Chapter IX: Mollusca99 (2)

[FIG. 113. THREE EMBRYOS OF A CEPHALOPOD WITH A VERY SMALL
YOLK-SACK. (After Grenacher.)

_a._ blastopore; _br._ branchiæ; _inf.1_ and _inf.2_ posterior and
anterior folds of the funnel; _g.op._ optic ganglion (?); _oc._
eye; _wk._ white body; _ac._ auditory pit; _os._ stomodæum; _an._
anus; _mt._ mantle; 1, 2, 3. 1st, 2nd, and 3rd pairs of arms.

Still later a fourth pair of arms is added as a bud from each of the posterior pair, and with the growth in length of the arms the suckers make their appearance. The mouth is gradually carried up so as to be surrounded by the arms. The ciliation of the surface becomes more extensive.

During the whole of the above development the interior of the embryo is filled with yolk, although no external yolk-sack is present. The internal yolk-sack falls into three sections; a cephalic section, a section in the neck, and an abdominal section. Of these, that in the neck is the first to be absorbed. The cephalic portion fills out the ventral protuberance already spoken of. The hinder section becomes occupied by the liver which exactly fits itself into this space as it absorbs the material previously there.

It will be convenient at this point to complete the account of the Cephalopoda by a short history of their germinal layers, and by a fuller description of the mantle, shell, and funnel than that given in the preceding pages.

[FIG. 114. LONGITUDINAL VERTICAL SECTION THROUGH A LOLIGO OVUM WHEN
THE MESENTERIC CAVITY IS JUST COMMENCING TO BE FORMED. (After
Bobretzky.)

_gls._ salivary gland; _brd._ sheath of radula; _oe._ oesophagus;
_ds._ yolk-sack; _chs._ shell-gland; _mt._ mantle; _pdh._
mesenteron; _x._ epiblastic thickening between the folds of the
funnel.]

It has already been shewn that in the region of the germinal disc a thick layer of cells becomes interposed between the epiblast and the yolk membrane. This layer (fig. 115 _m_) is mainly mesoblastic, but also contains the elements which form the lining of the alimentary tract. Its cells first become differentiated into mesoblast and hypoblast after the shell-gland has become a fairly deep pit. The mode of differentiation is shewn in fig. 114. On the posterior side of the mantle, at the point marked in fig. 111 B, _an_, a cavity is formed between the yolk membrane and the mesoblast cells (fig. 114, _pdh_). This cavity is the commencement of the anal extremity of the mesenteron, and the columnar cells lining it constitute the hypoblast. The remainder of the lower layer cells are the mesoblast. The mesenteron gradually extends itself till it meets the stomodæum (fig. 127). The proctodæum is formed as a shallow pit close to the first formed part of the mesenteron.

The mesoblast gives rise not only to the organs usually formed in this layer, but also to the nervous centres, etc.

The mantle and shell. The mantle first arises as a thickening of the epiblast on the dorsal surface of the embryo. The thickened integument, with the subjacent mesoblast, soon forms a definite projection, in the centre of which appears a circular pit (figs. 114 _chs_ and 115 _shs_). This pit, which has already been spoken of as the shell-gland, resembles very closely the shell-gland of other Mollusca. The fold around the edge of the shell-gland grows inwards so as gradually to circumscribe its opening, which before long becomes completely obliterated; and the gland forms a closed sack lined by epiblast which grows in an anterior direction (figs. 114 and 127 _cch_).

[FIG. 115. DIAGRAM OF A VERTICAL SECTION THROUGH THE MANTLE REGION
OF AN EMBRYO LOLIGO. (From Lankester.)

[This figure is turned the reverse way up to fig. 114.]
_ep._ epiblast; _y._ food-yolk; _m._ mesoblast; _m´._ cellular yolk
membrane; _shs._ shell-gland.]

The edges of the mantle now begin to project, especially on the posterior side (fig. 127), and within the cavity formed by this projecting lip there are placed the anus (_an_), gills, etc. The projecting lip of the mantle is formed both of epiblast and mesoblast. The whole of the anterior side of the mantle is filled by the elongated shell-sack (_cch_), within which the shell or pen soon becomes secreted.

There are certain difficulties in comparing the shell-gland of the Cephalopoda with that of other Mollusca which will best be rendered clear by the following quotation from Lankester[105]:

[105] "Development of Pond-Snail." _Quart. J. of Micro.
Science_, 1874, pp. 371-374.

"The position and mode of development of the shell-gland of the Cephalopoda exactly agree with that of the shell-gland as seen in the other Molluscan embryos figured in this paper. We are therefore fairly entitled to conclude from the embryological evidence that the pen-sack of Cephalopoda is identical with the shell-gland of other Mollusca.

"But here--forming an interesting example of the interaction of the various sources of evidence in genealogical biology--palæontology crosses the path of embryology. I think it is certain that if we possessed no fossil remains of Cephalopoda the conclusion that the pen-sack is a special development of the shell-gland would have to be accepted.

"But the consideration of the nature of the shell of the Belemnites and its relation to the pen of living Cuttle-fish brings a new light to bear on the matter. Reserving anything like a decided opinion as to the question in hand, I may briefly state the hypothesis suggested by the facts ascertained as to the Belemnitidæ. The complete shell of a Belemnite is essentially a straightened nautilus-shell (therefore an external shell inherited from a nautilus-like ancestor), which, like the nautiloid shell of _Spirula_, has become enclosed by growths of the mantle, and unlike the shell of _Spirula_, has received large additions of calcareous matter from those enclosing overgrowths. On the lower surface of the enclosed nautilus-shell of the Belemnite--the phragmacone--a series of layers of calcareous matter have been thrown down forming the guard; above, the shell has been continued into the extensive chamber formed by the folds of the mantle, so as to form the flattened pen-like pro-ostracum of Huxley.

