The Molluscs of the Great African Lakes. — II. The Anatomy of the Typhobias, with a Description of the New Genus (Batanalia)

No entire specimen of Typhobia has hitherto been described, and we have consequently remained entirely in the dark as to the real morphological character of what is probably the most remarkable fresh-water Gastropod at present known.

Presumably from the characters of its empty shell this genus has been classed by the conchologists¹ with the Melanias, as a new sub-section of that group.³ But into what serious error determinations of this sort may lead, when based on conchological evidence alone, the present paper, which contains the first anatomical description of the mollusc, will suffice to show.

It will be seen that the structural features of the Typhobias, so far from establishing the above conchological anticipations, in every way confirm the conclusions at which I arrived respecting the marine origin of these molluscs from a study of the distribution of the African lake fauna in general.³ Hence the actual facts of anatomy are, as I anticipated from the facts of distribution that they would be, directly in conflict with all those geological speculations respecting the interior of Africa that have been hitherto more or less generally held. In the paper to which I have referred¹ it was seen that the Typhobias belong to, and are, one of the most remarkable constituents of the quasi-marine or Halolimnic section of the Tanganyika fauna. They consequently share, along with the other members of this group, the strange geographical isolation which is its distinctive mark.

Like nearly all the Halolimnic animals, Typhobia is found pretty abundantly in Tanganyika, occurring in some places in the most astonishing profusion, but, so far as it is at present known, the mollusc is found living nowhere else in the world. I first obtained the empty shells of T. Horei on the long sandy beaches near the south-west corner of the lake, and subsequently on the southern shore of the deep Kituta Bay. They were readily recognised by the head men of the villages, who told me they had never seen the Gastropod alive, but only the shells when washed up empty along the beach. From this statement of the natives, and from the spinous character of the shells, I thought it probable that they would be found living on mud, but I was unable to find them in the muddy reaches among the Kinyamkolo Islands, in depths of fifty to one hundred feet, nor indeed in any portions of the lake that were of similar depth. It was not until I had extemporised a primitive deepwater dredging apparatus that I obtained the Gastropod alive.²

In June, 1896, we were on the west coast of Tanganyika and on the southern shores of Cameron Bay, and here I was able to obtain the strong bark rope used by the Wafipa fishermen for their nets. As these nets are hauled by rows of men on the ropes at either end, the ropes themselves are strong enough to drag a heavy net with all its weights and stretchers over several hundred yards of ground; and with them I was consequently enabled to dredge in water that varied from 500 to 850 feet in depth. Eventually in this manner, during the months of June and July, about a hundred Typhobias were obtained alive. Of these some were examined on the spot, some preserved in various ways, stored in spirit, and eventually brought back. The living Typhobias were associated with another deep-water Gastropod, also alive, a brief description of which will be found at the end of the descriptive part of this paper.

Except in the characters of the shell, this new genus is almost identical anatomically with T. Horei, consequently it is unnecessary that I should do more than point out in what it differs from the form already known.

External Characters.—The general appearance of a living Typhobia is seen in fig. 1. They are always very active when brought up from the deep water they inhabit, probably being uncomfortable through the decreasing pressure. The tentacles are very long and slender, and the eyes completely at their base ; the snout is wrinkled and very much pigmented on the upper surface, and it is so long and slender as to suggest the ordinary protrusible snout or introvert of Prosobranchs ; on dissection, however, it is seen to be simply elongated externally, very retractile, but, like that of Pterocera, not introvertible in any sense. The foot is very broad, and of the same pale semi-transparent yellow as that of Anodonta. The mantle is prolonged into the well-marked anterior and posterior siphons (fig. 2).

