In a recent paper in the ‘Arbeiten ans dem zoologischen Instituten der Universität Wien,’ Sassi (16) has given a full account of the anatomy of Anomia ephippium, and has elucidated several hitherto obscure points relating to the kidneys and the reno-pericardial orifices in that species. Before Sassi’s paper had come into my hands I had made some observations on the allied genus Ænigma, a tropical member of the Anomiacea, the principal outcome of which has been to confirm Sassi’s results. At the same time I have found certain differences between the genera Anomia and Ænigma which are of sufficient importance to make it worth while publishing a short account of the latter genus.

With Plates 15—17, and 2 Text-figures.

I am indebted to Mr. R. Shelford, of Emmanuel College, Cambridge, and of the Hope Department of Zoology at Oxford, for the specimens of Ænigma ænigmatica, Chemnitz, which form the subject of this memoir. They were found living under conditions described by previous authors, on the roots and branches of Nipa (a palm of the family Phytelephantinæ) at Sarawak, and Mr. Shelford tells me that the specimens he collected were uncovered not only at low tides, but also at high neap tides, so that they remained exposed for days together to the rays of a tropical sun, but nevertheless always remained moist and fresh. The power of resisting desiccation for so long a time is unusual among lamellibranchiate molluscs, and indicates some structural adaptations enabling the animal to survive in such a habitat. I shall show in the course of this paper that the thickened mantle lobes and the presence of certain irregular passages and channels connected with the mantle cavity may be regarded as modifications for the retention of a sufficient supply of moisture, but in the essential features of its anatomy Ænigma is very similar to Anomia ephippium.

Ænigma, like other Anomiacea, is firmly fixed by a stout byssus passing through a large sinus or circular perforation in the right valve of the shell. The specimens given me by Mr. Shelford have been carefully preserved for histological purposes, and the right valve has in every case been wholly or partially removed. I am therefore unable to give au exact account of it, but from the fragments remaining in several specimens, I can confirm the descriptions given in conchological works, namely, that the upper’ extremities of the sinus through which the byssus passes curve round above the byssus and meet one another above the ligament, but do not fuse together as in adult specimens of Anomia ephippium.

The left valve varies very much in size and shape, and its form seems to be moulded to some extent upon the substratum to which the animal is attached. It is generally more or less elongate-oval in shape, of a dark purple colour, nacreous and iridescent internally. It is very thin and more or less translucent, especially when wet, this latter character being apparently of some importance to the economy of the animal, as will appear later on.

There are certain points of difference between the left valve of Ænigma and that of Anomia which require some little explanation. In the latter genus the left valve is more or less symmetrical ; the umbo is close to the dorsal margin and is more or less median in position and the ligament is immediately within and below the umbo. In Ænigma, as is clearly shown in the text-figure 1, the left valve is asymmetrical ; the umbo is prominent and obliquely curved, so that it points toward the anterior and upper margin, and it is situated at a distance of some 3—5 millim. from that margin, but is connected with it by a narrow slit or fissure. Internally this fissure extends as far as the ligament, and is more or less at right angles to it. The ligament, which is 3—4 millim. long in large-sized specimens, is situated obliquely above the umbonal fossa, and a line, A B, drawn at right angles to it indicates the original dorso-ventral axis of the shell. It is obvious that an inequality of growth, clearly indicated by the growth lines on the outer surface of the valve, has produced a secondary symmetry, and that the apparent dorso-ventral axis has been rotated through an angle of 40° in a postero-anterior direction. Sassi has shown that in Anomia ephippium the animal has undergone a similar rotation to an extent of 90°, that is to say through an angle twice as great as that in Ænigma, but there is also this difference between the two genera ; that whereas in Anomia the rotation has affected the position of the animal with respect to its shell, the latter retaining its symmetry, in Ænigma the inequality of growth has affected the shell as well as the contained animal, and both are asymmetrical to the same degree, as may be seen by a comparison of pl. 15, fig. 1, with the text-figure. Notwithstanding the change of symmetry, it will be convenient to describe the upper and nearly straight margin of the shell as dorsal, the lower convex margin as ventral, and the two ends as anterior and posterior respectively.

The impressions of the posterior adductor muscle, the anterior and posterior’ extractor muscles of the foot (here serving as retractors of the byssus) and of the large byssus muscle are shown in the text-figure and do not require special description. But it should be noticed that there is a fifth muscular impression, pl.m., which has been overlooked in previous descriptions of this genus, but Woodward (17) describes and figures a similar muscle under the name of the branchio-pallial muscle in Placuna placenta. This impression marks the point of attachment of a specialised band of palliai muscles running forward and downward in the left mantle lobe.

Fig. 1, pl. 15, is a representation of the animal lying in the left valve of the shell. The right mantle lobe has been largely cut away, and the left mantle lobe is obviously contracted by the action of preservative fluids; otherwise the animal is shown in its natural position. As has been explained above, Ænigma has not undergone so great a degree of rotation as Anomia, and retains its original symmetry with regard to the shell. We find accordingly that palliai cavity or sinus lying dorsad of the (posterior) adductor muscle, and containing the posterior recurved free ends of the gills, is smaller than in the latter genus, and the dorsal sutural union of the mantle edges is considerably longer, extending from the posterior edge of the ligament to the ventricle between the points marked x, x; in fig. 1.

Before proceeding to a description of the different organs it is necessary to point out that, in addition to the postero-anterior rotation described above, the anterior half of the body of Ænigma (and of all other Anomiacea) has been twisted round from left to right in connection with the peculiar development of the byssus and its retractor muscles. The nature and effect of this torsion has already been described by de Lacaze Duthiers (8) and Sassi (16) ; but a repetition is not out of place, if only to save the reader the trouble of reference to their papers. In a normal Lamelli-branch the byssus cavity and groove are situated on the posterior margin of the foot, and the retractor muscles of the byssus are paired and symmetrical, passing dorsally to their insertions on the right and left valves on either side of the hinge. In the Anomiacea the byssus, instead of passing out between the valves, passes through the well-known hole or sinus in the right valve, and drags the posterior margin of the foot over to the right. The retractor muscle of the byssus, instead of being paired and attached near the hinge line of both valves, is single and is attached near the centre of the left valve. As a consequence of these displacements, the whole of the anterior half of the body is twisted over to the right in such a way that the paired organs of the right side come to lie above the byssus muscle, and those of the left side below it; the visceral nerve commissures and the kidneys being specially affected. When this torsion of the foot and the lower part of the visceral mass is kept in mind, much that is puzzling in the anatomy of the adult is made clear.

The Mantle.—The edges of the mantle are free for nearly the whole of their extent, and are only united dorsad of the visceral mass between the points marked x, x in fig. 1. The left mantle lobe is entire, and covers in the whole left side of the animal, its continuity being broken only by the surfaces of attachment of the adductor muscle, the retractor muscles of the foot and byssus, and the special slip of palliai muscle referred to above. The edges of both mantle lobes are thickened and bear small palliai tentacles covered by a columnar epithelium. There are also a few much larger tentacles on the posterior lower margin of the right mantle lobe. There are no marginal palliai eyes or pigment spots, but at the bottom of the groove formed in the thickened palliai margin a track of columnar cells, supplied by twigs of the circumpallial nerve, can be distinguished. In the region of the visceral mass, that is to say in its upper half, the left palliai lobe is very thin ; but in its lower and posterior half, where it covers the large recurved gills, it is greatly thickened by the abundant development of lacunar tissue. The slip of palliai muscle, attached to the left valve of the shell, is shown at pl.m. in fig. 2. It coincides in position with the attachment of the axis of the left gill to the mantle lobe, and must function as a retractor of the gill. The left gonad extends backward to the level of the anus in the left palliai lobe, and in the same lobe there is a large anterior extension of the left gonad, forming a conspicuous sausage-shaped swelling in front of the mouth (figs. 1 and 2, go.a.).

The most remarkable feature in the left palliai lobe is the presence of a number of deeply pigmented spots, arranged in an irregular oval at a considerable distance from the palliai margin. These eye-spots, as I must call them, vary in number in different individuals. In the specimen shown in fig. 2 there are twenty-three. They are, however, very constant in position: the most anterior eye-spot is always situated just behind the surface of attachment of the anterior retractor pedis muscle, and the most posterior close to the attachment of the palliai muscle. Each eye-spot consists of a ring of black pigment surrounding a central opaque white area. The histology of these organs will be described in the latter part of this paper ; but I may say here that their minute structure leaves little doubt that they are sensory in function and adapted for visual functions. As they are situated at a greater or less distance from the edge of the mantle, they must always be covered, by the shell, and the existence of visual organs in such a position is somewhat extraordinary. But, as I have pointed out, the left valve of the shell is thin and translucent enough to allow a considerable amount of light to pass through. As Ænigma spends a large part of its existence uncovered by the sea, with its valves tightly closed to prevent evaporation, it is probable that these eyes are efficient in informing the animal of the duration of day-light, or, at any rate, of the incidence of direct sunlight. It is probable enough that after sunset the valves of the shell are slightly opened to admit of the aeration of the water contained in the palliai chamber, and are kept tightly closed to prevent evaporation during the heat of the tropical day.

