Chapter IX: Mollusca99 (1)

[99] The classification of the Mollusca adopted in the present
chapter is shewn in the subjoined table:

I. ODONTOPHORA. II. LAMELLIBRANCHIATA.
1. Gasteropoda. _a._ Dimya.
_a._ Prosobranchiata. _b._ Monomya.
_b._ Opisthobranchiata.
_c._ Pulmonata.
_d._ Heteropoda.
2. Pteropoda.
_a._ Gymnosomata.
_b._ Thecosomata.
3. Cephalopoda.
_a._ Tetrabranchiata.
_b._ Dibranchiata.
4. Polyplacophora.
5. Scaphopoda.

Although the majority of important developmental features are common to the whole of the Mollusca, yet at the same time many of the subdivisions have well-marked larval types of their own. It will for this reason be convenient in considering the larval characters to deal successively with the different subdivisions, but to take the whole group at once in considering the development of the organs.

_Formation of the layers and larval characters._

ODONTOPHORA.

Gasteropoda and Pteropoda. There is a very close agreement amongst the Gasteropoda and Pteropoda in the general characters of the larva; but owing to the fact that the eggs of the various species differ immensely as to the amount of food-yolk, considerable differences obtain in the mode of formation of the layers and of the alimentary tract.

The spheres at a very early stage of segmentation[100] become divided into two categories, one of them destined to give rise mainly to the hypoblast, the other mainly to the epiblast. According as there is much or little food-yolk the hypoblast spheres are either very bulky or the reverse. In all cases the epiblast cells lie at one pole, which may be called the formative pole, and the hypoblast cells at the opposite pole. When the bulk of the food-yolk is very great, the number of hypoblast spheres is small. Thus in Aplysia there are only two such spheres. In other cases, where there is but little food-yolk, they may be nearly as numerous as the epiblast cells. In all these cases, however, as was first shewn by Lankester and Selenka, a gastrula becomes formed either by normal invagination as in the case of Paludina (fig. 107), or by epibole as in _Nassa mutabilis_ (fig. 105). In both cases the hypoblast becomes completely enclosed by the epiblast. _The blastopore is always situated opposite the original formative pole._ In the large majority of cases (_i.e._ Marine Gasteropoda, Heteropoda, and Pteropoda) the blastopore becomes gradually narrowed to a circular opening which eventually occupies the position of the mouth. It either closes or remains permanently open at this point. In some cases the blastopore remains permanently open and becomes the anus. The best authenticated instance of this is _Paludina vivipara_, as was first shewn by Lankester (No. 263).

[100] The reader is referred for the segmentation to pp. 98-102,
and to the special description of separate types.

In some instances the blastopore assumes before closing a very narrow slit-like form, and would seem to extend along the future ventral region of the body from the mouth to the anus. This appears, according to Lankester (No. 262), to be the condition in Lymnæus, but while Lankester believes that the closure proceeds from the oral towards the anal extremity, other investigators hold that it does so in the reverse direction. Fol (No. 249) has also described a similar type of blastopore. In an undetermined marine Gasteropod, with an embolic gastrula, observed by myself at Valparaiso, the blastopore had the same elongated form as in Lymnæus, but the whole of it soon became closed except the oral extremity; but whether this finally closed could not be determined. It is probable that the typical form of the blastopore is the elongated form observed by Lankester and myself, in which an unclosed portion can indifferently remain at either extremity; and that from this primitive condition the various modifications above described have been derived[101].

[101] Rabl (No. 268) describes a blastopore of this form in
Planorbis which closes at the mouth.

Before the blastopore closes or becomes converted into the oral or anal aperture, a number of very important embryonic organs make their appearance; but before describing these it will be convenient to state what is known with reference to the third embryonic layer or mesoblast.

This layer generally originates in a number of cells at the lips of the blastopore, which then gradually make their way dorsalwards and forwards, and form a complete layer between the epiblast and hypoblast. The above general mode of formation of the mesoblast may be seen in fig. 107, representing three stages in the development of Paludina.

In some cases the mesoblast arises from certain of the segmentation spheres intermediate in size between the epiblast and hypoblast spheres. This is the case in _Nassa mutabilis_, where the mesoblast appears when the epiblast only forms a very small cap at the formative pole of the ovum; and in this case the mesoblast cells accompany the epiblast cells in their growth over the hypoblast (fig. 105).

In other cases the exact derivation of the mesoblast cells is quite uncertain. The evidence is perhaps in favour of their originating from the hypoblast. It is also uncertain whether the mesoblast is bilaterally symmetrical at the time of its origin. It is stated by Rabl to be so in Lymnæus[102].

[102] Rabl (No. 268) has quite recently given a more detailed
account than previous observers of the origin of the mesoblast in
Planorbis. He finds that it originates from the posterior one of
the four large cells which remain distinct throughout the
segmentation. By the division of this cell two 'mesoblasts' are
formed, one on each side of the middle line at the hinder end of
the embryo. Each of these again divides into two, an anterior and
a posterior. By the division of the mesoblasts there arise two
linear rows of mesoblastic cells--the mesoblastic bands--which
are directed forwards and divided transversely into two parts, an
anterior continued from the front mesoblast, and a posterior from
the hinder mesoblast.

If Rabl's account is correct, there is a striking similarity
between the origin of the mesoblast in Mollusca and in Chætopoda.
It appears to me very probable that the mesoblastic bands are
formed (as in Lumbricus) not only from the products of the
division of the mesoblasts, but also from cells budded off from
one or both of the primary germinal layers.

