, owing to their inability to escape
from the vessel in which they were confined, had been obliged
to fix on the Coralline. After preservation with corrosive
sublimate and decalcification of the Alga, sections were easily
prepared. In this manner, I succeeded in obtaining numerous
individuals of various ages, fixed under perfectly normal con-
ditions.

148 SIDNEY F. HARMHR.

My study of the metamorphosis of Pedicellina has led me
in the main to a complete confirmation of the account already
given by Barrois (No. 3), and summarized on pp. 312 and 313
of my previous paper on Loxosoma (No. 4), where I have
ventured on a criticism of Barrois’ conclusions which I do not
find to be jusiified by my own investigation of the subject.
In opposition to my previous opinion, 1 must now conclude
that the post-larval changes consist in a remarkable metamor-
phosis, and that the first bud is formed after the primary
individual has acquired its adult characters. Barrois has
published no figures illustrative of his statements, the actual
details of the process being difficult to understand from his
very short description, whilst the morphological nature of the
changes remains entirely obscure. The subject appears to me,
therefore, to deserve further consideration.

The structure of the Jarva is well kuown from tlie researches
of Haischek,! and it will be unnecessary to describe it in more
than a few of its details.

In the swimming attitude of the larva, the ciliated ring is
everted to the exterior, whilst from the oral face project two
prominent structures ;—the epistome, with iis tuft of long cilia,
and the anal cone, on which opens the anus. During the re-
tracted condition, however, the ciliated ring is reflected to the
interior of the large vestibular cavity, whose outer walls are
formed by the fold of skin which bears tlie ciliated ring itself
(cf. Pl. XVI, fig. 1). The floor of the vestibule is constituted
by the ventral or oral surface of the larva, being specially de-
pressed between the base of the epistome and the anal cone,
and at the sides of the latter.

As Barrois has correctly stated, fixation takes place by the
oral surface, the larva being meanwhile in its “retracted” condi-
tion. Pl. XVI, fig. 1, a median longitudinal section, will
serve to illustrate the method of fixation. It will be noticed
that the long axis of the stomach is approximately parallel
to the surface of attachment.

1 Vide the summary of Hatschek’s results in Balfour, ‘Comp. Emb.,’ vol
pp- 242—246.

ON THE LIFE-HISTORY OF PEDICELLINA. 149

Fig. 3 represents a horizontal section of a larva not long
after its fixation: the occurrence of brain (= “ dorsal organ,”
v. No. 4), cesophagus, and rectum in the figure sufficiently de-
fines the level of the section. The epistome is cut in the region
of its greatest thickness, whilst at the summit of the anal
cone is seen the depression into which opens the anus. By
comparing this with fig. 1, it will be observed that the anus
has already altered its position, since it is now directed some-
what forwards, the rectum being more nearly parallel to the
stomach than before. The cells of the vestibular epithelium
are very high at the sides of the anal cone, and are character-
ized by the special readiness with which they take up colouring
matters.

Fig. 5 represents a horizontal section through the apices of
the epistome and anal cone of another individual of the same
age. The epistome is here seen to be continuous, at each side,
with a fold of vestibular epithelium; epistome and folds
together forming (as seen in this section) a horseshoe-shaped
ridge partially embracing the sides of the anal cone, in which
region the two lateral folds become evanescent. The result of
this arrangement is the formation of a somewhat deep ciliated
groove (0.g.) running round the greater part of the vestibule,
and passing in front into the transversely elongated, funnel-
shaped mouth. Posteriorly, however, owing to the disappear-
ance of the lateral folds, the oral grooves fade away at the
sides of the anus, where vestibule and oral grooves conse-
quently appear continuous in such a section as that represented
in fig. 5. The relations of these structures will become more
clear on reference to fig. 4, a larva somewhat older than those
previously described, the section passing transversely through
the region of the anal cone, in the plane a B in fig. 1. At the
sides of the anal cone are the two lateral portions of the vesti-
bule (/.v.), these structures being separated from the oral
grooves by the folds already mentioned. Inthe more anterior
sections of the series, the lateral folds become continuous with
the epistome, and the oral grooves with the mouth. Further
back, on the contrary, the folds become lower, and finally dis-

150 SIDNEY F. HARMER.

appear, so that the oral grooves are not distinguishable in the
post-anal region of the vestibule. The above description,
together with a reference to fig. 5, will thus show that the deep
post-anal groove (m.v.) of fig. 1 is continuous equally with the
oral grooves and with the general vestibular cavity. For
further clearness, dotted lines in the same figure indicate the
position and relations of the right lateral fold as it would appear
by looking at the wall of the vestibule from the inside of the
latter. The relations of half of the ciliated ring and of the
right oral groove are also shown in the figure.

Fig. 2 represents a longitudinal section of a recently-fixed
larva, passing in the direction of the line cp in figs. 3 and 4.
One of the lateral folds, owing to its projection inwards into
the vestibule, separates the latter into two portions, containing
respectively the mouth (and oral groove) and part of the epis-
tome. The latter portion obviously corresponds to one of the
lateral regions of the vestibule (/. v.) in fig. 3. Fig. 2 further
explains the continuity of the tip of the epistome with the
lateral folds (cf. figs. 1 and 5). In more median sections of
the same series the latter are not seen, the epistome being per-
fectly free at its apex, whilst the separation of the vestibular
cavity into two parts is not apparent.

A considerable portion of the base of the epistome and of
the sides of the anal cone is formed of a remarkable tissue,
composed of large cells, with transparent contents, hardly
staining with colouring matters (fig. 2, 7). The nature of
this tissue (which atrophies during the metamorphosis) is
unknown to me.

The revolution (about to be described) of the alimentary
canal was obviously well understood by Barrois, although I
did not formerly succeed in making out his exceedingly con-
cise statements on this head.

Figs. 8 and 9 represent two sections of an obliquely longitu-
dinal series through a more advanced stage. Fig. 9 involves
the rectum, whilst fig. 8 shows the mouth and esophagus. In
the latter figure is seen one of the deep portions of the vesti-
bule lying at the sides of the rectum, which is itself of course

ON THE LIFE-HISTORY OF PEDICELLINA. 151

not visible. The dorsal organ and the sucker have both dege-
nerated, and are represented merely by the “ globules” de-
scribed by Barrois in various parts of the larva after its
metamorphosis. These “globules” are rounded nucleated
cells, which do not stain readily with reagents, their general
form being shown in fig. 8, &c.

It is obvious, from an inspection of the two sections figured,
that the stomach has now taken up a position inclined to the
surface of attachment, the concavity of the alimentary canal
being directed somewhat backwards.

Remarkable changes, already described in part by Barrois,
have by this time occurred.!

Fig. 9 shows that the aperture of the vestibule has closed,
so that this cavity has no longer any communication with the
exterior. The vestibule is partially divided into three por-
tions, which do not, however, quite correspond with those
described by Barrois. The most ventral portion (v. v. in
fig. 9) corresponds to the region near the previous vestibular
aperture, and is destined to atrophy completely. The next
portion (v. 07.) is in connection with the mouth (fig. 8), whilst
the most dorsal portion (v. an.) contains the anal cone, and is
at this stage and later the largest and most important part of
the vestibule. The second or oral division still communicates
with the ventral portion, whilst it is almost separated from the
dorsal or anal division by the growth of the epistome and of
the lateral folds.

In another section of the series it is seen that the oral and
anal divisions of the vestibule still communicate by a small
aperture, as in the diagram, fig. 16 (a. v. v.).

The anal portion of the vestibule is very large, and is grow-
ing, at the previously posterior end of the larva, away from
the surface of attachment. The cells lining this part of the
vestibule are obviously engaged in active growth and multi-

1 The following statements will be more readily understood with the
assistance of Pl. XVII, fig. 16, representing in a diagrammatic form a
median longitudinal section through an individual of the same age as figs.
8 and 9.

152 SIDNEY F. HARMER.

plication, their protoplasm being finely granular and staining
readily with colouring matters. The backward growth of the
vestibule occurs first in the regions at the two sides of the anal
cone (cf. fig. 3), but soon extends to the median portion be-
hind the cone (fig. 9), so that this part of the vestibule grows
towards the free end of the fixed larva, during the rotation of
the alimentary canal, as a single actively extending diverticu-
lum, in which the primary differentiation of median and lateral
regions is no longer marked.

Fig. 6 will serve to explain more clearly the relations of the
oral grooves and neighbouring structures at a stage very slightly
earlier than that of figs. 8 and 9. The section passes in a direc-
tion corresponding to the line kK 1 in fig. 16, and consequently
involves the apex of the epistome, the lateral folds, and the
oral grooves. The anal cone, visible in fig. 5, is, of course, not
involved by the section, which in other respects differs from the
former figure mainly in the facts that the diameter of this por-
tion of the vestibule has become lessened, and that by the par-
tial rotation of the alimentary canal the apex of the epistome
has come nearly into contact with the posterior wall of the
vestibule (the manner in which this happens being understood
by comparing fig. 1 with fig. 16), whilst the form of the lateral
folds is at the same time altered (cf. fig. 6 with fig. 5). By this
change of position of epistome and lateral folds, the oral and
anal sections of the vestibule communicate merely by a com-
paratively small round aperture. The oral grooves are no
longer continuous posteriorly with the anal portion of the
vestibule, although on the left side of the section at least, a trace
of the former continuity is distinguishable. During later stages
the growth of epistome and lateral folds completely separates
the oral from the anal division of the vestibule, the aperture
a.v.v. in fig. 6 being gradually constricted until it finally
disappears.

At the stage of figs. 8 and 9 a considerable amount of his-
tolysis is taking place. This process affects specially the
stomach, the epistome, the anal cone, and the ventral portion
of the vestibule. In the case of the stomach, portions of the

ON THE LIFE-HISTORY OF PEDICELLINA. 153

epithelial cells and some of their nuclei pass bodily into the
lumen of the organ (ef. figs. 8 and 9), where they are found
quite free at later stages. The more projecting parts of the
epistome and of the anal cone lose most of their component
cells. The cilia of the former become indistinct, the cell-sub-
stance itself obviously degenerating (fig. 9). Ultimately ciliated
portions of the cells are thrown off into the vestibule (figs. 9
and 12), in which they can be discovered until a very late stage
in the metamorphosis. They no doubt leave the vestibule
either by the mouth or by the (adult) vestibular aperture,
when the latter is formed.

The histolysis of the ventral portion of the vestibule (fig. 9,
v. v.) similarly results in the passage of fragments of cells into
its own cavity.

This process is again illustrated by fig. 12, a section passing
in the plane of the line EF in fig. 9. The permanent vestibule
is in this section (cf. fig. 16) completely separated from the
degenerating portion, its lumen, like that of the latter, con-
taining fragments of degenerating cells.

The ventral division of the vestibule (v. v.) in fig. 9 occupies
the position of the future stalk, and in later stages its cavity
becomes more and more reduced until it finally atrophies.
During this process, the cells previously found in its lumen
disappear. In sections parallel to the plane of attachment the
cavity (just before its atrophy) appears as a fine tube surrounded
by a series of elongated cells radiating from it towards the
body wall. It is very tempting to assume that these cells are
phagocytes, engaged in the destruction of the vestrbule. After
the atrophy of the latter, its place is occupied by numerous
*‘ lobules” (fig. 10), which will themselves be replaced by
ordinary connective-tissue corpuscles (fig. 13).

The same assertion may be made of other parts of the
‘‘ primary body cavity,” which is at the stage of fig. 9 almost
completely filled with ‘ globules,” resulting from the histolysis
of the brain, the sucker, the tissue at the base of the epistome
and anal cone, and other larval structures. When the primary
individual is mature the ‘‘ globules” have disappeared, aud are

12

154 SIDNEY F. HARMER.

replaced by a gelatinous matrix, in which lie connective-tissue
corpuscles. Are we not justified in assuming that the
“‘slobules” are the active agents in the histolysis, and that
they are in fact typical phagocytes ?

During the histolysis of portions of the anal cone, the latter
structure itself becomes much depressed. This feature of the
metamorphosis, although already obvious in fig. 9, may be
further illustrated by means of fig. 7, a section passing in a
plane corresponding to the line 1 g in fig. 16.

Owing to the further depression (occurring at a slightly later
stage) of the anal cone, the marked bilateral arrangement of
this part of the vestibule is, in part at least,lost. At the stage
of figs. 8 and 9, as cau be easily seen from these figures them-
selves, the posterior portion of the vestibule is no longer re-
duced in the median plane to a small slit between arial cone
and vestibular wall (as in fig. 1), but is, in this position also,
a spacious cavity lined by a columnar epithelium (fig. 9).

After the anal cone has reached the condition of the latter
figure the vestibule, in sections parallel to the long axis of the
stomach, will usually appear bounded posteriorly by a simple
uniformly curved wall, whilst its cesophageal side is floored by
the degenerating tissue of the epistome (fig. 7). In later
stages, however, the well-developed epithelium of the sides of
the vestibule extends inwards, so that the cavity is then en-
tirely bounded by its permanent, partially regenerated epi-
thelium.

In the next stage represented very considerable changes
have occurred, whereby the alimentary canal has taken up a
position not unlike that which it will ultimately retain.
Fig. 10 represents an actual section which passes in the median
longitudinal plane of a larva at this stage. Whereas in
fig. 16 the axis of the stomach is but slightly inclined to the
surface of attachment, in the present instance it has assumed
a position almost at right angles to this plane, and the con-
cavity of the gut is now directed towards the primitively pos-
terior end of the fixed larva. In the course of this rotation
of the alimentary canal the vestibule, owing to atrophy of one

ON THE LIFE-HISTORY OF PEDIOELLINA. 155

at least of the portions described in the last stage, has become
somewhat simplified. All the more ventral regions (situ-
ated in the neighbourhood of the surface of attachment) have
completely disappeared, and in their place is found a mass of
cells filling a cylindrical stalk, which obviously corresponds to
that of the adult Pedicellina. The anal division of the ves-
tibule has continued its backward growth and now lies almost
at the free end of the young animal. At about this stage it
acquires a secondary opening to the exterior on the side corre-
sponding to the posterior surface of the larva. This opening
is formed by a simple concrescence between the vestibular
epithelium and the external ectoderm of the body, accompa-
nied by a linear perforation formed at the point of junction
of these two distinct portions of ectoderm. My sections have
given me no indication of the occurrence of a “ labial invagi-
nation” (Barrois, q. v.) placing the above portion of the ves-
tibule in connection with the exterior.

The character of the vestibular aperture, immediately after
its formation, may be seen from fig. 1], a section passing in a
plane corresponding to eH in fig. 10. The vestibular aperture,
at the sides of which tentacles (¢.) are already developing, is
shown, by an examination of the remaining sections of the
series, to have the form of a slit elongated in the direction of
the median plane of the animal. Immediately before the for-
mation of the aperture the vestibular epithelium would appear,
in a section of this kind, quite unconnected with the external
ectoderm, but already extending towards it in the form of a
median groove, similar in appearance to the portion g. v. in
igi.

The mouth in fig. 10 has, at first sight, the appearance of
being closed. By a comparison, however, of fig. 10 with
fig. 16, it would seem that the apex of the epistome is really
represented (in the former) by the ectoderm closing the (per-
manent) mouth, and it is thus probable that the commence-
ment of the digestive tube in fig. 10 (v. ov.) is a part of the
oral division of the vestibule. This impression is strongly
confirmed by a section (not figured) similar to, but later than,

156 SIDNEY F. HARMER.

fig. 9. In the individual referred to, the stalk portion of the
vestibule is still present, but is small, and is connected with
the cesophagus very much as in the diagram fig. 16; i.e. at
some distance from the point where the apex of the epistome
ultimately meets the vestibular wall.

In somewhat later stages the permanent mouth is formed by
the perforation of the septum between the two portions of the
vestibule in fig. 10, and probably in the position of the aper-
ture a. v. v. in fig. 16.

In living individuals of the same age could usually be dis-
covered a small projection of the surface of the body in the
region marked ?s. in fig. 10. This represents the larval
“sucker,” which, as Barrois has correctly stated, disappears
during the metamorphosis. The region of the “ dorsal organ”
or brain of the larva is doubtless indicated by the marked angle
on the left side of the stalk of the individual just referred to.
None of the previous histological peculiarities of the organ re-
main at this stage, and it is in fact already almost impossible to
distinguish with certainty its position.

It appears to me that Barrois has suggested the real expla-
nation of the metamorphosis of Pedicellina, although he has
confined himself to one or two short statements, which are
given without any indication of the manner in which they are
to be interpreted. I quote below one or two passages from
Barrois’ note so many times referred to (3), the given quota-
tions reproducing, so far as I am aware, the whole of Barrois’
explanation of this complicated subject.

(i) ‘‘La premiére position” [corresponding, from the de-
scription, with my own fig. 10] “représente un état tout a
fait analogue au Loxosoma, avec anus en haut et csophage
en bas.”

(ii) “ L’inférieure”’ [portion du vestibule] “qui porte la
couronne, et dont les éléments viennent former la glande du
pied.”

(iii) ‘‘ Les deux organes énigmatiques de l’exoderme” [i. e.
sucker and dorsal organ] ... . “ne sont, suivant moi, que
des organes provisoires; tous deux sont rejetés sur la face

ON THE LIFE-HISTORY OF PEDICELLINA. 157

dorsale, ot ils finissent par disparaitre, peu 4 peu. Sans doute
il faut voir, dans les deux soies décrites par Salensky sur la
face dorsale du Loxosoma crassicauda, le reste de l’organe
des sens antérieur”’ [i. e. the dorsal organ] ‘‘ qui, d’aprés mes
recherches, vient occuper cette place.”

I have already (4) explained my reasons for the belief that
the dorsal organ at any rate, and perhaps the sucker, are im-
portant organs, which throw considerable light upon the mor
phology of the Polyzoa, so that I cannot accept Barrois’ con-
clusion that these structures have no particular significance.