"Whether in the Belemnites the folds of the mantle which thus covered in and added to the original chambered shell, were completely closed so as to form a sack or remained partially open with contiguous flaps must be doubtful.

"In _Spirula_ we have an originally external shell enclosed but not added to by the enclosing mantle sack.

"In _Spirulirostra_, a tertiary fossil, we have a shell very similar to that of _Spirula_, with a small guard of laminated structure developed as in the Belemnite (see the figures in Bronn _Classen u. Ordnungen des Thierreichs_).

"In the Belemnites the original nautiloid shell is small as compared with _Spirulirostra_. It appears to be largest in Huxley's genus _Xiphoteuthis_. Hence in the series _Spirula_, _Spirulirostra_, _Xiphoteuthis_, _Belemnites_, we have evidence of the enclosure of an external shell by growths from the mantle (as in Aplysia), of the addition to that shell of calcareous matter from the walls of its enclosing sack, and of the gradual change of the relative proportions of the original nucleus (the nautiloid phragmacone) and its superadded pro-ostracal and rostral elements tending to the disappearance of the nucleus (the original external shell). If this view be correct as to the nature of these shells, it is clear that the shell-gland and its plug has nothing to do with them. The shell-gland must have preceded the original nautiloid shell, and must be looked for in such a relation whenever the embryology of the pearly Nautilus can be studied. Now, everything points to the close agreement of the Belemnitidæ with the living Dibranchiata. The hooklets on the arms, the ink bag, the horny jaws, and general form of the body, leave no room for doubt on that point; it is more than probable that the living Dibranchiata are modified descendants of the mesozoic Belemnitidæ. If this be so, the pens of _Loligo_ and _Sepia_ must be traced to the more complex shell of the Belemnite. This is not difficult if we suppose the originally external shell the phragmacone, around which as a nucleus the guard and pro-ostracum were developed, to have finally disappeared. The enclosing folds of the mantle remain as a sack and perform their part, producing the chitino-calcareous pen of the living Dibranch, in which parts can be recognised as corresponding to the pro-ostracum, and probably also to the guard of the Belemnite. If this be the case, if the pen of _Sepia_ and _Loligo_ correspond to the entire Belemnite shell minus the phragmacone-nucleus, it is clear that the sack which develops so early in _Loligo_ and which appears to correspond to the shell-gland of the other Molluscs cannot be held to do so. The sack thus formed in _Loligo_ must be held to represent the sack formed by the primæval upgrowth of mantle folds over the young nautiloid shell of its Belemnitoid ancestors, and has accordingly no general significance for the whole Molluscan group, but is a special organ belonging only to the Dibranchiate stem, similar to--but not necessarily genetically connected with--the mantle fold in which the shell of the adult _Aplysia_ and its congeners is concealed. The pen, then, of Cephalopods would not represent the plug of the shell-gland. In regard to this view of the case, it may be remarked that I have found no trace in the embryonic history of the living Dibranchiata of a structure representing the phragmacone; and further, it is possible, though little importance can be attached to this suggestion, that the Dibranchiate pen-sack, as seen in its earliest stage in the embryo _Loligo_, etc., is fused with the surviving remnants of an embryonic shell-gland. When the embryology of _Nautilus pompilius_ is worked out, we shall probably know with some certainty the fate of the Molluscan shell-gland in the group of the Cephalopoda."

The funnel. The general development of the funnel has already been sufficiently indicated. The folds of which it is formed are composed both of epiblast and mesoblast. The mesoblast of the anterior part of each half of the funnel would appear to give rise to a muscle passing from the cartilage of the neck to the funnel proper. The posterior parts gradually approximate, but meet in the first instance ventrally. The two folds at first merely form the side of a groove or imperfect tube (fig. 113 C and 124 ff.), but soon the free edges unite and so give rise to a perfect tube, the primitive origin of which by the coalescence of two halves would not be suspected. In Nautilus the two halves remain permanently separate but overlap each other, so as to form a functional tube.

[FIG. 116.

I. CHITON WOSSNESSENSKII. (After Middendorf.)
II. CHITON DISSECTED to shew _o._ the mouth; _g._ the nervous ring;
_ao._ the aorta; _c._ the ventricle; _c´._ an auricle; _br._ the
left branchiæ; _od._ oviducts. (After Cuvier.)
III., IV., V. STAGES OF DEVELOPMENT OF CHITON CINEREUS. (After
Lovén.)
The figure is taken from Huxley.]

Polyplacophora. The external characters of the embryo of Chiton have long been known through the classical observations of Lovén (No. 285), while the formation of the layers and the internal phenomena of development have recently been elucidated by Kowalevsky (No. 284). The eggs are laid in April, May, and June, and are enclosed in a kind of chorion with calcareous protuberances. The segmentation remains regular till sixty-four segments are formed. The cells composing the formative half of the ovum then divide more rapidly than the remainder; there is in this way formed an elongated sphere, half of which is composed of small cells and half of larger cells. In the interior is a small segmentation cavity. From its eventual fate the hemisphere of the smaller cells may be called the anterior pole, and that of the larger cells the posterior. An involution of the cells at the apex of the posterior pole (though not of the whole hemisphere of larger cells) now takes place, and gives rise to the archenteron. At the same time an equatorial double ring of large cells appears on the surface between the two poles, which becomes ciliated and forms the velum. At the apex of the anterior pole a tuft of cilia, or at first a single flagellum, is established (fig. 116 III. and IV.).