As is the case with many other fresh-water Gastropods, the shells of Typhobia Horei vary to a remarkable degree; indeed, the extreme forms when isolated, and not linked together by the innumerable intermediate forms which actually occur in Tanganyika, differ so widely that the French concho-logist¹ Bourguignat regarded these differences as sufficient to split the genus up into four species, under the new name of Hylacantha ; his four so-called species being respectively H. H orei, H. Bourguignati, H. longirostris, and H. Jubertii. Had this author, however, been able to obtain large numbers of these shells on the spot, and to have made collections of the extreme and intermediate varieties, it is hardly conceivable that he would have ever regarded their variations as specifically distinct. Be this as it may, however, it is quite evident, when large numbers of these shells are studied, that their varieties cannot be regarded as specific forms. The shells of Typhobia Horei are, however, undoubtedly polymorphic, for there are about four well-marked varieties, into one or another of which the great majority of the specimens I collected tend to fall (figs. 13, 16, 23, 25). When the shells are very young, about the time of birth (figs. 27, 28), they are destitute of all but the merest trace of spines as well as of the pronounced rostral beak, which is a marked structural characteristic of the majority, but by no means of all the older shells. Now the fact that in some adult shells the rostral beak is entirely wanting (figs. 16 and 20) shows that, in this respect, such shells have not deviated from their embryological character. All the four extreme types of variation graduate off into forms which, in the more or less complete absence of a rostral beak, approximate to the shells represented in figs. 16 and 25. Such shells, therefore, may be said to represent the extreme of least specialisation. The remaining extreme polymorphs (figs. 13, 23) can be all traced through successive stages from the types represented in figs. 12, 17, 18, and 21. Thus between the extremes, figs. 13 and 16, there are intermediate forms, such as figs. 12 and 17 ; the extreme, fig. 23, has intermediates, such as figs. 18 and 21, while the extreme, fig. 25, is connected up by forms similar to that represented in fig. 20 and perhaps 21. Any number of intermediate stages could have been represented for each series ; but for obvious reasons those only have been selected which seemed to best express the transition from any one type to the next. It would thus appear that Bourguignat’s four species were formed in the absence of an adequate quantity ot material to work upon, and this conclusion is finally clinched by the fact that the anatomical characters of the soft parts of the extreme variations are indistinguishable from one another.

We may, therefore, conclude that the name Typhobia Horei, as given by Smith,¹ stands rightly for but one species after all.

The nervous system.—The nervous system of Typhobia, both in its general relationships and in the details of its different constituent parts, is almost entirely unique. Viewed from above, it is at once obvious that there is a great condensation and fusing together of the chief ganglionic masses, no commissure being visible externally between the cerebral ganglia (fig. 5,5). Each cerebral ganglion gives off anteriorly two sets of nerves, one obliquely above the other (fig. 34, 1,3) ; the upper and external arises from a prolongation of the ganglion comparable to the “saillie labiale.” These nerves are distributed to the tentacles and the eyes. From each cerebral ganglion below the “saillie” there arises another set of buccal nerves, the two innermost of which pass forward, enlarge into the buccal ganglia (fig. 34, 18, and fig. 5, 1), and unite again below the mouth. The remaining members of this set of nerves are distributed to the buccal mass, and to the parietes of the anterior portion of the head and snout. Ganglionic cells extend along the buccal nerve trunks as far as the buccal ganglia. Laterally each cerebral ganglion gives off a number of small nerves distributed to the head (figs. 34, 19). Towards the posterior upper surface of the cerebral ganglion the paired otocyst nerves arise, and pass obliquely backward over the pleuro-pedal commissures to the enormous otocysts (fig. 35, 6).

Below, the cerebral ganglia are connected with the pedal ganglia by the rather loug cerebro-pedal commissures (figs. 6, and 35, 20). Immediately behind these there is found on each side a second commissure, which at first sight appears to pass from the cerebral ganglion also (fig. 35, 11). In reality this is the pleuro-pedal commissure, the pleural ganglion being displaced forward, so as to lie closely applied to and immediately beneath the cerebral ganglia. In order to understand the relations of these ganglia it is necessary to examine several sections, as they cannot be seen by the ordinary methods of dissection. Fig. 9 represents a section taken through a point marked X in the general figure of the nervous system given in fig. 35. It shows the cerebral ganglion (a), separated from the pleuropedal commissure (b), while the fore-part of the pleural ganglion is seen as a continuation of this commissure (c). Fig. 10 is a little further back, and shows the pleural ganglion (a) and the posterior portion of the pleuro-pedal connective (b) But the section is still in front of the connection between the pleural and cerebral ganglia, both ganglia appearing separate. Fig. II is slightly further back again, and shows the pleuro-cerebral connective (a). There is visible also the posterior portion of the cerebral ganglion immediately before it passes downwards and is merged in the pleural ganglion itself. The pleural ganglia are thus seen to be displaced, and their real position is indicated by shading in the general arrangement of the nervous system (figs. 34 and 35). The pleuro-pedal connective is consequently shorter than it would be if the pleural ganglia were in their normal positions, while the pleuro-cerebral connective, as such, may be said to be almost entirely wanting.