The right mantle lobe is very irregular in shape, and presents a large anterior sinus corresponding with the sinus of the right valve of the shell, and serving for the passage of the byssus. In the anterior part of the body the right mantle lobe is attached by a very narrow band of tissue to the visceral mass, the line of attachment running above and nearly parallel to the upper edge of the byssus cavity. In this part of the body, indeed, the viscera are thrust over to the left side of the body and the visceral mass is adherent to the left palliai lobe. But in the hinder part of the body the rectum and cæcum of the crystalline style, passing respectively above and below the adductor muscle, cross over from the left to the right side, and are here adherent to the right mantle lobe and embedded in the mass of the right gonad and right kidney (compare figs. 11 and 13). The lower and posterior part of the right mantle lobe is exceedingly thick, and its inner surface is pitted and folded in a very irregular manner. As the animal lies with its right valve lowermost these folds and pits must serve for the retention of water during the long periods in which it is uncovered by the tides.

The mouth, as is shown in fig. 1, lies asymmetrically on the right side. It is concealed in a labial groove formed by two deep folds of the integument, which are nothing else than the greatly enlarged and modified labial palps. The extent and relations of these labial folds in Anomia were first described by de Lacaze Luthiers (8) and more fully by Sassi (16) ; they have very much the same relations in Ænigma as in that genus. The external labial fold passes round in front of the mouth forming a hood and the internal fold passes behind the mouth, the two enclosing between them a groove of varying depth lined by a high columnar ciliated epithelium, which contains numerous gland-cells in the region of the mouth. On the left side of the body the two folds, as they pass backward from the mouth, become very deep and prominent, and enclose between them a deep groove or gutter whose walls are thrown into numerous vertical folds covered by a ciliated epithelium. At a short distance behind the mouth the folds hang far down in the mantle cavity on the left side of the foot and are suspended from the visceral mass above by a thin sheet of tissue. Towards the posterior end of the foot the groove becomes shallower, and its walls are less folded, and eventually the external labial fold unites with both the direct and the reflected lamella of the left external demibranch, and the internal labial fold with the direct lamella of the left internal demibranch. Thus the left labial groove becomes continuous with the inter-branchial chamber of the left side (see text-figure 2, A and B). On the right side the two labial folds run back above the byssus cavity, parallel with and below the line of attachment of the right mantle lobe to the visceral mass. Dorsad of the byssus the labial folds are inconspicuous, the labial groove contained between them is shallow, and the epithelium lining is ciliated and glandular, but not ridged. Towards the posterior end of the byssus the groove turns downward and backward, and the labial folds enclosing it increase rapidly in vertical depth. At the same time the walls of the now very deep groove are thrown into numerous vertical ridges, and the epithelium of the ridges is richly ciliated but not glandular. The outer right labial fold, continuing to increase in vertical depth, eventually passes into the upturned reflected lamella of the right external demibranch, and the internal right labial fold becomes continuous with the thin membrane by which the direct lamella of the right internal demibranch is attached to the body-wall below the byssus muscle. The right labial groove thus becomes continuous with the inter-branchial chamber of the right side, but it is continued backwards as a cul-de-sac for some distance beyond the point of union with the gill, and in sections the right branchia appears to be suspended from the lower wall of this cul-de-sac, as is shown in text-figure 2, E.

The foot (fig. 1, f.) is reduced to a flat muscular projection at the anterior angle of the byssus cavity. In most of the spirit-preserved specimens it is contracted to a small, lanceolate, muscular mass, but in the individual figured it is unusually long and ribbon-shaped. The extremity of the foot is always pointed, and bears a single small tentacle, similar in all respects to the marginal tentacles of the mantle. There is no infundibuliform cavity at the end of the foot as in Anomia ephippium, but the right (morphologically the ventral) surface is grooved and covered with numerous transverse, ciliated ridges. The mass of the foot is highly muscular and contains numerous mucous glands, and it is evident that this organ is very extensile. It seems probable that it can be protruded some distance beyond the shell, and that it is auxiliary to nutrition, minute particles being swept by ciliary action along the groove on its right surface, and thence to the right labial groove.

The principal muscles connected with the foot have not undergone as much modification in Ænigma as in Anomia owing to the lesser degree of rotation in the former genus. The two tapering muscular bands running respectively forwards and backwards from the great retractor muscle of the byssus to their surfaces of attachment on the left valve, are clearly the homologues of the anterior and posterior retractors of the foot of other Lamellibranchia (fig. 2, a.r.p. and p.r.p.). But here they have shared in the torsion of the other organs of the anterior part of the body, have become twisted round to the right, have lost their primitive connection with the right valve, and are inserted on the left valve only, serving rather as accessory retractors of the byssus than as retractors of the foot. It will be observed that both the anterior and posterior retractors of the foot arise from double origins above and below (morphologically right and left) of the retractor muscle of the byssus, but that each is formed into a single strand shortly before its attachment to the shell.

The retractor muscle of the byssus is a large, coarsely-fasciculated muscle, nearly circular in section, and passes transversely from its attachment to the left valve to the byssus cavity. It probably represents the right and left retractors of the byssus of symmetrical Lamellibranchia, but betrays no sign of its primitive paired origin. De Lacaze Duthiers (8), however, considers that the right retractor byssi muscle is aborted in Anemia, but it should be observed (fig. 10) that the muscle is equally well developed above, that is on the morphological right, and below, that is on the morphological left of the byssus cavity.

The byssus cavity is large and shallow, more or less oval or lozenge shaped, and bordered above and below by muscular lips, which are really the posterior continuations of the right and left margins of the foot. The bottom of the cavity is lined by a number of close-set, parallel folds or lamellae running fore and aft in the direction of the long axis of the shell. These lamellæ form the byssus gland, the histology of which will be described in the latter part of this paper. But I may state here that the byssus of Ænigma is not calcified like that of Anomia, but consists of a number of parallel plates of the byssus substance, secreted by the epithelial cells covering the ridges of the byssus gland. These lamellae are fused externally into a plate which is directly and firmly fixed to the substratum to which the animal is attached.

The Gills.—The natural position of the gills, as seen from the right side is clearly shown in fig. 1. They are relatively even larger than in Anomia, and, as in that genus, their posterior ends are recurved and form the posterior boundary of the dorsal or supra-branchial chamber into which the anus opens, and in which the ventricle of the heart lies. In their microscopic as well as their macroscopic structure the gills of Ænigma are singularly like those of Anomia ephippium as described by Ridewood (15), to whose paper the reader is referred for details. Thus the reflected filaments of the right and left internal demibranchs are fused together and the whole series of filaments are in organic continuity along the line of union, and are traversed by a blood-vessel. The upper ends of the reflected filaments of the right and left external demibranchs are secondarily reflected downwards to form the so-called velar fold or flap, and at the angle of reflection the whole series of filaments are united in organic continuity, and are traversed from end to end by a blood-vessel (text-figure 2, b.v.). The lower ends of the velar filaments are, however, independent, and have an arrangement of cilia which appears to have been overlooked by previous observers. As is shown in fig. 6 these velar filaments are sub-triangular in section, and the chitinous lining of their cavities is thickest on the internal or morphologically ventral side, this being the reverse of what obtains in the direct and reflected limbs of the filaments. The outer (morphologically dorsal) side of the velar filament is broad and flat, and is covered by a cubical epithelium bearing a number of short stiff cilia. The opposite face of the filament is narrowed and covered by longer columnar epithelial cells bearing long fine cilia, continuous with the frontal cilia of the reflected or ascending limb of the filament. There can be little doubt that the short stiff cilia borne on the palliai faces of the velar filaments have the function of ciliated discs, and give a sufficient amount of friction against the inner surface of the mantle to prevent the whole of the reflected lamella from slipping down. One might expect to find a corresponding ciliated ridge, or row of ciliated discs, on the mantle, but I can find no trace of them, and am certain that they do not exist. It should be noted that the velar flap on each side abuts on the thickened and corrugated lower moiety of the mantle lobe, so that ciliated discs on that side are unnecessary.

As in Anomia ephippium, there are no ciliated discs on either the direct (descending) or reflected (ascending) limbs of the filaments, but there is a well-developed single row of discs at the angle of reflection, the details of which are shown in fig. 5. There is, as in A. ephippium, a very low inter-filamenter septum close to the angle, but beyond this no other interlamellar junctions whatever. A transverse section of two filaments, showing the frontal and lateral cilia and other details, is given in fig. 7. The attachments of the gills to the body-wall and mantle are affected to a considerable extent by the asymmetry of the anterior half of the body and by the fact, mentioned above, that the visceral mass is adherent to the left mantle lobe in the anterior part of the body but to the right mantle lobe in the posterior part. As these relations have not been sufficiently fully described in Anomia, I will enter into them somewhat closely in Ænigma. The posterior recurved ends of the gills lie free in the mantle cavity and are not attached to the mantle. In the more anterior part of their courses the right and left branchiae are differently affected by the asymmetry due to the twisting of the byssus and foot over to the right. The relations of the left branchia are on the whole those of a normal symmetrical Lamellibranch. At a point nearly vertically below the anus its axial fold becomes attached by a deep suspensory fold to the left mantle lobe. In this fold run the muscular fibres which have already been described as a specialization of the palliai musculature forming a retractor muscle of the branchia (figs. 2 and 13). Passing forward the axial attachment of the left branchia becomes shifted more and more towards the middle line, largely in consequence of the interposition of the fibres of the posterior retractor pedis muscle between it and the mantle (fig. 12). In front of the adductor muscle the axis of the left branchia is suspended in the mantle cavity by a deep suspensory fold attached to the lower surface of the visceral mass just below the lower limb of the left kidney, and has lost all connection with the left mantle lobe. These relations are continued forward as far as the posterior edge of the foot, where the direct and reflected lamellæ of the outer demibranch become continuous with the external left labial fold and the direct lamella of the inner demibranch passes into the internal left labial fold as described above (p. 260).