It is obvious that, however accurate Barrois’ conclusions
(quoted above) may be, they need further explanation. The
similarity between larva and adult in the Entoprocta, even
in the position of the buds in Loxosoma, is so striking that
some means of comparing the two stages is necessary. I
therefore suggest the following explanation of the relation
between larva and adult.

It does not seem to me that Caldwell’s theory of the sur-
faces of the Polyzoa receives any support from the metamor-
phosis of Pedicellina. The short line between mouth and
anus remains unchanged throughout the metamorphosis, and
in order to prove that it is not ventral, it still remains neces-
sary to show that the dorsal organ of the larva is not a brain,
and that the larval surfaces do not correspond with those of a
Trochosphere.

Figs. 17—19 (Pl. XVII) are diagrams representing a pos-
sible explanation of the metamorphosis of the Entoprocta,
but although founded on the history of Pedicellina, Loxo-
soma is the form which is actually (hypothetically) repre-
sented.

Fig. 17 explains a possible conception of one of the earlier
stages in the acquirement of the sessile habit by the free-
swimming Polyzoon ancestors. The form is, however, to all
intents and purposes, a Loxosoma larva, with brain, sub-
cesophageal ganglion (not discovered in Pedicellina until a
stage later than fig. 10), and a pair of buds, one of which is
shown. I believe there are no authentic instances of the fixa-

158 SIDNEY F. HARMER.

tion of a Polyzoon larva by any other than its oral surface,
and it may therefore be assumed that this method of fixation
was acquired at a very early stage in the phylogeny of the
group. Let us suppose, however, that this “Archi-Loxo-
soma,” on fixing itself by the edge of its vestibule, left an
aperture (for the entrance of food), surrounded by the ciliated
ring (vide fig. 17), leading from the exterior into the otherwise
closed vestibule, and situated behind the anus.

Subsequent development may be imagined to give rise to a
form like fig. 18, in which the vestibular opening is an elon-
gated slit, extending along the whole of the region formerly
occupied by the posterior side, and still surrounded by the
ciliated ring. The mouth, in order to obtain its food as con-
veniently as possible, now faces the posterior side (of the
former stage), and this has entailed a rotation of the entire
alimentary canal, in the manner shown in fig. 18.

By the growth of the proximal end of the Polyzoon, the
mouth would be thrust away from the point of support, and
the animal might thereby obtain an advantage in procuring
food by meaus of its ciliary currents. But during this process,
the proximal portions of the ciliated ring would become far
less efficient for obtaining food than the distal portions, and
would tend to atrophy. The final result would be the acquire-
ment of a form like fig. 19, representing in a very slightly
diagrammatic form, an adult Loxosoma. The ciliated ring is
here represented as consisting of two disconnected portions, cor-
responding (1) to the ring of tentacles ; (2) to the foot-gland (cf.
the second of Barrois’ conclusions quoted on p. 156). The foot-
gland has remained practically as an open groove, a series of
ciliated tentacles having been developed round the margin of
the permanent vestibule.

The position of the buds in the larval Loxosoma appears at
first sight fatal to the above hypothesis. That this larva does
actually develop buds normally can hardly be doubted, since
I have shown not only that these structures are developed
twenty-four hours after hatching (which might, however, be an
abnormal circumstance, due to the want of proper conditions

ON THE LIFE-HISTORY OF PEDICELLINA. 159

for fixation), but also that ectodermic thickenings, the com-
mencements of the buds, are to be detected.some time before
the embryo is ready to leave the maternal vestibule, the possi-
bility of the development having been influenced by abnormal
conditions being here out of the question.

In figs. 17,18, and 19, the position of the dorsal organ is
represented as not having been much altered during the rota-
tion of the alimentary canal, which has, so to speak, been
pulled through the loop formed by the dorsal organ and the
somewhat hypothetical subeesophageal ganglion. Assuming for
the moment this position for the dorsal organ, we find that
throughout the metamorphosis the buds retain their original
situation (in Loxosoma) between the dorsal organ and the
ciliated ring, and that their position with regard to the cso-
phagus is practically the same as that which characterised them
at their first appearance.

Is there, however, any reason for believing that the position
of the dorsal organ is correctly indicated in the diagrams? It
seems to me that this question must be answered in the affirma-
tive. In the first place, the degenerating dorsal organ of
Pedicellina does in reality occupy this position, and in the
second place (vide No. 3 of Barrois’ conclusions on p. 156),
the circumcesophageal commissures may be represented by the
strong ganglionated nerves passing from the ganglion to the
“posterior sense-organs”’ in L. crassicauda, as originally
described by Salensky (see also No. 4, Pl. xix, fig.1). Should
the metamorphosis of Loxosoma be proved to bear out this
suggestion of Barrois’, we must assume either that the whole
brain has atrophied, or that the adult possesses at most a small
portion of the brain at the ends of the two widely separated
cesophageal commissures.

With regard to the actual metamorphosis of Pedicellina,
I have to point out that Ihave not succeeded in demonstrating
the presence either of cesophageal commissures or of a sub-
cesophageal ganglion. The latter structure becomes distinct
only at a stage later than fig. 10, and it then has the position
which characterises the adult ganglion.

160 SIDNEY F. HARMER.

No. 1 of Barrois’ conclusions quoted on p. 156, appears to
me perfectly just. It is impossible in fact not to be struck
with the great resemblance between the solitary Pedicellina
shown in fig. 10 and an adult Loxosoma, and this similarity
is quite conspicuous even at much later stages. The obliquity
of the lophophore in Loxosoma is hence, on the view already
explained, another of the archaic features of this genus, the
lophophore having still a marked inclination to the “ anterior”
side of the animal (fig. 19).

It is unfortunate that the metamorphosis of Loxosoma,
which possesses a foot-gland, should be unknown, but we are
able to make certain inferences from the phenomena of budding.
Both vestibule and foot-gland originate as longitudinal groove-
like invaginations of the ectoderm of the “ anterior” face of
the bud. Fig. 15 is a reproduction of a drawing from Oscar
Schmidt, in which the foot-gland is represented as originating
from the two proximal cells of the ectoderm of the ‘ anterior ”
side of the bud, and in which it is further seen that these cells
are not in the least marked off from those which are taking
part in the formation of the vestibule. The relations of lopho-
phore and foot-gland in this figure are indeed exactly those of
the ciliated ring in the diagram (fig. 18).

The Metamorphosis of Pedicellina viewed in its
relation to the above Hypothesis.

I have no reason to believe that the position of the ciliated
ring shown in fig 1] is in any way altered during the subsequent
metamorphosis. ‘This structure in all probability degenerates
in situ.

The ciliary apparatus of an ordinary Trochosphere is not,
however, constituted entirely by the preoral circlet. In the
neighbourhood of the latter there occurs in Polygordius,
e. g., (cf. Hatschek, No. 2) a series of smaller cilia forming a
postoral circlet, whilst a third part of the apparatus is con-
stituted by “a ciliated groove running between the two ciliated
rings, and prolonging itself into the ciliated mouth.” This

ON THE LIFE-HISTORY OF PEDIOCELLINA. 161

last portion is obviously represented in Pedicellina by the
ciliated oral grooves, continuous, as in Polygordius, with
the mouth. The relations of these grooves during the
metamorphosis appear to me to deserve further consider-
ation.

We have found that the median postanal portion of the
vestibule is continuous with the oral grooves, of which it may,
indeed, be said to form a part. Acccording to Hatschek
(1) it is, like other portions of the vestibule, lined by ciliated
cells.

If we are justified in assuming that the oral groove—a part
of the typical Trochospheral ciliary apparatus—extends, poten-
tially at least, from the mouth completely round the vestibule
to the postanal region, it seems to me that considerable light
is thrown on the metamorphosis. The morphological position
of the oral groove will be in no way altered during the rotation
of the alimentary canal, and in fig. 16 it will continue to pass
from the mouth round the ab-anal side of the altered lateral
folds to the median post-anal portion of the vestibule, even
though it is no longer distinguishable in the persisting division
of the latter structure. In figs. 16 and 6 we observe, however,
the commencement of a separation of the oral groove into two
parts—one continuous with, and becoming indistinguishable
from, the “oral” section of the vestibule (v. or. in fig. 16), and
the other potentially passing from the free apex of the epistome
in fig. 16 to the end of the reference line m.v. in the same
figure. The position of this latter portion will be the median
line passing from a.v.v. to m.v. Owing to the fact that it
is situated behind the anal cone it is, of course, unpaired (cf.
fig. 5), and it appears to me that its situation may be very
fairly considered to be represented by the linear groove which
in fig. 11 has formed the permanent vestibular aperture. From
the margins of this groove are developed the tentacles, which,
if the above reasoning is legitimate, are formed from the region
of the oral groove.

The fact that the tentacles of the adult lophophore of the
oral side are on the ab-anal side of the mouth appears to me

162 SIDNEY F. HARMER.

to prove that the lophophore itself is developed from a mor-
phologically przoral portion of the oral groove.

The relation between the velum proper and the oral cilia has
become, in the Entoprocta, considerably complicated by the
formation of a fold of integument (vestibular wall), carrying
the former to some distance from the latter. When the
Pedicellina larva attaches itself, the distance between the two
structures becomes increased. The velar portion maintains its
position at fixation, and soon atrophies ; the oral groove, on
the contrary, growing away from the degenerated velum. Even
during the phylogenetic history of the process we may suppose
that the velum atrophied at fixation. This is par excellence
a locomotive structure, and would be useless in an attached
condition. The oral cilia would, however, continue (in the
hypothetical stage of fig. 18) to convey food to the mouth, and
the cells bearing them would, after a time, become prolonged
into tentacles, by which their range of activity would be
extended. :

During the abbreviated metamorphosis of Pedicellina it
has hence resulted (if the above be true) that the velum takes
no part in the change of position involved in the passage to the
adult condition.

Summarizing the above, I may express my conviction (1)
that the metamorphosis of Pedicellina is a simple modifica-
tion of a more archaic process, due to abbreviation of develop-
ment, (2) that the oral groove persists in part as the adult
lophophore, (3) that the vestibule closes at fixation, and under-
goes the whole of its alterations in the interior of the larva,
opening secondarily only when the adult condition is practi-
cally attained.

The adult form is reached by the elongation of the stalk of
fig. 10, and by the replacement of its contained “ globules ” by
characteristic connective-tissue and muscle-cells ; by the for-
mation of a stolon and a diaphragm, and by various alterations
in the calyx. The more important of these consist in the
complete (or almost complete) loss of the obliquity of the
lophophore, in the development of the permanent ganglion

ON THE LIFE-HISTORY OF PEDIOCELLINA. 163

and generative organs (if these are formed in the primary
individual, as is probably the case) and in the complete forma-
tion of the vestibular aperture and tentacles. I have made no
special observations on most of the above points, although on
the important question of the origin of the colony from the
primary individual, I am able to throw some light.

In the first place, it may be stated that adult colonies are
by no means restricted to one growing point, as stated by
Hatschek (1). Of very common occurrence is the develop-
ment of two growing points, one at each end of the unbranched
stolon: I have noticed this even before the formation of a
single secondary calyx. A third growing point may be deve-
loped as a lateral branch of the main stolon; the amount of
branching is, however, always slight in P. echinata, and
apparently in all cases the cesophagus of each calyx is on the
side directed to the growing point to which this calyx properly
belongs, as already indicated by Hatschek.

The formation of the stolon is shown in fig. 13, a longi-
tudinal section of the stalk of a completely developed but
still solitary individual. The young stolon, which is cut
medianly, is developed on the cesophageal side of the Pedicel-
lina. The base of the stalk (which is alone represented) con-
sists of a thick cuticle, underneath which occurs a layer of
ectoderm, surrounding a gelatinous matrix in which lie con-
nective-tissue and muscle-cells. The section, however,—an
extremely good preparation—is contradictory to the theory of
Hatschek, according to which the apex of the stolon is pro-
vided with a hypoblastic vesicle derived from the dorsal organ,
and engaged in the formation of the mid-gut of the secondary
calyces. I may at once state that I have entirely failed to
convince myself of the occurrence of any such vesicle, at any
period, in the stolon, and I am forced to believe that Hatschek
has been mistaken in assuming its existence. Neither in
sections nor in entire specimens (whether living or treated
with reagents) could I discover the slightest evidence of the
presence of Hatschek’s vesicle, although I have investigated
both adult and young stolons in this connection.

164 SIDNEY F. HARMER.

It appears to me probable that the growing point of the
stolon of Pedicellina (vide fig. 13) consists solely of an ecto-
dermic layer secreting a cuticle and of a mass of indifferent
mesodermic connective-tissue cells, embedded in a structure-
less jelly. If this is the case, the only organ derived from the
hypoblast of the embryo would appear to be the mesenteron of
the primary individual, all other parts of the colony being
devoid of any derivatives of hypoblast cells.

This conclusion can hardly be avoided unless we assume that
some of the stellate cells of fig. 16 are really hypoblastic in
nature, although indistinguishable from the mesoderm cells in
their appearance. Owing to the nature of the process by
which the dorsal organ degenerates, it is impossible to assert
that some of its cells do not become ameeboid wandering cells
which migrate into the growing point. It can, however, be
safely stated that no hypoblastic vesicle is formed from the
degenerating dorsal organ. It may further be pointed out that
the conclusion arrived at on a previous occasion as to the
nervous (epiblastic) nature of the dorsal organ, in Pedicellina
as in Loxosoma, is in opposition to the view that this struc-
ture plays any part in the budding.

The well-known fact that calyces of Pedicellina may fall
from their stalks, which thereupon develop new calyces, appears
to me in direct contradiction to Hatschek’s view of the bud-
ding. The loss of the calyces is probably a normal, periodically
occurring process, which is perhaps to be regarded as a means
of rejuvenescence, and which is at least analogous to the forma-
tion of the “ brown bodies ” in the Ectoprocta. Itis exceed-
ingly easy to discover individuals in healthy colonies in which
the calyx has been lost, and a new “ bud” (easily recognised
by its small size and immature condition) is being developed
just below the scar. Specimens kept in captivity seem inva-
riably to lose their calyces if the quantity of water is not very
large, the calyx falling off at the “diaphragm.” This struc-
ture, which is merely a constriction at the base of the calyx,
filled by a row of flat cells, is perhaps a special arrangement
by which the calyx can break away from the stalk, without

ON THE LIFE-HISTORY OF PEDICELLINA. 165

injury to the latter. I have been unable to show that calyces
which have thus left their stalks are able to become the starting-
points of fresh colonies. The specimens under observation have
invariably died after a day or two, even if kept in a tide-pool.

Calyces formed at the scars produced in the manner above
indicated, seem to me (from superficial examination of entire
specimens) to develop in exactly the same manner as those
produced at the true growing point. The occurrence of this
phenomenon is undoubtedly adverse to Hatschek’s theory of
budding; the whole of the stomach falls away with the calyx,
whilst the existence of a plug of cells filling up the diaphragm
appears to preclude the possibility of the migration of any
cells derived from the stomach to the proximal side of the
diaphragm. Unless, indeed, it is assumed that some of the
*“ connective-tissue”’ cells of the stalks as well as of the stolon
are endodermic in nature, it must be concluded that none of
the cells of the bud are descendants of any of the cells belong-
ing to the embryonic hypoblast.

With regard to the further history of the budding (whether
at the growing point or at the apex of an old stalk) I have very
little to say. The free end of the stolon (or stalk) before long
develops an ectodermic invagination (fig. 14) destined to give
rise to the lophophore and, according to my view, to the whole
of the alimentary canal of the bud. The latter is from the
first continuous with the lophophoral rudiment, and in other
sections of the series to which fig. 14 belongs, the stomach and
vestibular cavity are separated from one another by means of a
septum. The latter does not, however, cut off the whole of
the deepest part of the invagination, but, since it is not deve-
loped ‘in the position of the esophagus the vestibule and stomach
remain continuous with one another (as in fig. 14). By the
formation of a diaphragm and by other processes already
described by Hatschek, the bud attains its adult condition.
The continuation of the stolon is formed by a lateral outgrowth
from that region in the young bud which afterwards becomes
the base of its stalk, precisely as in fig. 13 with the exception of
the fact that the new growing point is formed long before the

166 SIDNEY F. HARMER.

bud is itself mature. It is worthy of remark that the young
vestibular invagination does not occur accurately at the apex
of the stolon, but on the side of the apex turned towards the
growing point. In this respect it exactly agrees with the
position of the vestibular invagination formed near the apex
of a stalk which has lost its calyx, and again with that of the
incompletely rotated vestibule in intermediate stages of the
metamorphosis. It may indeed be said that the young
vestibule of all the buds is inclined towards the growing point,
and that in all cases it subsequently undergoes a rotation in
the same direction (but to a less marked degree) as that occur-
ing at the metamorphosis.

The history of the Pedicellina-larva appears to me to point
to the existence of a fixation-period in Loxosoma also. In
this case, the buds observed by me in the larva of L. Lepto-
clini would probably have to undergo a change of position,
during the metamorphosis, similar to that represented in figs.
17—19. I am inclined to believe that the degeneration of the
larval stomach observed in the same species, after a free life of
one or two days, was abnormal, and was due to the absence of
the conditions necessary for fixation.

On the Nature of the “Brown Bodies” of the
Ectoprocta.

The above statements with regard to the life-history of the
Entoprocta may, perhaps, give some indication of the manner
in which the “ brown bodies” of the Ectoprocta have origi-
nated. There can probably be no longer any doubt whatever
that these structures are degenerated polypides, which are
subsequently replaced by new ones budded off from the walls
of the zccecia.

In the metamorphosis of Pedicellina the purely larval
organs degenerate and form a mass of cells, which subsequently
become connective-tissue cells. The degeneration is here
slight, and has not yet acquired sufficient importance to give
rise to a characteristic “ brown body.”

