The present paper forms the first instalment of the zoological report of an expedition which, through the generous support of the Royal Society and the British Association, I was able to make to Lake Tanganyika during 1895 and 1896.

The primary objects of this expedition were—

1. To study the unique fauna of Lake Tanganyika on the spot ;

2. To make what observations were possible in the Nyassa region while I was en route; and—

3. To bring back properly preserved material for the complete morphological investigation of the more remarkable lake organisms after I returned.

Before proceeding to the purely zoological matters with which I propose to deal, it is appropriate that I should here express my sincere thanks to Professor Ray Lankester, to whom I have been indebted for the primary suggestion of the whole inquiry, and for much kindly help since my return. I have also to thank Professor G. B. Howes for the use of the Huxley Laboratory and invaluable advice, without both of which I should never have been able to get the subject through, while I have been very materially indebted to Sir John Kirk, who procured for the expedition the necessary introductions to the administrative gentlemen through whose districts it had to pass. And last, but not least, I have to thank Sir Harry Johnston for the very effective support he lent the expedition, and without which it would have been impossible for me to attain the objects which I had in view.

Excluding the polar regions proper, there exists in the fresh waters of the different continents a type of fauna which in the character of its constituents is essentially the same. Certain forms are added and others are omitted as we pass from the more temperate to the equatorial zones, but beneath these changes there exists a substantial similarity, so easily recognisable and so marked that geologists have not hesitated to distinguish between fresh-water and marine fossiliferous deposits wherever they may be found. On the other hand, that there is a hard and fast demarcation between freshwater and marine faunas is not true, for there are many instances of animals—for example, of prawns and crabs, which in this country are purely oceanic—having made their way up the rivers into inland and elevated fresh-water lakes. Further, there are a number of animals that belong neither to salt nor fresh water, but are inhabitants of the brackish regions which lie between inland fresh waters and the sea. That the well-established and more permanent fresh-water organisms of the present day are descended from older phyla that were once marine, is an accepted truth. This view is necessitated by the theory of common descent, and it is supported, as in the case of the ganoid fishes, by the similarity of numerous living fresh-water organisms to older oceanic types. It is significant, however, that with few or no exceptions, all the well-established fresh-water organisms of to-day are not directly referable to the earliest oceanic forms, but rather to those which in their temporal distribution stand intermediate between then and now. It seems that the fresh-water molluscs of the present day first make their definite appearance in Tertiary times, for much doubt has recently been raised as to the genuineness of the so-called carboniferous Unio, Tichogonia, and Planorbis; these forms being now regarded as more nearly related to Anthracosia, Avicula, and Serpula respectively. The family Limnæidæ, which is now so universal in its distribution, does not certainly extend further back than the Jurassic period. The same is true of the fresh-water Melaniidæ and of the Paludinidæ, but I need scarcely point out that it is necessary to use the greatest caution in drawing any inferences respecting the date of origin of the true fresh-water forms from these apparent facts. It may, however, be taken as approximating to the truth to say that although the typical and universal fresh-water molluscs of the present do not appear upon the stage of life as such before the Jurassic period, they almost certainly originated from a series of marine types which had become completely differentiated from their oceanic associates long before this time. Some of these antecedent organisms are probably represented in the palæontological record by those extinct genera with which the earliest known modern fresh-water types are usually associated.1

The facts of morphology are themselves in harmony with such a view, for in their anatomy the living fresh-water molluscs do not approximate to any of the more modern marine genera; they hark back to those more permanent marine types which were in existence long before Jurassic times. They certainly bear no resemblance to the generalised conceptions or archetypes of the more modern marine genera which appeared during Tertiary and post-Tertiary times, such as Strombus, Pteroceros, Rostellaria, Conus, Mitra, Chenopus, and the like. It is this fact that is of first importance to us here ; for if it should be found that in some district at the present time there exists a fresh-water fauna which departs from the normal and universal type in the possession of genera which approximate to those that are undoubtedly modern and marine, we shall have very strong primâ facie evidence for regarding these organisms as recent importations from the sea. Now such a fauna is presented to us in that of Tanganyika at the present day, for in this lake there have been known to exist ever since 1859 what appear to be the shells of some six genera of Gastropods, which are entirely unlike any known fresh-water forms, while their shells at the same time simulate several modern oceanic types. The interest in these strange molluscs, which have been known hitherto only by their conchological characters, was greatly augmented when in 1883 Boehm found jelly-fish in the lake; and during my recent expedition I have been able to add deep-water crabs, prawns, sponges, 1 and Protozoa to this anomalous list of organisms, all of which appear to possess the same marine affinities.