In the succeeding developmental period the blastopore, which has so far had the form of a circular pore at the posterior extremity of the body, undergoes a series of very remarkable changes. In conjunction with a gradual elongation of the larva it travels to the ventral side, and is prolonged forwards to the velum as a groove. The middle part of the groove is next converted into a tube, which opens externally in front, and posteriorly communicates with the archenteron. The walls of this tube subsequently fuse together, obliterating the lumen, and necessarily causing at the same time the closure of the blastopore. The tube itself becomes thereby converted into a plate of cells on the ventral surface between the epiblast and the hypoblast[106].

[106] There is a striking similarity between the changes of the
blastopore in Chiton and the formation of the neurenteric canal
of Chordata; especially if Kowalevsky is correct in stating that
the pedal nerves are developed from the ventral plate.

While the above changes have been taking place the mesoblast has become established. It is derived from the lateral and ventral cells of the hypoblast.

After the establishment of the germinal layers the further evolution of the larva makes rapid progress. A transverse groove is formed immediately behind the velum, which is especially deep on the ventral surface; and the stomodæum is formed as an invagination of the anterior wall of the deeper section of the groove. Behind the stomodæum the remainder of the ventral surface grows out as a flattened foot.

The dorsal surface behind the velum constitutes the mantle, and becomes divided by six or seven transverse grooves into segment-like areas, which may be called mantle plates (fig. 116 IV.). These areas would seem (?) to correspond to so many flattened out shell-glands. Immediately behind the velum the eyes appear as two black spots (fig. 116 IV.).

While the above external changes take place the archenteron undergoes considerable modifications. Its anterior section gives rise, according to Kowalevsky, to a dorsal (?) sack in which the radula is formed; while the liver arises from it as two lateral diverticula.

From the above statements it would appear that Kowalevsky holds that the oesophagus and radula sack are both derived from the walls of the archenteron and not from the stomodæum. Such an origin for these organs is without parallel amongst Mollusca.

The larva becomes about this time hatched, and after swimming about for some time attaches itself by the foot, throws off its larval organs, cilia, etc., and develops the shell.

The shell appears first of all during larval life in the form of spicula on the middle and sides of the head, and later on the middle and sides of the post-oral mantle plates (fig. 116 V.). The permanent shell arises somewhat later as a series of median and lateral calcareous plates, first of all on the posterior part of the velar area, and subsequently on the mantle plates behind. The three calcareous patches of each plate fuse together and give rise to the permanent shell plates. The original spicula are displaced to the sides, where they partly remain, and are partly replaced by new spicula.

The nervous system is formed during larval life as four longitudinal cords:--two lateral--the branchial cords, and two ventral--the pedal. Paired anterior thickenings of the pedal cords meet in front of the mouth to form the oesophageal ring. The pedal cords and their derivatives are believed by Kowalevsky to be developed from the lateral parts of the plate formed by the metamorphosis of the blastopore. The median part of the plate is still visible after the formation of these parts.

The chief peculiarity of the larva of Chiton (apart from the peculiar ventral plate) consists in the elongation and dorsal segmentation of the posterior part of the body. The velum has the normal situation and relation to its mouth. The position of the eyes behind it is however abnormal.

The elongation and segmentation of the posterior part of the trunk is probably to be regarded as indicating that Chiton has early branched off from the main group of the Odontophora along a special line of its own, and _not_ that the remaining Odontophora are descended from Chiton-like ancestral forms. The shell of Mollusca on this view is not to be derived from one of the plates of Chiton, but the plates of Chiton are to be derived from the segmentation of a primitive simple shell. The segmentation exhibited is of a kind which all the trochosphere larval forms seem to have been capable of acquiring. The bilateral symmetry of Chiton, which is quite as well marked as that of the Lamellibranchiata, indicates that it is a primitive phylum of the Odontophora.

Scaphopoda. The external characters of the peculiar larva of this interesting group have been fully worked out by Lacaze Duthiers (No. 286).

The segmentation is unequal and conforms to the usual molluscan type. At its close the embryo becomes somewhat elongated, and there appears on its surface a series of transverse ciliated rings. As soon as these become formed the larva is hatched, and swims about by means of its cilia. Six ciliated bands are formed in all, and in addition a tuft of cilia is formed in a depression at the anterior extremity.

The larva thus constituted is very different in appearance to the larvæ already described, and its parts very difficult to identify; the next stages in the development shew however that the whole region of the body taken up by the ciliated rings is part of the velar area, while the small papilliform region behind is the post-velar part of the embryo. This latter part grows rapidly, and at the same time the ciliated rings become reduced to four; which gradually approach each other, while the region on which they are placed grows in diameter. The rings finally unite, and form a single ring on a projecting velar ridge. In the centre of this ring is placed the terminal tuft of cilia on a much reduced prominence.

By the time that these changes have been effected in the velum, the post-velar part of the embryo has become by far the largest section of the embryo, so that the velum forms a projecting disc at the front end of an elongated body. The mantle is formed as two lateral outgrowths near the hinder extremity of the body which leave between them a ventral groove lined by cilia; on their dorsal side is formed a delicate shell. The mantle lobes continue to grow, and by the time the above changes in the velum are effected they meet and unite in the ventral line and convert the groove between them into a complete tube open in front and behind. A stream of water is driven through this tube by the action of the cilia. The shell, which is at first disc-shaped like the shell of other molluscan larvæ, moulds itself upon the mantle and is so converted into a tube. At the front end of the mantle tube, which does not at first cover the velum, there is formed the foot. It arises as a protuberance of the ventral wall of the body, which rapidly grows forwards, becomes trilobed as in the adult, and ciliated.