The fold formed by the united upper ends of the reflected filaments of the right and left inner demibranchs acquires no attachment till it reaches the anterior end of the great extractor byssi muscle. Here it is pushed over to the left and unites with the body wall close alongside of the attachment of the direct lamella of the left inner demibranch, and its blood-vessel passes into the blood sinus lying below the recurrent limb of the left kidney, this sinus discharging its blood into the left auricle (text-figure 2, B and c, v).

As may be seen in fig. 1, the anterior end of the right outer demibranch curves round dorsally behind the byssus cavity. As explained above (p. 261) its reflected lamella becomes continuous with the right external labial fold, and its direct lamella with the right internal labial fold behind the byssus (see text-figure 2, D and E). The anterior filaments of both lamellæ of the right outer demibranch are very short, but those of the direct lamella of the right inner demibranch’ are very long and, as may be seen in fig. 1, d.b2, extend below and to the left of the byssus cavity as far forward as the foot. As is shown in text-figure 2, A to E, and in fig. 10, these elongated anterior filaments of the right inner demibranch are connected with the ventral surface of the body, below the byssus muscle, by a thin membrane, æ, which stretches across to the left side and eventually, as in A, becomes attached to the left inner’ labial fold. Thus, as a consequence of the displacement of organs several times referred to, the inner demibranch of the right side is carried across the body and attached to what is morphologically the left side of the foot.

The mouth leads into a narrow, and, for a Lamellibranch, relatively long, oesophagus, lined by a ciliated glandular epithelium, whose characters will be described in the latter part of this paper. The oesophagus passes into a capacious stomach occupying the greater part of the visceral mass dorsad of the foot. The roof of the stomach is thin, and lined by a few ciliated but non-glandular columnar epithelium, which extends down for some distance on the right wall, especially in the anterior half of the stomach. The right wall and floor, and in the posterior half of the stomach, the left wall are, on the contrary, lined by an epithelium consisting of very long attenuated ciliated cells, intermixed with which are numerous elongated claviform gland-cells filled with yellow granules. The floor of the stomach is also thrown into longitudinal folds, and is covered by a thick cuticular layer, the “flèche tricuspide” of Poli, which is apparently secreted by the yellow gland-cells. This question will be discussed more fully in the latter part of this paper. The stomach is embedded in the liver, which opens into it by several large ducts. One of these ducts is dorsal, and communicates with the superior lobe of the liver; the remainder are posterior and ventral, and some of them run far back in the posterior mass of the liver before breaking up into branches (figs. 9 and 10, li. d.). Posteriorly the stomach presents a dorsal cæcum, into which one of the largest of the posterior liver ducts opens. Ventrally it narrows in diameter, and gives off the intestine and sac of the crystalline style.

The intestine opens into the stomach by an aperture common to itself and the sac of the crystalline style. The entrance to the intestine is guarded by a number of prominent longitudinal ridges covered by ciliated epithelium. The intestine itself is a narrow tube running back for some distance close to and on the right of the sac of the crystalline style. In this part of its course its lumen is narrowed by the projection of four prominent ciliated ridges into its interior. Just anterior to the heart the intestine enlarges somewhat suddenly in diameter, the four internal longitudinal ridges disappear, and it turns sharply upwards, makes a single complete turn, bends up at a sharp angle towards the ventricle, and then runs a straight course below the ventricle to the anus. In the last section of its course there is a distinct typhlosole in the large intestine. The rectum is very short and funnel-shaped. Its epithelium differs entirely from that of the large intestine, consisting of clear, attenuated, ciliated cells with deeply-staining nuclei, indicating that it is a proctodæum. The histology of the alimentary tract will be described further on.

As in Anomia, the sac of the crystalline style is excessively long. At first nearly median in position (fig. 10), it passes over to the right and runs back, closely attached to the right mantle lobe, below the adductor muscle. In the posterior part of its course it lies parallel to and above the attachment of the right branchia, and finally it curves forward and ends blindly (figs. 10—13, cry.).

The ventricle of the heart, as in all Anomiacea, lies free in the dorsal bay of the mantle cavity, and is not enclosed in a pericardial sac. Situated dorsally to the intestine, it sits, so to speak, astride of the latter organ, the auricles passing down on either side of it like a rider’s legs. The walls of the ventricle are very thick and muscular, as also are the walls of the auricles, but the latter-not to.so great a degree as in Anomia. The aorta arises from the antero-ventral angle of the ventricle (its aperture is guarded by a valve), and, running towards the right dorsad of the rectum, it at once divides into three branches. The median branch penetrates the liver mass, and soon breaks up into branches and disappears. The left-hand branch constitutes the anterior aorta. This vessel runs forward to the right of and above the intestine, passes through the coil of the intestine to the left upper’ side of the visceral mass, and runs forward over the stomach to the oesophagus, where it divides into numerous branches. The right-hand branch of the aorta runs down on the right of the intestinal coil towards the cæcum of the Crystalline style, and then bends back to take a posterior course, supplying the gills and right mantle lobe with blood.

The course of the veins is similar to that described in Anomia ephippium by Sassi. The left auricle divides below the level of the intestine into two branches. Of these the posterior runs down to the outside of the large upper posterior lobe of the left kidney, and, turning back, may be traced as far as the visceral ganglia, where it receives several vessels from the gills and left mantle lobe. The anterior’ branch courses forward in close connection with the upper limb of the left kidney, and its outer wall is thickened by an abundant glandular tissue, which will be described in connection with the excretory organs. The right auricle dilates to form a large thin-walled sac lying to the right of and partially above the intestine. The walls of this sac are locally thickened by the same glandular tissue that accompanies the anterior branch of the left auricle, which, from its relations to the reno-pericardial funnels, must be regarded as the representative of the pericardial gland. The venous cavity may be traced back along the right side of the upper lobe of the right kidney, and in the posterior part of its course it receives numerous accessions from irregular sinuses and vessels bringing back blood from the right mantle lobe, the right branchia and the sac of the crystalline style.

The right and left auricles communicate with one another by a sinus passing ventrad of the intestine, immediately below the hinder end of the ventricle. A similar vessel occurs in Anomia ephippium, and was interpreted by Pelseneer (13) as a relic of the pericardium, a mistake which anybody might be excused for falling into, seeing how peculiar are the relations of the kidneys, venous channels, and remnants of the pericardium in that genus.

Sassi (16) is the first author who has given a correct and intelligible account of the kidneys in Anomia ephippium. Fig. 4, which is a reconstruction from one of my series of sections, shows that these organs are extremely similar in Ænigma. The chief difference between the two genera consists in the fact that, whereas in Anomia the left kidney forms a complete loop behind the byssus muscle, in Ænigma the upper and lower limbs are free from one another posteriorly. In the right kidney the upper limb extends much further back along the intestine in Ænigma than it does in Anomia. In all other respects I can fully confirm Sassi’s observations. In Ænigma, as in Anomia, the right and left kidney sacs communicate by a wide opening situated below the auricle (fig. 4, ap.). The left renal sac runs forward, dorsad of the byssus muscle, as far as the foot, and then tarns sharply hack, running just above the attachment of the axis of the left branchia as far as the visceral ganglion, where it turns upward and opens into the suprabranchial chamber’ somewhat in front of the right renal aperture. The left reno-pericardial canal is, as in Anomia, situated far forward near the anterior bend of the renal sac (l. rp. in fig. 4), and the left gonaduct opens into the renal sac by a distinct ciliated funnel situated close above and nearly opposite to the left reno-pericardial canal. The right renal sac is roughly crescentic in shape, the concavity of the crescent embracing the adductor muscle. The right reno-pericardial funnel (r. rp. in figs. 4 and 11) is situated immediately below the right auricle, and the right gonaduct opens into the right wall of the kidney sac nearly opposite to, but a little further back than, the reno-pericardial funnel. So far the only difference between the structure and position of these organs inÆnigm a, and those in Anomia, as described by Sassi, consists in the fact that whereas that author says there are no ciliated funnels to the gonaducts in the latter genus, there are very distinct gonaducal funnels, with a well differentiated ciliated columnar epithelium (fig. 15) in Ænigma.