ON THE LIFE-HISTORY OF PEDICELLINA. 167

Whilst in the adult Loxosoma nothing comparable to the
formation of “ brown bodies” is known, the adnlt Pedicel-
lina has developed a special arrangement—the constriction at
the base of the calyx—by which the latter may be lost without
material injury to the remainder of the colony.

In the adult Ectoprocta there seems to be the same
necessity for the rejuvenescence of some of the organs, but
here the occurrence of a thick ectocyst, usually intimately con-
nected with that of neighbouring individuals, in general pre-
vents the loss of any part of the body wall, asin Pedicellina.
In some of the stoloniferous Ctenostomata, however, the
entire zocecium is deciduous.

But even in Pedicellina one may almost speak of a
““7ocecium”’ in the same sense as in the Ectoprocta. It isa
well-known fact that septa occur at intervals across the stolon
of Pedicellina, and in most cases are developed in such a
manner that a piece of the stolon, connected with the base of
each stalk, is cut off from the remainder of the stolon by a pair
of symmetrically-placed septa. There are thus typically two
septa between the bases of each two stalks, and stalk-bearing
and stalkless sections of the stolon alternate regularly with one
another.

It is thus possible to consider stalk plus portion of stolon
connected with it, the representative of a zocecium. The distal
end of the zocecium is from time to time segmented off, carry-
ing with it the whole of the alimentary apparatus, whilst a new
polypide is developed within the remaining portion by a process
of budding. By the formation of a new constriction the distal
part of the zocecium—the calyx—becomes again differentiated
from the proximal part—the stalk.

In the Ectoprocta the occurrence of the same process is
usually obviously impossible, and the poly pide alone degenerates,
forming a “ brown body” which subsequently passes into the
new stomach, and is ejected by the anus. The occurrence of
this circumstance is already foreshadowed in two particulars in
Pedicellina. We find, in the first place, that a new polypide
is actually budded off by the ectoderm of the zocecium at or

168 SIDNEY F. HARMER.

before the loss of the calyx; and, in the second place, that the
tissues have already acquired, at the metamorphosis, the power
of disposing of degenerated structures.

In the Ectoprocta one may hence suppose that, owing
to the inconvenience of losing a portion of the zoccium at each
rejuvenescence, the new polypide is budded off near the pre-
ceding one, instead of from an entirely different part of the
zocecium, as in Pedicellina (below the diaphragm). The
degenerating alimentary canal and other structures are then
worked up by the “ Parenchymgewebe” (Vigelius), which has
inherited this kind of power from the larval tissues, into the
condition of a “brown body,” which passes into the new
stomach, and reaches the exterior by means of the anus.

In the development of the Ectoprocta an archenteron is
formed, in a large number of cases at least. The embryo is,
however, richly supplied with yolk; it develops within the in-
terior of the parent, and its alimentary canal is hence, in many
cases, functionless,

At its metamorphosis this larva possesses no functional ali-
mentary canal, and must hence form a new one. But since in
its previous phylogenetic history our Polyzoon has acquired
the power of developing new “ polypides” from various parts
of its ectoderm, a fresh gut could without difficulty be formed
within the body wall of the metamorphosed larva; since the
latter is now in the same condition as an adult zocecium whose
polypide has just become a “ brown body.”

This, indeed, is what actually happens. The larva passes at
once into the condition of a zoccium containing a “brown
body,” the remains of its larval organs. The complicated me-
tamorphosis of Pedicellina has been given up, the larval
structures now degenerating by the method employed during
the atrophy of the polypides in adult individuals, and finally
leaving the zocecium by passing as the first ‘‘ brown body”
into the alimentary tract of the primary polypide, and thence
to the exterior.

The metamorphosing Ectoproctan larva is probably in
the same condition (irrespective of the difference pointed out

ON THE LIFE-HISTORY OF PEDICELLINA. 169

in the methods by which the alimentary canal is lost in the
two cases) as the primary individual of a Pedicellina colony
would be immediately after the loss of its calyx, supposing that
it had not meanwhile developed a stolon and secondary
calyces.

Unless I am mistaken in my views with regard to the meta-
morphosis of Pedicellina, it appears to me necessary to con-
clude that in the Entoprocta the ventral line of the body
extends from a. v.? in figs. 10 and 19, down the right sides of
the figures, as far as a. v.1 The median dorsal line will in con-
sequence be represented by the entire left sides from a. v.' to
a.v.2 These surfaces are most clearly expressed in the young
Loxosoma bud, in which the whole of the surface turned
away from the parent (characterised by the possession of ves-
tibule and foot-gland) is ventral, whilst the opposite surface of
the bud is, conversely, dorsal.

I hope to be able before long to publish some account of the
development and metamorphosis of the Ectoprocta. ‘Till that
time I prefer to withhold any further expression of opinion
with regard to the surfaces and relations of the larve of this
group of the Polyzoa.

List or Parers REFERRED TO.

1. B. Hatscnex.— Embryonalentwicklung und Knospung der Pedicellina
echinata,” ‘ Zeits. f. wiss. Zool.,’ Bd. xxix, 1877, S. 502.

2. B. Hatscuex.—* Studien zur Entwicklungsgeschichte der Anneliden,”’
© Arb. a. d. Zool. Inst. zu Wien,’ Bd. i, 1878, 8. 277.

3. J. Barrots.—‘ Métamorphose de la Pédicelline.” ‘Comptes rendus de
P Acad. des Sci.,’ T. xcii, 1881, p. 1527.

4. S. F. Harmer.—* On the Structure and Development of Loxosoma,”
‘Quart. Journ. Mic. Sci.,’ vol. xxv, 1885, p. 261.

13

170 SIDNEY F. HARMER.

EXPLANATION OF PLATES XVI & XVII,

Illustrating Mr. S. F, Harmer’s Paper on “ The Life-history
of Pedicellina.”

Reference Letters.

an. Anus. an.c. Analcone. a. v.'and a.v.? Hypothetical morphologically
anterior and posterior ends, respectively, of the vestibular aperture. a. v. v.
Aperture between oral and anal divisions of vestibule (in position of permanent
mouth). &. Bud. ér. Brain (= “dorsal organ”). c.c. Fragments of
ciliated cells. cc. p. Ciliated pit of brain. c. +r, Ciliated ring. d.s. Dorsal
sense-organ (of Loxosoma). pst. Epistome. f dr. Fibrous part of brain.
J. g. Foot-gland. ga. Ganglion of adult. gy. p. Growing point of stolon.
g. v. Median groove of permanent vestibule, ultimately becoming the vesti-
bular aperture (in position of part of oral groove of larva?). zt. Intestine.
1. f. Lateral fold of vestibular wall. 7. v. Lateral portions of anal division of
vestibule. m. Mouth. mes. Mesoderm. m.v. Median postanal portion of
the anal division of the vestibule. @. Csophagus. 0.9. Oral groove.
rec. Rectum. s. Sucker. st. Stomach. 7. Tentacle. v. Vestibule. ». a.
Its aperture. v. az. “ Anal” division of vestibule. v. or. “Oral” division.
v. v. Ventral division. 2. Large-celled tissue at base of epistome and anal
cone.

PLATE XVI.
Pedicellina echinata.

Fre. 1.—Median longitudinal section of a larva quite recently fixed (on
Coralline).

Fic. 2.—Obliquely longitudinal section (in the plane C D in figs 3 and 4)
of a similar larva.

Fic. 3.—Horizontal section of a slightly older larva, passing through brain
(= dorsal organ), cesophagus, epistome, and anal cone.

Fie. 4.—Obliquely transverse section (in the plane A B in fig. 1), at a stage
very soon after fixation.
ee

1 In describing one section as passing in a plane indicated in the figure of
another, it is to be understood that the details in the two individuals do not
always exactly correspond. This is due, partly to a difference in age between
the two larve figured, and partly to variations in the position of the internal
structures, owing to varying conditions of muscular contraction.

ON THE LIFF-HISTORY OF PEDICELLINA. 171

Fic. 5.—Horizontal section, at an early stage in the metamorphosis, passing
through the tip of the epistome, the lateral folds and oral grooves, and the
apex of the anal cone.

Fires. 6 and 7.—Two sections of a considerably older individual, passing
respectively in the planes K L and I J in Fig. 16.

Fies. 8 and 9.—Two sections of an individual of the age of Fig. 16, passing
in an obliquely longitudinal direction. Fig. 8 cuts the mouth and one of the
lateral portions of the permanent vestibule, Fig. 9 passing through the rectum
and the degenerating vestibule of the stalk. In another section of the same
series the two parts of the vestibule are continuous, exactly as in the diagram,
Fig. 16.

Fic. 10.—Median longitudinal section of an advanced, but still solitary,
individual.

Fie. 11.—Horizontal section (in the plane GH in Fig. 10) through a
similar specimen.

Fie. 12.—Section of an individual of the age of Figs. 8 and 9, passing in
the plane Ei F in the latter figure.

Fie. 13.—Median longitudinal section through the stalk of a solitary
individual with commencing primary stolon. The arrow indicates the position
of the oral side of the calyx.

Fig. 14.—Obliquely transverse section of a young bud, developed at the
growing point.

PLATE XVII.

Fic. 15.—Young bud of Loxosoma, from the ventral side. Copied from
O. Schmidt, ‘ Arch. f. mik. Anat.,’ Bd. xii, 1876, Pl. III, fig. 17.

Fic. 16.—Diagrammatic longitudinal section of a metamorphosing Pedi-
cellina at the stage of Figs. 8, 9, &c.

Fics. 17—19.—Diagrams illustrating the supposed morphological nature of
the metamorphosis of the Entoprocta. A full explanation is given in the
text.

ic SGA J
- 5 = 7 par
7 a= ie yf aM -
‘ -~>? 7 = , = es
- = ‘ ame c@
a es hee “eo
af — a
! ws lS he i
_ 7 at eee : P t
ce fs Boa ae “we hres lies e 7 : Y, | s 3
> ro : } - — A ue
- rs Wis ; ee ‘ Ns ere a ai ik
ae Yang ag Rak At Ee See ac
Te at bd ite agen) «TSP fond
a ay, ‘vail 4 pa > hat . 4
3 a fs wy i” @ ‘ i + ae Dw ig
ee Le gman Tal Vi ae <3 Sel.) Negeeal
a Spy eae ae eae: ial ME pull ba tsb Bai Sere eae
= ND ATL f. pe tes r ty iy i ue | ni ee
aa, a te CG: yo - eh ' 5
mad ; = i Fi Poe - +,
ee ee EWES Lah sob Al 0 Ul
. ~ ie a i - : ~ ‘
f ie? Vs & ‘ Ey ;
ae ea > oe ; a S J a
Ae ioe hee j j pee it iar jk ul : be * Pre
Se oy dahon (AR Sk aly bb aa ee
yen | alo aia iS. 7 Soul
‘ wa f "cae 4 ‘ _ J a“ 2
Ee a ee ee Th gt
p> sity hy i bs i. ye oa
Y \e- i : 7 hoa r »
3 - erie Fe ite Fat ih (=P pets
; oe a DAR s, oT Per pe alten Ce : oem aS
a A. 4, Ce as ‘ j te a a os
Boe ee Se le a ea

* y
: oa I ‘ae - ‘ : A a on a
Le.” my Baby Use eb ae Wun ean eS a ee > Rayos

> i -
ie

2
* 9 pf :
j Pn, tig e
aa
ow! . =
>
’ var nt
Sa | = : a
— 7
’ +
- =d L =
5 oe
4 1
D - ‘ ® :
pile '
# - hs | f i

§.F. Harmer del

Studies M.L. Vol. Ill, Pl. AVI.

F Huth, Lith® Edin?

Studies M.L.VolUIIL, PL XVI.

S.F.Harmer del. F Huth, lath? Edint

On Some Points in the Development of
Petromyzon fluviatilis.

By
Arthur E. Shipley, B.A.,
Christ’s College, Cambridge, Demonstrator of Comparative Anatomy in the

University.

With Plates XVIII, XIX, XX, and XXI.

Tur development of the Lamprey has occupied the attention
of many embryologists during the last fifty years. Of these we
owe the most complete accounts of the changes through which
the egg passes to Max Schultze, Owsjannikow, Calberla,
Scott, Balfour, and Dohrn. I have recently worked through
the development of Petromyzon again, and worked out the
origin of several organs which have hitherto been incompletely
known. In many of the most important points my researches
confirm those of the earlier observers, and to these I have only
referred at such length as would make the account intelli-
gible; in others, such as the persistence of the blastopore, the
origin of the ventral mesoblast, &c., I differ from previous
descriptions ; and some points, such as the development of the
heart, of the parts of the brain and cranial nerves, are worked
out for the first time.

The material for this article was obtained by artificially

1 The differences between Petromyzon planeri and fluviatilis are so
slight, and the intermediate forms so common, that I am disposed to follow
Anton Schneider, and to consider them as varieties of the same species. This

species may conveniently retain the name fluviatilis, as opposed to the
larger form Petromyzon marinus.

174 ARTHUR E. SHIPLEY.

fertilising the eggs of the ripe female Lampern, hatching the
larve out, and rearing them in confinement. The breeding
time is during the latter half of April and the beginning of
May.

The generative products of both male and female were
squeezed into glass vessels containing fresh water, and the
contents slightly stirred. The eggs at once adhered to the
bottom and sides of the vessel, and were left undisturbed for
three or four hours. The water was then poured off and a
fresh supply added. This was kept thoroughly aerated by
means of Semper’s aerating apparatus. The number of eggs
fertilised were about 70 per cent. of the total, though some
hatches were much more successful than others. The rate of
segmentation and development also varied greatly, being
influenced by the temperature and manner of aeration. The
unfertilised eggs very soon ‘could be distinguished from the
fertilized ; they developed great cavities or craters and were
soon attacked by fungi. The fungus, however, rarely affected
the developing eggs.

The spermatozoa have elongated heads, pointed at their free
end, but thicker at the end from which the tail arises (fig. 1),
Their length is from 35 to 40 micro. mm., of which the head
forms 8 micro.mm. They move actively about in the water
until they come into contact with an ovum. They enter the
egg through a micropyle, and Calberla states that the head
only enters the protoplasm of the ovum, the tail remaining
fixed in the micropyle, thus hindering the entrance of other
spermatozoa.

The eggs are almost spherical, with a diameter of about a
millimetre. On contact with water the outer cell-membrane
swells up and forms a gelatinous coating, by means of which
the eggs adhere to the bottom and sides of the vessel. This
gelatinous envelope is of considerable thickness ; it ultimately
disappears shortly before the embryo is hatched. Sections
through unfertilised eggs show the protoplasm crowded with
oval yolk granules, which stain deeply. These yolk granules
vary in size, and this is very evident in the segmenting eggs,

DEVELOPMENT OF PETROMYZON FLUVIATILIS. 175

where the yolk granules in the more quickly dividing upper
pole are much smaller than those in the more inert lower pole.
An attempt has been made to show that those parts of the
_ unsegmented egg containing the smaller granules is destined
to form the epiblastic parts of the embryo (16).! This view
seems to me to need confirmation. The small size of the yolk
granules in the epiblast might be due to the more rapid
division of these cells, causing a more rapid consumption of the
food-yolk.

The unusually deep staining which the yolk granules assume
very materially increases the difficulty of observation. Espe-
cially in the earlier stages of development the cell limits and
nuclei were rendered obscure by the masses of deeply stained
yolk granules.

As previous observers have stated, there are two polar bodies
extruded one after the other. After fertilization the egg con-
tracts, leaving a cavity between it and the egg membrane.

The first furrow appears about the fourth hour; it appears
first in the upper pole and spreads round the egg on each side.
Calberla states that the micropyle becomes at first oval, then
slit like, and finally passes over into the primary furrow. I
have not been able to observe this process in my eggs. He
further states that the first furrow divides the egg into two
unequal parts, a large epiblastic and a small hypoblastic ; the
smaller of these divides subsequently more rapidly than the
latter. Thus, according to him, the first furrow would cor-
respond with the first equatorial one in the Frog’s ovum.
Scott, although he had no fresh material to work with, was
able to correct this, and, as the latter suggests, Calberla was
probably misled by cases of abnormal segmentation. Many
of the eggs which apparently had not been fertilized divided
by one, two, and sometimes three furrows, and when this
took place the furrows were nearly always abnormal in
position.

The second furrow is vertical and at right angles to the
first, and also appears first in the upper pole. The third is

1 The figures in brackets refer to the list of papers at the end.

176 ARTHUR BE. SHIPLEY.

equatorial, but nearer the upper than the lower pole. After
its appearance the epiblastic half is separated from the hypo-
blastic or yolk-bearing half (fig. 2).

The external phenomena of segmentation have been accu-
rately described by Max Schultze, with the exception of
the next stage. After the first equatorial furrow he describes
two more in the same plane, but in my eggs the equatorial
furrow was followed by two vertical lines, which appear at
first in the upper pole exactly as they do in the Frog’s ovum
(fig. 3). These are followed by two more equatorial furrows
which divide the egg into thirty-two segments. After this the
segments of the epiblastic pole divide more rapidly than those
of the lower.