On the left, the pleural ganglion is continued below and very slightly across the subœsophageal space (figs. 5, 6, 6, 3, 34,16), as an enormous ganglionic trunk, in the course of which the subintestinal ganglion is superficially quite indistinguishable, but the locus of this ganglion is marked by the great pallial nerve (fig. 5, 7). The ganglionic character of this left cord continues, as is shown by the presence of ganglionic cells, for a long distance, the appearance it presents at the point marked x’ (fig. 35) being represented in section at fig. 37, 1. The right pleural ganglion gives off a relatively small nerve, which, after passing obliquely over the œsophagus, carries the supra-intestinal ganglion, from which nerves branch to the gill, osphradial ganglion, and to the left pallial anastomosis (figs. 34, 35). This anastomosis is formed in the usual way by a rather large nerve passing out from the left pleural ganglion, and meeting the branch from the supraintestinal ganglion near the angle of the gill (fig. 35, 16). The nervous system is, therefore, dialyneurous on the left, and the relations of the nerves here give no indication of the extraordinary state of asymmetry which is encountered in the same region on the right. On this side the pleural ganglion gives off a nerve (fig. 34, 8) which appears at first sight as if it would form the right pallial anastomosis, either in the region of the subintestinal ganglion or along the course of the right pallial nerve. This nerve, however, after passing directly outwards for some distance bends sharply forwards, branches once, and each of the twigs diminishes rapidly, and dies out in the parietes of the mantle and the body-wall. The great mantle nerve which is given off from the subintestinal ganglionic trunk (fig. 34, 10) passes outwards, and is also distributed, without forming any connection with the pleural branch, to the mantle of the right side. Neither is there any connection between the pleural and subintestinal ganglion beneath the œsophagus. We have thus on the right side a condition of things which is almost unique among all the Streptoneurous Prosobranchiata which have hitherto been investigated. It is neither zygoneurous nor yet dialyneurous, and the condition of these parts finds its only analogy in the rather unsatisfactory descriptions given by Bouvier ¹ of the nervous system of Solarium and the Scalarids. In general arrangement, and apart from the above singular feature, the character and arrangement of the cerebral, pleural, and intestinal ganglia with their nerves and connections show a marked and indisputable similarity to those of the corresponding parts in such forms as Strombus, Pteiocera, Cancellaria, Voluta, and their associates. The wide distribution of the ganglionic cells in the nerve-cords of Typhobia is a remarkable and undoubtedly primitive feature ; while the fact that the nervous system of Typhobia foreshadows and is similar to those of several rather widely separated modern marine genera is direct and incontrovertible evidence, so far as it goes, that these molluscs are old and modified marine forms.

The pedal ganglia project forwards and are curiously extended in front by two colossal nerves (fig. 35, 13), which at their points of origin possess ladder-like connections one with another (fig. 35, 14), and thus approximate to the primitive type of pedal nerves possessed by the Helicinidæ. The pedal ganglia themselves are united together by a great transverse commissure (fig. 35) which contains ganglionic cells ; and on their postero-lateral surfaces they give off four or five large nerves which pass into the foot. The pedal ganglia are connected with the cerebral, and the pleural ganglia by the cerebropedal and pleuro-pedal connectives already described.

The otocysts of the Typhobias are relatively immense, and each is innervated by two fine nerves springing from the upper portion of the cerebral ganglia (fig. 35, 6). The position occupied by the otocysts is very anomalous, being completely above and separated from the pedal ganglia, close to the cerebropleural ganglionic mass (fig. 35,10). The otoliths are numerous and small (fig. 36), and the otocyst is lined by a well-marked sensory epithelium, from the free internal surfaces of the cells composing which there are given off fine “sensory processes” projecting into the cavity of the sac (fig. 36, 1). The position and character of the otocysts, and the prolonged pedal ganglia, are features in keeping with the generally primitive characters which the other parts of the nervous system seem to possess.