In respect of the relations of the reno-pericardial funnels to the remnants of the pericardium, there is, however, a great difference between the two genera. Sassi has shown that in Anomia ephippium the reno-pericardial canals lead into a system of short and slightly branched tubules ending in glandular diverticula, and these he identifies, no doubt correctly, as the remnants of the pericardium. I have already shown that Ænigma is in several respects a less specialised form than Anomia, and in this particular matter of the reno-pericardial ducts and their connections it has clearly undergone less modification than the latter genus. Fig. 14 is a drawing of a section through the left reno-pericardial canal of Ænigma. The canal is lined by a columnar epithelium, each of whose constituent cells bears a long flagellum. The upper end of the canal is lined by a flatter, non-ciliated epithelium, and passes into a mass of reticulate glandular tissue containing many inter-cellular spaces. Though it is not particularly well shown in the section figured, the communication between these spaces and the reno-pericardial canal can be easily traced in the series of sections. Some few sections further back the canal appears to open, by an aperture so distinct that it might almost be described as a ciliated funnel, into a central lumen in the mass of glandular tissue, and this as it is traced backwards breaks up into a number of irregular branches or channels communicating with the spaces in question. From its relation to the reno-pericardial canal there can be no doubt that this lumen together with the irregular channels opening into it represents the pericardial cavity, nor can there be any doubt that the branching canals are homologous with the structures described by Sassi in Anomia ephippium. If the lumen represents the pericardial cavity, the glandular tissue clearly is the homologue of the pericardial gland, and the chief point of difference between Ænigma and Anomia is that in the former the pericardial gland is well developed, and retains not only its function as an excretory organ, but also its counection with the heart. As is shown in figs. 4 and 14, and in text-figure 2 A and B, l. pg., the left pericardial gland is of considerable vertical extent in the immediate neighbourhood of the left reno-pericardial, canal. Posteriorly it narrows to a thin band of glandular tissue, which may be traced backwards, lying to the outside of and above the upper limb of the left renal sac, and in close connection with the anterior of the two veins into which the left auricle divides, as far as the muscular wall of the left auricle itself. As it approaches the heart the pericardial gland spreads out over the walls of the auricle and becomes intimately fused with them, extending, as I shall show further on, into the thickened walls of the ventricle.

On the right side the right reno-pericardial canal (figs. 4 and 11, r. rp.) leads into the lumen of a similar but larger mass of glandular tissue, occupying the right-wall and floor of the dilatation of the right auricle described on p. 269. This dilatation of the blood-vessel has no apparent analogue in Anomia. The glandular sac forming part of the thickness of its walls is of course the right pericardial gland, and its histological structure as well as its relations to the upper posterior lobe of the right renal sac and to the right auricle are analogous to those of the left pericardial gland. The extent of the two glands and their relations to the kidneys and heart are indicated by the spaces enclosed between the black lines in fig. 4.

The structure and distribution of the pericardial glands of the Lamellibranchia has been worked out in great detail by Grobben (6). Without entering info the details of his careful and voluminous research on this subject, I may say briefly that he has shown that in a large number of Lamellibranchs belonging to different families the epithelial walls of the pericardium are glandular and have an excretory function. In many species, and among the Filibranchia in the Arcidæ, Mytilidæ, and Pectinidæ, the glandular tissue is localised on the auricles and in some other forms, e.g. Venus verrucosa (see Grobben, loc. cit., fig. 15), it extends to the ventricle of the heart. The characteristic histological elements of. these glands are oval or somewhat irregular cells with a spherical nucleus, alveolar protoplasm containing a few granules, and in the latter a distinct brown concretion. Though the shape, size, and appearance of these cells vary in the various species examined by Grobben, they are present in the pericardial glands of all, and are very distinct in character from the concentric concretions found in the kidney.1 The structure of the pericardial gland in Ænigma is shown in fig. 14. The tissues have doubtless undergone contraction, and are otherwise altered by the action of spirit, but it is clear enough that the bulk of the gland in the vicinity of the reno-pericardial canals is formed by a mass of branching cells, whose processes unite to form a reticulum. Or the structure might be otherwise described as a mass of vacuolated protoplasm containing numerous oval nuclei, smaller, and with a denser chromatic network than the nuclei of the kidney cells. This tissue may be traced along the upper margin of the kidney from the reno-pericardial funnel to the roots of the auricles on either side of the body, and on arriving at the thickened muscular walls of the auricles below the intestine it seems to thin out and disappear. In the spaces or vacuoles of this tissue are olive-brown concretions, which vary in size and appearance with the metabolic condition of the animal. In some of my series of sections the concretions are small and are contained in oval cells, as shown in fig. 16, a. These cells are almost identical with the pericardial cells of Venus verrucosa and Cardium edule figured by Grobben (loc. cit., figs. 53 and 54). In another of my series the concretions are much larger, and are either surrounded by a thin cell-envelope with the nucleus lying to one side, as in fig. 16, b, or the cell structure is no longer distinguishable. Similar conditions are figured by Grobben for divers Lamellibrauchs.

The presence of these highly characteristic cells and concretions not only enables us to identify the above-described tracts of tissue as pericardial glands, but also to trace the latter beyond their apparent limits. The pericardial glands appear to thin out on the auricular walls, and to stop short of the heart. But an examination of the walls of the heart with high powers of the microscope reveals the fact that they are penetrated by the glandular tissue. The two auricles, embracing between them the intestine, lie, like the ventricle, in the palliai cavity, and have relatively thick muscular walls, covered externally by a columnar epithelium continuous with the external epithelium of the body.

The muscle-fibres, both of the auricles and ventricle, cross one another in all directions, forming a sort of sponge-work with numerous spaces. A high power of the microscope shows that the tissue of the pericardial glands runs through these spaces in the muscular walls of the auricles, and extends into the ventricle. The characteristic cells containing olive-brown concretions can be distinguished even with a low power in the inter-muscular spaces of both auricles and ventricle (fig. 19). Furthermore, a careful examination of the columnar epithelium of the heart shows that the inner ends of its component cells do not rest on a basement membrane, but are prolonged internally into fine processes which run in between the muscular fibres. In other words, the external epithelium of the auricles and ventricle is fused to and partly immersed in the subjacent muscular tissue, much as the epidermal cells of many Platyhelmia are immersed in the subderinal muscular and connective tissue.

With these facts before us we have a ready explanation of the hitherto unsolved problem of the fate of the pericardium in the Anomiacea.

In a typical Filibranch with a well-developed pericardium we should find the following layers in transverse section :— 1. The external epithelium. 2. The outer epithelial wall of the pericardial cavity. 3. The inner glandular wall of the pericardial cavity adhering to the auricles or ventricle. 4. The muscular wall of the auricles or ventricle. It is clear that in Ænigma the pericardial cavity has disappeared in the vicinity of the heart, and that the external epithelium, the outer and inner walls of the pericardium, and the muscular wall of the heart itself have become more or less intimately fused together; in particular the glandular tissue of the inner pericardial wall has become incorporated with the muscular wall of the heart. None the less all these forms of tissue—epidermic, glandular, and cardiac muscular—can be recognised by careful microscopical examination. It is obvious that the disappearance of the pericardial cavity and the fusion of its walls with those of the heart is in some way correlated with the torsion of the body produced by the excessive development of the byssus muscle, and the attachment of the latter to the centre of the left valve of the shell. The result of this is, that the left kidney has been dragged forward to pass round the byssus muscle and the left reno-pericardial canal has been shifted forward far from its typical position, involving a great anterior extension of the left pericardial cavity. The corresponding structures on the right side have not suffered so much displacement, but the strain to which the whole system of organs must have been subjected during the passage from the larval to the adult condition is sufficient to account for the obliteration of the pericardial cavity, except in the immediate proximity of the reno-pericardial canals. Sassi has suggested that the whole of the venous vessel from the ventricle to the left reno-pericardial aperture is the representative of the left auricle, and if we regard all that vessel as auricle which is covered with pericardial glandular tissue, his suggestion is correct. But as the muscular coat of the afferent vessels of the heart does not extend as far in Ænigma as in. Anomia, and as the left afferent vessel divides close below the intestine into an anterior and a posterior vessel, I have preferred to restrict the term auricle to those muscular vessels which embrace the intestine.

he gonads

—The sexes are separate. The ovaries and testes occupy the same position in the two sexes, so a description of one will apply equally well to the other. The left gonad is much the larger of the two, and extends for a long distance in front of and behind the gonopore. Its anterior moiety has the same relations as in Anomia, that is to say, it runs forward below the stomach and liver, and just above and to the right of the left labial groove to the mouth, where it passes to the right of the attachment of the anterior retractor pedis muscle, and passing into the mantle forms the large præoral mass which is so conspicuous a feature when the animal is opened. The posterior moiety runs back as a fine canal, bearing a few slightly branched diverticula, below the cæcum of the crystalline style and above the byssus muscle; its course being here on the right rather than on the left side of the body. Behind the byssus muscle it passes over to the left side of the body, and divides into two branches running respectively along the upper and lower edges of the posterior retractor pedis muscle. Behind the attachment of this muscle the gonad passes into the left mantle lobe, and there forms a large follicular mass extending back for some distance beyond the level of the anus, as shown in figs. 2, 12, 13. The left gonad does not extend into the mantle in Anomia, and Pelseneer (14) has stated that it does not do so in any of the Anomiacea, a somewhat rash generalisation which must now be corrected. The right gonad in the anterior part of its course lies dorsad of the stomach and liver on the upper side of the visceral mass, where it forms a large lobe above and to the right of the intestinal loop. From this lobe two main branches pass backward. The one passes to the dorsal side of the cæcum of the crystalline style, and accompanies the cæcum for the whole of its course in the right mantle lobe, and forming a large mass above its extremity, as shown in fig. 13. The other branch, into which the gonaduct opens, maintains a more dorsal position, and, passing above the adductor muscle, forms a considerable follicular mass to the right of and above the posterior part of the intestine.