Fig. 5 represents a transverse section through an egg thirty-
six hours after fertilisation. In this stage it is a blasto-
sphere, with a segmentation cavity enclosed by a single layer
of cells except along the line where the epiblastic and hypo-
blastic cells joi. Here the layer is two cells thick. The
nuclei of the large cells appear small, but it must be recollected
that the amount of protoplasm is very small compared to the
yolk. The latter has been omitted for the sake of clearness.
Fig. 6 is taken from an egg twelve hours later. Here both
the roof and floor of the segmentation cavity are many cells
thick. A similar stage is found in the Frog’s ovum, but there
is this difference between the two. In the Frog’s egg the
whole of the roof of the segmentation cavity forms epiblast ;
in the Lamprey it is only the outermost layer. The following
stages are accompanied by a thinning out of the roof of the
segmentation cavity, and are represented in figs. 7 and 8.
On this point my observations tend to confirm those of
Calberla, and are opposed to those of Schultze, who found
a many-layered roof to the segmentation cavity just before
invagination. The thinning out appears to be brought about
by the inner cells of the roof passing round to the sides and
floor of the segmentation cavity. Just before the invagination
which forms the gastrula the roof of the segmentation cavity
consists of a single layer of cells; the segmentation cavity is

DEVELOPMENT OF PETROMYZON FLUVIATILIS. 177

large and occupies the whole of the upper hemisphere, whilst
the lower hemisphere is solid and consists of larger cells, which
we may speak of as yolk-cells. The most external layer of
these consists of rather columnar cells. These latter cells
soon become smaller than the inner yolk-cells, and about the
time of invagination the whole egg is enclosed by a layer of
small columnar cells, the epiblast. This is brought about by
the conversion of the outermost row of yolk-cells into small
columnar cells. As Balfour has shown, this takes place
latest in the region of the blastopore.

The invagination which forms the mesenteron commences
about 1380 hours after fertilisation ; it commences at one side
of the equator of the egg, in the region where the single layer
of epiblast cells passes into the yolk-cells (fig. 9). The invagi-
nation at first has a wide-arched slit-like opening, but this
soon narrows into a small circular pore (fig. 4). The segmen-
tation cavity is gradually obliterated by the invaginated cells.
These from the first enclose a cavity, the mesenteron. In this
respect the formation of the gastrula is like that of Amphioxus,
and differs from that of the Amphibia, where the mesenteron
appears later as a splitting underneath the invaginated cells.
The presence of a large amount of food-yolk causes the invagi-
nated cells to be pushed dorsalwards. The mesenteron
extends as a tubular cavity about two thirds round the
embryo. Its dorsal wall is composed of columnar cells resem-
bling those of the general epiblast ; the cells forming the floor
have the same characters as the yolk-cells (fig. 12). The
dorsal side of the mesenteron lies in immediate contact with
the under surface of the epiblast throughout its entire length.
In this respect again the Lamprey differs from the Frog, where
the invaginated hypoblast cuts off a mass of cells on its dorsal
side, which subsequently forms the mesoblast. °

The mesoblast now appears by the differentiation of two
bands of these yolk-cells, which lie in the angles formed by
the mesenteron and the epiblast (fig. 12). This differentiation
commences in front and is continued backward. The two
bands of mesoblast are separated dorsally by the juxtaposition

178 ARTHUR E. SHIPLEY.

of the dorsal wall of the mesenteron and the epiblast, and
ventrally by the hypoblastic yolk-cells which are in contact
with the epiblast over two thirds of the embryo. Subse-
quently, but at a much later date, the mesoblast is completed
ventrally by the downgrowth on each side of these mesoblastic
plates. This takes place at a comparatively early stage in the
head and that part of the trunk lying in front of the liver.
In the posterior part, which remains swollen with yolk, the
ventral completion of the mesoblast is delayed.

The first formation of the mesoblastic plates appears to
take place by a differentiation of the hypoblastic yolk-cells in
situ, and not from invaginated cells (figs. 12 and 13). The
subsequent downward growth is brought about by the cells
proliferating along the free ventral edge of the mesoblast,
these cells then growing ventralwards, pushing their way
between the yolk-cells and epiblast (fig. 11).

This account of the origin of the mesoblast differs from that
given by Scott. He describes the mesoblast as arising from
two sources—(1l) cells which are derived from the invagina-
tion of the blastoderm, (2) the outermost layer of the hypo-
blastic yolk-cells, which, according to Scott, split off from
the remainder, and form a ventral sheet which completes the
mesoblast in that side of the body. The mesoblast in the
head is derived only from the first source, as by the time it is
completed ventrally the head is raised above the yolk-con-
taining parts.

Shortly before the development of the head fold raises the
head from the yolk-bearing part of the embryo, the neural
plate becomes evident in the exterior. It extends as a low
ridge from the anterior lip of the blastopore to just in front of
the blind anterior end of the mesenteron, over two thirds of
the circumference of the embryo.

The blastopore is always visible at the posterior end of the
neural plate. Schultze has given a very complete set of figures
of the exterior of the embryo. As his figures show, with the
elongation of the embryo the anterior end curves round and
overlaps the posterior, thus obscuring the blastopore. Fig. 10

DEVELOPMENT OF PETROMYZON FLUVIATILIS. 179

is a section taken through the blastopore and the head soon
after the head is raised above the general level of the egg.

From his observations of the embryo as a whole, Schultze
came to the conclusion that the blastopore persisted and gave
rise to the anus, and he was supported in this view by Calberla.
Later observers, however, who have studied the development
of the Lamprey by means of sections, have maintained
Benecke’s view that the blastopore disappears. Scott
describes the neural canal enclosing the blastopore and figures
the neurenteric canal thus formed. He describes the formation
of the anus, from a protuberance of the alimentary canal
which approaches the epidermis and breaks through about the
twentieth day. Balfour also states that the blastopore closes
and does not form the permanent anus.

My observations of the embryo as an opaque object lead me
to the belief that the blastopore remained open. In this I
have been confirmed by sections taken through a series of
embryos preserved at intervals of a few hours. Primarily the
blastopore lies at the posterior dorsal end of the embryo
(fig. 4), but by the growth of the dorsal surface and the forma-
tion of the tail it comes to occupy a position in the ventral
surface. What was the anterior lip in the first position comes
to be the posterior in the latter.

Fig. 4 is a view of the embryo twelve days old, as an opaque
object, showing the blastopore at the posterior end of the
neural ridge. Fig. 16 is an oblique section through an embryo
about two days older, showing the nervous cord just separated
from the skin, and the notochord both continuing behind the
blastopore.

Scott was of opinion that the lumen of the invaginated
mesenteron persisted only in the fore-gut. Soon after the in-
vagination is completed this part of the alimentary canal lying
in the head and neck becomes raised from the rest of the
embryo. It is thus separated off from the yolk-cells, and the
hypoblastic cells in this region soon assume a definite columnar
appearance, though they continue to contain yolk granules for
some days. ‘This region extends to where the liver appears

180 ARTHUR E. SHIPLEY.

in older embryos. A similar change in the cells lining the
mesenteron takes place at its posterior end. The cells lining
the blastopore and extending for some distance into the ali-
mentary canal assume very early a columnar appearance and
appear perfectly continuous with the columnar epiblast (figs.
10, 14, and 16.) The cells lining the hind-gut retain the
character of the yolk-cells for a long time, but the lumen of
the mesenteron in this region never disappears, as Scott and
Calberla thought. The lumen of the alimentary canal, with
the exception of the mouth, is derived directly from the inva-
gination which forms the gastrula, and no part of it is ever
obliterated in the course of development.

A similar persistence of the blastopore to form the anus appears
to be common in the Amphibia. It has been shown to occur
in the Newt by Miss Johnson, in the Frog by Spencer,
and in Alytes by Gasser. Its occurrence in the Cyclos-
tomata seems to point to the fact that it is a primitive
feature retained in those eggs whose development is not greatly
modified by the presence of a large mass of yolk. Renewed
observations in the development of Amphioxus would pro-
bably throw some light on this point.

The Central Nervous System.

The early development of the central nervous system has
been so fully described by Calberla, Balfour, and Scott,
that little is left to be added to their account. But the origin of
the neural canal, the relationship of the posterior end of the neural
cord to the blastopore, and the later development of the parts
of the brain and the cranial nerves present points of interest.

Calberla was the first to show that the central nervous
system of the Lamprey arises by a delamination and not by
an involution of the epiblast. He described a similar origin for
the nervous system of the Teleostei, and Balfour and
Parker found the same to be the case in Lepidosteus.

The first trace of theneural plate appears about the eighth day
after fertilization, just after the invagination is completed. A

DEVELOPMENT OF PETROMYZON FLUVIATILIS. 181

shallow groove is seen running forward from the blastopore,
round about two thirds of the embryo and passing a little in
front of the blind end of the mesenteron. The groove isa very
temporary structure and is soon replaced by a ridge. This
arises by the epiblastic cells lining the groove, which are of a
columnar shape, budding off cells from their under surface.
The result of this is that a keel of cells is formed which forms
the neural ridge externally (fig. 12), and internally presses in
between the mesoblastic plates. The keel arises solely by the
epiblast cells budding off cells in their under surface only. It
is much deeper in the anterior third of its course, which region
ultimately forms the brain.

The keel in the course of two or three days loses its connec-
tion with the epidermis ; this occurs at first anteriorly and ex-
tends backward, and as Scott has pointed out, it does this of
itself and not by an ingrowth of the mesoderm in each side as
Calberla described.

Figs. 13, 15, and 16 show the solid neural cord lying above
the notochord, which by this time is separated off from the
hypoblast. It is important to notice that the neural canal
does not arise until after the connection between the neural
cord and epidermis is severed. It is about the origin of this
neural canal that my observations and those of Calberla and
Scott are at variance. They described the epidermic layer
of epiblast passing down into the nervous, in such a way that
the canal, when it does appear, is lined by this layer. I have
not been able to see any trace of this. The cells forming the
nervous system appear to me to be all split off from the under
surface of the epidermis in the dorsal middle line, and the
continuity of the epidermis in this region never seems to be
broken by any such invagination as they suggest. Balfour
was also doubtful on this point; but in his and Parker’s
work on the development of Lepidosteus, they state that
there is no evidence of the epidermic layer being concerned in
the formation of the canal.

The canal seems to arise as a split between the cells in the
axis of the solid cord, and not by the absorption of the central

182 ARTHUR E. SHIPLEY.

cells, as has been suggested in the case of the Teleostei. It
appears at first anteriorly and extends backward, and for some
little time the walls of the lumen are by no means sharply
defined. Processes from the cells lining the canal project into
its cavity and suggest the idea that they have been torn out
from between the cells of the other side.

The neural cord remains solid at its posterior end for some
time, and here it becomes fused with the surrounding struc-
tures in a somewhat remarkable way. It does not fuse round
the blastopore as Scott describes, indeed it is not easy, con-
sidering its mode of origin, to see how it could; and there is
no hollow neurenteric canal. Figs. 14 and 15 represent two
sections taken through a larva just after hatching. Fig. 14 is
through the region of the blastopore. It shows the neural
cord with its canal already formed ; beneath this lies the noto-
chord, and beneath this again asolid rod of cells which is con-
tinuous with the subnotochordal rod and the dorsal hypoblast.
This latter structure is the solid postanal gut. The mesoblastic
plates are seen separating off from the hypoblast yolk-cells
which occupy the remaining space with the epidermis. Dor-
sally this is produced to form the dorsal fin. Fig. 15 repre-
sents a section through the tail a little posterior to the blasto-
pore. Here the neural cord, notochord, and postanal gut have
fused into a rod-like mass of tissue which is ventrally con-
tinuous with the hypoblast cells; a few sections posterior to
this none of the three embryonic layers are distinguishable
except the epidermal portion of the epiblast. A longitudinal
median section through the tail is represented in fig. 20. This
shows the mass of indifferent tissue which lies in the tail and
which by internal differentiation gives rise, as the tail grows,
to mesoblastic somites, neural cord and postanal gut. This
mass of tissue, which in many respects reminds one of the
growing point in a plant, may be called the primitive streak.
It is perhaps worth while to point out that it lies at what was
originally the anterior lip of the blastopore.

A similar mass of tissue formed by the fusion of the pri-
mary layers has been described by Balfour and Parker in

DEVELOPMENT OF PETROMYZON FLUVIATILIS. 183

Lepidosteus, Spencer in the Frog, and Miss Johnson
in the Newt.

The further development of the central nervous system will
be described later after some of the details connected with the
mesoblast and hypoblast have been considered.

The Mesoblast.

The origin of this layer from the yolk-cells situated in the
angle between the epiblast and the invaginated endoderm has
been described above. For some little time the mesoblast re-
mains in the condition of two triangular masses of cells,
separated from one another dorsally by the notochord and ner-
vous system, ventrally by the yolk-cells which lie in contact with
the ventral epiblast. In the anterior end of the embryo the
mesoblast soon unites ventrally by lateral downgrowths; in
the trunk, however, which remains crowded with yolk-cells for
a week or ten days after hatching, this takes place much
later.

Scott has described the formation of the muscle-plates very
accurately, and it will therefore be unnecessary to give more
than a short résumé in order to make the following account
intelligible. About the twelfth or thirteenth day the meso-
blastic somites appear by the segmentation of the dorsal part
of the lateral mesoblastic plates. These appear at first ante-
riorly, and the segmentation extends backwards. The most
anterior one lies close behind the auditory sac. The ventral
unsegmented mesoblast has split into the splanchopleure and
the somatopleure on each side, and in the region just behind the
posterior gill-cleft these have met ventrally, forming a ventral
mesentery, connecting the alimentary canal with the ventral
body wall.

The mesoblast somites are shown in fig. 17, which repre-
sents a horizontal section through an embryo fourteen days
old. They are cubical masses of cells enclosing a small cavity,
often entirely obliterated, which represents part of the body
cavity. The cells surrounding this are at first uniform in size,
and each side is only one cell thick. Like the other cells of

184 ARTHUR EH. SHIPLEY.

the embryo they contain yolk granules, which are gradually
absorbed. In the tail region these mesoblastic somites con-
tinue to be segmented off from the primitive streak till five
or six days after the larva is hatched.

In transverse sections the mesoblastic somites appear trian-
gular, having a median side against the nervous system and
notochord, an external one against the epididymis and a ventral
one. Besides these there are the anterior and posterior sides.
The cells composing all these, except those of the external
layer, develope into longitudinal muscles. Whilst this is taking
place the dorsal surface of the embryo has become raised above
the general level, so that the embryo in section is no longer
round but pear-shaped.

As Stannius, Grenacher, and Langerhans have
shown, the muscles of the Lamprey fall into two groups, which
differ in structure as well as in their disposition. The first of
these form the myomeres, and are derived directly from the
mesoblastic somites ; the second comprise the muscles of the
eye, those belonging to the respiratory system, and those con-
nected with the upper and lower lip and mouth generally.
These seem to arise exclusively from the ventral unsegmented
parts of the mesoblast, and perhaps, in some cases, from wan-
dering mesoblast cells. The muscles of the heart resemble the
latter in many points.

Each myomere in the Lamprey or Ammoceete consists of a
number of plates of muscle-substance, lying one on the top of
another. Each plate is flat, and more less square in outline.
It is bounded anteriorly and posteriorly by the myotomes ex-
ternally by a connective-tissue layer closely connected with the
skin, and internally by a similar layer. Above and below, or
dorsally and ventrally, it is in contact with a similar muscle-
plate. In some myomeres which have become modified, such
as the anterior one which extends far forward over the ear, the
shape of the muscle-plate has lost its square outline and be-
come oblong, but in one of the myomeres of the trunk they
are almost square in longitudinal section.

From the above description it will be seen that each muscle-

DEVELOPMENT OF PETROMYZON FLUVIATILIS. 185

plate or “ Kastchen” of Stannius occupies the horizontal
space between two myotomes, and that they lie one on another,
so that in a horizontal section we see only one, in a transverse
or vertical section we see one lying on another like sheets of
paper. Each “ muscle-plate” contains several nuclei, which
stain more deeply than the muscle-substance. It is trans-
versely striated, and faint longitudinal striz can also be de-
tected; these correspond with fibrillz, into which the muscle-
substance easily breaks up. These latter are especially large,
and can be easily recognised in transverse sections near the
most external part of the ‘ muscle-plate.”’

The development of these muscle-plates is as follows :—
The outermost layer of cells forming the mesoblastic somite
does not appear to be converted into muscles. For along time
it persists as a definite layer of cubical cells with large nuclei
lying between the skin and the myomere; this is the case till
long after the other cells of the mesoblastic somite have deve-
loped into muscles. Finally, this layer seems to disappear,
but remains of it can still be distinguished lying just within
the skin, even when the myomere has assumed the appearance
characteristic of the full-grown Ammocete. This view that
the somatic layer does not take part in the formation of the
myomeres, is not in agreement with what Balfour has de-
scribed in the Elasmobranchs, where both the inner and
outer layer become muscular; but, on the other hand, the
muscles of the myomeres in Amphioxus appear to be de-
rived from the splanchnic layer only, and the same view is
supported by Gotte and the Hertwigs.

The remaining cells of the mesoblastic somite begin to
grow in between one another, and between each neighbouring
somite an intermuscular septum is deposited. The process of
growing in between one another is carried on until each cell
occupies the whole length from one myotome to the next, and
at the same time, each cell becomes somewhat flattened, so
that their transverse section, which was at first round, become
oval (fig. 24). At the same time longitudinal thickenings
occur in the cortical part of the cell, the medullary portion

14

186 ARTHUR E. SHIPLEY.

remaining clear and staining very slightly. The nucleus lies
in this medullary portion. The longitudinal thickenings occur
at intervals, so that in transverse section the cortex of the cell
appears beaded ; these fine fibrille stain fairly well so that
they can easily be distinguished from the medulla. The flat-
tening of the cell goes on until the cell occupies the whole
space between two myotomes, not only longitudinally but also
transversely (fig. 25). The original nucleus of each cell
divides into two or three, so that in each of these plates of
muscle-substance two or three nuclei can be seen and an occa-
sional yolk granule, which is, however, soon absorbed. In
addition to the longitudinal striation caused by the thread-like
thickenings in the cortex, a transverse striation appears. Each
plate of muscle-substance remains in this condition, with a
clear unstained medulla containing two or three deeply stained,
large, flat, oval nuclei (fig. 18), with a well-marked nucleolus ;
enclosed by a cortex, for about two weeks after hatching. The
cortex consists chiefly of its dorsal and ventral walls, and each
of these is thickened at regular intervals by the above-men-
tioned fibrille. Each fibrilla runs the whole length of the
myomere and is inserted into the intermuscular septa behind
and in front. About a fortnight after the young Ammoccete
is hatched, the substance of the fibrilla increases at the
expense of the medullary part, and this goes on until each
plate of muscle-substance consists exclusively of fibrillar sub-
stance. The nuclei have increased in number, but instead
of lying loose in medulla they become squeezed in be-
tween the fibrilla, lose their regular shape and can only
be recognised as small flattened bodies which stain deeply.
The whole plate of muscle-substance now consists of fibrillar
substance which stains uniformly with here and there a more
deeply stained nucleus (fig. 29). The whole appears homo-
geneous, the fibrilla cannot as a rule be recognised, though in
some cases they are seen in transverse section as dots. Each
“‘ Kastchen ” now resembles fundamentally the muscle-plate of
the adult Lamprey ; and it will be noticed that each is a deve-
lopment of what was a single cell.