The digestive system presents points which, like those appertaining to the nerves, are at once interesting and new. The buccal mass is exceedingly small, and the radular sac is very short. There are no horny jaws, and the radular dentition is unique and peculiar in the extreme. A single transverse row of teeth is represented in fig. 43 ; also in the upper figure on page 189. The salivary glands are long and branched, the œsophagus being long, slender, and longitudinally folded. Internally it is lined by ciliated and glandular cells. Posteriorly the œsoph agus opens into the right side of the stomach, which is divided into an anterior and posterior chamber, the œsophagus opening into the latter. This posterior chamber of the stomach is traversed by several marked folds, the most conspicuous of which extends longitudinally (fig. 44, 7). On the floor of the stomach, to the right of this fold, are found the openings of the “bile-ducts” (fig. 44,2, 3). The anterior chamber is separated from the posterior chamber of the stomach by a constricted annulus, and the anterior chamber encloses and is almost filled by a large crystalline style, represented in fig. 44, 5). The whole arrangement of the stomach, the position of the folds and apertures, the separation into a posterior and an anterior chamber or cæcum, and the presence in the latter of a crystalline style are all similar to the condition of things obtaining in Pterocera. The style and its sac are undoubtedly homologous with the structures described by Collier,¹ Huxley,² Haller,³ and others, and the morphological conclusions which can be drawn from the nature of the stomach are in harmony with those which I have pointed out in reference to this Gastropod’s ganglia and nerves (see p. 187).

On leaving the stomach the intestine bends twice in the manner represented in fig. 42, and towards its rectal extremity it is considerably enlarged (fig. 49, 1). This enlargement contains the curious glandular fold represented in fig. 46, i. The anus is carried on a slight projection of the rectum from the mantle wall, and during life is slightly in advance of the margin of the mantle (fig. 2, 7).

The “Liver” occupies the lower portion of the upper whorls of the shell (fig. 2, 6), and has the usual characters of a digestive gland. The “bile-ducts” open by two orifices in the floor of the stomach, behind the pyloric aperture.

The Kidney occupies the region behind and to the left of the heart (fig. 3, 6), and opens by a single minute pore, quite at the posterior extremity of the mantle cavity.

The Heart and Gills.—The heart is simple, lying rather obliquely at the end of the mantle cavity. There is a large pericardial chamber (fig. 3, 5). The ventricle tapers from before backwards, and is surmounted by a large, rather thinwalled auricle, which in turn receives the pulmonary vein. There are well-formed valves between the auricle and the ventricle (fig. 48, 2), and between the ventricle and the aortic trunk (fig. 48, 4)-From the aortic trunk the anterior and posterior aortæ diverge in the usual way (fig. 48, 7, 8).

The gill in Typhobia (fig. 3, 7) is very long, extending from the base to the margin of the mantle cavity. It is composed of simple broad-based triangular leaves, the apices of which are elongated. The osphradium lies at the base of a groove ; it is long and simple, not fimbriated or gill-like, in fact a mere ridge (fig. 3, 8). This ridge is innervated by a nerve which springs from the small osphradial ganglion. Externally the ridge is covered with ciliated and glandular cells, the relations of which and the characters of the osphradial nerve are shown in section in fig. 8. There is nothing peculiar about the gills except their great length.

The Reproductive Apparatus.—In Typhobia the sexes are distinct and the female viviparous, the whole reproductive apparatus being simple but somewhat peculiar. The ovaries and testes occupy the upper surface of the last two whorls of the spire, and in the female the eggs, with their bright green yolk, pass directly into the simple oviduct (fig. 47, 2). From this they reach the lower expansion of the oviduct (fig. 47, 2) ; and in this sac, which functions as a brood chamber or uterus, they go through the greater part of their development. The walls of the sac are very thin, and while the animal is alive the bright green yolk of the eggs is distinctly seen through the delicate semi-transparent shell, so that the sexes can be distinguished at a glance. The sac opens near the rectum, at the junction of the mantle and the body wall (fig. 3, 3). The mollusc breeds during the months of June and July.