In the case of Paludina the mesoblast becomes two layers thick, and then splits into a splanchnic and somatic layer, of which the former attaches itself to the hypoblast, and gives rise to the muscular and connective-tissue wall of the alimentary tract, and the latter attaches itself to the epiblast, and forms the muscular and connective-tissue wall of the body and other structures. The two layers remain connected by protoplasmic strands, and the space between them forms the body cavity (fig. 107). In most instances there would appear to be at first no such definite splitting of the mesoblast, but the layer has the form of a scattered network of cells between the epiblast and the hypoblast. Finally certain of the cells form a definite layer over the walls of the alimentary canal, and constitute the splanchnic mesoblast, and the remaining cells constitute the somatic mesoblast.

[FIG. 101. DIAGRAM OF AN EMBRYO OF PLEUROBRANCHIDIUM. (From
Lankester.)

_f._ foot; _ot._ otocyst; _m._ mouth; _v._ velum; _ng._ nerve
ganglion; _ry._ residual yolk spheres; _shs._ shell-gland; _i._
intestine.]

We must now return to the embryo at the time when the blastopore is becoming narrowed. First of all it will be necessary to define the terms to be applied to the various regions of the body--and these will best be understood by taking a fully formed larva such as that represented in fig. 101. The ventral surface I consider to be that comprised between the mouth (_m_) and the anus, which is very nearly in the position (_i_) in the figure. As a great protuberance on the ventral surface is placed the foot _f_. The long axis of the body, at this period though not necessarily in the adult, is that passing through the mouth and the shell-gland (_shs._): while the dorsal surface is that opposite the ventral as already defined.

Before the blastopore has attained its final condition three organs make their appearance, which are eminently characteristic of the typical molluscan larva. These organs are (1) the velum, (2) the shell-gland, (3) the foot.

The velum is a provisional larval organ, which has the form of a præoral ring of cilia, supported by a ridge of cells, often in the form of a double row, the ventral end of which lies immediately dorsal to the mouth. Its typical position is shewn in fig. 101, _v_. There are considerable variations in its mode and extent of development etc., but there is no reason to think that it is entirely absent in any group of Gasteropoda or Pteropoda. In a few individual instances, especially amongst viviparous forms and land Pulmonata, it has been stated to be absent. Semper (No. 274) failed to find it in Vitrina, Bulimus citrinus, Vaginulus luzonicus, and Paludina costata. It is very probably absent in Helix, etc.

In some cases, _e.g._ Limax (Gegenbaur), Neritina (Claparède), Pterotrachæa (Gegenbaur), the larva is stated to be coated by an uniform covering of cilia before the formation of the velum, but the researches of Fol have thrown very considerable doubt on these statements. In some cases amongst the Nudibranchiata (Haddon) and Pteropoda there are one or two long cilia in the middle of the velar area. In many Nudibranchiata (Haddon) there is present a more or less complete _post-oral_ ring of small cilia, which belongs to the velum.

The cilia on the velum cause a rotation of the larva within the egg-capsule. Cilia are in most cases (Paludina, etc.) developed on the foot and on a small anal area.

The shell-gland arises as an epiblastic thickening on the posterior and dorsal side. In this thickening a deep invagination (fig. 101, _shs._) is soon formed, in which a chitinous plug may become developed (Paludina, Cymbulia? etc.), and in abnormal larvæ such a chitinous plug is generally formed.

The foot is a simple prominence of epiblast on the ventral surface, in the cavity of which there are usually a number of mesoblast cells (fig. 101, _f_). The larval form just described has been named by Lankester the trochosphere larva.

Before considering the further external changes which the larva undergoes, it will be well to complete the history of the invaginated hypoblast.

[FIG. 102. EMBRYO OF A HETEROPOD. (From Gegenbaur; after Fol.)

_o._ mouth; _v._ velum; _g._ archenteron; _p._ foot; _c._ body
cavity; _s._ shell-gland.]

The hypoblast has after its invagination either the form of a sack (fig. 102) or of a solid mass (fig. 101). Whether the mouth be the blastopore or no, the permanent oesophagus is formed of epiblast cells, so that the oesophagus and buccal cavity are always lined by epiblast. When the blastopore remains permanently open the outer part of the oesophagus grows as a prominent ridge round the opening.

The mesenteric sack itself becomes differentiated into a stomach adjoining the oesophagus, a liver opening immediately behind this, and an intestine. The cells forming the hepatic diverticula and sometimes also those of the stomach may during larval life secrete in their interior peculiar albuminous products, similar to ordinary food-yolk.

The proctodæum, except when it is the blastopore, arises later than the mouth. It is frequently developed from a pair of projecting epiblast cells symmetrically placed in the median ventral line behind the foot. It eventually forms a very shallow invagination meeting the intestine. Its opening is the anus. The anus, though at first always symmetrical and ventral, subsequently, on the formation of the pallial cavity, opens into this usually on the right and dorsal side.

In the cases where the hypoblast is not invaginated in the form of a sack the formation of the mesenteron is somewhat complicated, and is described in the sequel.

From the trochosphere stage the larva passes into what has been called by Lankester the veliger stage (fig. 103), which is especially characteristic of Gasteropod and Pteropod Mollusca.

The shell-gland (with a few exceptions to be spoken of subsequently) of the previous stage flattens out, forming a disc-like area, on the surface of which a delicate shell becomes developed, while the epiblast of the edges of the disc becomes thickened. The disc-like area is the mantle. The edge of the area and with it the shell now rapidly extend, especially in a dorsal direction. Up to this time the embryo has been symmetrical, but in most Gasteropods the shell and mantle extend very much more towards the left than towards the right side, and a commencement of the permanent spiral shell is thus produced.

[FIG. 103. LARVÆ OF CEPHALOPHOROUS MOLLUSCA IN THE VELIGER STAGE.
(From Gegenbaur.)

A. and B. Earlier and later stage of Gasteropod. C. Pteropod
(Cymbulia). _v._ velum; _c._ shell; _p._ foot; _op._ operculum; _t._
tentacle.]