DEVELOPMENT OF PETROMYZON FLUVIATILIS. 187

The second variety of muscle-fibre met with in the Lamprey
seems to be exclusively derived from the ventral unsegmented
mesoblastic plate, and from the walls of the head cavities.
The muscles with this origin are those which serve to move the
lips, the velum and the other structures of the mouth, and
certain muscles connected with the gill apparatus, and prob-
ably the muscles of the eye. These latter have the same
histological structure, but owing to the fact that the eye does
not develope until the Lamprey stage, no eye muscles appear
till very late in the life of the Ammoceete and I have conse-
quently been unable to follow their development.

The muscle-fibres of this second variety of muscle tissue,
consist of long tubular cells, cylindrical in shape, with a
medulla of clear substance which does not stain, and a cortex
which is thickened at intervals by longitudinal rods. These
give the cortex a beaded appearance in transverse section.
The medulla contains the nucleus, which stains deeply. This
is at first single, but subsequently divides until a row of nuclei
occupy the axis of the muscle-fibre, in some cases so closely
packed as almost to touch. It will be noticed that these
muscle-fibres resemble in the minute structure the first stage
in the development of the muscles forming the myomeres.
These muscle-fibres are transverse striated.

The fibres of the heart belong to this second variety, and are
developed from the same part of the mesoblast. They, however,
possess certain peculiarities which will be described after the
formation of the heart has been considered.

The Heart.

The first appearance of the body cavity as a space takes
place in the region behind the posterior gill-cleft and in front
of the liver. The part of the embryo lying in front of this
region is at an early stage raised from the posterior half by the
backward growth of the head fold, and the embryo les within
the egg-shell bent in half, the angle of the bend being just in
that region where the heart is subsequently formed. By this
means all those parts in front of the liver are free from the
yolk-bearing cells, and the lining cells of the mesenteron all

188 ARTHUR K. SHIPLEY.

become columnar. In this anterior region the mesoblast soon
unites ventrally. In the posterior region the ventral union of
the mesoblast is delayed, the lateral plates of mesoblast lying
between the yolk-cells and the epiblast end in a free edge, and
until these edges unite, the yolk-cells are in contact with the
epidermis ventrally.

In the region between the liver and the last gill-slit the
mesoblast splits at about the fifteenth day into a somatic and a
splanchnic layer; between the two a well-developed body
cavity appears. The former layer lines the body wall, the
latter envelopes the alimentary canal. It forms a dorsal mesen-
tery supporting that structure, and a well-marked ventral
mesentery of considerable depth connecting the ventral wall of
the intestine with the body wall. It is in this ventral mesen-
tery that the heart is developed. The two layers forming the
mesentery fuse dorsally and ventrally, but separate from one
another in their middle, forming a cavity which is the lu-
men of the heart (fig. 24). Subsequently both the mesentery
connecting the heart with the alimentary canal—the meso-
cardium—and the ventral one connecting the heart with the
ventral body wall, atrophy and the heart lies as a tube uncon-
nected with the surrounding structures (fig. 25).

From the fact mentioned above that the mesoblast behind
the heart has not split into somatic and splanchnic layers nor
united ventrally, it will be seen that the cavity of the heart
communicates posteriorly with the space between the veutral
yolk-cells and the epidermis. Such a space would be equiva-
lent to part of the segmentation cavity. Soon after the heart
is formed such a space arises, and at once becomes crowded
with cells destined to form blood-corpuscles (fig. 26). At first
I was inclined to think that these cells were budded off from
the yolk-cells, but more careful observation has led me to
believe that they originate from the free edge of the lateral
plates of the mesoblast, which as I mentioned above are
growing down between the yolk-cells and the epiblast. These
corpuscles are oval with large nuclei, and they usually contain
at first one or two yolk granules which they soon absorb.

DEVELOPMENT OF PETROMYZON FLUVIATILIS. 189

The cavity in which the corpuscles lie in great numbers is subse-
quently shut off by the mesoblast as it grows downwards and
becomes the subintestinal vein. It is from the first continuous
with the posterior end of the heart, and the corpuscles soon
pass from it into that organ. From the first appearance of
the heart in the ventral mesentery its walls have been double ;
‘the splanchnopleure having split into two layers, of these the
outer is at first much the thicker consisting of cubical cells ;
the inner layer is composed of comparatively flattened cells.
The heart at first is a straight tube of the same length as the
section of the body cavity in which it lies. Very soon, how.
ever, it increases in length, and thus becomes slightly twisted ;
at the same time two constrictions appear, dividing it into three
chambers. The most posterior of these is the sinus venosus ;
it is directly continuous with the space in which the corpuscles
are developing. By this time this space has acquired definite
" walls by the downgrowth of the mesoblast in this region, and
it may now be spoken of as the subintestinal vein.

_ The liver which developes as a ventral outgrowth of the
intestine first makes its appearance in this space, and when the
latter gets closed off as a vein, the liver has hecome a branched
gland projecting into it, so that the blood returning from the
alimentary canal passes between the tubuliof the liver. Thus,
from the very first an hepatic portal system is present. The
tubuli of the liver do not appear to have any continuous meso-
blastic coating, though here and there a flattened cell can be
distinguished in the outside of a tubule.

The venous sinus communicates by a narrow opening with
the auricle or second chamber of the heart. This in its turn
opens by a similar narrow opening into the ventricle. This
latter opening is guarded by a pair of valves, which appear by
the tenth day after hatching; they effectually prevent any
regurgitation of the blood into the auricle. The walls of the
ventricle have undergone a considerable change. From the
cells of the inner lining a number of branched muscle-cells
have been developed (fig. 36). These cells stretch across the
cavity of the ventricle from side to side, and fuse and anas-

190 ARTHUR HE. SHIPLEY.

tomose with one another in a very complex manner. They
contain numerous nuclei, and show a longitudinal striation
though not a transverse one. The centre of the ventricle is
comparatively free from them, but at the sides they form a
spongy reticulum in the meshes of which corpuscles abound.

The ventricle passes anteriorly into the ventral aorta, and
at the point where the aorta passes into the solid tissue
between the gills there is another pair of valves resembling
the auriculo-ventricular ones. The ventral aorta, like the
other vessels, arises by a split in the mesoblast which subse-
quently acquires a definite wall. It passes forward as a single
vessel in the ventral median line until it reaches the thyroid
gland, and here it splits in two branches. Each branch then
passes forward on one side of this body, and ends in the most
anterior gill vessel. From the single part of the ventral aorta
three pairs of vessels are given off, passing in front of the
fifth, sixth, and seventh gill-slits respectively. The posterior
wall of the seventh cleft bears no gill filaments, and has no
vessel. From each side of the double part of the ventral aorta
five vessels are given off, the four posterior of these pass in
front of the first, second, third, and fourth gill-slits. The
most anterior is the vessel which in the earlier stages passes in
front of a gill-slit which subsequently disappears. In the
older embryos, when the mouth is fully formed it runs along
the base of the velum.

The vessels after traversing the gills unite in the dorsal
middle line to form the dorsal aorta; this runs backward to
the posterior end of the body, lying just underneath the noto-
chord. From its first appearance it gives off two transverse
vessels in the neighbourhood of the pronephros; these supply
the glomerulus, Anteriorly it gives off a pair of vessels to
supply the upper lip, the carotids. In the older larve the
aorta gives off a vessel which passes dorsally up one myotome,
then along the dorsal surface of the myomere behind it, and
hence the blood is collected by a vein which returns it to the
posterior cardinal down the next myomere. The larve are
fairly transparent, and in each myotome these two opposite

DEVELOPMENT OF PETROMYZON FLUVIATILIS. 191

currents can be seen, and along the top of each myomere a
backwardly directed stream. In the tail the aorta splits, and
one branch passes each side of the cloaca ; they unite ventrally,
and are continued forwards as the subintestinal vein. Before
it splits it gives off a vessel which runs back along the base of
the notochord to supply the tail; this may be termed the
caudal artery. The blood from this is returned by a caudal
vein which soon splits into the two posterior cardinal veins.
These large veins run forward, one each side of the aorta:
the duct of the pronephros runs in their wall. Anteriorly
they unite with the anterior cardinals, and form two ducts of
Cuvier which open into the sinus venosus. The anterior
cardinals bring back blood from the head. The tubuli of the
pronephros lie in their cavity, so that the pronephros, like the
kidney of the Ampbibia, has a double blood supply. The car-
dinal veins do not appear till after the subintestinal vein, which
for some little time is the only vein in the body. Later still a
vessel appears in the right side of the intestine, opposite the
subintestinal vein in the spinal fold; this, like the last named,
passes through the liver. In my latest stages also there is
an impaired vessel bringing blood back to the heart from the
ventral region of the gills; this is mentioned by Balfour.
The blood-corpuscles are of only one kind, large oval disc-like
structures, with a well-marked nucleus. The protoplasm
scarcely stains, but the nucleus assumes a deep colour.

Owing to the transparency of the larva, the circulation can
be watched with great ease. The walls of the vessels at first
possess no elasticity, hence great regurgitation takes place,
and the blood advances by a series of jerks. The valves at the
anterior end of the ventricle and between the auricle and the
ventricle prevent this affecting the blood in the heart.

The heart begins to. beat long before the cells exhibit any
histological differentiation into muscles.

The Pronephros.

The first origin of the larval excretory system is by no means
easy to make out, as it arises at a period when the embryo is

192 ARTHUR E. SHIPLEY.

crowded with yolk. Scott has described it fully, and in most
respects my observations confirm his. As he describes, the
first structure to appear is the segmental duct which is at first
solid. The cells forming this are derived from the mesoblast
cells which lie between the already segmented dorsal part of
the mesoderm and the ventral unsegmented portion. These
cells form a solid cord lying between the mesoblast and the
epiblast ; the cord continues to grow backward by a differen-
tiation of the cells in situ. A few hours later a lumen appears
in the centre of the cord by the separation of the cells; this
soon becomes elliptical in section (fig. 11). It opens into
the posterior part of the alimentary canal.

From this account it will be seen that at first the segmental
duct is between the mesoblast and epiblast ; it, however, soon
comes to occupy a deeper position by the growth of the sur-
rounding tissue. So far we have only considered the duct in
that part of its course where the body cavity is not yet deve-
loped ; but in the region of the heart, where the body cavity
has already appeared, its origin seems to be somewhat different.
The lumen of the segmental duct here becomes continuous
with a groove in the parietal peritoneum, lying near the angle
where the somatopleure and splanchnopleure diverge. When
this groove closes it leaves four or five openings which persist
as the openings of the ciliated funnels. This account of the
origin of the ciliated funnels agrees with that of Fiirbringer,
but differs from Scott’s, who describes the funnels arising
as blind projections of the segmental duct which acquire an
opening into the body cavity. Each funnel soon acquires
cilia, which extend for some distance down its lumen, and are
usually directed downwards towards the tubuli. The funnel
is composed of large cubical cells with a large nucleus, at its
lip it passes suddenly over into the flat cells of the peritoneal
epithelium. At its base it is continuous with a duct which
soou becomes elongated and coiled, and ultimately joins the
segmental duct. The walls of the tubuli are composed of
large clear glandular cells. The posterior end of the seg-
mental duct opeus into the cloaca.

DEVELOPMENT OF PETROMYZON FLUVIATILIS. 193

The segmental duct throughout its course runs in close
connection with the post-cardinal vein, lying in contact with
it, almost in its wall in the under and inner side. In
the anterior region this vein has so grown round the pro-
nephros that the tubuli really lie inside it (fig. 29). The
tubuli are covered by a few flattened cells whose presence
becomes more obvious about the twenty-fifth day by a deposit
of dark brown pigment. The tubuli have thus a venous
blood supply. The glomerulus on the other hand is supplied
from the aorta. There is only one glomerulus on each side,
stretching each side of the alimentary canal and extending
through about the same space as the glandular part of the
kidney. Each glomerulus is a diverticulum of the peritoneum,
which generally becomes sacculated ; it receives its blood by a
single vessel on each side directly from the aorta.

Since the time of Bowman it has been known that the
kidneys of Fishes, Frogs, and Snakes have a double blood
supply, the tubuli uriniferi being surrounded by a capillary net-
work of vessels which receive their blood from the renal portal
veins, and the glomerulus which is supplied with blood from
the aorta by the renal artery. It is an interesting fact to find
that a similar blood supply is present from the very first in
such a temporary organ as the pronephros of the Lamprey.

In the great majority of cases I found fine ciliated funnels
in each pronephros. The whole gland did not extend over a
greater space than that occupied by three myomeres, although
in some cases the ciliated funnels, which were of some length,
overlapped into a fourth myomere, but I was unable to confirm
the relationship alleged to exist between the number of ciliated
funnels and the number of somites through which the pro-
nephros extended.

The Skeleton.

The skeletons of the oldest larva at my disposition consisted
of the notochord derived from the endoderm, and of certain carti-
Jages in the head and branchial region derived from the lateral
mesoblast. The origin of the notochord has been completely

194 ARTHUR E. SHIPLEY.

described by Calberla, Scott, and others, and I have nothing
to add to their account. In the histological differentiation of
the chord froma solid string of more or less cubical cells, to
the vacuolated cylinder which forms the permanent notochord,
there is a stage which is perhaps worth mentioning. In the early
stages a transverse section of the chord shows portions of three
or four cells, a little later these cells have pushed their way
between one another and arranged themselves in such a way
that they occupy the whole room inside the sheath of the
notochord. Whilst in this condition vacuoles appear in the
substance of the cells and for a day or two the notochord pre-
sents very much the same structure as the notochord of Amphi-
oxus. This is, however, soon replaced by the vacuolated
appearance characteristic of the notochordal tissue of the higher
Vertebrata (figs. 18 and 23).

The posterior end of the notochord passes into the indifferent
mass of tissue described in the tail. The anterior end is
slightly curved downwards apparently by the increased vertical
height of the brain. It ends just behind the infundibulum,
its end being in contact with the posterior end of the nasal in-
vagination. There is no trace that it has ever passed in front
of this point, although in the young stages it reaches relatively
almost as far forward as the nervous system. ‘The relation of
its anterior end to the brain hence appears to be due to the
overgrowth of the nervous system anteriorly.

The cartilage which composes the rest of the skeleton is
characterised by the small amount of intercellular substance.
This stains very deeply. ‘Lhe cells are large with usually only
one nucleus, though sometimes two. I have endeavoured to
represent this structure in fig. 19. The branchial bases are
the first part of the skeleton to appear. They arise about the
twenty-fourth day as straight bars of cartilage lying external
and slightly posterior to the branchial vessel. In their relation
to the vessel they correspond with the extrabranchial bars of
the Tadpole, and the Sharks. The true branchial bars run
internal to the branchial vessel.

The bars run behind the gill-slit to which they belong, and

DEVELOPMENT OF PETROMYZON FLUVIATILIS. 195

there is no bar in front of the first persistent cleft. They are
slightly curved inwards towards the median linein the middle part
of their course where they bend round the external opening of
the cleft. About the thirtieth day they fuse with one another
ventrally and so two rods are formed which lie close together
in the posterior half of their course but diverge round the
thyroid. About the same time each bar sends forward two
processes, one above and the other below the opening of the
gill to which it belongs; these ultimately fuse with the pos-
terior edge of the gill bar next in front. The processes of the
most anterior bar fuse with each other. Dorsally the last six of
the bars also become continuous (fig. 42), and form two longi-
tudinal bars which run parallel and close to the notochord.
The most anterior bar does not join this rod but sends a process
inwards, serving to support the auditory capsule, which lies
just in front of it directly over the first persistent gill-cleft.

The first traces of the basi-cranial skeleton appear on the
thirtieth day as two rods of cartilage, the trabecule (figs. 40).
They lie close against the notochord for their posterior two
thirds, anteriorly, however, they diverge and surround the
pituitary space. About six days after their first. appearance
the trabeculze send out laterally a transverse bar of cartilage
which passes out on each side in front of the auditory capsule,
lying between the ganglia of the fifth and seventh nerves. Pro-
fessor Parker has identified this as the rudiments of the pedicle
and pterygoid. They lie in the tissue of the bar which is in
front of the first gill-cleft which has long ago disappeared.

Immediately beneath the trabecule the carotid artery runs
forward as an anterior continuation of the dorsal aorta. The
trabeculz have become continuous with the dorsal end of the
most anterior branchial bar, which is not united with the longi-
tudinal bar formed from the fused dorsal end of the other six.
The connection is very slight but is quite evident in sections.
between this and the dorsal end of the second bar some little
space exists, the latter when it commences lies at a slightly
lower level than the trabecule.

The above description represents the condition in my oldest

196 ARTHUR E. SHIPLEY.

larva, fifty-two days (fig. 43). The further development of
the Lamprey’s skull has been described by Professor Parker in
his great work on ‘The Skeleton of the Massipobranch
Fishes.’

The Mesenteron.

The cavity of the alimentary is formed by the invagination
of the endoderm described in the first section of this article,
when once found it does not disappear again, although in the
region of the intestine it may be reduced to a slit by the
pressure of the surrounding yolk-cells.