The testis (fig. 54, 6) opens by several small collecting channels into the simple vas deferens (fig. 54, 4)-This tube becomes somewhat but not much convoluted on its way, and ultimately expands into the curious enlargement represented in fig. 54, 2. On opening this it was seen in every case to bear about six singular parallel folds (fig. 49, 6). Beyond this expansion there is a curious finger-like outgrowth, extending from the duct into the mantle wall ; this process contains a muscular mass, which has all the appearance of, and probably is, an introvertible penis (figs. 45 and 46, 4). The lower extremity of the male genital duct opens by an elongate slit (fig. 49, 5). The possession of a penis in the mantle wall is a most curious fact, which would seem to indicate that that organ is analogous to, though very likely not homologous with, the penis of the Ampullariæ and other pulmonate Prosobranchs. In Typhobia the male genital gland is extremely interesting from a cytological point of view, as this genus is one of those Prosobranchs which, like Murox and Paludina, possess two forms of spermatozoa.

The small normal variety appear to arise one division after the heterotype which terminates the synaptic phase, and the cells out of which these normal spermatozoa are directly formed are, as in most cases, extremely small when the actual characters of the spermatozoa are taken on. On the other hand, the cells which directly metamorphose into the large spermatozoa, or megasperms, are very large, being similar in size and character to the synaptic (growing-cells) themselves. As in a former paper¹ I have advanced the view that after the synapsis any cellular generations which may exist are to be considered as potentially ova or spermatozoa, as the case may be, this fact that during the course of the spermatogenesis in Typhobia the small spermatozoa appear to arise two divisions after the formation of the synapsis, while the megasperms appear to be produced directly from the synaptic cells themselves, is extremely interesting.

Bathanalia.

The new generic type among the Gastropoda for which I propose the name of Bathanalia is at present represented by one specific form, Bathanalia Howesi, Pl. 12, figs. 29, 30, 31, 33. As the name implies, this species is found in association with T. Horei in the deep water of Lake Tanganyika, while in its anatomical features it is so similar to Typhobia that no special anatomical description is required.

There is only one point in Bathanalia that needs mention. I found after great difficulty, and by the help of sections, that in this genus there is a very slight but quite distinct pallial anastomosis on the right side.

It is entirely on account of the remarkable characters of the shell that I have thought it necessary to separate Bathanalia as a distinct genus from Typhobia.

The shell, as will be seen from fig. 29, is conical, composed of eight angular whorls, which from apex to base carry numerous short spines. The whorls are strongly sculptured, and the columella is open (fig. 33).

The mouth is ovoid, somewhat angular, and the mantle during life is prolonged into the last spine, forming a kind of false siphon (fig. 30, 1). The radula is shown in the process block, p. 189, and also in Pl. 12, fig. 32.

Comparative.

In establishing by comparison the true affinities of the Typhobias, I have purposely left undeveloped all those questions respecting the validity of the current systems of classification which such a comparison will inevitably raise. This course has been taken because I am now confident that the Halolimnic molluscs are among the few remaining indications of an ancient sea that once extended to or near the Tanganyika region of the present day. The anatomical features of the Halolimnic molluscs when studied together should therefore throw a most important light on the inter-relationships of those more modern marine genera and families which it has hitherto been so hard to solve. It would be premature and most unsatisfactory to view these questions from the anatomy of a single Halolimnic type ; for this reason I have reserved a discussion of these wider matters until I have had time to publish anatomical descriptions of the remaining Halolimnic forms.

The Typhobias have been classed by the conchologists among the Melanias, being regarded by Fischer¹ as a section of this group equivalent to Faunus or Melanopsis. Judged by their conchological characters alone it is by no means easy at first sight to understand why such a classification was ever made, as the Typhobia shell is almost as unlike any known Melania as that of a Pteroceras or a Cone. In assigning a systematic position to any animal concerning which the morphological study is incomplete, investigators are, however, always influenced, and often rightly influenced, by whatever collateral evidence respecting the habitat or modes of occurrence of such a form may be to hand. It was thus well known when the strange Typhobia shells were first described, that they came from a great equatorial fresh-water lake, and it appears to me that the early investigators only did the best they could with the purely conchological material they had before them, in concluding that although the shell of Typhobia had few characters in common with those of the Melaniidæ, they probably belonged to this group all the same. The Typhobias, however, happen to be one of those rare organisms in dealing with which, unless there is ample morphological material to draw upon, common sense anticipations such as the above are almost certain to be wrong. There was no reason when the Typhobias were originally described to suppose that Tanganyika, the great fresh-water lake in the centre of the African continent, had ever been connected with the sea. It was not known then that jelly-fish inhabited the lake, or that the Typhobias were only one in a long series of Gastropods which are not known to be living anywhere else in the world. Progressive zoological exploration has completely changed our views. The study of the distribution of the molluscs in the great African lakes points strongly, as I have shown, in the direction of the marine origin of the Halolimnic group of animals. We might, therefore, now with reason tend to be prejudiced in the opposite direction, i.e. in favour of the marine affinities of all the Halolimnic forms. Such a conception will, however, as I pointed out in my paper on the distribution of these forms,¹ require the very strongest morphological support, since it comes into the most uncompromising conflict with all those geological speculations respecting the character of the interior of Africa which were started by Murchison, and which affirm that the African interior has never been beneath the sea, at least since the period of the New Red Sandstone.² It is necessary, therefore, to use the greatest caution in determining what the affinities of the Typhobias and the other members of the Halolimnic group may really be.