On the completion of these changes the larva mainly differs in appearance from the adult by the projection of the velum beyond the edge of the shell. The velum soon however begins to atrophy; and the larva sinks to the bottom. The mantle tube and shell grow forward and completely envelop the velum, which shortly afterwards disappears. The mouth is formed on the ventral side of the velum at the base of the foot; at its sides arise the peculiar tentacles so characteristic of the adult Dentalium.

LAMELLIBRANCHIATA.

The larvæ of Lamellibranchiata have in a general way the same characters as those of Gasteropods and Pteropods. A trochosphere stage with a velum but without a shell is succeeded by a veliger stage with a still more developed velum, a dorsal shell, and a ventral foot.

The segmentation is unequal, and in a general way like that of Gasteropoda, but the specially characteristic Gasteropodan type with four large yolk spheres is only known to occur in Pisidium, and a type of segmentation similar to that of Anodon (p. 100) appears to be the most frequent.

There is an epibolic or embolic gastrula, but the further history of the formation of the germinal layers has been worked out so imperfectly, and for so few types, that it is not possible to make general statements about it. What is known on this head is mentioned in connection with the description of the development of special types.

The blastopore in some cases closes at the point where the anus (Pisidium), and probably in other cases where the mouth, is eventually formed. In Anodon it is stated to close at a point corresponding neither with the mouth nor the anus, but on the dorsal surface!

The embryo assumes a somewhat oval form, and in the free marine forms there appears very early in front of the mouth a well-developed velum. This is formed according to Lovén from two papillæ, and takes the form of a circular ridge armed with long cilia. In the centre of the velar area there is usually present a single long flagellum (fig. 117 B and C). The velum never becomes bilobed.

[FIG. 117. THREE STAGES IN THE DEVELOPMENT OF CARDIUM. (After
Lovén.)

_hy._ hypoblast; _b._ foot; _m._ mouth; _an._ anus; _V._ velum;
_cm._ anterior adductor muscle.]

In the later stages, after the development of the shell, the velum becomes highly retractile and can be nearly completely withdrawn within the mantle by special muscles. It forms the chief organ of locomotion of the free larva.

In some fresh-water forms, which have no free larval existence, the velum is very much reduced (Anodon, Unio, Cyclas) or even aborted (Pisidium). In these forms as well as in Teredo and probably other marine forms (_e.g._ Ostrea) the central flagellum is absent. It has been suggested by Lovén, though without any direct evidence, that the labial tentacles of adult Lamellibranchiata are the remains of the velum. The velar area is in any case the only representative of the head. In some marine forms a general covering of cilia arises before the formation of the velum; and in Montacuta and other types there is developed, as in many Gasteropoda, a circumanal patch of cilia.

A shell-gland appears at a very early period on the dorsal surface in Pisidium, Cyclas and Ostrea, and probably in most marine forms (fig. 118, _shs_). It is somewhat saddle-shaped, and formed of elongated non-ciliated cells bounding a groove. It flattens out and on its surface is formed the shell, which appears usually to have the form of an unpaired saddle-shaped cuticle, on the two sides of which the valves are subsequently formed by a deposit of calcareous salts. In Pisidium the two valves are stated by Lankester to be at first quite independent and widely separated, and it has been suggested by Lankester, though not proved, that the ligament of the shell is developed in the median part of the groove of the shell-gland.

The mantle lobes are developed as lateral outgrowths of the body: they usually have a considerable extension before they are covered by the shell. In Anodon and Unio the larval mantle lobes are, however, formed in a somewhat exceptional way, and are from the first completely covered by the valves of the larval shell. The larval mantle lobes and shell in Anodon and Unio are subsequently replaced by the permanent structures.

[FIG. 118. AN EMBRYO OF PISIDIUM PUSILLUM. (From Lankester.)

_f._ foot; _m._ mouth; _ph._ pharynx; _gs._ bilobed stomach; _pi._
intestine; _shs._ shell-gland.]

The adductor muscles are formed soon after the appearance of the shell. The posterior sometimes appears first, _e.g._ Mytilus, and at other times the anterior, _e.g._ Cardium.

The foot arises in the usual way as a prominence between the mouth and anus. In comparison with Gasteropoda it is late in appearing, and in many cases does not become prominent till the shell has attained a considerable size. In its hinder part a provisional paired byssus-gland is developed from the epidermis in Cyclas and other forms. In other cases, _e.g._ Mytilus, the byssus-gland is permanent. The byssus-gland occupies very much the position of the Gasteropod operculum, and would appear very probably to correspond with this organ. The anterior part of the foot is usually ciliated.

The gills appear rather late in larval development along the base of the foot on either side, between the mantle and the foot (fig. 120, _br_). They arise as a linear row of separate ciliated somewhat knobbed papillæ. A second row appears later. The two rows give rise respectively to the two gill lamellæ of each side.

The further history of the development of the gills has been studied by Lacaze Duthiers (No. 297) in Mytilus. The first row of gill papillæ formed becomes the innermost of the two lamellæ of the adult. The number of papillæ goes on increasing from before backwards. When about eleven have been formed, their somewhat swollen free extremities unite together, the basal portions being separated by slits.

The free limb is formed by the free end of the gill lamella bending upon itself towards the inner side and growing towards the line of attachment of the lamella. The free limb is at first not composed of separate bars, but of a continuous membrane. Before this membrane has grown very wide, perforations are formed in it corresponding to the spaces between the bars of the attached limb.

The outer gill lamella develops in precisely the same way as, but somewhat later than, the inner. The rudiments of it appear when about twenty papillæ of the inner lamella are formed. Its first papillæ are formed near the hind border of the inner lamella, and new papillæ are added both in front and behind. Its free limb is on the outer side.