The most anterior section, including the branchial region
and that part of the intestine in front of the liver, is now
separated from the rest by the raising of the head and neck
from the remaining part of the embryo. The lining cells of
this portion at once assume a columnar character; the hypo-
blastic cells in the region of the blastopore, or as it may now
be termed the anus, also assume a similar form. But the cells
in the middle part of the intestine still retain the features
of the yolk-cells, those forming the roof of the enteron being
however, rather more columnar than those of the floor and
sides.

In the head region almost the whole of the space inside the
epiblast is taken up with the brain, which has a great depth,
and with the notochord and the alimentary canal, which ends
blindly in front. A small band of mesoblast lies on each side
of the nervous system and notochord. This segments dorsally
into a series of myomeres, the first lying close behind the ear.
Ventrally the mesoblast has not grown down between the en-
doderm, so that along the sides and under surface the hypoblast
and epiblast are in contact. The first gill-slit appears, as Scott
has described, about the twelfth or thirteenth day, the others
arise during the next three or four days, the most posterior
being the last formed. The gill-slits appear to me to be the
result of the ventral downgrowth of mesoblast taking place
only at certain places, these forming the gill-bars. Between
each downgrowth the hypoblastic lining of the alimentary

DEVELOPMENT OF PETROMYZON FLUVIATILIS. 197

canal remains in contact with the epiblast, and here the
gill opening subsequently appears about the twenty-second
day.

Huxley was the first to point out that the embryo
Lamprey possesses eight gill-slits, and his account has been
confirmed by Scott and Dohrn, who, however, point out
that the first slit remains closed, and does not open to the ex-
terior, as Huxley described. Dohrn has further shown that the
first or rudimentary gill-slit becomes converted in the ciliated
groove encircling the mouth, which was first described by
Anton Schneider in Ammocetes.

Fig. 27 represents a longitudinal horizontal section of the
head of a twenty-one days’ old embryo. The eight primitive
gill-slits are here shown lined by columnar epithelium, which
in the posterior seven is most flattened at those points where
the opening will subsequently appear. The corresponding area
in the first cleft, however, will be seen to be lined with very
high columnar cells. These cells afterwards acquire cilia and
come to lie in a deep groove.

The branchial vessels have only appeared in the first gill-
bars, but the cells which will be converted into the cartilagi-
nous gill arches have already become distinct (dr. 4.). About
the twenty-second day a process begins to grow backward from
the middle of each gill-bar into the gill-slit behind. This re-
duces the slit to a <-shaped opening. After the opening to
the exterior has been established the gill-bars overlap each
other, the passage from the cavity of the mouth to the exterior
being directed outwards and backwards. Each gill-bar
acquires a few gill filaments, into which the blood courses.
The whole is covered by a layer of thick columnar epithelium
continuous with that lining the rest of the mouth, except cer-
tain small areas, mostly at the end of the short filaments,
where the epithelium has become suddenly thin, thus putting
the blood into closer communication with the surrounding
water.

The columnar glandular-looking cells which line so much of
the cavity of the mouth contain a number of very fine gran-

198 ARTHUR E. SHIPLEY.

ules, which stain deeply with hematoxylin, giving the cell a
very characteristic appearance. I have been unable to form
any opinion as to the nature or fate of these granules.

The ciliated ring mentioned above is shown in section
in fig. 41,c. g. It lies close in front of the most anterior gill-
bar; ventrally its two halves converge and run back as two
parallel grooves to the opening of the thyroid gland in the
ventral median line. The grooves here unite, and after receiv-
ing the opening of the thyroid they continue as a single groove
running in the ventral median line as far,as the most posterior
gill arch. Dorsally the grooves unite and become continuous
with a median dorsal ridge, which is covered by high columnar
cells, also ciliated. This ridge extends from the first gill arch
to the commencement of the esophagus. Anton Schneider
describes a band of cilia running from this dorsal ridge on
each side along each gill arch. This is not present in my
oldest larva, but is no doubt formed later.

Dohrn (28) has recently described the development of the
thyroid so fully, and his paper is so beautifully illustrated,
that it appears to me to be superfluous to describe again the
origin of this organ. I can only confirm his results. He
deals at length with the homology of the thyroid of Ammo-
ceetes, with the endostyle of Ascidians, and the hypobranchial
ridge in Amphioxus. And the homology of the circumoral
ciliated ring in Ammoceetes and Ascidians is also pointed
out. To these homologies we may add, I think, that of the
dorsal ciliated ridge of the young larval Lamprey to the dorsal
lamella of Ascidians, and the hyperpharyngeal groove of
Amphioxus. It is a curious fact, however, that in the last
animal the form of the structure is reversed. We find ven-
trally a ridge and dorsally a groove, whereas in Ammoceetes
and Ascidians we have the ridge dorsal and the groove ventral.
In spite of this, I thins Dohrn’s arguments fully support the
homology of the ventral organs, and the same reasoning holds
good for the dorsal.

The alimentary canal behind the branchial region may be
divided into three sections. Langerhans has termed these

DEVELOPMENT OF PETROMYZON FLUVIATILIS. 199

the stomach, mid-gut, and hind-gut, but as the most anterior
of these is the narrowest part of the whole intestine, it would
perhaps be better to call it cesophagus. This part of the ali-
mentary canal lies entirely in front of the yolk, and is, with
the anterior region which subsequently bears the gills, raised
from the rest of the egg when the head is folded off. In my
later larvee it is composed of a single layer of very high co-
lumnar cells, and is ciliated throughout. Round this is a thin
layer of cells, which, I imagine, give rise to the muscular
coats. The whole is supported by a dorsal mesentery, each
side of which lies the head kidney (fig. 25). The ciliated
columnar cells are directly continuous with those covering the
dorsal ridge of the branchial region, but not with those of the
ventral groove; this later connection must arise subsequently,
as Anton Schneider describes it in the fully-grown
Ammocete.

The mid-gut which follows the cesophagus is, in the
younger stages, crowded with yolk granules. The cells of the
roof soon acquire a columnar shape, whilst the ventral part
consists of a mass of cubical cells, each crowded with yolk.
By degrees the yolk is absorbed, and the cells assume the same
character as those lining the csophagus. The lumen of the
mid-gut is very much larger than that of the cesophagus, the
alimentary canal expanding suddenly at the commencement
ofthe former. The absorption of yolk takes place from before
backward, so that lumen and walls of the fore part of the mid-
gut assume their permanent size and form, whilst the posterior
half is choked with yolk. The lining high columnar cells are
ciliated and quite continous with those of the cesophagus.

By the time the yolk is all absorbed a longitudinal invagi-
nation of the wall of the mid-gut takes place. This occurs
anteriorly on the left side, but twisting through a quarter of
circle it comes to lie in the ventral side posteriorly. The
ridge thus formed reduces the lumen of the alimentary canal
from around to a reniform shape in section. In this ridge or
spiral valve runs the subintestinal vein, which has become
quite small and has lost its median ventral position. Around

200 ARTHUR HK. SHIPLEY.

this vessel, filling up the space between the two sides of the
spiral valve, is a quantity of fatty tissue. The cilia on the
inner face of the spiral valve are very evident.

The lumen of the mid-gut is so large that almost the whole
of the body cavity in that region of the Ammoceete is taken
up by this part of the intestine; consequently the liver, the
only gland opening into the mid-gut, is pushed forward and
lies on each side and below the cesophagus. This gland has its
origin at a very early stage, about the fourteenth day, as an
evagination of the mid-gut, whilst the latter is still crowded
with yolk. The diverticulum thus produced grows out in the
ventral side of the alimentary canal into that space between
the hypoblast and epiblast which was mentioned above as
being crowded with blood-corpuscles. This space subsequently
becomes enclosed by definite walls by the downgrowth of the
mesoblast in this region. It becomes the subintestinal vein
which still continues to supply the liver with venous blood.
The single diverticulum soon begins to branch, and at an early
stage one of the branches becomes differentiated from the
others, acquires a large lumen, and forms the gall-bladder.
The cells forming the liver are cubical with large nuclei, they
do not appear to have a definite outer layer of flattened cells,
though occasionally such a cell is present. In the older larve
the gall-bladder has a great relative size. It lies embedded in
the liver on the right side of the esophagus. The bile-duct
runs from it above the mid-gut, bending down to enter the
mid-gut in the spiral valve on the left side.

The hind-gut is smaller than the mid-gut, its anterior limit
is marked by the termination of the spinal valve, which does
not extend into this region. The two segmental ducts open
into it just where it turns ventrally to open to the exterior by
a median ventral anus. Its walls are in this region slightly
puckered. The cells lining it are not so high as in the other
parts of the intestine, but more cubical.

Its lumen is from an early stage lined with cells which
have lost their yolk, and it is in wide communication with the
exterior from the first. This condition seems to be, as Scott

DEVELOPMENT OF PETROMYZON FLUVIATILIS. 201

suggests, connected with the openings of the ducts of the pro-
nephros, for this gland is completed and seems capable of
functioning long before any food could find its way through
the mid-gut, or indeed before the stomodzum has opened.

The stomodzeum has a very early origin ; it commences on
the fifteenth day as an invagination of ectoderm against the
blind anterior end of the fore-gut. This gradually deepens
and attains a very large size, partly due to great development
of the upper lip, which grows forward and downward to con-
stitute the large hooded structure which is so characteristic of
the Ammoceete. The greater part of this hood consists of
simple muscle-fibres which interlace and cross one another in
a diagonal direction. The lower lip does not reach so far for-
ward as the upper (figs. 34 and 35). About the twentieth day
the velum begins to appear in the posterior angle of the sto-
modeum. This structure is formed by two grooves which
gradually deepen and cut off a flap of tissue on each side of the
middle line. These two grooves, shown in fig. 27, are not very
deep. The tissue between them is broken through the next
day so that the two lateral folds that remain are covered on
their anterior face by epiblast, and on the greater part of their
posterior face by hypoblast (fig. 28). Subsequently the meso-
blast in these two flaps develope into muscle-fibres, and in the
young larva a constant current is kept up by them, passing in
at the mouth and out at the gill-clefts. This current is easily
demonstrated by the aid of a little Indian ink suspended in
the water.

On the twenty-third day two tentacles begin to grow out
from the under surface of the upper lip, one each side of the
middle line; a little later two more appear on the sides, but
placed more posteriorly ; later still two more appear behind the
level of the last ; these are situated at the junction of the lower
lip with the upper. Finally, a median tentacle appears in the
ventral middle line. This last is far longer than the others
and from its base a ridge, which is at first low, but increases in
height posteriorly, extends back between the ventral portion of
the ciliated ring (figs. 40 and 41). The number of tentacles

15 ;

202 ARTHUR E. SHIPLEY.

is afterwards increased by a pair of new ones arising between
each of those already formed. The tentacles subsequently
become branched (fig. 39).

With regard to the mesoblast of the head I have little to
add to the descriptions of Balfour and Scott. The area
over which the gills extend at their first appearance extends to
the posterior boundary of the sixth myomere. The most
anterior myomere is situated close behind the ear, and the ear
lies above the hyobranchial or first persistent gill-cleft. So
that at their first appearance the six posterior gill-clefts cor-
respond in their extent with the six anterior myomeres. As
the larva grows the gill region appears to elongate with rela-
tion to the muscular myomeres, so in my latest larva there are
about nine myomeres over the area of the six gills (fig. 43).
These anterior myomeres become V-shaped with the open
angles directed forwards; turned the opposite way to those of
Amphioxus.

The mesoblast between the gills arranges itself into head
cavities (fig. 21), and as Balfour and Scott have already
shown, there are two head cavities in front of the hyomandi-
bular cleft. These are at first continuous, but with the for-
mation of the stomodzum they separate. One becomes pre-
oral and obviously corresponding with the premandibular
head cavity of Elasmobranchs; the other with the mandi-
bular (fig. 21). The walls of these cavities ultimately form
the skeleton of the gill arches, the muscles of which are all
of the tubular kind. Owing to the rudimentary condition of
the eye in Ammocetes, no eye-muscles are present and conse-
quently it is impossible to say whether or no they are derived
from the walls of the head cavities, but the researches of
Stannius and Langerhans have shown that they possess
the same histological characters as the muscles of the gills and

upper lip.

The Central Nervous System.

The development of the central nervous system has been
described above up to the stage when the central canal has

DEVELOPMENT OF PETROMYZON FLUVIATILIS. 203

first appeared. The lumen is at first circular in outline, and
the walls of the canal of uniform thickness (fig. 11). Ulti-
mately in the region of the body the lumen becomes elon-
gated and slit like (fig. 24); in the anterior end the lumen
widens into the variously shaped cavities which form the
ventricles of the brain. The cells forming the walls of the
canal are primarily more or less cubical, but they soon
become spindle shaped, except those which form the roof
and the floor of the central canal. These are formed of
a single layer of short columnar cells. The canal is in
the youngest stages proportionately very much larger than in
the later; its size is diminished and its form altered by the
thickenings which take place in different parts of the brain.

The white matter first makes its appearance on the eighteenth
day as two thin bands, one on each side of the brain and
spinal cord (fig. 37). Later these unite in the ventral side
and form an anterior commissure. After the appearance of
the white matter the ganglion cells lose their spindle-shaped
outline and become again circular.

The cranial flexure is very slight; the anterior end of the
brain is, however, slightly bent down, and with it the anterior
end of the notochord (fig. 23).

About the sixteenth day considerable changes take place in
the brain; from the anterior and ventro-lateral angles of the
fore-brain two diverticula are given off; these are the optic
vesicles (fig. 30). They continue to grow upwards and back-
wards till their blind end reaches a position behind and above
the anterior end of the notochord.

At the blind end of the diverticulum a knob is formed by
the outer face proliferating cells, which form a multicellular
retinal layer. The posterior face later on developes pigment
in its cells. The lens is budded off from the inside of the
single layer of epidermis, and lies as a flattened mass of cells
close against the retinal layer (fig. 40). The stalk of the
primary vesicle becomes solid by its walls coalescing on all
sides, and forms the optic nerves. At their origin these nerves
form a commissure projecting into the cavity of the fore-brain

204 ARTHUR BE. SHIPLEY.

on its ventral side; by the twenty-second day this optic
chiasma is covered in by a single layer of ganglion cells. It is
this body that Dohrn has by mistake figured as the Tuber
cinereum (21). The commissure is shown in transverse
section in fig. 39; the lumen of the infundibulum is seen
below it, the cavity of the fore-brain above.

About the same time that the optic vesicles commence to be
given off from the anterior end of the brain a median dorsal
evagination also appears. It was mentioned above that in the
median line, both dorsally, ventrally, and in front, the central
canal is enclosed by a single layer of more or less columnar
cells, whilst the lateral walls are thick. This single layer is
interrupted ventrally by the formation of the optic chiasma.
Dorsally it is produced on the sixteenth day by the evagina-
tion in question, which is the rudiment of the pineal gland
(fig. 31). The walls of the pineal gland then consist at first
of a single layer of cells forming a hollow sac which pushes its
way between the brain and the epidermis, spreading out on all
sides (fig. 31). At first its lumen is continuous with that of
the fore-brain, but ultimately, by the folding of its walls, its
cavity is obliterated and the communication with the lumen of
the fore-brain is shut off.

The eighteenth day, two days after the first appearance of
the optic vesicles and the pineal gland, is the earliest date
on which I have been able to recognise the appearance of any
division into fore-, mid-, and hind-brain. On this day the
single layer of cells roofing the central canal becomes folded in
the manner indicated in fig. 23. This takes place at about
the level of the attachment of the velum, a little in front of
the ear. In larva of fifty-two days, this groove has not
changed its form, but has become deeper.

The division between the fore- and hind-brain is by no
means so well marked; indeed, I have been unable to find any
external groove, although it has been described by previous
writers. Longitudinal horizontal sections through the brain
show, however, that just behind the infundibulum and pineal
gland the walls thin out so that the lumen appears diamond

DEVELOPMENT OF PETROMYZON FLUVIATILIS. 205

shaped. This thin wall I conclude makes the division between
the optic thalami and the crura cerebri.

The hind-brain and mid-brain resemble each other closely
in structure, the mid-brain being only a trifle larger. Their
cavity, which is at first slit like, becomes triangular by the
lateral growth of the roof which pushes the side walls apart
dorsally (figs. 40 and 41). This thin roof extends back as far
as the second gill-cleft, after which it disappears and the
nervous system has the structure represented in fig. 42.

About the forty-fifth day a median longitudinal fold appears
in the thin roof; this is the first of the numerous folds found
in the roof of the mid- and hind-brain of the adult (fig. 41).

The fore-brain still has its thick side walls, the optic thalami.
Just in front of the stalk of the pineal gland a commissure of
transverse fibres is found which runs from side to side on about
the twenty-third day. This commissure corresponds with the
Commissura tenuisima, described by Ahlborn in his
exhaustive work on the brain of the adult Lamprey. It also
probably corresponds with the commissure found by Balfour
in Scyllium situated just in front of the base of the pineal
gland. Osborn has recently described a similar commissure
in the brain of the Amphibia, Menopoma, Meno-
branchus, Amphiuma, and Rana, and I have adopted the
name he proposes for it, the Superior Commissure. The com-
missure of the pineal stalk in the Mammalian brain seems to
occupy the same relative position. This superior commissure
is at first covered with but a few ganglion cells, but these
afterwards increase until two bodies are formed, the Ganglia
Habenule. The left one is very small (fig. 39), but the right is
a structure of considerable size, projecting downwards and back-
wards, and reducing the lumen of the fore-brain to a Y-shaped
slit. These bodies have been fully described by Ahlborn ia
the adult; it is interesting to note that the curious asymmetry
they possess is present from their first appearance. No other
commissure has made its appearance by the fifty-second day.