It will have been seen from the foregoing anatomical description that unless the family of the Melaniidæ¹ is to be considered as an utterly heterogeneous group, the Typhobias are structurally near, if they are not at, the opposite end of the whole Tænio-glossate series. The Melanias, as they stand at present, are certainly by no means homogeneous, and as Bouvier very justly remarks, “La famille est une des plus mal etablies dans tout le groupe des Prosobranches,” but they do contain a substratum, possibly a majority of naturally associated forms, and although it will be most important, when dealing with other Halolimnic molluscs, to set limits to this group, the question of its heterogeneity does not obtrude upon the present discussion, the Typhobias being sufficiently distinct to be at once dissociated from all those Melanian forms which have up to the present time been anatomically examined. Whether they may have relations among those numerous so-called Melanias, the anatomy of which is utterly unknown, need not be discussed.

The unique and characteristic nervous system of Typhobia at once dissociates this form from all the ordinary fresh-water types. The great subintestinal ganglionic cord presents no analogy even to the zygoneurous types of nervous system, such as those of Potamides, Cerithidea obtusa, and Pyrgus sulcatus, which have been rightly regarded by Bouvier and others as representing the transitional links between the Paludina, Bythinia (?), and true Melanian types of nervous system on the one hand, and Haller’s generally marine “longicommissurate” families on the other. In Typhobia Horei the pleuro-subintestinal cord is in a most extraordinary condition, at once primitive, specialised, and unique. It is specialised in having lost the left pallial anastomosis, being thus neither zygoneurous nor dyaloneurous on the left side, a condition of things which finds its only parallel in the rather doubtful descriptions given by Bouvier¹ of the nervous systems of the Scalarids and Solarium. It is unique in the enormous development of the pleuro-subintestinal cord, the whole of this side of the nervous system being so disproportionate to the other as to distinctly foreshadow the secondarily acquired orthoneury of the Helicinidæ. The almost complete fusion of the cerebral ganglia in Typhobia, and the reduced and shortened-up pleuro-cerebral connectives, are conditions undoubtedly analogous to those obtaining in the Strombi, the Pteroceras, the Cancellaridæ, and other forms. The displacement of the pleural ganglia and the almost complete disappearance of the cerebro-pleural commissure on both sides, appear to be characters peculiar to Typhobia alone, while the position of the otocysts in the head, and not in the foot, is most primitive, but may have been accentuated by the forward displacement of the pleural ganglia, and the consequent necessity for the otocyst nerves to pass over these ganglia before they reach the otocysts. On the other hand, the otocyst nerves are very short, and even if the pleural ganglia were in their normal position, the otocysts would still be very high up in the head. The complete fusion between the pedal ganglia, and the presence of ganglionic cells in what remains of the pedal commissure, lead to the same inferences as do the characters of the cerebral ganglia. The great forward prolongation of the pedal ganglia, and the ladder-like connections between the proximate portions of the great anterior pedal nerves, are far more primitive characters. These do, in fact, suggest that the approximation in the posterior portion of the nervous system to that condition, of secondarily acquired orthoneury witnessed in the Helicinidæ, may not be altogether illusory after all. The characters of the nervous system of Typhobia show thus in a manner which does not appear to be capable of serious disputation, that this Gastropod has no relation to, nor indeed any but the most remote phylogenetic connection with, the hitherto recognised fresh-water forms. Nor has the nervous system any of those characters which could be regarded as possibly possessed by the forerunners of the Melanias, the Paludinas, the Bythinias, or indeed any of the recognised fresh-water types. Therefore, so far as the nerves go, the anatomy of the Typhobias gives a flat contradiction to the view that these Gastropods may be the survivors of any extinct fresh-water stock.¹ The nervous system of Typhobia exhibits, on the other hand, the characters of some ancient but more especially of several modern marine genera. Therefore the evidence which can be gathered from the anatomy of the nerves is exactly in accord with the deductions which were drawn from the study of the distribution of these Gastropods, the Typhobias appearing to be among the survivors of some old, but not geologically ancient, marine types.² What is true of the nervous system is, however, true of the remainder of the soft parts. Beginning with the digestive organs, it will be seen on reference to fig. 43 that the Typhobia radula, although very singular and self-contained, is still comparable to that of several marine Tænioglossa. Thus in the massive characters of the admedian teeth, the long slender character of the laterals, as well as the form of the median tooth, this radula approaches to those of Chenopus, Zenophora, Trochiformis, Pteroceras, Strombus, and Pustularia, while it has many characters in common with the radulæ of Crepidula, Trochita, Hy ponix, Turritella, and Cassis, and it resembles in a less degree those of Vermetus, Triton, Ranella, and Natica.