It is my object in the present paper, to ascertain to what conclusions as to the nature and origin of this anomalous series the facts of distribution lead ; while in those which follow I shall deal with the morphological affinities of the hitherto unknown individual forms, and thus determine whether the conclusions to which these facts of distribution seem to point are really sound. Almost no definite observations have been hitherto available for the study of this subject, and consequently the material contained in this and the following papers will be mostly new. I would, therefore, invite particular attention to the positive character of the evidence which I shall bring forward in support of the recent marine origin of a number of the animals contained in Tanganyika, as compared with the wholly negative character of that upon which the geological speculations of Murchison respecting the “permanence of terrestrial conditions “in the African interior at present rest.

To believe that the marine animals of Tanganyika are among the few remaining indications of a sea that once extended to the very heart of the African continent is to come into the most uncompromising conflict with the theory put forward by Sir Roderick Murchison in 1852. Yet such a view is now supported by the strongest kind of zoological evidence it is possible to get.

In order to arrive at a sufficiently complete comprehension of the general type of the molluscan fauna which characterises the individual lakes, it is quite unnecessary and even prejudicial to discuss the question of the occurrence or non-occurrence of many of the so-called specific forms, since a large number of these are merely geographical varieties, while others have been based on minute and often purely fanciful conchological distinctions. The four species of Hylacantha, described by Bourguignat in 1890, for example, are certainly nothing more than the rather remarkable polymorphs of the original Typhobia Horei described by Smith in 1881. But what is true of the polymorphic Typhobia is equally true of the polymorphic Paramelania and Neothauma. I shall therefore consider the distribution of the genera alone, or the main issue will become lost in the pursuit of really non-existent types.

In Nyassa, which was the first great lake I reached, and which has hitherto been better known than all the rest, there have been recorded some sixteen genera of molluscs, namely,Limnæa, Isodora,, Physa, Physopsis, Planorbis, Ancylus, Ampullaria, Lanistes, Vivipara, Cleopatra, Bythinia, Melania, Spatha, Iridina, Corbicula, and U nio.

In the smaller lakes occurring in this district, such as Shirwa, there have been found a fewer number of the same genera. In Kela, which is within twenty miles of the south end of Lake Tanganyika, I found Planorbis and Limnæa. In Mwero there have been recorded Unio, Ampullaria, Lanistes, Vivipara,Cleopatra,Bythinia, and Melania; while in Bangweolo, according to Lieutenant Weatherly, there are no shelled molluscs, but it is hardly credible that their absence in this lake will be maintained. It is thus certain that the generic forms occurring in Nyassa completely cover the mollusca in a very large number of African lakes indeed. In the Victoria Nyanza all the Nyassan genera have been recorded, and more or fewer of the constituents of the same Nyassa list constitute the faunas of the remaining members of this more northern group of lakes,—as, for example, the Albert Nyanza, the Albert Edward, Beringo, and the like.

To facilitate comparison I have arranged the names of all the lakes about which anything definite is known in the tabular form given on p. 166. On the left-hand side will be found a list of all the genera hitherto known to be contained in each. From this table it will be seen that more or fewer members of the Nyassa list of genera are contained in every lake, but that there is a curious reduction of the number of the genera as we pass from the greater lakes to the less. This is probably due to the impermanence of the conditions in the smaller lakes, for we find in Shirwa, which is salt, and in Kela, which has periodically dried up, only those forms which can stand a wide amount of change. If we pass momentarily, however, from the study of the genera among the lakes to that of the specific forms, it will be found that there is a certain amount of variation in the specific representation in the genera contained ; and that when the lakes are widely separated—as, for example, Nyassa and Lake Mwero—such specific variations are often strongly marked. Judged, however, by the genera alone, it will be seen that there is a remarkable uniformity in the character of the African fresh-water molluscs over an immense area of ground. To this rule of uniformity in type which characterises the molluscan fauna of all the lakes about which anything is known, Tanganyika seems to form a solitary and striking exception. But the differences which this lake presents are in one sense more illusory than real, for on inspection of the table it will be seen that Tanganyika does contain, and fully represented, the great lake list of molluscs found in the Nyassa to the south, and the Victoria Nyanza to the north. It differs from the other lakes in there being here added to the otherwise universal list a number of entirely new forms. The genera which compose this superadded series comprise among others the six genera of Gasteropoda which have been known hitherto only by their empty shells, namely, Typhobia, Paramelania, the so-called Lythoglyphus of Tanganyika, Syrnolopsis, Nasopsis, and Limnotrochus. I was, during my recent expedition, enabled to add to this isolated, series at least two entirely generic forms, for which I have proposed the names Bathanalia and Bythoceras. We have, therefore, now in Tanganyika some eight genera of Gasteropoda which are not found in any of the other lakes, and to this isolated list of molluscs there should probably be further added among the Lamellibranchiata the so-called Unio Burtoni, and one of the Tanganyika Spathas. Consequently there are now known to exist in Tanganyika ten genera of molluscs, which appear to be restricted to the lake in which they were originally found.