The cerebral hemispheres show some signs of appearing as
lateral outgrowths in my oldest larve, but no trace of paired

206 ARTHUR TH. SHIPLEY.

lateral ventricles are to be seen. The lateral outgrowths of
the hemispheres embrace between them a mass of tissue formed
at the back of the olfactory pit, which resembles in every way
nerve matter. This structure is shown in figs. 33, 34, and 35,
drawn from a series of sections taken through the head of a
fifty-two days’ larva. This tissue in question appears to con-
sist of ganglion cells. It is traversed by a canal which ends
blindly behind and opens by the median nasal pit in front.
Posteriorly it is continuous with a sheet of tissue which is de-
scribed by Dohrn and Scott as giving rise to the pituitary
body (fig. 39). Unfortunately my larve were not sufficiently
old to enable me to determine whether this mass of tissue
comes into closer relation with the brain and forms the olfac-
tory lobes, or whether, as seems more probable from what we
know of the development of these structures in other animals, it
forms only the peripheral portion of the olfactory apparatus.

About the twenty-fifth day some of the ganglion cells in the
postero lateral angle of the grey matter become much larger
than the surrounding ones. These cells are particularly fre-
quent in that part of the hind-brain lying between the audi-
tory capsule. They probably develope into the “outer large
cells” of Reissner.

With regard to the development of the cranial nerves, I
have no observations on the origin of the olfactory nerve, as
this apparently does not arise till a much later stage than that
attained by my oldest larve. The origin of the optic nerve as
an outgrowth of the brain has been described above. Owing
to the rudimentary condition of the eye, the muscles of that
organ are not developed, and consequently the third, fourth,
and sixth nerves do not arise till a much later stage. This
leaves the fifth, seventh, eighth, ninth, and tenth nerves to be
considered.

The origin of these nerves is much obscured by the yolk
which crowds the cells of the embryo at the time they first
appear. On the seventeenth day the first origin of the ganglia
in the fifth and seventh nerve is seen. The ganglia arise as
proliferations of the epiblast. By this means a knob of cells

DEVELOPMENT OF PETROMYZON FLUVIATILIS. 207

is formed, which arises at about the level of the notochord
(fig. 82). This heap of cells arises close behind the lens of the
eye, but seems to be distinct from it. It is divided into a
larger anterior part, which belongs to the fifth nerve, and a
smaller posterior portion, which forms the ganglion of the
seventh. The roots of the nerves seem to me—though it is
difficult to be certain on this point—to arise as outgrowths
from a neural ridge in the lateral surface of the brain; these
grow down and fuse with epiblastic thickening. This origin
of the roots of the nerves corresponds with that described by
Balfour, Marshall, Van Wijhe, and Beard, in the
Elasmobranchs, and differs from what occurs in the Am-
phibia as described by Spencer, where the nerve also is
derived from the inner layer of epiblast. As Spencer sug-
gests, this is probably due to the presence of a double layer of
epiblast, the epidermic and nervous, in the Amphibia.

By the nineteenth day the ganglion of the fifth nerve has
completely separated off from the skin. It has now divided
into two portions, which have, however, a common root taking
its origin from the hind-brain just in front of the ear. The
most anterior part forms a large ganglion on the root of a
nerve which runs over the eye (fig. 22). This is the oph-
thalmic ganglion, and the nerve is the ophthalmic branch of
the trigeminus; it probably corresponds with the _portio-
profunda of the ophthalmicus superficialis of the Elasmobranchs.
Immediately behind the ophthalmic ganglion, but quite dis-
tinct from it, lies the ganglion of the other half of the fifth
nerve. From this a mandibular nerve proceeds to run close
behind the mouth, and later a maxillary branch appears pre-
orally. In the angle between these ganglia the eye lies. The
nerve connecting the ophthalmic with the main ganglion of
the fifth nerve, described by Ahlborn in the adult, is not found
at this stage, and both the ganglia are of approximately equal
size.

The seventh nerve arises behind the fifth and enters its
ganglion, which, when separated off from the epiblast, lies close
in front of the ear capsule (fig. 38). In early stages

208 ARTHUR E. SHIPLEY.

whilst the most anterior gill-cleft—spiracle—is still present,
the nerve can be seen passing from the ganglia between the
rudimentary gill-cleft and the first persistent one—the hyo-
branchial. Later on the ganglion increases in size, and ex-
tends round the under and inner face of the auditory sac
towards the ganglion of the ninth nerve, but it never quite
reaches it, and the connection between the ganglion of the
seventh and of the tenth nerves must be of later origin.
Neither does the ganglion of the seventh fuse with that of the
fifth, though they are close together, and the root of the
seventh does not enter the ear capsule to leave it again, as is
the case in the adult. After the appearance of the ciliated
ring in the place of the first gill-cleft, the seventh nerve sup-
plies this structure.

A few fibres from the brain enter the recessus labyrinthi
of the ear; these arise close to the root of the seventh, and
constitute the eighth nerve.

The ganglia of the ninth and tenth nerves would seem to
arise from a mass of cells split off from the epiblast close
behind the ear. Ata little later stage the ninth nerve has its
ganglion lying close against the posterior boundary of the ear ;
the nerve is continued along the posterior wall of the first
persistent cleft, the hyobranchial. The ganglion seems to be
still connected with the ganglion of the tenth nerve. This is
a very large structure; it lies more dorsally than the others and
it is in close connection with the mid-brain, having as yet deve-
loped noroot. Behind it and connected with it lies a ganglion
which is situated dorsally above the second persistent gill-cleft ;
from this chord the main branch of the vagus is continued
backward, lying just external to the anterior cardinal vein
(fig. 42). In front of each remaining cleft the chord bears a
large ganglion, so that, counting the first, there are six distinct
ganglia borne on the vagus. Ihave not been able to trace the
fibres of this nerve beyond the last gill-cleft, but my friend
Mr. Ransom, of Trinity College, tells me he has traced the
vagus into the heart in the adult Petromyzon. Each of the
ganglia in the yagus supplies the gill-cleft behind which it lies.

DEVELOPMENT OF PILTROMYZON FLUVIATILIS. 209

There is no trace of the ramus lateralis of the vagus even in
my oldest larve.

The ganglion on the ninth nerve lies in front of the first
myomere, between that and the ear, whilst that of the vagus
lies between the first and second. The first dorsal root of the
spinal nerves with its ganglion lies between the third and
fourth myomere. Behind this there is a dorsal ganglion lying
opposite each myotome.

Sagemehl (17) has described very correctly the origin of
the spinal nerves. The dorsal roots with their ganglia arise
from a neural ridge which is at first of the same size all along.
From this the ganglia begin to grow out about the eighteenth
day, intersegmentally, that is opposite the myotomes. The
ganglia are in connection with one another for some time by a
longitudinal commissure. This commissure appears to consist
of the remains of the neural ridge; it ultimately disappears, as
in Elasmobranchs. The dorsal nerves, after leaving the
ganglia, run into the myotomes and eventually, I believe, reach
the skin, though on this point I cannot be quite certain. On
the other hand the ventral roots consist of nerve-fibres only,
and run straight into the myomeres. They appear, according
to Sagemehl, very soon after the first appearance of white
matter in the chord, and they never have any connection with
the dorsal roots. The resemblance between the distribution of
the spinal nerves of this larva with those of Amphioxus as
described by Rohon is very striking.

The ear is formed, as Scott has described, from an invagina-
tion of the epiblast. This appears very early about the four-
teenth day. It soon deepens and becomes completely shut off,
consisting then of an oval vesicle with a dorsally placed stalk,
the recessus labyrinthi. This last is the remains of the duct
leading to the exterior. The ear is in the same condition in
my oldest larve. No signs of the semicircular canals have
appeared. The epithelium lining the vesicle is high and
columnar ; about the twenty-second day certain patches of the
epithelium become higher than the others and the cells develope
each a very large cilium which projects into the cavity and

210 ARTHUR E. SHIPLEY.

bears a knob at its free end (fig. 41). About the same time a
number of small concretions appear in the ear. These form
the numerous spherical otoliths.

Summary.

I have now described the structure of the chief organs in
my oldest larva, and I propose to conclude this paper by a brief
summary of the results obtained.

In the first place the mesoblast is not completed ventrally by
a layer of cells split off from the hypoblastic yolk-cells, as
Scott has described. But the ventral mesoblast is formed by
the downgrowth of the mesoblastic plates, which ultimataly
meet and unite in the ventral middle line.

The blastopore does not close up, as later observers have
maintained, but, as Max Schultze described thirty years ago,
it persists as the anus. There is no neurenteric canal, though
a solid strand of tissue proceeds back from the alimentary
canal and fuses with an indifferentiated mass of cells, into which
the nervous system and mesoblast also pass.

The lumen of the alimentary canal is that of the mesenteron ;
it does not become obliterated during larval life. In its anterior
end the hypoblast remains in connection with the epiblast at
certain points, and here the gill-clefts arise ; between these the
mesoblast grows down and forms the gill-bars. The origin of
the ciliated ring and the hypopharyngeal groove and hyper-
pharyngeal bar are also described, and the ciliated condition
of the cesophagus and stomach.

The ‘‘ muscle- plates,” whose structure is so peculiar in the
Lamprey, arise each from a single cell of the mesoblastic
somites. This increases in size, slides in between the neigh-
bouring cells, and ultimately occupies the whole of the space
between two myotomes. Its nucleus divides until each cell
contains several nuclei. Striated fibrils then appear and in-
creases till the whole “ muscle-plate ” consists of little else be-
sides these fibrils, squeezing between them a few nuclei. These
«muscle-plates ” arise from the segmental half of the meso-
blast ; the muscles of the gills, lips, and probably of the eye,

DEVELOPMENT OF PETROMYZON FLUVIATILIS. 211

have a different structure and arise from the ventral unseg-
mented part.

The blood-corpuscles arise from the ventral free edges of the
mesoblast, before they unite in the ventral middle line, they
collect in a large sinus just behind the heart. The heart
appears in the ventral mesentery, formed by the union of the
lateral mesoblastic plates ; at first its lumen is continuous with
the sinus just mentioned. This sinus lies between the hypo-
blastic yolk-cells and the epiblast; it subsequently acquires
walls and forms part of the subintestinal vein.

The ciliated funnels of the pronephros are left as apertures
by the segmental duct which in its anterior end is formed from
agroove. The groove closes up at intervals, leaving four or five
openings which become the funnels. They do not arise as
blind projections from the duct, which subsequently, acquire
ciliated openings. From the first the pronephros has a
double blood supply, pure blood from the aorta passing to the
glomerulus, and impure blood in the cardinal veins surrounding
the tubuli.

The early development of the skeleton is described up to the
stage where Professor Parker commenced his researches.

The canal of the central nervous system developes after the
neural chord has separated off from the epidermis; it does not
appear to be lined by any invaginated epidermis, as Calberla
and Scott maintained.

The first sign of differentiation of the parts of the brain is
the formation on the sixteenth day of the optic vesicles and
pineal gland. The division into fore-, mid-, and hind-brain
appears soon after, but the fore- and mid-brain are not sepa-
rated by any well-marked groove. The first transverse com-
missure to appear is situated just in front of the stalk of the
pineal gland. It forms the superior commissure of Osborn.
Afterwards the ganglion cells thicken round it and form the
asymmetrical ganglia habenule.

The ganglia on the fifth, seventh, ninth, and tenth nerves
are derived from epiblastic thickenings. Their roots probably
arise as outgrowths from the neural ridge. The ganglion of the

212 ARTHUR KE. SHIPLEY.

fifth divides into two parts, the ophthalmic and mandibular ;
these have a common root.

The seventh nerve at its first appearance supplies the first
or spiracular gill-cleft ; when this is converted into the ciliated
ring it continues to be supplied by the seventh nerve.

The connection between the fifth, seventh, and tenth nerve
ganglia does not exist and must be of later origin.

The tenth nerve has a large ganglion on its root and bears a
ganglion above each of the last six gill-clefts. No trace of the
ramus lateralis is to be seen.

The origin of the ganglia on the cranial nerves has no
relation to the sense-organs of the skin; these have not
appeared even in my oldest larva.

LITERATURE REFERRED TO.

(1) 1836. Jon. Mttrer.—‘ Vergleichende Anatomie der Myxinoiden, der
Cyclostomen mit durchbohrten Gaumen,’ Berlin.

(2) 1851. Srannrus.—* Ueber den Bau den Muskeln bei Petromyzon
fluviatilis,” ‘ Gottinger Nachrichten,’ 1851.

(3) 1856. Auc. Miitrer.—“ Ueber der Entwicklung der Neunaugen,”
© Miiller’s Archiv,’ 1856.

(4) 1856. Max Scuutrz.—‘ Die Entwickelungsgeschichte von Petromy-
zon Planeri,’ Haarlem.

(5) 1864, Aue. Mitter.—“ Beobachtungen tiber die Befruchtungserschein-
ungen im Hi der Neunaugen,”’ ‘ Verhandl. d. Konigsberger
Phys.-dkonom. Gesellsch.’

(6) 1867. Grenacnrr.— Beitrage zur Erkenntniss der Muskeln der Cyclo-
stomen und Leptocardier,” ‘ Zeit. f. wiss. Zool.,’? Bd. xvii.

(7) 1870. Owssanntkow.—‘ The Development of Petromyzon fluvia-
tilis’ (Russian).

(8) 1873. Paut Lancrrnans.— Untersuchungen iiber Petromyzon
Planeri,” Freilung, i B., 1873.

(9) 1878. Wity. Mtiter.—“ Ueber die Hypobranchialrinne der Tunika-
ten und deren Vorhandsein bei Amphioxus und den Cyclo-
stomen,” ‘Jen. Zeit. f. Med. u. Naturwiss.,’ Bd. vii.

0) 1875. Witu. Miitter.—* Ueber das Urogenitalsystem des Amphioxus
und der Cyklostomen,” ‘ Jen. Zeit. f. Med. u. Naturwiss.,’ Bd. ix.

DEVELOPMENT OF PETROMYZON FLUVIATILIS. 213

(11) 1877.

(12) 1877.

(13) 1878.

(14) 1879.
(15) 1880.
(16) 1881.
(17) 1882.
(18) 1882.
(19) 1883.

(20) 1883.

(21) 1883.
(22) 1884.

(23) 1885.

E, Catperta.— Der Befruchtungsvorgang beim Petromyzon
Planeri,” ‘ Zeit. f. wiss. Zool.,’ Bd. xxx.

HE. Catperta.—‘ Zur Entwicklung des Medullarrohres u. der
Chorda dorsalis der Teleostier und der Petromyzonten,” ‘Morph.
Jahrbuch,’ Bd. iii.

KuprrerR UND Brnrcke.— Der Vorgang der Befruchtung am
Hi der Neunaugen,” ‘ Festschrift zur Feier von Th. Schwann,’
KGnigsberg.

Anton ScHNEIDER.—‘ Beitrage zur vergleichenden Anatomie und
Entwicklungsgeschichte der Wirbelthiere,’ Berlin, 1879.

W. B. Scorr.—“ Vorlaufige Mittheilung. ib. d. Entwicklungs-
geschichte d. Petromyzonten,” ‘ Zool. Anzeiger,’ Nos. 63 and 64,

Nouert.—“ Recherches sur le développement du Petromyzon
planeri,” ‘Archives de Biologie,’ T. ii.

SacEmMpHL.—‘ Untersuchungen iiber die Entwicklung der Spinal-
nerven,’ Dorpat, 1882.

W. B. Scorr.—“ Beitrage zur Entwicklungsgeschichte der Petro-
myzonten,” ‘Morph. Jahrbuch,’ Bd. vii.

W. K. Parker, “On the Skeleton of the Marsipobranch Fishes,”
© Phil. Trans.,’ Part ii, 1888.

AutBorn.— Untersuchungen tiber das Gehirn der Petromyzon-
ten,” ‘Zeit. f. wiss. Zool.,’ Bd. xxxix.

Doury.— Die Entstehung der Hypophysis bei Petromyzon
Planeri,” ‘ Mitth. aus der Zool. Stat. zu Neapel.,’ Bd. iv.
AutBorn.— Ueber den Ursprung und Austritt der Hirnnerven

von Petromyzon,” ‘ Zeit. f. wiss. Zool.,’ Bd. xl.

Dourn.— Die Thyroidea bei Petromyzon, Amphioxus und Tuni-

caten,” * Mitth. aus der Zool. Stat. zu Neapel.,’ Bd. vi.

214 ARTHUR E. SHIPLEY.

EXPLANATION OF PLATES XVIII, XIX, XX, and
XXI,

Illustrating Mr. Arthur E. Shipley’s Paper on “Some Points
in the Development of Petromyzon fluviatilis.”