The gonaducts, as has been described above, open by distinct ciliated funnels into the renal sacs close to the reno-pericavdial apertures of the sides of the body to which they belong. A gonaducal funnel of a male is shown in fig. 15. There is no difference in the histological characters of the funnel in the two sexes. The gonaducts, that is to say those sections of the ovarian or testicular tubules that are lined by a low cubical instead of a germinal epithelium, are extremely short. The histology of the ovaries and testes does not call for special description, but in one of my series of sections the ovary was penetrated throughout by a number of green filaments, whose nature I could not satisfactorily determine. They have the appearance of filamentous algæ, and may possibly be symbiotic or parasitic within the mollusc. But they are not of constant occurrence, for I could find no trace of them in two other females examined.

The Nervous System

—As may be seen in fig. 3, the nervous system is of the usual lamellibranchiate type, and the modifications it has undergone are attributable only to the torsion which has affected these in common with all the other organs of the anterior moiety of the body. Thus the right cerebro-pleural ganglion lies above and somewhat behind the left. The cerebro-pedal connectives are relatively short, and the right pedal lies above the left pedal ganglion. The right visceral connective passes above, and the left visceral connective below, the byssus muscle and the right visceral ganglion is somewhat above and in advance of the left. But beyond this distortion which it shares with the other organs of this region, the nervous system presents few features requiring special description. The otoliths lie a short distance behind the pedal ganglia, and are connected with the latter by short nerves. The two stout nerves passing from the pedal ganglia, that from the right ganglion distributed to the upper and that from the left ganglion, distributed to the lower surface of the byssus muscle are worthy of mention, because de Lacaze Duthiers (8) describes a large nerve passing from the visceral ganglia to the byssus muscle in Anomia. No such nerve is present in Ænigma, the byssus muscle being wholly innervated from the pedal ganglia. The branchial and palliai nerves issuing from the right and left visceral ganglia should be noted. The branchial nerve of either side runs straight down into the suspensory fold of the gill axis, and there enters a distinct branchial ganglion (fig. 3, br. g.), from which fine nerves run forwards and backwards in the gill axis. The palliai nerve issues separately from the branchial nerve on the right side, but the two have a common origin from the left visceral ganglion. The branches of the two viscero-pallial nerves diverge in the mantle lobes and unite in each mantle lobe with a distinct circumpallial nerve, which forms a complete ring in the thickened margin of the mantle, and forms a connection anteriorly with nerves issuing from the cerebro-pleural ganglia.

Histology

—The specimens of Ænigma collected by Mr. Shelford were so well preserved that I have been able to work out the histology of some of the organs in some detail, but I do not propose here to do more than give an account of some of the more striking features that came under my notice. A detailed account of the histology of various members of the Lamellibranchia, including the Anomiacea, is indeed a desideratum, but for such a task fresh specimens are necessary, and where they are not available it is inexpedient to attempt more than a description of such characters as can be accurately studied in sections.

Palliai Organs

—The lobes of the mantle were so much contracted in spirit that I have been unable to make out the details of the histology of the marginal tentacles to my satisfaction. The tentacles are covered by a high columnar epithelium, and receive an abundant nerve supply from the circumpallial nerve. There are some specially large tentacles on the lower posterior edge of the right palliai lobe (fig. 13, t.), and near the bases of some of these there are ganglionic enlargements on the circumpallial nerve. The thickened edges of the mantle present the three reduplications commonly occurring in the Lamellibranchia, and in the bottom of the groove formed by the two outermost reduplications there is, in the hinder part of both mantle lobes, a tract of very definite columnar cells supplied with twigs from the circumpallial nerve. These, cells are probably sensory in function.

As may be seen in the sections, figs. 9—13, the lower moieties of the mantle lobes overlapping the gills are very much thickened, and their inner surfaces are thrown into a number of folds, which are specially prominent in the right lobe. These foldings are no doubt exaggerated by contraction in spirit, but there can be little doubt that they exist in the fresh state, and serve to retain water when the animal is uncovered for long periods by the tide. The thickened lobes of the mantle are nearly entirely composed of a lacunar tissue, traversed by very few branching muscle-fibres. The greater part of the musculature of the mantle lies immediately below the external layer of epithelium, and all within this is a spongy lacunar tissue, which must serve as a reservoir for the retention of fluid.

In this connection it may be observed that the asymmetrical arrangement of the organs, and in particular the narrow attachments of the mantle lobes to the visceral mass except in the regions of the attachments of the muscles to the valves of the shell, induces a certain laxity and independence of the organs, which can be better understood by an examination of figures 11—13 than by any description. As a consequence there are numerous irregular, spaces communicating with the suprabranchial or the pallial cavities, and some of them, such as those marked z in figs. 11 and 13, end in veritable culs-de-sac. The entrances to these culs-de-sac is guarded by two ridges, whose thickened borders are covered by a columnar’ epithelium furnished with stiff cilia. A transverse section of one of these ridges is shown in fig. 8. There can be little doubt that when the thickened borders of these ridges are in apposition the cilia interlock, and thus close the entrance to the cul-de-sac, which then serves as a reservoir for water. The largest of these reservoirs are situated between the hinder part of the right labial groove and the visceral mass (fig. 11, z.), between the right kidney and the right mantle lobe, below the level of the heart (text-figure 2, F, z.), and between the hinder part of the intestine and the right mantle lobe (fig. 13, z.). It is evident, then, that Ænigma is well provided against the danger of desiccation when exposed for days together to the rays of the sun.

The most remarkable of the palliai organs are the pigment-spots or eyes of the left palliai lobe, whose position has been described on p. 258. The structure of these organs is altogether peculiar in that two epidermic layers, namely, those of the outer and inner face of the mantle, share in their formation. As is shown in fig. 20, each eye comprises a cornea, a lens, a vitreous body, and a layer of deeply-pigmented cells forming a rudimentary retina. The cornea is formed by a local modification of the external epithelium of the mantle, which, elsewhere formed of columnar cells with an admixture of glandular cells, here becomes flatter; its component cells are transparent and vacuolated, and the cell outlines are scarcely distinguishable, but the nuclei are distinct and oval, with scattered chromatin granules. The lens (fig. 21) is a biconvex lenticular mass lying below the cornea, but separated from it by a thin sheet of tissue continuous with the vitreous body. The lens, in spirit specimens, is of an opaque white colour, and no definite structure can be discerned in it. It is a vacuolated mass of finely granular substance containing a few nuclei, and as the latter are identical with the nuclei of the cornea, it is probable that the lens is formed by a proliferation of the epithelial cells. The lens is embedded in a vitreous body consisting of a mass of finely granular, vacuolated protoplasm containing numerous nuclei which are smaller and stain more deeply than the nuclei of the lens or cornea. This vitreous body, in which no cell outlines are distinguishable, is evidently a local modification and concentration of the sub-epidermal tissue. The retinal layer, as is clearly shown in fig. 20, is formed by the epithelium lining the inner face of the mantle. Some of the eyes are borne on very thin parts of the mantle, and in these, as the vitreous body occupies the whole thickness of the mantle, the retina simply consists of a modification of the cells lining the inner face of the mantle. They become very large and columnar (fig. 21), and are thickly loaded with black pigment granules, but there are no chitinous rods or rhabdomes, such as are commonly found in retinal cells. On the other hand they bear a close resemblance to the retinal cells of the visual organs of the siphons of Mya arenaria, Solen vagina, and Dreissensia poly-morph a, figured by Sharp (17). In the case of those eyes situated on thickened portions of the mantle, the internal epithelium is deeply invaginated, as is shown in fig. 20, and the extremity of the invagination spreads out below the vitreous body to form a two-layered optic cup which in section bears some resemblance to the optic cup of a developing vertebrate eye. The cavity of the optic cup is in free communication with the palliai cavity by the “optic stalk,” if we may give this name to the stalk of invagination. In most cases the whole of the invaginated cells are pigmented, those of the stalk as well as those of the cup, and both layers of the cup are always pigmented, but it is only the cells of the anterior layer of the cup, that is to say, those in contact with the vitreous body, that are enlarged and columnar. The hinder wall of the cup and the walls of the stalk are composed of low cubical or flat epithelial cells and the transition between these and the large columnar pigmented cells at the margin of the cup is well shown in fig. 21. The preservation of my specimens was not good enough to allow of my working out such delicate details as the nerve supply of these organs. In every case a nerve derived from the circumpallial nerve could be detected in close proximity to an eye, and in fig. 20 two such nerves are seen in section, one on each side of the optic stalk. But I have been unable to trace nerve-fibres running into those columnar pigmented cells which, because of their characters and position I have called retinal. Further investigations are necessary before the exact nature and function of these “eyes” can be determined with certainty. But their structure points to their being photoscopic, or possibly thermoscopic. I have already alluded to their somewhat paradoxical position and suggested that they may serve to warn the animal to keep the valves of the shell closed during the day-time.