The edge of the mantle forms a projecting lip separating the dorsal visceral sack from the head and foot. An invagination appears, usually on the right in Gasteropods, and eventually extends to the dorsal side (fig. 103 B). It gives rise to the pallial or branchial cavity, and receives also the openings of the digestive, generative and urinary organs. In most Pteropods it is also formed to the right, and usually eventually extends afterwards towards the ventral surface (fig. 103 C). In the pallial cavity the gills are formed, in those groups in which they are present, as solid processes frequently ciliated. They are coated by epiblast and contain a core of mesoblast. They soon become hollow and contractile.

The velum in the more typical forms loses its simple circular form, and becomes a projecting bilobed organ, which serves the larva after it is hatched as the organ of locomotion (fig. 103 B and C). The extent of the development of the velum varies greatly. In the Heteropods especially it becomes very large, and in Atlanta it becomes six-lobed, each lateral half presenting three subdivisions. It is usually armed on its projecting edge with several rows of long cilia, and below this with short cilia which bring food to the mouth. It persists in many forms for a very long period. Within the area of the velum there appear the tentacles and eyes (fig. 103 B). The latter are usually formed at the base of the tentacles.

The foot grows in most forms to a very considerable size. On its hinder and dorsal surface is formed the operculum as a chitinos plate which originates in a depression lined by thickened epiblast, much in the same way as the shell (fig. 103 B and C, _op_). In the typical larval forms it is only possible to distinguish the anterior flattened surface of the foot for locomotion and the posterior opercular region, but special modifications of the foot are found in the Pteropods and Heteropods, which are described with those groups. The foot very often becomes richly ciliated, and otic vesicles are early developed in it (fig. 101, _ot_).

All the Gasteropods and Pteropods have a shell-bearing larval form like that first described, with the exception of a few forms, such as Limax and perhaps some other Pulmonata, in which the shell-gland closes up and gives rise to an internal shell.

The subsequent metamorphosis in the different groups is very various, but in all cases it is accompanied by the disappearance of the velum, though in some cases remnants of the velum may persist as the subtentacular lobes (Lymnæus, _Lankester_) or the lip tentacles (Tergipes, _Nordmann_). In prosobranchiate Gasteropods the larval shell is gradually added to, and frequently replaced by, a permanent shell, though the free-swimming veligerous larva may have a long existence. In many of the Opisthobranchiata the larval shell is lost in the adult and in others reduced. Lankester, who has especially worked at the early stages of this group, has shewn that the larvæ are in almost every respect identical with those of prosobranchiate Gasteropods. They are all provided with a subnautiloid shell, an operculated foot, etc. The metamorphosis has unfortunately been satisfactorily observed in but few instances. In Heteropods and Pteropods the embryonic shell is in many cases lost in the adult.

The following sections contain a special account of the development in the various groups of Gasteropoda and Pteropoda which will complete the necessarily sketchy account of the preceding pages.

Gasteropoda. To illustrate the development of the Gasteropoda I have given a detailed description of two types, viz. _Nassa mutabilis_ and _Paludina vivipara_.

[FIG. 104. SEGMENTATION OF NASSA MUTABILIS. (From Bobretzky.)

A. Upper half divided into two segments. B. One of these has fused
with the large lower segment. C. Four small and one large segment,
one of the former fusing with the large segment. D. Each of the four
segments has given rise to a fresh small segment. E. Small segments
have increased to thirty-six.]

Nassa mutabilis. This form, the development of which has been very thoroughly worked out by Bobretzky (No. 242), will serve as an example of a marine Gasteropod with a large food-yolk. The segmentation has already been described, p. 102. It will be convenient to take up the development at a late stage of the segmentation. The embryo is then formed of a cap of small cells which may be spoken of as the blastoderm resting upon four large yolk cells of which one is considerably larger than the others (fig. 104 A). The small and the large cells are separated by a segmentation cavity. The general features at this stage are shewn in fig. 105 A, representing a longitudinal section through the largest yolk cell and a smaller yolk cell opposite to it. The blastoderm is for the most part one cell thick, but it will be noticed that, at the edge of the blastoderm adjoining the largest yolk cell, there are placed two cells underneath the edge of the blastoderm (_me_). _These cells are the commencement of the mesoblast._ In the later stages of development the blastoderm continues to grow over the yolk cells, and as it grows the three smaller yolk cells travel round the side of the largest yolk cell with it. As they do so they give rise to a layer of protoplasmic cells (fig. 105, _hy_) which form a thickened layer at the edge of the blastoderm and therefore round the lips of the blastopore. These cells form the hypoblast. The whole of the protoplasmic matter of the yolk cells is employed in the formation of the hypoblast. The rest of them remains as a mass of yolk. A longitudinal section of the embryo at a slightly later stage, when the blastopore has become quite narrowed, is represented in fig. 105 C. The greater part of the dorsal surface is not represented.

[FIG. 105. LONGITUDINAL SECTION THROUGH THE EMBRYO OF NASSA
MUTABILIS. (After Bobretzky.)

A. Stage when the mesoblast is commencing to be formed.
B. Stage when the yolk is half enclosed. The hypoblast is seen at
the lips of the blastopore.
C. Stage when the blastopore (_bp_) is nearly obliterated.
D. The blastopore is closed.
_ep._ epiblast; _me._ mesoblast; _hy._ hypoblast; _bp._ blastopore;
_in._ intestine; _st._ stomach; f. foot; _sg._ shell-gland; _m._
mouth.]

Two definite organs have already become established. One of these is a pit lined by thickened epiblast on the posterior and dorsal side (_sg_). This is the shell-gland. The other is the foot (_f_) which arises as a ventral prominence of thickened epiblast immediately behind the blastopore. The hypoblast forms a ring of columnar cells round the blastopore. On the posterior side its cells have bent over so as to form a narrow tube (_in_), the rudiment of the intestine.