Reference Letters.

a. Anus. a.c. Anterior cardinal. ao. Aorta. aw. Har. aur. Auricle.
6. c. Body cavity. 4. c. Blood-corpuscles. dp. Blastopore. 6r.1-d7.8 First
to eighth gill-clefts.. 67.4. Skeleton of branchial bars. 6r.v. Vessels of bran-
chial bars. c. Cerebral hemispheres. c.g. Ciliated groove. d./ Dorsal fin.
d. l. Dorsal lamella. d.m. Dorsal mesentery. ¢. Hye. e.g. Egg membrane.
ep. Hpiblast. 7. 6. Fore-brain. fg. Fore-gut. g. Groove between mid-
and hind-brain. g. 4. 7. Left ganglion habenule. g. 4. 7. Right ganglion ha-
benule. g/. Glomerulus. g.z. Ganglion cells at base of olfactory invagination.
Ah. Heart. hd. Head. Ad. c. Head-cavities. 4.4. Hind-brain. Ay. Hypo-
blast. 7. Iter a tertio ad quartum venticulum. if Infundibulum. Ud. ¢.
Liver tubules. 7.7. Lower lip. 7. ¢. Lamina terminalis. m. Mesenteron.
m.6. Mid-brain. m. dr. Muscle of branchial bar. mes. Unsegmented mesoblast.
mes. som. Mesoblastic somites. m./. Muscle-fibre of heart. m.g. Mid-gut. m.
Muscle-plate. my. Myomere. x. Notochord. a. Olfactory invagination.
n.r. Neural ridge. zw. Nucleus of muscle-plate. 0. e. Ciliated epithelium
lining nasal invagination. op. ch. Optic chiasma. oph. Ophthalmic ganglion.
op. th. Optic thalami. op.v. Optic vesicle. p.g. Postanal gut. pix. Pineal
gland. pit. Pituitary body. pr. Primitive streak. r. 7. Recessus labyrinthi.
s. c. Segmentation cavity. s. cm. Superior commissure. s.d. Segmental duct.
sm.pl. Somatopleure. sp. c. Spinal cord. sp. gi. Spinal ganglion. sp. pl.
Splanchnopleure. sf. Stomodeum. s.v. Sinus venosus. ¢. Tentacles. 7h.
Thyroid gland. ¢r. Trabecule. ¢wb. Tubule of pronephros. w. 7. Upper lip.
v. Velum. v. ao. Ventral aorta. ven. Ventricle. v./f. b. Cavity of fore-
brain. v. 4. 6, Cavity of hind-brain, ».7. Ventral ridge in mouth. vv.
Valves of the heart. y.c. Yolk-cells. V.g. Ganglion of fifth nerve. V.g.e.
Epiblastic ingrowth to form ganglion of fifth nerve. VZJI. g. Ganglion of
seventh nerve. X.g. Ganglion of tenth nerve.

PLATE XVIII.

Fic. 1.—Spermatozoa of Petromyzon fluviatilis.

Fic. 2.—Segmenting ovum at the completion of the third or equatorial fur-
row. ¢.g. Egg membrane.

Fie. 3.—Segmenting ovum, showing the next two vertical furrows which
have divided the upper cells and are extending into the lower.

DEVELOPMENT OF PETROMYZON FLUVIATILIS. 215

Fic. 4.—Ovum after the invagination is complete, twelve days old, showing
the blastopore, dp., at posterior end of the neural ridge, z. r.

Fic. 5.—Transverse section through ovum of thirty-six hours. ep. Hpiblast.
s.c. Segmentation cavity. y.c. Yolk-cells.

Fig. 6.—Transverse section through ovum of forty-eight hours. ss. c. Seg-
mentation cavity. ep. Hpiblast. y.c. Yolk-cells.

Fic. 7.—Transverse section through ovum of sixty-seven hours.

Fic. 8.—Transverse section through ovum of eighty-six hours, showing
epiblast gradually thinning out.

Fie. 9.—Longitudinal section through commencing gastrula, 136 hours.
bp. Blastopore. Ay. Hypoblast. y.c. Yolk-cells. m. Mesenteron. s.c. Seg-
mentation cavity.

Fic. 10.—Section through embryo of about the same stage as Fig. 4.
bp. Blatsopore. y.c. Yolk-cells. Ad. Head.

Fic. 11.—Transverse section through the body of an embryo just before
hatching, seventeenth day. sp.c. Spinal cord. 2. Notochord. m. Me-
senteron. mes. Mesoblast. s. d. Segmental duct. Zeiss’s A, oc. 2,
cam. luc.

Fie. 12.—Transverse section through embryo of thirteenth day. sp. c. Spinal
cord. 2. Notochord. mes. Mesoblast. m. Mesenteron. y. c. Yolk-cells.
Zeiss’s A, oc. 2, cam. luce.

Fic. 13.—Transverse section through embryo of fourteen days. Letters
as in Fig. 12. Zeiss’s A, oc. 2, cam. luc.

Fic. 14.—Transverse section through tail of larva twenty days old. sp. e.
Spinal cord. . Notochord. p.g. Solid postanal gut. mes. Mesoblast. dp.
Blastopore. d./. Dorsal fin. Zeiss’s A, oc. 3, cam. luc.

Fic. 15.—Transverse section from the same series as Fig. 14, but posterior
to blastopore. d./. Dorsal fin. mes. Mesoblast. pr. Fused tissue of noto-
chord, spinal cord, and postanal gut, or primitive streak. Zeiss’s A, oc. 3,
cam. luc.

Fie. 16.—Transverse section of embryo just before hatching, seventeen
days, through region of blastopore. 4p. Blastopore. sp.c. Spinal cord. x.
Notochord. y.c. Yolk-cells.

Fic. 17.—Longitudinal section of embryo, showing formation of somites.
n. Notochord. mes. som. Mesoblastic somites. sp.c. Spinal cord. d./.
dorsal fin.

Fic. 18.—Longitudinal section of embryo just before hatching. sp.c. Spinal
cord. my. Myomere. sm. pl. Somatopleuric layer of somite. sp. p/. Splanch-
nopleuric layer. 2. Notochord. Zeiss’s A, oc. 3, cam. luc.

Fie. 19.—A piece of the cartilage of a branchial bar.

Fic. 20.—A longitudinal vertical section through the tail of a larva twenty-

216 ARTHUR E. SHIPLEY.

one days old. a. Anus. p.g. Solid postanal gut. 2. Notochord. sp. c.
Spinal cord. yr. Primitive streak. y.c. Yolk-cells.

Fie. 21.—A longitudinal section through side of head of seventeen days’
embryo, showing the first three evaginations to form gill-clefts, and the true
head-cavities. aw. Har. Ad. c’. and hd. c''. The first and second head-eavity.
br\., br®., and br’. The first rudiments of gill-clefts. 7. v. The vessels of
gills. s¢. Stomodeum. Zeiss’s A, oc. 3.

PLATE XIX.

Fic, 22.—A longitudinal section through side of head of a larva twenty-
one days old. aw. Har. e. Hye. dr'., br?., dr>., drt. The first to fourth
primary gill-clefts. #6. Hind-brain. oph. Ophthalmic ganglion. V7. g.
Ganglion in main branch of fifth nerve.

Fic. 23.—A median longitudinal section through the head of a larva twenty-
one days old. jin. Pineal gland. op.ch. Optic chiasma. z#/. Infundibulum.
m. Notochord. s¢. Stomodeum, 77. Second primitive gill-cleft. ¢%. Thyroid
gland. za. Olfactory invagination. pz¢. Pituitary invagination. m. b. Mid-
brain. 4.4. Hind-brain. g. Groove between mid- and hind-brain. J. ¢.
Lamina terminalis.

Fic. 24.—Transverse section through the body of a larva of twenty days.
sp.c. Spinal cord. fg. Fore-gut. . Notochord. som. pl. Somatopleure.
sp.pl. Splanchnopleuric layers of myomeres. J. c. Body cavity. 4. Heart.
c. f. Ciliated funnel. s.d. Segmental duct. Zeiss’s A, oc. 3, cam. luc.

Fic. 25.—Transverse section through trunk of larva about twenty-four days.
Letters as in Fig, 24, and ao. Aorta. a.c. Anterior cardinal. d. m. Dorsal
mesentery. sp. gl. Spinal ganglion. gl. Glomerulus. Zeiss’s C, oc. 1,
cam. luc.

Fic. 26.—Section through embryo, one day before hatching, seventeen days
old, cut whilst in the egg-shell. 4. Heart. sp.p/. Splanchnopleure. sm.pi.
Somatopleure. 77. and dr*. Seventh and eighth gill-clefts. Ad. Head-cavities
behind these. y.c. Yolk-cells. m.g. Mid-gut. 4. c. Body cavity. Zeiss’s
A, oc. 3, cam. luc.

Fic. 27.—Longitudinal horizontal section through a larva about twenty-two
days. br!—tir®. The eight primary gill-clefts. dr. v. Vessels of gills. dr. b.
Branchial bars. fg. Fore-gut. ¢ub. Tubule of pronephros. s¢. Stomodeeum.
v. Velum. g.z. Ganglion cells at base of nasal invagination. op. ch. Optic
chiasma. mf. Infundibulum. v.f. b. Cavity of fore-brain. 7%. Notochord.
Zeiss’s B, oc. 1, cam. luc.

Fic. 28.—Longitudinal horizontal section through larva of thirty-six days,
u.l. Upper lip. »v. Velum. 74. Thyroid gland. », ao. Ventral aorta. ven.

DEVELOPMENT OF PETROMYZON FLUVIATILIS. 217

Ventricle. aur. Auricle. vv. Valves. s.v. Sinus venosus. J. ¢. Liver
tubules. m.g. Mid-gut. 47.6. Branchial bars. v.7. Ventral ridge. my.
Myomere. Zeiss’s A, oc. 1, cam. luc.

Fic. 29.—Transverse section through proneplhros of larva of forty-seven
days. 2. Notochord. m.p. Muscle-plates. xu. Nucleus. ao. Aorta. a. ¢.
Anterior cardinal. g/l. Glomerulus. ub, Tubules. s.d. Segmental duct.
61. c. Blood-corpuscles. /.g. Fore-gut. d.m. Dorsal mesentery. Zeiss’s D,
oc. 1, cam. luc.

Fie. 30.—Transverse section through fore-brain of embryo, seventeen days.
na. Olfactory epithelium. op. v. Optic vesicle. v.f.6. Cavity of fore-brain.

Fre. 31.—Transverse section through thalamencephalon of larva of eighteen
days. pi. Pineal gland. op. th. Optic thalmi. v.f. 6. Cavity of fore-brain.
na. Olfactory epithelium.

Fic. 32.—Transverse section through region of mid-brain of larva of sixteen
days. s¢. Stomodial epithelium. V. gy. e. Epiblastic origin of ganglion of fifth
nerve. 2. Notochord. m.6. Mid-brain.

Fics. 33, 34, and 35.—A series of sections through the anterior end of head
of a larva fifty-two days old, to show the ganglia cells at base of olfactory
epithelium. w. 7. Upper lip. JZ. 7. Lower lip. ¢. Tentacles. g. x. Ganglion
cells at base of nasal invagination. o.e. Ciliated epithelium lining nasal in-
vagination. ¢. Cerebral hemispheres. v.f. J. Cavity of fore-brain.

PLATE XX.

Fic. 36.—Branched muscle-fibres of heart of larva forty-nine days old.
61. c. Blood-corpuscles. m.f. Muscle-fibre cut across.

Fic. 37.—Transverse section through the hind-brain, showing appearance of
white matter and ganglion of fifth nerve. #6. Hind-brain. s¢. Stomodeum.
V. g. Ganglion of fifth nerve. This section is rather oblique.

Fie. 38.—Transverse section through hind-brain, showing origin of ganglion
of seventh nerve from epiblastic ingrowth. VJJ.g. Ganglion of seventh
nerve. az. Auditory vesicle. fg. Fore-gut.

Fig. 39.—Transverse section through fore-brain of larva forty-nine days
old, to show superior commissure. iz. Pineal gland. v.f. 6. Cavity of fore-
brain. s. cm. Superior commissure. g.4./. Left ganglion habenule. g. h. 7.
Right ganglion habenule. op. ch. Optic chiasma. pit. Pituitary body. inf.
cavity of infundibulum. w./. Upper lip. 7.7. Lower lip. 7. Tentacles.
Zeiss’s C, oc. 1, cam. luc.

Fic. 40.—Transverse section through mid-brain of larva of forty-nine days.
i. Iter a tertio ad quartum ventriculum. e. Hye. ¢r. Trabecule. v. 7.
Ventral ridge. Zeiss’s C, oc. 1, cam. luc.

16

218 ARTHUR E. SHIPLEY.

Fie. 41,—Transverse section through hind-brain of larva of fifty-two days.
v. h. 6, Cavity of hind-brain. aw. Har. 7.7. Recessus labyrinthi. VII. g.
Ganglion of seventh nerve. d./. Dorsal lamella. c.g. Ciliated groove. v.7.
Ventral ridge. v. Velum. ao. Aorta. 4ér.v. Branchial vessels. Zeiss’s A,
oc. 3, cam. luc.

Fie, 42.—Transverse section through region of sixth gill-bar of fifty-two
days’ larva. Gr°. Sixth gill-bar. sp. g/. Spinal ganglion. ao. Aorta. a. c.
Anterior cardinal. 4dr. v. Branchial vessels. @o.v. Ventral aorta. X. 9.
Ganglion in tenth nerve. d./. Dorsal lamella. dr. d. Skeleton of branchia
bars. m. br, Branchial muscles.

PLATE XXI.

Fic. 43.—Drawing of larva of fifty-two days. The notch in the liver,
behind the heart, is due to the large gall-bladder, through whose walls the
cesophagus is seen, This drawing was made by Mr, E, Wilson from the living
specimen.

~
is

AS

io

&
fey

ie}

a
ie

pons
yes

5c.

ox

Bae

ot

[F

Studies M.L.Vol. lll, Pl. XVIII.

RaeBos

SS
Ye

F. Huth, Litht Edin™

Ve

(2)

S Boor Bags Socal

fepe2 pono ones

Studies M.L-Vol.IIl,P] XIX.

Tea “a
CA Brac gHgoooOoS

ers
ce

i

Sy ASS
a REN
a 1

airy
[sTo] Elles { «191

ufo

F Huth, Lith® Edin’

Studies M.L.Vol. III, P1.XX.

&
i,

ogo

Sodda,

Roos

Corus
Creer

F Huth, Lith Edin*

Studies M.L.Vol. IIT, Pl. XX1.

4 Dag

peo —S— sams wie
Sse aess aa ene aaa ai eeuey
Thy bead

¥F Auth, Lith™ Edin*

Discs F. Wc oe

STUDIES

FROM THE

| MORPHOLOGICAL LABORATORY

=

IN THE

!

UNIVERSITY OF CAMBRIDGE.

EDITED BY

‘ADAM SEDGWICK, M.A, ERS.

‘ . FELLOW AND LECTURER OF TRINITY COLLEGE, CAMBRIDGE,
Vol. Ill. Part 1.

London :

C. J. CLAY AND SONS,
CAMBRIDGE UNIVERSITY PRESS WAREHOUSE,
AVE MARIA LANE.

1886

Price Seven Shillings and Sixpence.

CONTENTS.

PAGE
W. Batson. On the morphology of the © Hnteropneuta Plates
I.—XII. Parts I. and II. ; : 1

W. Bateson. On the Ancestry of the Chordata. . : : 67

The above are reprinted from the Quarterly Journal of Microscopical Science.

STUDIES

FROM THE

MORPHOLOGICAL LABORATORY

IN THE

UNIVERSITY OF CAMBRIDGE.

From the Batrour Lisrary, New Muszums, CAMBRIDGE.

The Balfour Library will be glad to receive publications in

exchange for the “Studies.”

Londo:
¢. J. CLAY AND SONS,
CAMBRIDGE UNIVERSITY PRESS WAREHOUSE,
AVE MARIA LANE.
1886

1 JIN | NT.

STUDIES

FROM THE

MORPHOLOGICAL LABORATORY

IN THE

UNIVERSITY OF CAMBRIDGE.

EDITED BY

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

FELLOW AND LECTURER OF TRINITY COLLEGE, CAMBRIDGE.
Vol Nie Part t.

London:
©. J. CLAY AND SONS,
CAMBRIDGE UNIVERSITY PRESS WAREHOUSE,
AVE MARIA LANE.
1886

‘ pee a a
a ~ ee .
* :
' é i
.

tie eh
i ‘ oN = -_ ‘
pe P oe ie Capt 4 B _
ae Seb Ne aa
e 7 ‘ 2 2 7h :
’ 5 aes er ¥ ; :
eo Ai a ee a ie Oo ‘si
an Waa py &. ae
2 Le at a : a
Tie OSLER ERG
, t a ve sustin
MS hoe
i i oes | masa: ee MGT
io rey : 4 " Nie wire ia grt Boer ec
a oa ag ar
“ “ar > DRA io
: it oe Fs a -
pos arte ee :
7 " oR . 4 t
| oreo lee
Roos! as ‘e orn a ener AL rege s
t t desis tT he ae BY . =
ae me ad
7 ae ~ Nyitg
- *

yeoth

~~ el ~
ere -

soak ata

STUDIES

FROM THE

MORPHOLOGICAL LABORATORY

UNIVERSITY OF CAMBRIDGE.

EDITED BY

ADAM SEDGWICK, M.A., F.R.S.

FELLOW AND LECTURER OF TRINITY COLLEGE, CAMBRIDGE.

Vol. Ill. Part 2.

London:

Cc. J. CLAY AND SONS,
CAMBRIDGE UNIVERSITY PRESS WAREHOUSE,
AVE MARIA LANE.

1888

Price Seven Shillings and Sixpence.

CONTENTS.

PAGE
Water Hears. The Development of the Mole ee mie
Stages Eto J. Plates XIII, XIV,XV. . : 105

Sipnry F. Harmer. On the Life History of Pedicellina. Plates
XVI, XVII. ; . : f : 5 : PRs 2) LCL

ArtHuR E. Suretey. On some points in the development of
Petromyzon fluviatilis. Plates XVIII, XIX, XX, XXI. . 2 nha

The above are reprinted from the Quarterly Journal of Microscopical Science.

STUDIES

FROM THE

MORPHOLOGICAL LABORATORY

IN THE

UNIVERSITY OF CAMBRIDGE.

EDITED BY

ADAM SEDGWICK, M.A., F.R.S.

FELLOW AND LECTURER OF TRINITY COLLEGE, CAMBRIDGE,
Vol. Ill. Part 2.

London:

C. J. CLAY AND SONS,
CAMBRIDGE UNIVERSITY PRESS WAREHOUSE,
AVE MARIA LANE.

1888

ye

rho

4
2

i

ae
Buh

Aiea ;
rier ney vy :
aoe ahi tachi
PA tire z Dat We : .
tite ; ; |
; re peat ty BARC tbls" ;
yi 4 sme ith fn
oir hed iy
Hi ata it
afte é

i

Tia eS

(Nis

}
ti