Although I am concerned here primarily with the distribution of the molluscs in these lakes, it must be clearly understood that the marine organisms, such as jelly-fish, crabs, prawns, sponges, and Protozoa, with which the above molluscs are associated, share equally the same geographical limitation. The ten quasi-marine molluscan genera being, in fact, only one section of a complete fauna, containing widely separated types, which in Tanganyika exists along with the normal fresh-water stock the lake contains. The fauna of Tanganyika is thus a double series, and to distinguish its apparently marine constituents from the more normal lake animals I shall speak of them in future as the Halolimnic group.

Now it can be confidently affirmed that there are no Halolimnic animals in Nyassa, Shirwa, or Kela, all of which I visited and dredged ; and they are certainly not present in Bangweolo or Mwero. Yet these organisms are so conspicuous and common, when they do occur, that they would certainly have been recorded from the Victoria Nyanza, the Albert Edward, and the Albert, if any Halolimnic animals had existed in these more northern lakes. So far as is at present known, then, the Halolimnic fauna is entirely restricted to the confines of Tanganyika, in which lake it was originally found. The remarkable isolation and independence of the Halolimnic fauna which these facts disclose is of the first importance when we attempt to ascertain from what source they may have sprung; it is extremely important, therefore, that the conclusions which result from the study of the geographical distribution of these forms can be corroborated from another point of view.

TABLE I.

Showing the Molluscan Genera which have hitherto been recorded in the Principal African Lakes.

Showing the Molluscan Genera which have hitherto been recorded in the Principal African Lakes.
Showing the Molluscan Genera which have hitherto been recorded in the Principal African Lakes.

When comparatively examined, the observations which I was enabled to make respecting the bathymetric distribution of the molluscs in the lakes will be seen to show quite as clearly as the facts of their geographical distribution that the Halolimnic series is something entirely distinct from the normal fresh-water population of the lake. Through the kindness of Sir Harry Johnston I was enabled to attach myself to one of the Nyassa gunboats, and thus to become acquainted with the facts of molluscan distribution in Nyassa before I reached Tanganyika on my way north ; and the observations made on this voyage have been of the utmost value as material for comparison with what I subsequently saw.

Nyassa is a relatively narrow lake of great and unknown depth ; soundings of 300 fathoms, no bottom, having been obtained throughout a great proportion of its area. In total length it covers some 340 miles, and it varies from 20 to 40 miles across. The shores are of the most varied description, steep and precipitous in some places, in others bounded by extensive flats. The lake has a free outlet down the Shire River and the Murchison cataracts, the water being consequently clean and fresh. Owing to the lake’s great extent, the shores are often swept by a heavy surf, and the fairly strong currents which are observed are probably produced by the trade winds, for they seem generally on the surface to set from south to north. The fauna of such a lake is exposed to the same conditions as those to which it would be on an open oceanic coast. The enormous depth of the lake in many places rendered it impossible to dredge, and whether with an efficient deep-water apparatus anything further could be obtained from the abysmal recesses it is impossible to say. But it was very soon apparent that the molluscan portion of the population rapidly thinned out with increasing depth and distance from the coast, and that beyond 100 feet one could often dredge for miles over rocks and sand and mud without securing a single shell.

In many places the lake was floored with compact drifted masses of shells and shell-fragments, consisting chiefly of the Nyassa Viviparas. No sponges grew upon these shells, and heyond an occasional Melania such stretches of the lake were without life of any sort or kind. The curious diminution of the molluscs of Nyassa beyond the immediate coast-line is a striking feature of the lake throughout, and a tolerably correct idea of the bathymetric distribution of the individual genera will be gathered from the accompanying table, which contains an epitome of the observations made.

TABLE II.

Bathymetric Distribution of Molluscs in Nyassa.

Bathymetric Distribution of Molluscs in Nyassa.
Bathymetric Distribution of Molluscs in Nyassa.