In Mytilus the two limbs (free and attached) of each bar of the gill are joined at wide intervals by extensile processes, the 'inter-lamellar junctions,' and the successive bars are attached together by ciliated junctions. In many other types the concrescences between the various parts of the gills are carried much further; the maximum of concrescence being perhaps attained in Anodon and Unio[107].

[107] R. H. Peck, "Gills of Lamellibranch Mollusca." _Quart. J.
of M. Science_, Vol. XVII. 1877.

Large paired auditory sacks seem always to be developed in the foot; and clearly correspond with the auditory sacks in Gasteropoda.

Eyes are frequently present in the larva, though they disappear in the adult. In Montacuta and other types a pair of these organs is formed at the base of the velum on each side of the oesophagus, not far from the auditory sacks. They are provided with a lens.

A row of similar organs is present in the larva of Teredo in front of the foot.

Cardium. As an example of a marine Lamellibranchiate I may take _Cardium pygmaeum_, the development of which has been studied by Lovén (No. 291). The ova, surrounded by a thickish capsule, are impregnated in the cloaca. The segmentation takes place much as in Nassa (_vide_ p. 101), and the small segments gradually envelop the large hypoblast spheres; so that there would seem to be a gastrula by epibole. After the hypoblast has become enveloped by the epiblast, one side of the embryo is somewhat flattened and marked by a deepish depression (fig. 117 A). From Lovén's description it appears to me probable that the depression on the flattened side occupies the position of the blastopore, and that the depression itself is the stomodæum. At this stage the embryo becomes covered with short cilia which cause it to rotate within the egg-capsule.

Close above the mouth there appear two small papillæ. These gradually separate and give rise to a circular ridge covered with long cilia, which encircles the embryo anteriorly to the ventrally-placed mouth. This structure is the velum. In its centre is a single long flagellum (fig. 117 B). Shortly after this the shell appears as a saddle-shaped structure on the hinder part of the dorsal surface of the embryo. It is formed at first of two halves which meet behind without the trace of a hinge (fig. 117 C). The two halves rapidly grow and partially cover over the velum, and below them the mantle folds soon sprout out as lateral flaps.

The alimentary tract has by this time become differentiated (fig. 117 C). It consists of a mouth (_m_) and ciliated oesophagus probably derived from the stomodæum, a stomach and intestine derived from the true hypoblast, and an hepatic organ consisting of two separate lobes opening into the stomach. The anus (_an_) appears not far behind the mouth, and between the two is a very slightly developed rudiment of the foot (_b_). The anterior adductor muscle (_cm_) appears at this stage, though the posterior is not yet differentiated.

The larva is now ready to be hatched, but the further stages of its development were not followed.

Ostrea. The larvæ of Ostrea, figured by Salensky (No. 293), shew a close resemblance to those of Cardium. The velum is however a simple ring of cilia without a central flagellum. The proctodæum would appear to be formed later than the stomodæum, and the earliest stage figured is too far advanced to throw light on the position of the blastopore.

Pisidium. The development of Pisidium has been investigated by Lankester (No. 239). The ovum is invested by a vitelline membrane and undergoes development in a brood-pouch at the base of the inner gill lamella.

The segmentation commences by a division into four equal spheres, each of which, as in so many other Mollusca, then gives rise by budding to a small sphere. The later stages of segmentation have not been followed in detail, but the result of segmentation is a blastosphere. An invagination, presumably at the lower pole, now takes place, and gives rise to an archenteric sack.

[FIG. 119. THREE VIEWS OF AN EMBRYO OF PISIDIUM IMMEDIATELY AFTER
THE CLOSURE OF THE BLASTOPORE. (After Lankester.)

A. View from the surface.
B. Optical section through the median plane.
C. Optical section through a plane a little below the surface.
_ep._ epiblast; _me._ mesoblast; _hy._ hypoblast; _p._ cells
apparently budding from the hypoblast to form mesoblastic elements.]

The embryo now rapidly grows in size. The blastopore becomes closed and the archenteric sack forms a small mass attached at one point to the walls of the embryonic vesicle (fig. 119, _hy_). In the space between the walls of the archenteron and those of the embryonic vesicle stellate mesoblast cells make their appearance, derived in the main from the epiblast, though probably in part also from the hypoblastic vesicle (_vide_ fig. 119 C, _p_). The cavity between the hypoblast and epiblast, which contains these cells, is the body cavity. Fig. 119 represents three views of the embryo at this stage. A is a surface view shewing the epiblast; B is an optical section through the median plane shewing the hypoblast and some of the mesoblast cells; and C is an optical section shewing the mesoblast cells. A prominence on one side of the embryo now develops which forms the commencement of the foot, and the archenteric sack grows out at its free extremity into two lobes, but remains attached to the epiblast by an imperforate pedicle. The next organ to appear is the stomodæum. It arises as a ciliated epiblastic ingrowth which meets the free end of the archenteric sack, fuses with it, and shortly afterwards opens into it (fig. 118, _ph_). Between the mouth and the attachment of the enteric pedicle is placed the foot (_f_), which becomes ciliated. On the dorsal side of the enteric pedicle there appears a saddle-shaped patch of epiblast cells bounding the sides of a groove (_shs_). This is the rudiment of the shell-gland.

The enteric pedicle, or intestine as it may now be called, soon acquires a lumen, though still imperforate at its termination where the anus is eventually formed. Ventral to the intestine is placed a mass of cells--the rudiment of the organ of Bojanus. It is stated to be developed as an ingrowth of the epiblast.

In a slightly later stage the shell-gland rapidly increases in size and flattens out, and on the two sides of it there appear the rudiments of the two valves, which are at first quite distinct, and separated by a considerable interval (fig. 120). Before the appearance of the valves of the shell, the mantle folds have already grown out from the sides of the body.