In the next stage (fig. 105 D) the blastopore completely closes, but its position is marked by a shallow pit (m) where the stomodæum is eventually formed. The foot (_f_) is more prominent, and on its hinder border is formed the operculum. The shell-gland (not shewn in the figure) has flattened out, and its thickened borders commence to extend especially over the dorsal side of the embryo. A delicate shell has become formed. In front of and dorsal to the mouth, a ciliated ring-shaped ridge of cells, which is however incomplete dorsally, gives rise to the velum. On each side of the foot there appears a protuberance of epiblast cells, which forms a provisional renal organ. The hypoblast now forms a complete layer ventrally, bounding a cavity which may be conveniently spoken of as the stomach (_st_), which is open to the yolk above. Posteriorly however a completely closed intestine is present, which ends blindly behind (_in_).

The shell and with it the mantle grow rapidly, and the primitive symmetry is early interfered with by the shell extending much more towards the left than the right. The anus soon becomes formed and places the intestine in communication with the exterior.

With the growth of the shell and mantle the foot and the head become sharply separated from the visceral sack (fig. 106). The oesophagus (_m_) becomes elongated. The eyes and auditory sacks become formed.

[FIG. 106. LONGITUDINAL SECTION THROUGH AN ADVANCED EMBRYO OF NASSA
MUTABILIS. (After Bobretzky.)

_f._ foot; _m._ mouth; _ce.v._ cephalic vesicle; _st._ stomach.]

With further growth the asymmetry of the embryo becomes more marked. The intestine takes a transverse direction to the right side of the body, and the anus opens on the right side and close to the foot in the mantle cavity which is formed by an epiblastic invagination in this region. The cavity of the stomach (fig. 106, st) increases enormously and passes to the left side of the body, pushing the food-yolk at the same time to the right side, and the point where it communicates with the intestine becomes carried towards the posterior dorsal end of the visceral sack. The walls of the stomach gradually extend so as to narrow the opening to the yolk. The part of it adjoining the oesophagus becomes the true stomach, the remainder the liver; its interior is filled with coagulable fluid.

Paludina. Paludina--Lankester (No. 263) and Bütschli (No. 244)--is a viviparous form characterised by the small amount of food-yolk. The hypoblast and epiblast cells are distinguished very early, but soon become of nearly the same size.

In the later stages of segmentation the epiblast cells differ from the hypoblast cells in the absence of pigment. The segmentation cavity, if developed, is small. A perfectly regular gastrula is formed (fig. 107 A and B), which is preceded by the embryo assuming a flattened form. The blastopore is at first wide, but gradually narrows, and finally assumes a slightly excentric position. _It becomes not the mouth, but the anus._

When the blastopore has become fairly narrow, mesoblast cells (B, _me._) appear around it, between the epiblast and hypoblast. Whether they are bilaterally arranged or no is not clear; and though coloured like the hypoblast, their actual development from this layer has not been followed.

[FIG. 107. FOUR STAGES IN THE DEVELOPMENT OF PALUDINA VIVIPARA.
(Copied from Bütschli.)

_ep._ epiblast; _hy._ hypoblast; _me._ mesoblast; _bl._ blastopore;
_an._ anus; _st._ stomodæum; _sh._ shell-gland; _V._ velum; _x._
primitive excretory organ.]

The velum appears about the same time as the mesoblast, in the form of a double ring of ciliated cells at about the middle of the body (B and C, _V_). The mesoblast rapidly extends so as to occupy the whole space between the epiblast and hypoblast, and at the same time becomes divided into two layers (C). Shortly afterwards a space--the body cavity--appears between the two layers (D) which then attach themselves respectively to the epiblast and hypoblast, and constitute the somatic and splanchnic layers of mesoblast. The two layers remain connected by transverse strands.

By a change in the relations of the various parts and especially by the growth of the posterior region of the body, the velum now occupies a position at the end of the body opposite the blastopore. Immediately behind it there appear two organs, one on the dorsal and one on the ventral side. That on the dorsal side (_sh_) is a deep pit--the shell-gland--which is continuous with a layer of columnar epiblast which ends near the anus. The other organ (_st_), situated on the ventral side, is a simple depression, and is the rudiment of the stomodæum. Between it and the dorsally placed anus is a slight prominence--the rudiment of the foot. On the two sides of the body, between the epiblast and hypoblast on a level with the shell-gland are placed two masses of excretory cells, the provisional kidneys (D, _x_). These are probably _not_ homologous with the provisional renal organ of Nassa and other marine Prosobranchiata. At a later period a ciliated cavity appears in them, which probably communicates with the exterior at the side of the throat.

In the later stages the foot grows rapidly, and forms a very prominent mass between the mouth and the anus. An operculum is developed somewhat late in a shallow groove lined by thickened epiblast.

A provisional chitinous plug is formed in the shell-gland which soon becomes everted. The shell is formed in the usual way on the everted surface of the shell-gland. The thickened edge of this part becomes the edge of the mantle, and soon projects in the neighbourhood of the anus as a marked fold.

With the rapid growth of the larva the invaginated mesenteron becomes relatively reduced in size. In its central part yolk-spherules become deposited, while the part adjoining the blastopore (anus) becomes elongated to give rise to the intestine. The stomodæum grows greatly in length and joins the dorsal part of the archenteron which then becomes the stomach. The part of the mesenteron with yolk-spherules forms the liver. With the development of the visceral sack the anus shifts its position. It first passes somewhat to the left, and is then carried completely to the right.

The development of _Entoconcha mirabilis_ (Joh. Müller, No. 265), a remarkable Prosobranchiate parasitic in the body cavity of Synapta, which in the adult state is reduced to little more than an hermaphrodite generative sack, deserves a short description. It is viviparous, and the ovum gives rise to a larva which from the hardly sufficient characters of the foot and shell is supposed to be related to Natica.

There is nothing very striking in the development. The food-yolk is scanty. The velum, as might be anticipated from the viviparous development, is small. The tentacles are placed not within, but behind the velar area. There is a nautica-like shell, a large mantle cavity, and a large two-lobed foot.