The lines representing the bathymetric extent of each genus are thickened so as to show where it is found in the, greatest abundance, and approximately at what depth it ceases to exist.

From this table it will be seen that the purely fresh-water molluscan population of Nyassa is more or less completely restricted to a littoral band along the shores, and that the great majority of the genera do not extend in depth beyond 200 feet.

Thus Lake Nyassa, so far as molluscs go, is thinly peopled, and great extents of its shore and deep bottom are altogether uninhabited. The above facts of distribution, indeed, give the impression that the molluscan population of Nyassa retains its character of an importation from the ponds and streams which in the vast lake is, as it were, completely out of place.

Let us now turn to Lake Tanganyika, and compare the distribution of the molluscs in this lake with the observations I have just described in relation to Nyassa. But let me first point out that the physiographical features of Lake Tanganyika are slightly different from those obtaining in the great Nyassan valleys. Tanganyika is 2700 feet instead of 1500 feet above the level of the sea; but notwithstanding this greater elevation, the climate of Tanganyika is appreciably hotter than that of Nyassa.

The shores of Tanganyika are, perhaps, on the whole more precipitous, and are certainly less extensively fringed with the broad coast belts of modern alluvium so characteristic of the margin of Nyassa. The southern half of Lake Tanganyika is not nearly so deep as Nyassa, the water being generally little more than from 900 to 1200 feet in depth ; and it is not so pure, being always impregnated with an appreciable taint of several salts.

These slight physiographical differences which exist between Nyassa and Tanganyika cannot, however, be considered as having any potency to modify the individual fauna which the lakes contain. For there is a far greater difference between the physiographical features of Nyassa and the Victoria Nyanza than between Nyassa and Tanganyika. Yet the molluscan population of both the Victoria Nyanza and Nyassa are essentially the same. We cannot, therefore, regard the less amount of difference which is perceptible between Nyassa and Tanganyika as in any way responsible for the wide faunistic differences which exist between these lakes. This view is also in accordance with the very important fact that all the genera found in Nyassa exist in Tanganyika also. There is no more difference between the normal, non-Halolimnic molluscan population of Tanganyika and Nyassa than there is between that of Nyassa and Victoria Nyanza—a state of things which seems to indicate clearly that the difference in physiographical features which exists between these lakes are incapable of greatly modifying the forms they contain. Therefore whatever wide difference in type and mode of distribution is apparent between the faunas of Nyassa and Tanganyika must be the expression of the difference of the animals themselves, and not of the slight differences of condition under which they live.

In confirmation of this view, we find that all the Nyassan molluscan genera are distributed both in Tanganyika and Nyassa in a similar way, i. e. they are more or less restricted to the coast-line and the shallow, sheltered places, such as creeks and bays. For comparison my observations on the bathymetric distribution of the molluscs in Tanganyika have been epitomised in the same tabular form I used while speaking of Nyassa. From this table it will be clearly seen that if we exclude the Halolimnic fauna altogether, the bathymetric distribution of the molluscs both in Nyassa and Tanganyika is approximately the same. In both cases the fresh-water molluscs are restricted to the sheltered, shallow portions of the lakes. But directly we pass from the consideration of the normal fauna to that of the Halolimnic forms the most striking changes are at once observed. Instead of the Halolimnic molluscs being restricted to the shallow creeks and bays about the coast, they swarm on the rough surf-swept rocks and on the open beach. And what is more remarkable than this, they extend in great profusion to the deepest portions of the lake. Thus, dredging in water which varied in depth from 800 to 1200 feet, I always obtained plenty of Typhobia, Paramelania, Bathynalia, and Bythoceras among the Gastropods, as well as the so-called Unio Burtoni among the Lamellibranchiata ; and how far these genera extended beyond these depths I cannot say, but they showed no signs of dying out, but rather the reverse. On the lake floors which were not so deep as this, from 200 to 300 feet below the surface, but which were yet deep enough to have yielded nothing by dredging in Nyassa, there was an abundance of Limno-trochus, Syrnolopsis and Neothauma, together with those varieties of Melania which inhabit Tanganyika. It is thus rendered apparent by these observations that the Halolimnic molluscs are all either surf-swept rock dwellers, or entirely deep-water forms. Unfortunately we are as yet entirely ignorant of the distribution of the molluscs in any of the great lakes besides the two which I have named. But as the normal fresh-water fauna of Nyassa and Tanganyika have the same bathymetric distribution, it is probable that these same genera inhabiting the remaining lakes which have not yet been investigated will be found to be similarly disposed.