[FIG. 120. DIAGRAMMATIC VIEW OF ADVANCED LARVA OF PISIDIUM. (Copied
from Lankester.)

_m._ mouth; _a._ anus; B. organ of Bojanus; _mn._ mantle; _f._
foot.]

At a somewhat later stage the gills appear as a linear series of small independent buds within the folds of the mantle behind the foot (fig. 120, _br_). The anterior adductor also becomes differentiated.

The alimentary tract in the meantime has undergone considerable changes. The primitive lateral lobes dilate enormously and become ciliated. At a still later stage their walls undergo peculiar changes, the nature of which is somewhat obscure, but they appear to me to be of the same character as those in many Pteropods and Gasteropods, where the cells of the hepatic diverticula, to which the lobes of Pisidium apparently correspond, become filled with an albuminous material.

The later stages in Pisidium have not been followed.

It is remarkable that in Pisidium a veliger stage does not occur. This is probably due to the development taking place within the brood-pouch. The late development of the otocysts is also remarkable. A byssus-gland was not formed up to the stage observed. In Cyclas calyculata (Schmidt), a byssus-gland also appears to be absent.

Cyclas. The development of Cyclas as described by Von Jhering is very unlike that of Pisidium, and the differences would seem to be too great to be accounted for except by errors of observation.

The segmentation of Cyclas is similar to that of Anodon (_vide_ p. 82), and a mass of large cells enclosed by the smaller cells gives rise to the hypoblast. In the interior of this mass there appears a lumen, and a process from it grows towards and meets the epiblast, and gives rise to the oesophagus and mouth,--a mode of development of these parts without parallel amongst Mollusca. A very rudimentary velum would appear, according to Leydig (No. 290), to be developed at the cephalic extremity. A shell-gland is formed of the same character as in Gasteropods. According to Leydig the shell appears as a single saddle-like structure on the dorsal surface; the lateral parts of this become calcified, and give rise to the two valves, but are united in the middle by the membranous median portion. At the two sides of the body the mantle lobes are formed, as in Pisidium.

Very shortly after the formation of the shell the byssus-gland appears as a pair of small follicles in the hinder part of the foot. It rapidly grows larger and becomes a paired pyriform gland, in which are secreted the byssus threads which serve to attach all the embryos at a common point to the walls of the brood-pouch.

The foot is large, and ciliated anteriorly. Otolithic sacks and peda ganglia are developed in it very early.

Unio. The ovum of Anodonta and Unio is enveloped in a vitelline membrane, the surface of which is raised into a projecting trumpet-like tube perforated at its extremity (fig. 12). This structure is the micropyle. The micropyle disappears in Anodonta piscinalis when the egg is ripe, but in Unio persists during the whole development. The ova are transported, in a manner not certainly made out, into the space between the two limbs of the outer gills of the mother, and there undergo their early development. The animal or upper pole of the egg is placed at the pole opposite to the micropyle.

The segmentation is unequal (__vide__ p. 100) and results in the formation of a blastosphere with a large segmentation cavity. The greater part of the circumference of the egg is formed of small uniform spheres, but the lower (with reference to the segmentation) pole is taken up by a single large cell. The small spheres become the epiblast, and the large cell gives rise to hypoblast and mesoblast[108].

[108] The account of the remainder of the development till the
larva becomes hatched is taken from Rabl, No. 292.

The single large cell next divides into two, and then four, and finally into about ten to fifteen cells. These cells form an especial area of more granular cells than the other cells of the blastosphere. Most of them are nearly of the same size, but two of them (according to Rabl), in contact with each other, but placed on the future right and left sides of the embryo, are considerably larger than the remainder. These two cells soon pass into the cavity of the blastosphere, while at the same time the area of granular cells becomes flattened out, and then becomes involuted as a small sack with a transversely elongated opening, which does not nearly fill up the cavity of the blastosphere. This involuted sack is the archenteron.

The two large cells, which lie in immediate contact with what, following Rabl, I shall call the anterior lip of the blastopore, next bud off small cells, which first form a layer covering the walls of the archenteron, but subsequently develop into a network filling up the whole cavity of the primitive blastosphere. The space between these cells is the primitive body cavity. For a long time the two primitive mesoblast cells retain their preponderating size[109]. At the hinder end of the body, and at the end opposite therefore to the two mesoblast cells, are placed three especially large epiblast cells.

[109] In this description I follow Rabl's nomenclature. According
to his statements the ventral part of the body is the original
animal pole--the dorsal the lower pole; the anterior end the
mesoblastic side of the opening of invagination.

In Anodonta and Unio tumidus there appears at this period a patch of long cilia at the anterior end of the body. These cilia cause a rotation of the embryo and would appear to be the velum. In Unio pictorum they do not appear till much later.

Immediately following this stage the changes in the embryo take place with great rapidity. In the first place a special mass of mesoblast cells appears at the hinder end of the archenteric sack; and becoming elongated transversely gives rise to the single adductor muscle. On the subsequent formation of the shell the muscle becomes inserted in its two valves. The blastopore next becomes closed, and the small archenteron grows forwards till it meets the epiblast anteriorly, and at the same time detaches itself from the epiblast in the region where the blastopore was placed. Where it comes in contact with the wall of the body in front a small epiblastic invagination arises, which meets and opens into the archenteric sack and forms the permanent mouth.