The characters of the salivary glands, the relation of the stomach to the œsophagus, of the intestine to the stomach, the position of the apertures in the stomach as well as the character of the pronounced median fold in the posterior stomachic chamber, are all characters which are strictly analogous to those obtaining in the Strombi and Pteroceras. The crystal-line style which I found in Typhobia and in certain other Tanganyika Gastropods, and more especially the pyloric cæcum, in which this style is contained, requires more attention than it has hitherto received. The existence of these structures in connection with the alimentary canal has long been known in the Lamellibranchiata, and it was formerly supposed to be confined to them. It is further well known that in the Lamellibranchiata the style has by no means the same relations in them all. In Anodonta and Mutela it is free in the intestine and not contained in a pyloric cæcum as in Lutraria and many other forms, nor is the cæcum present in the former types. When, however, the style and the cæcum are both present, the latter structure apparently has invariably the relations represented in fig. 53, the cæcum being a long stomachic appendiculum. This cæcum does not necessarily contain a style, and thus of the Lamellibranchiata it may be said that in some the style has indefinite relations, and there is no pyloric cæcum present ; while in others the cæcum is present, but does not necessarily contain a style. It was long ago pointed out—first, I believe, by Collier¹ in 1829—that in several Gastropods, the Strombidæ, also in species of Trochus and of Murex “there is an organ, the crystalline styletto, confined erroneously by a celebrated naturalist (Cuvier) to the bivalves. It is enclosed in a sheath that passes parallel to and by the side of the œsophagus to the stomach, into which the styletto enters, leaving its coverings.”

This interesting observation was subsequently confirmed and extended by Huxley² to Pterocera, and the organ was again more completely described by M. F. Woodward, the relations of the style and cæcum to the stomach as this author describes them being shown in fig. 52. Haller³ has confirmed Collier’s observation especting the existence of this structure in Strombus, and has extended them to Rostellaria, but he did not recognise the significance of the structure in relation to that of the Lamellibranchs ; nor does he appear to have been aware of Collier’s and Huxley’s observalions on this point. Lastly, I have found the style and its sac to be present in Typhobia, where it has exactly the same relations to a stomachic cæcum as in Strombus, Pterocera, or in those Lamellibranchiata in which this structure is present. I may also remark that the crystalline style and its cæcum are present in the so-called. Lithoglyphus of Tanganyika, the affinities of which Gastropod have been entirely misinterpreted.

From the complete similarity of the style, and more especially of the stomachic cæcum, in those Lamellibranchiata which possess it, and in those Gastropods where it is also present, there can be little doubt, as Collier, Huxley, and Woodward have already clearly seen, that the structures are in reality strictly homologous throughout. But the great importance and suggestive character of this conclusion has been much obscured by Fischer¹ and others who confuse the true crystalline style in its sac with the doubtful structure known as the “Flêche tricúspide.” There is little doubt that the “Flêche” has in the majority, if not in all cases been merely the cuticular lining of the stomach which has become detached, as it most readily does. With the appearance thus produced are to be classed the bodies described by Fischer in the stomachs of Cyclostoma and Paludina. Young² also describes in Helix pomatia a cuticular lining to the intestine, which he erroneously compares with the crystalline style of the Lamellibranchs. There appears to be no similarity between the cæcum described by Cuvier and Keferstein³ in Buecinum and that in Strombus and the Typhobias. Further investigation is, however, undoubtedly required.⁴