In Buccinum, and Neritina only one out of the many ova included in each egg-capsule develops. The rest atrophy and are used as food by the one which develops.

Opisthobranchiata. It will be convenient to take a species of Pleurobranchidium (Aplysia), observed by Lankester (No. 239), as a type of Nudibranchiate development. The ovum first divides into two segments, and from these small segments are budded off, which gradually grow round and enclose the two large segments. The small segments now form the epiblast.

At the aboral pole the epiblast becomes thickened and invaginated to form the shell-gland, and shortly afterwards the velum and foot are formed in the normal way, and a stomodæum appears close to the ventral edge of the velum (fig. 101). The two yolk cells (_ry_) still remain distinct, but a true hypoblastic layer (probably derived from them, though this has not been made out) soon becomes established. Prominent cells early make their appearance at the base of the foot, which become at a later period invaginated to form the anus. Otolithic sacks (_ot_) become formed in the foot, and the supraoesophageal ganglia from a differentiation of the epiblast (_ng_).

At a later period the shell-gland becomes everted, and a nautiloid shell developed. The alimentary tract becomes completed, though the two yolk cells long retain their original distinctness. The shell-muscle is developed, and peculiar pigmented bodies are formed below the velum. The foot becomes prominent and acquires an operculum.

The metamorphosis of Tergipes has been more or less completely worked out by Nordmann and by Schultze (No. 271).

In _Tergipes Edwardsii_ worked out by the former author, the larva when hatched is provided with a large velum, eyes, tentacles, an elongated operculated foot, and mantle. In the next stage both shell and operculum are thrown off, and the body becomes elongated and pointed behind. Still later a pair of gill-processes with hepatic diverticula becomes formed.

The velum next becomes reduced, and two small processes, which give rise to the lip tentacles and a second pair of gills, sprout out. An ecdysis now takes place, and leads to further changes which soon result in the attainment of the adult form.

In _Tergipes lacinulatus_, observed by Schultze, the velum atrophies before the shell and operculum are thrown off.

Pulmonata. The development of the fresh-water Pulmonata appears from Lankester's observations on the pond-snail (Lymnæus) to be very similar in all important particulars to that of marine Branchiogasteropoda. The velum is however less developed than in most marine forms. The shell-gland, etc. have the normal development. In Lymnæus the blastopore has an elongated form and it is still a matter of dispute whether it closes at the mouth or anus.

In the Helicidæ there is a gastrula by epibole. The shell-gland, as may be gathered from Von Jhering's figures, has the usual form, and an external shell of the usual larval type is developed. There is a ciliated process above the mouth, which extends into the lumen of the mouth. This process is often regarded as a rudimentary velum, but probably has not this value. There is no other organ which can be homologous with the velum.

The development of Limax presents some peculiarities. The yolk-spheres (hypoblast) form a large mass enclosed by the epiblast cells. A shell-gland is formed in the usual situation, which however, instead of being everted, as in ordinary forms, becomes closed, and in its interior are deposited calcareous plates which give rise to the permanently internal shell. The foot grows out posteriorly, and contains a large provisional contractile vesicle, traversed by muscular strands which contract rhythmically.

Although an external shell is present in Clausilia in the adult, the shell-gland becomes closed in the embryo as in Limax, and an internal plate-like shell is developed. The shell is at first covered by a complete epithelium, which eventually gives way in the centre, leaving covered only the edges of the shell. It thus comes about that the original internal shell becomes an external one. It is very difficult to bring this mode of development of the external shell into relation with that of other forms. Clausilia like Limax develops a large pedal sinus.

In both Limax and Clausilia cilia are early developed and cause a rotation of the embryo, but how far they give rise to a distinct velum is not clear.

Heteropoda. The Heteropod embryos present in their early development the closest resemblance to those of other Gasteropods. The segmentation takes place according to the most usual Gasteropod type; (vide p. 99) and after the yolk cells have ceased to give origin to epiblast cells they divide towards the nutritive pole, become invaginated, and line a spacious archenteron. The epiblast cells at the formative pole gradually envelop the yolk (hypoblast) cells, and the blastopore very early narrows and becomes the permanent mouth.

Simultaneously with the narrowing of the blastopore, the shell-gland is formed at the aboral pole, and the foot on the ventral side. The velum appears as a patch of cilia on the dorsal side, which then gradually extends ventrally so as to form a complete circle just dorsal to the mouth.

The larva, after these changes have been completed, is represented in fig. 102.

In later stages the shell-gland becomes everted, and a shell is developed in all the forms both with and without shells in the adult. The foot grows very rapidly, and an operculum is in all cases formed behind. A bilobed invagination in front gives rise to the mucous gland. The velum enlarges and becomes bilobed.

Though the blastopore remains permanently open as the mouth, the oesophagus is formed as an epiblastic ingrowth. The rudiment of the proctodæum appears as two epiblastic cells symmetrically placed behind the foot, which subsequently pass to the right side, and give rise to a shallow invagination which meets the mesenteric sack. In the latter structure the cells of part of the wall develop a peculiar nutritive material, and form a nutritive sack which eventually becomes the liver. The part of the sack connected with the epiblastic oesophagus becomes constricted off as the stomach. The remainder, which unites with the proctodæum, forms the intestine.

The structural peculiarities of the adult are formed by a post-larval metamorphosis. The caudal appendage of Pterotrachea and Firoloidea is formed as an outgrowth of the upper border of the hind end of the foot. The so-called fin arises as a cylindrical process in front of the base of the foot, which is eventually flattened laterally. In the Atlantidæ it is in some cases at first vermiform, and in other cases attains directly its adult structure. The embryonic foot itself gives rise in Pterotrachea, Firoloidea and Carinaria to the tail, on the dorsal and posterior side of which the operculum may still be seen in young specimens. In Atlanta it forms the posterior part of the foot on which the operculum persists through life.