While these changes have been taking place the shell is formed as a continuous saddle-shaped plate on the dorsal surface. From this plate the two valves are subsequently differentiated. On the dorsal surface they meet with a straight hinge-line. Each valve is at first rounded, but subsequently becomes triangular with the hinge-line as base. The valves are not quite equi-sided, but the anterior side is less convex than the posterior. At a later period a beak-shaped organ is formed at the apex of each valve in the same manner as the remainder of the shell. This organ is placed at about a right angle with the main portion of the valve. It is pointed at its extremity and bears numerous sharp spines on its outer side, which are especially large in the median line (_vide_ fig. 121 A). It is employed in fixing the larva, after it is hatched, on to the fish on which it is for some time parasitic. The shell is perforated by numerous pores.

After the shell has become formed a new structure makes its appearance which is known as the byssus-gland. It is developed as an invagination of the epiblast at the hinder end of the body: Rabl was unable to determine whether it was formed from the three large epiblastic cells present there or no. It subsequently forms an elongated gland with three coils or so round the adductor muscle on the left side of the body, but opening in the median ventral line. It secretes an elongated cord by which the larva becomes suspended after hatching.

For some time the ventral portion of the body projects behind the ends of the valves of the shell, but before these are completely formed a median invagination of the body wall takes place, which obliterates to a large extent the body cavity, and gives rise to two great lateral lobes, one for each valve. These lobes are the mantle lobes.

Before the mantle lobes are fully formed peculiar sense organs, usually four in number, make their appearance on each lobe. Each of them consists of a columnar cell, bearing at its free end a cuticle from which numerous fine bristles proceed. Covering the cell and the parts adjoining it is a delicate membrane perforated for the passage of the bristles. The largest and first formed of these organs is placed near the anterior and dorsal part of the mantle. The three others are placed near the free end of the mantle (_vide_ fig. 121 A). These organs probably have the function of enabling the larva to detect the passage of a fish in its vicinity, and to assist it therefore in attaching itself. When the embryo is nearly ripe there appears immediately ventral to and behind the velum a shallow pit on each side of the middle line, and the two pits appear to be connected by a median transverse bridge. These structures have been the cause of great perplexity to different investigators, and their meaning is not yet clear. According to Rabl the median structure is the somewhat bilobed archenteron, and according to his view it is not really connected with the laterally placed pits. The cilia of the velum overlie these latter structures and make them appear as if their edges were ciliated. They are regarded by Rabl as the rudiments of the nervous system.

With the development of the shell, the mantle, and the sense organs, the young mussel reaches its full larval development, and is now known as a Glochidium (fig. 121 A).

If the parent, with Glochidia in its gills, is placed in a tank with fish, it very soon (as I have found from numerous experiments) ejects the larvæ from its gills, and as soon as this occurs the larvæ become free from the egg membrane, attach themselves by the byssus-cord, and when suspended in this position continually close and open their shells by the contraction of the adductor muscle. If the mussels are not placed in a tank with fish the larvæ may remain for a long time in the gills.

[FIG. 121.

A. GLOCHIDIUM IMMEDIATELY AFTER IT IS HATCHED.
_ad._ adductor; _sh._ shell; _by._ byssus cord; _s._ sense organs.
B. GLOCHIDIUM AFTER IT HAS BEEN ON THE FISH FOR SOME WEEKS.
_br._ branchiæ; _au.v._ auditory sack; _f._ foot; _a.ad._ and
_p.ad._ anterior and posterior adductors; _al._ mesenteron; _mt._
mantle.]

Before passing on to state what is known with reference to the larval metamorphosis, it may be well to call attention to certain, and to my mind not inconsiderable, difficulties in the way of accepting in all particulars Rabl's account of the development.

In all Gasteropod Molluscs the lower or vegetative pole of the ovum is ventral, not dorsal as Rabl would make it in Unio. The blastopore in other Molluscs always coincides either with the mouth or anus, or extends between the two. The surface on which the foot is formed is the ventral surface. On the dorsal surface are placed, (1) the velum near the mouth, (2) the shell-gland near the anus. In Anodon the velum is placed just dorsal to the mouth, then according to Rabl follows the blastopore, and in the region of the blastopore is formed the shell. The blastopore is therefore dorsal in position. It occupies in fact the ordinary place of the shell-gland, and looks very much like this organ (which is not otherwise present in Anodon and Unio). Without necessarily considering Rabl's interpretations false, I think that the above difficulties should have been at any rate discussed in his paper. More especially is this the case when there is no doubt that Rabl has made in his paper on Lymnæus a confusion between the mouth and the shell-gland.

Investigations on the post-embryonic metamorphosis of Glochidium have been made by Braun (No. 287), and several years ago I made a series of observations on this subject, the results of which agree in most points with those of Braun. I was however unsuccessful in carrying on my observations till the young mussel left its host.

The free Glochidia very soon attach themselves to the gills, fins, or other parts of fish which are placed in the tank containing them; after attachment they become covered by a growth of the epidermic cells of their host, and undergo their metamorphosis.

The first change that takes place is the disappearance of the byssus and the byssus organ. This occurs very soon; shortly afterwards all traces of the velum and sense organs also become lost.

At the time of the disappearance of these bodies, at the point of the projection from which the byssus cord arose, and very possibly from this very projection, the foot arises as a rounded process which rapidly grows and soon becomes ciliated (fig. 121 B, _f_).

The single adductor muscle begins to atrophy very early, but before its entire disappearance rudiments are formed at the two ends of the body, which at a later period can be distinctly recognised as the anterior and posterior adductor muscles (fig. 121 B, _a.ad_ and _p.ad_).

After the formation of these parts the gills arise as solid and at first somewhat knobbed papillæ covered with a ciliated epidermis, on each side of, but somewhat in front of (!) the foot (fig. 121 B, _br_). In the foot there soon appear the auditory sacks (_au.v_), and the foot itself becomes a long tongue-like ciliated organ projecting backwards[110].