The Byssus and Byssus-gland

—In looking through the literature of the subject one notes with surprise the controversy about the nature of the byssus in Anomia ephippium. The calcareous plate or “ossicle” was a stumbling-block to many of the older authors, who regarded it as a third valve of the shell, and, though de Lacaze Duthiers gave cogent reasons for regarding the ossicle as a calcified byssus, and Moore (12) described the fixation of the young forms and the modification of the left valve produced by the asymmetrical attachment of the byssus, neither of these authors gave any account of the microscopical structure of the byssus gland, and as recently as 1878 von Jhering emphatically denied any homology between the “Falten-organ” of Anomia and the byssus-cavity of other Lamelli-branchia.

The question was finally decided by Barrois (1), who showed that the macroscopic and microscopic structure of the “Faltenorgan” is in all essential particulars identical with the byssus cavity of Arca tetragona. But Barrois contented himself with two very diagrammatic woodcuts of the byssus cavity of the two genera, and I have been unable to find any accurate drawing or description of the minute structure of the byssus of the Anomiacea. The “ossicle” being absent in Ænigma, the byssus and byssus-gland of this genus is a more suitable object for microscopical study than the corresponding organ in Anomia, and its resemblance to the byssus gland of Area tetragona as described by Boutan (4) is obvious. In the latter genus the foot is very small, and has a groove on its hindei’ or posterior surface; the byssus cavity is very large, and the byssus is a stout oval structure consisting of a number of lamella) which, as Boutan says, overlie one another like the coats of an onion. The byssogenous gland consists of some twenty to twenty-five parallel epithelial folds or laminae, which traverse the interior of the byssus cavity, and the chitinoid lamellae of the byssus are formed from, and their inner ends are contained between, these laminæ. Following Boutan, I will describe the internal divisions of the byssus itself as “lamellæ” and the epithelial folds of the byssogenous gland as “laminæ.”

The extent of the byssus-cavity in Ænigma and its relation to the massive byssus-muscle are shown in the small scale drawing fig. 10, and a transverse section through both byssus and byssus-gland as seen under a higher’ magnification in fig. 17. The left-hand part of the latter figure shows that the lamellae of the byssogenous gland are folds of the epithelium lining the byssus cavity, each fold having in its centre a core of connective tissue. The figure, though drawn under a magnification of 600, is on too small a scale to show histological details clearly, but it may be seen that the edge of each lamina is covered by large columnar cells filled with granules. To the right of the figure the byssus itself is seen in situ. It consists of a number of chitinoid lamellæ lying between the byssogenous epithelial laminae. The dark line down the centre of each lamella shows its double origin from the walls of adjacent laminae. The outer ends of the lamellae are united to form a plate, and it is evident that this plate increases in thickness by the addition of material secreted by the large glandular cells on the edges of the laminae. This plate is firmly attached to the substratum on which the animal rests. There is no trace of calcification in it, and in several of my specimens the bark of the root to which it was fixed remains adherent, showing that there is no question of the existence of an “ossicle,” which has been torn off when the animal was detached. A comparison of this drawing with Boutan’s figures (loc. cit., pl. 14, figs. 18, 21, and 22) leaves no doubt as to the identity, in all essential particulars, of the byssus and byssus-gland of Ænigma with the corresponding structures in Area tetragona. The laminae are much more numerous and the byssus cavity is relatively wider and shallower in the former’ genus, that is all.

Fig 18 is a very highly-magnified drawing of the outer end of a single byssogenous lamina lying between two lamellae of the byssus of Enigma. The sides of the laminae are clothed by a clear, generally-cubical, ciliated epithelium. I have no doubt that this is a ciliated epithelium, and that it corrresponds with the ciliated epithelium lining the byssus cavity of other Lamellibranchia, as, for instance, in Cyprina islandica (Carrière, 5), Dreissensia polymorpha (Horst, 7), and Jousseaumia (Bourne, 3). Boutan, however, is of a very different opinion. He says, of similar cells in Area, “Au-dessus de l’epithelium, en contact avec le produit sécrété, on aperçoit une striation très nette qu’on serait tenté de prendre, an premier abord, pour des cils vibratiles; en réalité, ce ne sont que des petits bâtonnets de matière sécrétée, absolument immobiles.” Immobile they may possibly be, as I suspect that their function is to afford sufficient surface friction to prevent the byssus lamellae from slipping out of place, but that they are true cilia is shown by their insertion on a striated border of each epithelial cell, by their correspondence with the cilia undoubtedly borne by similar cells in other Lamellibranchia, and by the fact that they are present where the secretory activity is in abeyance, but absent where it is still in progress, which is the reverse of what would be the case if they were “little rods of secreted material.”

Between the epithelial walls of the lamina is a plexus of connective-tissue cells, among which there are elongate pyriform or spindle-shaped masses of granules in which no nuclei can be distinguished. Deeper down in the lamince a few glandular cells loaded with granules are scattered through the connective-tissue core, but there is no compact mass of byssogenous cells, such as is usually to be found in other Lamellibranchia. The elongated strings and globules of granules must be identified with the streams of granules which I have described in Jousseaumia (3) as travelling by intercellular paths from the byssus gland to the byssus cavity. I am of the opinion that the byssogenous cells break up, and that the secretion travels between the irregular spaces of the connective tissue, and that there are not definite ducts as described by Horst (7) in Dreissensia. Boutan figures irregular branching ducts in Area tetragona (4, pl. 13, fig. 12), but he does not enter into histological details, and his figure might equally well be interpreted according to Horst’s views or my own. Perhaps the most remarkable feature in Ænigma is the cap of granular columnar cells on the edge of each byssogenous lamina. These cells are clearly continuous with the ciliated epithelium of the sides, but they are not ciliated, and are filled with byssogen granules. It may be inferred that they have taken up these granules from the intercellular channels of the connective tissue, and that they secrete them again at their free surfaces, thus adding to the thickness of the byssus plate. The concentric lines in the latter (fig. 18) clearly indicate that there has been a continuous addition of fresh matter from the large granular cells capping the edge of the lamina. There is no possibility of confusing the byssogenous with the mucous cells in Ænigma; the latter are, indeed, numerous in the foot and in the lips of the byssus cavity, but they never penetrate into the laminae, and are easily distinguished by their oval or polygonal shape small nuclei and clear contents.

I will conclude with a few remarks on the histology of the labial grooves and alimentary tract.

The right and left labial grooves pass without any distinct line of demarcation into the mouth. The right groove as stated on p. 260 is shallow and smooth for a large part of its course posterior to the mouth ; the left groove, on the contrary, soon becomes deep, and is thrown into numerous vertical ridges. In both grooves the vertical ridges are covered by a very high, ciliated, columnar epithelium, in which no gland cells can be distinguished. But the smooth portions of both grooves are lined by a characteristic epithelium shown in fig. 23. The ciliated columnar cells are very distinct, and have a doubly refractive border. Between them are two kinds of gland cells, elongated granular, and ovoid clear cells. The former are elongated and occupy the spaces between the ciliated cells, their free ends reaching to the surface. They are filled with fine yellow granules, and their nuclei are to be found in the inner third of their length. As these nuclei are identical with those of the ciliated cells it is probable that the granular gland cells are modifications of ciliated cells. The ovoid clear cells are very large, with clear contents staining pink in picro-indigo carmine; their nuclei stain uniformly dark red in borax carmine. From their staining properties these ovoid cells appear to be mucous cells, and they are similar in size and appearance to the mucous cells of the foot, but, unlike the latter, are not rendered polygonal by mutual pressure.

The epithelium of the labial grooves passes into the œsophagus, but the finely-granular cells soon disappear, and their place is taken by large coarsely-granular gland cells. The mucous cells at the same time disappear. The œsophagus is surrounded by a very distinct layer of subepithelial muscular fibres. The epithelial lining of the œsophagus passes gradually into that of the stomach. In this cavity the glandular cells, as has been already described on p. 267, are restricted to the side walls and floor; the roof is thin and lined by moderately long ciliated cells only. It is noticeable that the thick, cuticle-like lining of the stomach, the “flêche tricuspide” of Poli, corresponds exactly in extent to the area in which the yellow glandular’ cells occur, and is not present in the roof and upper portion of the left wall where these cells are absent. A series of transverse sections shows that the thick glandular tract of the stomachal epithelium is continued posteriorly into the cæcum of the crystalline style, while the ciliated non-glandular tract of the roof and left side passes into the small intestine, and at the entrance into the latter is thrown into a number of stout, ciliated ridges which form a straining apparatus, and are continued into the four prominent ridges projecting into the lumen of the intestine. The transition from the glandular epithelium of the floor of the stomach into the characteristic epithelium of the cm cum of the crystalline style is a gradual one; the ciliated cells of the stomach gradually become shorter and stouter, the yellow gland cells gradually become scarcer, until shortly after its origin from the stomach the caecum is lined exclusively by the epithelium shown in fig. 22. I have shown (3) that in Jousseaumia there is a similar localised tract of glandular cells in the stomach passing through a similar transitional epithelium into the caecum of the crystalline style, and I have suggested that the last-named structure is secreted by the gland cells in question.