It will be seen from all this that the Typhobias and other Tanganyika Gastropods possess crystalline styles and cæca which have identically the same relations, and are structures which are undoubtedly homologous with the similar formations present in numerous Lamellibranchiata, and in a few other Gastropods as well. Now the practical importance of these facts to the present inquiry is this, that the particular Gastropods in which as yet the cæcum has been indubitably recorded, are Strombus, Pterocera, Rostellaria, Murex vertagus, Trochus turritus, the two species of Typhobia at present known, and the so-called Tanganyika Lithoglyphus. It may also possibly be present in Bythinia. Once more, then, the Typhobias in the characters of their stomachs and their related cæca are structurally near to those marine families with which by the character of their nerves and radulæ they were seen to be akin. In the possession of a style and its sac they further exhibit anatomical features possessed by the Lamellibranchs on the one hand, and by the connecting link between the Lamellibranchs and the Prosobranchs, the diotocardiate Trochi, on the other.

The gills in Typhobia are very similar to those in Strombus and Pterocera, and the osphradium resembles completely the same structure in all those Strombi which I have examined. The heart, as will be seen from reference to page 190, possesses the characters which are exhibited by nearly all the Tænioglossa.

The siphon possessed by the Typhobias is a structure of doubtful value from a classificatory point of view, and even in its narrower application it is by no means to be trusted, as both Bouvier and Haller have already shown. An interesting example of the impossibility of separating the holostomous from the siphonostomous Prosobranchs has come before me during the present investigation, for while examining one of the Melanias which the authorities of the British Museum generously placed at my disposal for comparison, I found in one, the exact species of which was doubtful, and which had been collected by Mr. Cumingin the Philippine Islands, the small but quite apparent siphonal extension of the mantle represented in fig. 4. This Melania had in every other respect the true characters of the group, but from the existence of the siphon it would, according to the old arrangement, have to be removed from the Melaniidæ and associated with those families of the Tænioglossa to which it most certainly does not belong. The distinct but small anterior prolongation of the mantle in Typhobia (fig. 3, 1), does not therefore appear to be of primary morphological importance, but its existence is undoubtedly another indication of the general similarity of the Typhobias to the forms which I have named.

The reproductive apparatus in the Typhobias has been probably much modified through changed conditions, and the peculiar position of the penis is possibly more the result of extreme specialisation than the retention of any primitive condition.

From all this it will be seen that the Typhobias can hardly be said to be archaic forms, but they do, as in the character of the nerves and the otocysts, possess some undoubtedly archaic characters. They are far less specialised in the characters of the foot and mantle than Strombus and Pterocera, to which in other respects they appear to be closely allied. They certainly possess none of those characters which would suggest that they can by any possibility be regarded as the persistent representatives of an old fresh-water stock. They do, however, simulate and retain the characters of the nerves of the Solarium and the Scalarids, and they probably indicate the road by which the more modern marine genera of the Strombidæ and their associates have been evolved. But to my mind the most remarkable features which they present are those which I have pointed out as indicating an approximation to several forms which have been generally regarded as recent productions; that is, they distinctly bridge the gap between several twigs which are well up in the phylogenetic tree.

Lastly it will have been seen that in many ways the Typhobias are self-contained, and have undoubtedly undergone individual specialisation of their own. It will therefore be most expedient, most natural, and most expressive of the actual anatomical facts, to separate these two genera of Typhobias as a family by themselves, the members of which have affinities with, and stand in the relation of forerunners of, those more modern forms which group themselves about the Strombidæ. They have been seen also to exhibit more or fewer of the characters of a wider range of forms, more especially of the Aporrhaidæ, Xenophoridæ, Cypræidæ, and that ill-defined group the Ptenoglossa.

For this family I propose the name Typhobiidæ; Typhobia Horei and Bathanalia Howesi represent the two generic forms at present known.

The Typhobias are intensely interesting forms ; their affinities show that they have without doubt been cut off from an exclusively marine stock at what is, geologically speaking, no very remote period of time.