The embryonic shell is completely lost in Pterotrachea and Firoloidea, and the shell is rudimentary in Carinaria. With its atrophy the mantle region also becomes much reduced.

The velum is enormously developed in many Heteropods. In Atlanta it is six-lobed, each of the two primitive lateral lobes being prolonged into three processes, two in front, and one behind. As in all other cases, it atrophies in the course of the post-larval metamorphosis.

[FIG. 108. EMBRYO OF CAVOLINIA (HYALEA) TRIDENTATA. (After Fol.)

_m._ mouth; _a._ anus; _s._ stomach; _i._ intestine; sigma.
nutritive sack; _mb._ mantle; _mc._ mantle cavity; _Kn._ contractile
sinus; _h._ heart; _r._ renal sack: _f._ foot; _pn._ epipodia; _q._
shell; _ot._ otolithic sack.]

Pteropoda. The early larval form of the Pteropods is closely similar to that of marine Gasteropods. There are usually only three hypoblastic spheres at the close of the segmentation in the Thecosomata, and a somewhat larger number in the Gymnosomata. The blastopore closes at the oral region, on the nutritive side of the ovum, and the shell-gland is placed at the original formative pole. The velum, shell-gland and foot have the usual relations. Although many of the adult forms are symmetrical, there is very early an asymmetry visible in the larva, shewing that the Pteropods are descended from asymmetrical ancestors. In the Gymnosomata there is a second larval stage after the loss of the shell when the larva is provided with three rings of cilia (fig. 109). In most forms of Pteropods the dorsal part of the body, covered by the mantle, is produced into a visceral sack like that of the Cephalopoda (fig. 108).

The velum varies considerably in its development in different forms. In the Hyaleidæ it is comparatively small and atrophies early; while in Cymbulia (fig. 103) and the Gymnosomata it is large and bilobed, and persists till after the foot has attained its full development.

The free edge of the velum is provided with long motor cilia, and its lower border with small cilia which bring the food to the mouth. In Cleodora there is a median bunch of cilia in the centre of the velum like that in the Lamellibranchiata, Nudibranchiata, etc.

The shell-gland forms a pit at the aboral end of the body, and in Cymbulia a chitinous plug appears to be normally formed in this pit. The pit afterwards everts itself. The edge of the everted area becomes thickened and gradually travels towards the anterior end of the body. On this everted area a small plate is developed, which forms the commencement of the embryonic shell with which the larvæ of all Pteropods are provided.

The remainder of the embryonic shell is secreted in successive rings by the thickened edge of the mantle, and grows with this till it reaches the neck (fig. 108). The permanent shell is added subsequently, usually on a very different model to the larval shell. The fate of the embryonic shell is very various in different forms. In the Hyaleidæ the animal withdraws itself from the larval shell, which becomes shut off from the permanent shell by a diaphragm. The larval shell then becomes detached.

In the Styliolidæ the permanent shell becomes twice the size of the embryonic shell while the animal is still in an embryonic condition, but the larval shell persists for life. In the Cymbulidæ there is an embryonic and secondary shell, which persist together during larval life. They are eventually cast off at the same time and replaced by a permanent shell.

[FIG. 109. FREE-SWIMMING PNEUMODERMON LARVÆ. (After Gegenbaur,
copied from Bronn.)

The velum has atrophied in both larvæ.
In A three ciliated bands are present, and the auditory vesicles are
visible.
In B the tentacles with suckers and the epipodia have become
developed.
_an._ anus.]

In the Gymnosomata an embryonic shell is developed, and a secondary shell added to it during embryonic life. Both are cast off before the adult condition is attained. After the shell has been cast off three ciliated rings are developed (fig. 109). The anterior of these is placed between the velum and the foot, and the two hinder ones on the elongated posterior part of the body.

The ciliated rings give to these larvæ a resemblance to Chætopod larvæ; but there can be no doubt that this resemblance is a purely superficial one. The anterior ring atrophies early (fig. 109 B), and the second one soon follows suit. It is probable that the hindermost one does not persist through life, although it has been observed in forms with fully developed sexual organs. Most of these larvæ have not been traced to their adult forms. They have been referred to Pneumodermon, Clio, etc.

The most characteristic organ of the Pteropods is the foot, which is prolonged into two enormous lateral wings, the epipodia. These develop at different periods in different larvæ, but are always distinct lateral outgrowths of the foot.

In the Hyaleidæ the foot is early conspicuous, and soon sends out two lateral prolongations (fig. 108 _pn._) which develop with enormous rapidity as compared with the medium portion, and give rise to the epipodia. The whole of the foot becomes ciliated.

In the Cymbulidæ, though not in other forms, an operculum is developed on the hinder surface of the foot (fig. 103 C). The epipodia are late in appearing.

In the Gymnosomata the foot is developed very early, but remains small. The epipodia do not appear till very late in larval life (fig. 109 B).

In Pneumodermon and some other Gymnosomata there appear on the hinder part of the head peculiar tentacles with suckers like those of the Cephalopoda (fig. 109 B). It is not certain that these tentacles are genetically related to the arms of the Cephalopoda.

Cephalopoda. The eggs of the Cephalopoda are usually laid in special capsules formed in the oviduct, which differ considerably in the different members of the group.

In the case of Argonauta each egg is enveloped in an elongated capsule provided with a stalk. By means of the stalk the eggs are attached together in bunches, and these again are connected together and form transparent masses, which are placed in the back of the shell. In octopus the eggs are small and transparent: each of them is enclosed in a stalked capsule. In Loligo the eggs are enveloped in elongated sack-like gelatinous cords, each containing about thirty or forty eggs. The cords are attached in bunches to submarine objects. In Sepia each egg is independently enveloped in a spindle-shaped black capsule, which is attached to a stone or other object.