Our knowledge of the structure and functions of the “flêche tricuspide” and the crystalline style is due to the researches of Barrois (2), List (10) and Mitra (11). Thanks to the last author, we know that the style consists of a proteid material belonging to the globulin class, and that it contains an active amylolytic ferment. He supposes, without giving any very cogent reasons for his conclusions, that the substance of the style is secreted by the liver, and is stored up as a flexible solid in the caecum, or in some forms, in a special compartment of the stomach or intestine. Barrois has carefully investigated the structure of the “fleche tricuspide” of the stomach and the crystalline style in Donax trunculus, and has entered much more fully into the histology of the tissues in contact with these structures than has Mitra.1 He describes the “flêiche tricuspide” of Don ax as forming a complete lining of the cavity of the stomach, and as being more or less adherent to the anterior end of the crystalline style. He does not recognise any gland cells in the epithelium of the stomach, but suggests that the “flêche tricuspide” is formed from a granular mass detached from the ends of the epithelial cells, and gives some not very satisfactory figures in support of his statement. He further suggests that the function of the “flêche tricuspide” is to protect the walls of the stomach from injury. In his description of the caecum of the crystalline style of Donax, Barrois calls attention (and as far as I can determine he is the only author who has done so) to the existence of a groove lined by a modified epithelium, running the whole length of the right side of the caecum of the crystalline style, from its origin from the stomach to its extremity. There is a similar groove along the right side of the caecum of /Enigma, showing the same histological characters as those described and figured by Barrois in Donax. I have given a careful drawing of this groove and the adjacent parts of the wall of the caecum in fig. 22. As regards the characters of the epithelial cells, it corresponds so exactly with Barrois’ description that I need give no further account of it, except to call attention to the extremely long cilia borne by the short cells lining the bottom of the groove, and the short and fine cilia borne by the tract of modified columnar cells on the upper side of the groove. These are not described by Barrois, who says on the contrary, “toute la surface épitheliale est tapissée de cette épaisse et forte couche de oils vibratiles d’ont j’ai parlé à. maintes reprises, et sur d’excellentes preparations au carmin aluné, j’ai pu la suivre aussi bien au niveau de 1’ épithelium modifié que sur la reste de la section.” I think Barrois must have been mistaken on this point, but however that may be, the modified cilia are very conspicuous in Ænigma. Their arrangement leaves no doubt that the groove has a function analogous to that of the endostyle of an Ascidian, and that its cilia sweep a current of liquid or viscous matter (I could find no solid particles in it) along the length of the cæcum.

From a consideration of these histological details, I suggest that the material of the crystalline style is secreted by the yellow glandular cells of the epithelium of the stomach, and that the so-called cuticular lining of the stomach, or “flêche tricuspide “of Poli, is nothing more than the coagulated viscous secretion of these glands. This viscous secretion, I suggest, is swept into the caecum by the action of the cilia of the ciliated groove, and is there moulded and solidified into the substance of the crystalline style which, as Mitra has proved, is a globulin containing an amylolytic ferment. During the process of digestion, the style as a whole is moved forward into the cavity of the stomach by the action of the stiff brushlike cilia of the normal cæcal epithelium, and is gradually dissolved, liberating the amylolytic ferment. It is true that several authors, including Barrois, have asserted that the anterior end of the crystalline style is continuous with the so-called “flêche tricuspide,” but this confirms rather than contradicts my suggestion. These same authors assert, and I agree with them, that the substance forming this cuticular coat or “flêche tricuspide” is identical with the crystalline style. When one examines this substance carefully, one finds that it forms a lining to the wall of the stomach, thin in some places, thicker in others, and where it is thicker its inner surface (that is the surface farthest from the epithelium) passes insensibly into a mass of granular coagulum, in which the anterior end of the style appears in some cases to be imbedded. But there is no real continuity between the two. The anterior end of the style is in all cases much reduced in diameter, and is evidently undergoing dissolution. It cannot at one and the same time be losing material and receiving additions from the substance which has been called the cuticle or the “flêche tricuspide.” My interpretation is that the crystalline style is added to at its hinder end, and that the material for its renewal is carried down the caecum by the ciliated groove. List (10) has given an interesting account of the formation of the crystalline style in specimens of Mytilus fed with Indian ink, which is not inconsistent with my suggestion. He does not, however, appear to have subjected the style to the same careful chemical analysis as Mitra.

The numerous follicles of the liver are formed of two kinds of cells. Large, coarsely-granular, clear cells with distinct nuclei, and wedged in among the outer ends of these a smaller number of deeply-staining, finely-granular cells resembling the demilune cells of mixed salivary glands of mammals.

The intestine, in the first part of its course where it is of narrow calibre, with four ridges projecting into its lumen, is lined by an epithelium consisting of attenuated ciliated cells with deeply-staining, densely-crowded nuclei, among which are a few goblet cells, smaller, and with more finely-granular contents those occurring in the stomach. In the loop of the intestine the ridges die out, the goblet cells disappear, and the attenuated ciliated epithelium alone remains. In the straight part of the intestine this is replaced by a columnar-ciliated epithelium with a clear, somewhat granular cytoplasm and pale oval nuclei containing a sparse chromatic reticulum (fig. 24). There are no gland cells in this region, but the cytoplasm of the epithelial cell stains bluish-green in picroindigo-carmine, and is very distinct from the section of the intestine preceding it and from the proctodæum. It is usual to call this part of the intestine the “rectum.” But it is so sharply marked off from the terminal portion of the alimentary tract that I prefer to describe it as the large intestine, and to restrict the name rectum to the short, somewhat enlarged section of the gut which opens to the exterior by the anus This rectum, which is clearly proctodæal in origin, is lined by an epithelium composed of clear and very attenuated ciliated cells with deeply-staining nuclei.

The whole of the large intestine, and a section of the small intestine preceding it, is infested by sporozoan parasites, whose characters are shown in fig. 24. They are not sufficiently well preserved to admit of careful description, but they are evidently the trophozoites of a Coccidian, some of which are forming cysts containing sporoblasts. They have some likeness to the genus Klossia, but as I have not been able to discover the spores, or to trace the various stages of the life-history of this parasite, it will be better to record its existence without conferring upon it a new and probably a misleading name.

  1. Ænigma, though modified in the same direction as Anomia, has undergone a less degree of torsion, and has retained more of the typical features of a normal Lamellibranch.

  2. There is, on the left side, a specialised pallial muscle, attached to the left valve, and acting as a retractor of the left gill.

  3. A ring of eye-spots, of peculiar structure, is found on the left mantle lobe, at a considerable distance from the edge of the mantle.

  4. Adaptations for resisting desiccation during long exposure to the sun and air are found in the thickening and corrugation of the lower moieties of the mantle lobes and in the existence of cæcal extensions of the pallial cavity, which can be closed by the apposition of the ciliated edges of ridges developed on the mantle and body-wall.

  5. The structure of the byssus gland is of the same type as that of Area tetragona and Auomia ephippium. There is no calcified ossicle.

  6. The inner demibranch of the right gill is attached to what is morphologically the left side of the foot. The minute structure of the gills is curiously similar to that of Anomia ephippium. The velar filaments of the external demibranchs bear special cilia.

  7. The intestine is coiled.

  8. The kidneys and the openings of the reno-pericardial ducts and gonaducts into the kidneys are similar to those of Anomia ephippium. The gonopores have ciliated funnels.

  9. There are extensive remnants of the pericardial gland. The wall of the pericardial cavity is shown to be incorporated with the walls of the ventricle and auricles.