[110] The position of the foot and gills in the larva represented
in Fig. 119 B would be more normal if the convex and not the
flatter side of the shell were the anterior. I have followed Rabl
and Flemming in the determinations of the anterior and posterior
end of the embryo, but failed to rear my larvæ up to a stage at
which the presence of the heart or some other organ would
definitely confirm their interpretation. I originally adopted
myself the other view, and in case they are mistaken, the
so-called velum would be a circumanal patch of cilia, while the
position of the primitive mesoblast cells as well as of the
byssus would better suit my view than that adopted in the text on
the authority of the above observers.

The mantle lobes undergo great changes, and indeed by Braun the mantle lobes are stated to be formed almost entirely _de novo_. The permanent shell is (Braun) formed on the dorsal surface of the still parasitic larva in the form of two small independent plates. I have not followed the changes of the alimentary canal, etc., but at an early stage there is visible, dorsal to the foot, a simple enteric sack.

By the time the larva leaves its host all the organs of the adult, except the generative organs, have become established.

The post-embryonic development of the organs of Glochidium is similar in the main to that of other Lamellibranchiata. This fact is of some importance on account of the peculiarities of the earlier developmental stages.

The byssus organ, the toothed processes of the shell, and the sense organs of the Glochidium can hardly be ancestral rudiments, but must be organs which have been specially developed for the peculiar mode of life of the Glochidium. Whether the single muscle is to be counted amongst such provisional organs is perhaps a more doubtful point, but I am inclined to think that it ought to be so.

If however the single muscle is an ancestral organ, it is important to observe that it entirely disappears as development goes on and the two adductor muscles in the adult are developed independently of it.

_General review of the characters of the Molluscan larvæ._

The typical larva of a Mollusc, as has been more especially pointed out by Lankester, is essentially similar to the larva of a number of invertebrate types, and especially the Chætopoda, with the addition of certain special organs characteristic of the Mollusca.

It has a bent alimentary tract, with a mouth on the ventral surface and a terminal or ventral anus. The alimentary tract is divided into three regions: oesophagus, stomach, and intestine. There is a variously developed præoral lobe with a ring of cilia--the velum, and a peri-anal lobe, often with a patch of cilia (Paludina, etc.). In all these characters it is essentially similar to a Chætopod larva. The two characteristic molluscan organs are (1) a foot between the mouth and anus, and (2) an invagination of the epiblast on the dorsal side at the hinder end of the body, which is connected with the formation of the shell.

The larvæ of most Gasteropoda, Pteropoda, and Lamellibranchiata present no features which call for special remark; but the larvæ of the Gymnosomata amongst the Pteropoda, and of the Scaphopoda, Polyplacophora and Cephalopoda present interesting peculiarities.

The larvæ of the Gymnosomata are peculiar in the presence of three transverse ciliated rings, _situated behind the velum_ (Fig. 109). These rings might be regarded as indications of a rudimentary segmentation; but, as already indicated, this view is not satisfactory. There is every reason for thinking that these rings have been specially acquired by these larvæ.

At first sight the larvæ of the Gymnosomata might be supposed to resemble those of the Scaphopoda, which are also provided with transverse ciliated rings; but, as shewn above, the rings of the Scaphopoda are merely parts of the extended velar ring.

Thus, the ciliated rings of the two larvæ--so similar in appearance--are in reality structures of entirely different values, being in the one case parts of the velum, and in the other special developments of cilia behind the velum.

The great peculiarity of the early larva of the Scaphopoda is the enormous development of the præoral lobe, which gives room for the development of the ciliated rings. In the presence of a central tuft of cilia, at the anterior extremity, the larva of the Scaphopoda resembles that of the Lamellibranchiata, etc.

The larva of the Polyplacophora resembles that of Lamellibranchiata in its anterior flagellum, and that of the Scaphopoda in the large development of the præoral lobe; but is quite peculiar amongst Mollusca in the transverse segmentation of the mantle area.

The embryo of the Cephalopoda agrees very closely with that of normal Odontophora in the formation of the mantle and (?) of the shell-gland, but is quite exceptional (1) in the almost invariable presence of a more or less developed external yolk-sack, (2) in the absence of a velum, (3) in the absence of a median foot, and in the presence of the arms.

The presence of a yolk-sack may most conveniently be spoken of in connection with the foot, and we may therefore pass on to the question of the velum.

The velum is one of the most characteristic embryonic appendages of the Mollusca, and its absence in the Cephalopoda is certainly very striking. By some investigators the arms have been regarded as representing the velum, but considering that they are primitively placed on the posterior and ventral side of the mouth, and that the velum is essentially an organ on the dorsal side of the mouth, this view cannot, in my opinion, be maintained with any plausibility.

Various views have been put forward with reference to the Cephalopod foot. Huxley's view, which is the one most generally adopted, is given in the following quotation[111].

[111] _The Anatomy of Invertebrated Animals_, p. 519.

"But that which particularly distinguishes the Cephalopoda is the form and disposition of the foot. The margins of this organ are, in fact, produced into eight or more processes termed arms, or _brachia_; and its anterolateral portions have grown over and united in front of the mouth, which thus comes, apparently, to be placed in the centre of the pedal disk. Moreover, two muscular lobes which correspond with the epipodia of the Pteropods and Branchiogasteropods, developed from the sides of the foot, unite posteriorly, and, folding over, give rise to a more or less completely tubular organ--the funnel or _infundibulum_."

Grenacher, from his observations on the development of Cephalopoda, argues strongly against this view, and maintains that no median structure comparable with the foot is present in this group: and that the arms cannot be regarded as taking the place of the foot, but are more probably representatives of the velum.