In a decapod form with pelagic larvæ, described by Grenacher (No. 280), the eggs were enclosed in a somewhat cylindrical gelatinous mass. In each mass there were an immense number of eggs arranged in spirals. Each ovum was enclosed in a structureless membrane, within which it floated in a colourless albumen.

The ovum itself within the capsule is a nearly homogeneous granular mass, without a distinct envelope. Development commences by the segregation, at the narrow pole of the ovum opposite the egg-stalk, of the greater part of the protoplasmic formative material[103]. This material forms a disc equivalent to the germinal disc of meroblastic vertebrate ova. The germinal disc in Sepia and Loligo does not, however, undergo a quite symmetrical segmentation (Bobretzky, No. 279). When eight segments are present, two of them close together are much smaller and narrower than the remainder; and when, in the succeeding stages small segments are formed from the inner ends of the large ones, those derived from the two smaller segments continue to be smaller than the remainder: so that throughout the segmentation one pole of the blastoderm is formed of smaller segments, and the blastoderm exhibits a bilateral symmetry[104]. The partial segmentation results in the formation of a blastoderm covering one pole of the egg, but, unlike the vertebrate blastoderm, formed of a single row of cells. This blastoderm very soon becomes two or three cells deep at its edge, and the cells below the surface constitute the layer from which the mesoblast and hypoblast originate (fig. 110 _ms_). The origin of the mesoblast at the edge of the blastoderm is a phenomenon equivalent to its origin at the lips of the blastopore in so many other types. The external layer forms the epiblast.

[103] In Octopus and Argonauta (Lankester) as soon as the
blastoderm is completed the egg reverses its position in the
egg-shell; the cleavage pole taking up a position nearest the
stalk.

[104] I do not know the relation of this axis of symmetry to the
future embryo.

[FIG. 110. SECTION THROUGH THE BLASTODERM OF A LOLIGO OVUM AT THE
BEGINNING OF THE FOURTH DAY. (After Bobretzky.)

_ms._ mesoblast; _d._ cell at the edge of the blastoderm; _c._ one
of the segmentation cells.]

The whole blastoderm does not take its origin from the segmentation spheres, but, as was discovered by Lankester (282), a number of nuclei arise spontaneously in the yolk outside the blastoderm, around which cell bodies become subsequently formed. They make their appearance near to, but not at the surface, extending first in a ring-like series in advance of the margin of the blastoderm, but subsequently appearing indiscriminately over all parts of the egg. They take no share in forming the epiblast, but would seem, according to Lankester, to assist in giving rise to the lower layer cells, and also to a layer of flattened cells which eventually completely encloses the yolk, and may be called the yolk membrane. The cells of the yolk membrane first of all appear at the thickened edge of the blastoderm. From this point they spread inwards under the centre of the blastoderm (fig. 115 _m´_), and, together with the epiblast cells, outwards over the yolk generally; so that before long (on the tenth day in Loligo) the yolk becomes completely invested by a membrane of cells.

In the non-germinal region the blastoderm is formed of two layers, (1) a flattened epiblast, and (2) the yolk membrane. In the region of the original germinal disc the epiblast cells become columnar, and below them is placed a ring of lower layer cells, which gradually extends towards the centre so as finally to form a complete layer. Below this again comes the yolk membrane just spoken of.

Before describing the further fate of the separate layers it is necessary to say a few words as to the external features of the embryo. In the adult Cephalopod it is convenient, for the sake of comparison with other Mollusca, to speak of the narrow space enclosed in the arms, which contains the mouth, as the ventral surface; the aboral apex as the dorsal surface; and what is usually called the upper surface as the anterior and the lower one as the posterior.

Employing this terminology the centre of the original blastoderm is the dorsal apex of the embryo. In the typical forms with a large yolk-sack the whole embryo is formed out of the original germinal disc; the part of the blastoderm which is continued as a thin layer over the remainder of the egg forms a large ventral yolk-sack appended to the head of the embryo. The following description applies especially to two types, which form the extremes of the series in reference to the development of the yolk-sack. The first of these with a large yolk-sack is Sepia, of which Kölliker in his classical memoir (No. 281) has published a series of beautiful figures. The second, with a small yolk-sack, is the pelagic larva of an unknown adult described by Grenadier (No. 280).

In a young blastoderm of Sepia viewed from the dorsal surface, a series of structures appear which are represented in fig. 111 A. In the middle is a somewhat rhomboid prominence which forms the rudiment of the mantle (_mt_). In its centre is a pit which forms the shell-gland. On each side of the mantle is a somewhat curved fold (_f_). These folds eventually coalesce to form the funnel. They are divided into two parts by a small body which forms the cartilage of the funnel. The smaller part of the fold behind this body gives rise to the true funnel, the part in front becomes (Kölliker) the strong muscle connecting the funnel with the neck-cartilage. In front and to the sides are two kidney-shaped bodies (_oc_), the optic pits. Behind the mantle are two buds (_br_), the rudiments of the gills.

[FIG. 111. TWO SURFACE VIEWS OF THE GERMINAL DISC OF SEPIA. (After
Kölliker.)

_mt._ mantle; _oc._ eye; _f._ folds of funnel; _br._ branchiæ; _an._
posterior portion of alimentary tract; _m._ mouth. 1, 2, 3, 4, 5,
arms; _p._ cephalic lobe.]

In the somewhat later stage rudiments of the two posterior pairs of arms make their appearance outside and behind the rudiments of the funnel. The head is indicated by a pair of lateral swellings on each side, the outer of which carries the eyes. The whole embryo now becomes ciliated, though the ciliation does not cause the usual rotation. At a slightly later stage the second, third, and fourth pairs of arms make their appearance slightly in front of those already present. The posterior parts of the funnel rudiments approach each other, and the anterior meet the rudiments of the neck-cartilage. The gills begin to be covered by the mantle-edge, which now projects as a marked fold. At a slightly later period two fresh rudiments may be noted, viz. the oral (fig. 111 B, _m_) and anal invaginations, the latter of which is extremely shallow and appears at the apex of a small papilla which may be spoken of as the anal papilla. These invaginations appear at the two opposite poles (anterior and posterior) of the blastoderm. Shortly after this the rudiment of the first pair of arms arises considerably in front of the other rudiments, at the sides of the outer pair of cephalic swellings (fig. 111 B, 1).