  10. An internal ciliated groove runs along the whole length of the caecum of the crystalline style.

  11. The left gonad extends far back into the left lobe of the mantle.

Barrois
, Th. —“
Sur la structure de 1’Anomia ephippium
,”
‘Bull. Sei. Dép. Nord’ (2)
,
ii
,
1879
.
Barrois
, Th. —“
Le stylet crystallin des Lamellibranches
,”
‘Rev. biol. Nord France,’
1
,
1890
.
Bourne
,
G. O.
—“
On Jousseaumia, a new genus of Eulamellibranchs
,”
‘Report on the Pearl Oyster Fisheries and Marine Biology of the Gulf of Manaar,’ part v, Roy. Soc
.,
1906
.
Boutan
,
L.
—“
Recherches sur le Byssus des Lamellibranches
,”
‘Arch. Zoo). exper. et gdn.’ (3)
,
iii
,
1895
.
Carriere
,
J.
—“
Die Drusen im Fusse der Lamellibranchiaten
,”
‘Arb. Zool. Inst. Wurzburg,’
v
,
1879
.
Grobben
,
C.
—“
Die Pericardialdruse der Lamellibranchiaten
,”
‘Arb. zool. Inst. d. Univ. Wien.,’
vii
,
1888
.
Horst
,
R.
—“
1st die Byssus eine Cuticularbildung ?
‘Tijdschr. Ned. Dierk. Vereen.’ (2)
,
ii
,
1889
.
Lacaze
Duthiers
,
H.
de
. —“
Mémoire sur 1’organisation de 1’Anomie (Anomia ephippium)
,”
‘Ann. des Sei. Nat.,’ 4e ser
.,
ii
,
1854
.
Letellier
. —“
Étude sur le fonction urinaire chez les Mollusques acéphales Lamellibranches
,”
‘Arch, de Zool. expér. et gén. (2)
,
v
bis,
1887
.
List
, Th. —“
Die Mytiliden
,”
‘Fauna u. Flora des Golfes von Neapel,’
xxvii
,
1902
.
Mitka
,
S. B.
—“
The Crystalline Style of Lamellibranchia
,”
‘Quart. Journ. Mier. Sei.,’ N. S
.,
xliv
,
1901
.
Morse
. —“
Remarks on the Relations of Anomia
,”
‘Proc. Boston Soe. Nat. Hist.,’
xiv
,
1871
.
Pelseneek
,
P.
—“
Contribution a 1’etude des Lamellibrauches
,”
‘Arch. de Bio).,’
xi
,
1891
.
Pelseneek
,
P.
—“
Mollusca
,”
Lankester’s ‘Treatise on Zoology,’
part v,
1906
.
Ridewood
,
W. G.
—“
On the Structure of the Gills of the Lamelli branchia
,”
‘Phil. Trans. B.,’
cxcv
,
1903
.
Sassi
,
M.
—“
Zur Anatomic von Anomia ephippium
,”
‘Arb. zool. Inst. d. Univ. Wien.,’
xv
,
1903
, p.
81
.
Sharp
,
B.
—“
On the Visual Organs in Lamellibranchiata
,”
‘Mitlb. Zool. Stat. Neapel,’
v
,
1884
.
Woodward
,
S. P.
‘Ann. Mag. Nat. Hist.,’
xvi
,
1855
.

EXPLANATION OF PLATES 15—17,

Illustrating Prof. Gr. C. Bourne’s paper on “The Structure of Ænigma aanigmatica, Chemnitz: a Contribution to our Knowledge of the Anomiacea.”

Lettering in all the figures.

ad. Adductor muscle, an. Anus. ap. Aperture between the right and left renal sacs, a.r.p. Anterior retractor pedis muscle, br.g. Branchial ganglia, by.c. Byssus cavity, by.l. Lamella of byssus. by.lm. Lamiuto of byssogenous gland, by.pl. Byssus plate, c.cm. Cerebral commissure, c.d. Ciliated discs, c.jl. Latero-froutal cilia of gill filaments, c.fr. Frontal cilia, c.s. Chitinous skeleton of gill filaments, c.v. Ventral cilia on velar filaments. cn. Cornea, cry. Caæum of crystalline style. db1. Right external deniibranch, db2. Right internal demibranch. db3 Left internal demibranch. dbx. Left external demibranch. ep.ext. External epithelium, e.s. Pallial eye spots, f. Foot. gl.c. Gland cells, go.a. Anterior lobe of left gonad, go.l. Posterior lobe of left gonad, go.r. Right gonad, int. Small intestine, ini1. Large intestine, k.cn. Concretions in cavity of kidney, l.cp. Left cerebropleural ganglion, l.ep. Left excretory pore, l.f.r. Right labial folds or palps, l.gd. Left gonoducal opening into kidney, la.l. Left labial groove. Ig.r. Right labial groove, li. Liver, li.d. Hepatic ducts, l.k. Left kidney. In. Lens of eye. l.p.g. I<eft pericardial gland, l.rp. Left reno-pericardial aperture, m. Mouth, m.f. Muscle fibres, mt. Left mantle lobe. mt’. Right mantle lobe. a>. (Esophagus, p.c.c. Pericardial cells with endoplastic concretions. pd.g. Pedal ganglia, pl.m. Branchio-pallial muscle, pr.p. Posterior retractor pedis muscle. r. Rectum, r.cp. Right cerebro-pleural ganglion. r.ep. Right excretory pore. r.k. Right kidney, r.p.g. Right pericardial gland, r.rp. Right reno-pericardial aperture, rt. Retina, st. Stomach. v. Ventricle of heart, v.cm.l. Left visceral connective. v.cm.r. Right visceral connective, vet. Velar fold of gill filaments, v.g. Visceral ganglia. vi. Vitreous body. v.m. Visceral mass. x. The dorsal pallial suture, z. Cavities between various parts of the body, ending blindly and serving for storage of water. I. Reflected lamella of left external demibranch. n. Direct lamella of left external demibranch, in. Direct lamella of left internal demibranch, iv. Reflected lamella of left internal deinibranch. v. Union between the reflected lamella of the right and left internal demibranchs. VI. Reflected lamella of right internal demibranch. vil. Direct lamella of right internal demibranch. vm. Direct lamella of right external demibranch. IX. Reflected lamella of right external demibranch.

PLATE 15.

PIG. 1.—View of the animal lying in the left valve of the shell, after removal of the right mantle lobe. × about .

FIG. 2.—The animal removed from its shell and viewed from the left side, showing the ring of pallial eye-spots. The principal organs of the left side are seen through the transparent left mantle lobe.

FIG. 3.—A reconstruction from a series of sections to show the course of the alimentary canal and the nervous system.

FIG. 4.—A reconstruction from a series of sections to show the renal organs and the extent and relations of the pericardial glands. The right kidney is represented in a darker, the left kidney in a lighter tone. The spaces enclosed by the black lines represent the pericardial glands.

FIG. 5.—A horizontal section through the angle of the direct and reflected lamellre of one of the demibranchs, showing the single row of ciliated discs, c.d. Highly magnified.

FIG. 6.—A transverse section through three velar filaments, showing the stiff cilia, c.v., on the ventral faces of the filaments. Zeiss’ horn. imm. comp. oc. 4.

FIG. 7.—A transverse section through two adjacent gill filaments. Zeiss* horn, imm., comp. oc. 4.

FIG. 8.—A transverse section through a pallial fold guarding the entrance to one of the water reservoirs, showing the ciliated ridge, e.r., on the edge of the fold. n. A branch of the pallial nerve.

ON THE STRUCTURE OF /ENIGMA 2ENIGMATI0A. 295

PLATE 16.

FIG. 9.—A transverse section through the body at the level of the foot fir. The “flechc tricuspide” of Poli.

FIG. 10.—A transverse section passing through the posterior half of the byssus muscle.

FIG. 11.—A transverse section passing through the ventricle of the heart.

FIG. 12.—A transverse section passing through the middle of the adductor muscle.

FIG. 13.—A transverse section passing about 2 millim. in front of the anus.

FIG. 14.—A section through the right reno-pericardial canal and the adjacent portion of the pericardial gland. Highly magnified.

FIG 15.—A section through the ciliated funnel of the opening of the left gonaduct of a male into the kidney. The columnar ciliated cells of the funnel are conspicuous, go. A follicle of the testis, sp.d. Sperm duct.

FIG. 16.—a. A group of pericardial cells containing small concretions. b. A group of larger concretions from the pericardial gland of another specimen of ænigma.

PLATE 17.

FIG. 17.—A transverse section through the byssus and byssogenous laminte. × 600.

FIG. 18.—A transverse section through the distal end of a byssogenous lamina and two adjacent lamellae of the byssus. oil. Ciliated epithelial cells at the sides of the lamina, grn. A mass of byssogen granules in the connective tissue core of the lamina). Zeiss’ hom. imm., comp. oc. 4.

FIG. 19.—Portion of a section through the wall of the ventricle of the heart, showing ep. ewt., the external epithelium; m.f., the branchial cardiac muscle fibres; pe.e., the characteristic pericardial gland cells with endoplastic concretions.

FIG. 20.—A section through a pallial eye, showing the deep invagination of the epithelium lining, the inner face of the mantle, and the modification of the deeper invaginated cells to form a retinal layer.

FIG. 21.—A portion of another section through a pallial eye. Zeiss’ hom. imm., comp. oc. 4.

FIG. 22.—A portion of a section through the ctecum of the crystalline style, showing the ciliated groove, gt., and the tract of modified epithelium on the dorsal side of it. The section is reversed in the drawing, so that the ventral side is uppermost.

FIG. 23.—A section through the right labial groove, close to the mouth.

FIG. 24.—A section through the epithelium of the large intestine, spor. Sporozoan parasites.

1

The concretions in the kidney are commonly said to consist of uric acid ; but Letellier (9) states that no uric acid is excreted by any Lainellibrancbiate, and that the urinary concretions consist of calcium carbonate and acid phosphates of lime and magnesium. According to this author the Lamellibranch kidney excretes urea, the Gastropod kidney uric acid.

1

Mitra appears to have been very imperfectly acquainted with the researches of his predecessors on this subject, and his quotations of literature are mostly derived from text-books. Had he read Barrois’ memoir, he would have found that as early as 1686 v. Heide suggested that the crystalline style was subservient to digestion: “aliquendo cogitavi hunc stylum suppeditari alimini fermentum.” Barrois gives a very interesting account of the various views that have been held on the origin and function of the crystalline style.