Fig. 111 B represents a view from the dorsal surface of an embryo at this stage. In the centre is the mantle with the shell-gland which is now very considerably raised beyond the general surface. Concentric with the edge of the mantle are the two halves of the funnel, the anterior half meeting the dorsal or neck-cartilage and the posterior halves approaching each other. The oral invagination is shewn at _m_ and the anal immediately in front of _an_. The gills, nearly covered by the mantle, are seen at _br_. At _p_ are the cephalic swellings, and the eye is seen at _oc_. The arms 1-5 form a ring outside these parts. The whole of the embryo, with the exception of the gills, the funnel, and the outer border of the blastoderm, is richly ciliated.

The embryo up to this time has had the form of a disc or saucer on the surface of the yolk. After this stage it rapidly assumes its permanent dome-like form, and becomes at the same time folded off from the yolk. The blastoderm is very slow in enveloping the yolk, and the whole yolk is not completely invested till a considerably later stage than that represented in fig. 111 B. As soon as the blastoderm covers the yolk-sack cilia appear upon it. The mantle grows very rapidly, and its free border soon projects over the funnel and gills. After the two halves of the funnel have coalesced into a tube, it comes to project again beyond the edge of the mantle.

On the completion of the above changes the resemblance of the embryo to a Cuttle-fish becomes quite obvious. Three of the stages in the accomplishment of these changes are represented in fig. 112.

[FIG. 112. SIDE VIEWS OF THREE LATE STAGES IN THE DEVELOPMENT OF
SEPIA. (After Kölliker.)

_m._ mouth; _yk._ yolk-sack; _oc._ eye; _mt._ mantle.]

To the ventral side of the embryo is attached the enormous external yolk-sack (_yk_), which is continuous with an internal section situated within the body of the embryo. The general relations of the embryo to the yolk will best be understood by reference to the longitudinal section of Loligo, fig. 127.

The arms gradually increase in length, and the second pair passes in front of the first so as eventually to lie completely in front of the mouth. The arms thus come to form a complete ring surrounding the mouth, of which the original second pair, and not, as might be anticipated, the first, completes the circle in front. The second pair develops into the long arms of the adult.

After the embryo has attained more or less completely its definite form (fig. 112 C) it grows rapidly in size as compared with the yolk-sack. The latter structure is at first four or five times as big as the embryo, but, by the time of hatching, the embryo is two to three times as big as the yolk-sack.

Loligo mainly differs from Sepia in the early enclosure of the yolk by the blastoderm, and in the embryo exhibiting the phenomena of rotation within the egg-capsule so characteristic of other Mollusca.

In Argonauta the yolk-sack is still smaller than in Loligo, and the yolk is early completely enclosed by the blastoderm. A well-developed outer yolk-sack is present during early embryonic life, but is completely absorbed within the body before its close. Cilia appear on the blastoderm very early, but vanish again when the yolk is about two-thirds enclosed. There is, during embryonic life, no trace of a shell, but the mantle and other parts of the body become covered by peculiar bunches of fine setæ. The shell-gland develops normally in both Octopus and Argonauta, but disappears again without closing up to form a sack (Lankester).

The pelagic Decapod larva described by Grenacher, which forms my second type, must be placed with reference to the development of the yolk-sack at the opposite pole to Sepia. Segmentation, as in other Cephalopods, is partial, but the blastoderm almost completely envelops the yolk before any organs are developed; and no external yolk-sack is present. At a stage slightly before the closure of the yolk-blastopore the mantle is formed as a slight prominence at the blastodermic pole of the egg, and even at this early stage is marked by the presence of chromatophores. The edge of the blastoderm is ciliated. At a slightly later stage the embryo becomes more cylindrical, the edge of the mantle becomes marked by a fold, which divides the embryo transversely into two unequal parts, a smaller region covered by the mantle, and a larger region beyond this. The yolk is still exposed, but rudiments of the optic pit and of two pairs of arms have appeared. The first-formed arms are apparently the anterior, and not, as in Sepia, the posterior.

At a still later stage, represented in lateral and posterior views in fig. 113 A and B, considerable changes are effected. The yolk-blastopore is nearly though not quite closed. The mantle fold (_mt_) is much more prominent, and on the posterior side on a level with its edge may be seen the rudiments of the gills (_br_). The funnel is formed as two independent folds on each side (_inf1_ and _inf2_), which apparently correspond with the two divisions of the funnel rudiments in Sepia. The eye has undergone considerable changes. Close to each rudiment of the funnel may be seen a fresh sense organ--the auditory sack (_ac_). The ventral (upper in the figure) end of the body now forms a marked protuberance, probably equivalent to the foot of other Mollusca (_vide_ p. 225), at the sides of which are seen the rudiments of the arms (1, 2, 3). To the two previously present a third one, on the posterior side, has been added. The blastopore is placed on the anterior side of the ventral protuberance, and immediately dorsal to this is an invagination (_os_) which gives rise to the stomodæum. The ciliation at the edge of the blastopore still persists, but does not lead to the rotation of the embryo.

In later stages (fig. 113 C) the blastopore becomes closed, and the mantle region increases in length as compared with the remainder of the body. The ventral halves of the funnel, each in the form of a half tube, coalesce together to form a single tube (_inf_) in the same manner as in Sepia. A shallow proctodæum (_an_) is formed between the two branchiæ. The eyes (_oc_) stand out as lateral projections, and the arms become much longer.