TABLE III.

Bathymetric Distribution of Molluscs in Tanganyika.

Bathymetric Distribution of Molluscs in Tanganyika.
Bathymetric Distribution of Molluscs in Tanganyika.

It is thus apparent that the Halolimnic molluscs are completely dissociated from the normal fresh-water forms, along with which they exist in Tanganyika, not only by their singular geographical isolation, but by their bathymetric distribution also ; the conclusions to which the facts of their geographical distribution seem to point being thus completely substantiated from another point of view.

There are, however, yet other ways in which the fact that the Halolimnic fauna is entirely distinct from, and unconnected with the more normal series becomes clear. For in many branches of biological inquiry we are often rightly guided by impressions which, like the types of human physiognomy, are real enough, but quite incapable of definite expression. Impressions of this character are at once produced on reaching Tanganyika, as I did, after studying the fauna of several neighbouring lakes. For there is a singular and oceanic profusion of life in Tanganyika, which is quite peculiar, and it quickly becomes evident that this numerical increase in the aquatic population does not affect the normal fresh-water stock, it is solely produced by the astonishing abundance of the members of the Halolimnic group. In contrast with the shallows of Nyassa, the creeks and bays of Tanganyika swarm with crabs and prawns, and the open sandy beaches are strewn with empty Halolimnic shells ; dead detached fragments of the deepwater sponges are tossed up by hundreds on the shore. And on the extensive rocky coasts the barely submerged stones are covered with the so-called Lithoglyphus and Nasopsis, just as the half-tide rocks, swarm with Natica and Litorina on an English beach. Further, on putting out into the lake itself, the deep open water is filled and discoloured with clouds of pelagic Protozoa (chiefly Peridinia and Condylostoma); and during the dry season swarms of the lake jelly-fish are seen pulsating at all depths.

Recapitulating, it may be said, then, that the facts of the geographical and bathymetric distribution of the great lake molluscs lead to the following results :—That among all the fresh-water lakes of the African continent which have hitherto been explored there exists a type of fauna which is curiously similar throughout. It differs only in the specific representation of the same genera which these lakes contain. This generalised African lake fauna contains only those families and genera of molluscs which would be regarded as typically fresh-water, lake, river, and pond dwellers, in whatever continent the fresh water might occur. In one African lake, however, but in one lake only, there have been found to exist, superadded to this normal lacustrine stock, a number of Gastropods which do not closely resemble any other forms either living or extinct ; these molluscs are also completely dissociated from the remaining normal series of the lake in which they occur by their modes of life. Together these molluscs constitute the molluscan section of a whole faunistic series, which in Tanganyika is added to the normal fresh-water stock the lake contains. This fauna forms what I have called the Halolimnic group, and the tout ensemble of all the Halolimnic genera is marine.

To account for the presence of the Halolimnic organisms in Tanganyika, only three hypotheses which are even temporarily tenable can be found. It may be supposed—

1. That they have arisen as modifications of the ordinary fresh-water fauua through prolonged isolation in the lake;

2. That they are the surviving representatives of an extinct fresh-water stock ; or—

3. That they are comparatively recent importations from the sea.

Let us examine each of these three possible explanations in the light of the new facts of distribution which have just been detailed. Unless the conditions affecting the fauna of Tanganyika have been permanent for a greater period of time than has been the case with any of the other lakes, they could not have produced the Halolimnic fauna which this lake now presents. Unless we make the further suppositions (1) that the conditions in Tanganyika have been permanent, while those affecting the fauna in all the other lakes have changed so much as to kill off the Halolimnic forms they once possessed ; or (2) that all the other lakes are much younger than Tanganyika, and that therefore the Halolimnic fauna has not had time to develop in them yet. There is no evidence for either of these views, and there is direct evidence to show that Nyassa has been a fresh-water lake longer than Tanganyika. On the shores of Nyassa there are old raised beaches, forming white limestone cliffs which contain the fossilised remains of the shells now living in the lake. But in these old lake beds there are no traces of any Halolimnic forms, and this is all the more conclusive as the shells of the Halolimnic molluscs are much more solid and durable than those of the fossilised fresh-water forms.

The second hypothesis, that which suggests that the Halolimnic fauna may be the surviving representative of an ancient fresh-water stock which has become extinct, has great attractions, as it conforms to a famous geological speculation, and has at first sight the appearance of a certain modicum of positive support. For the shells of the Paramelanias of Tanganyika have been independently supposed by White and by Tausch to be identical with the extinct estuarine or brackish Pyrguliferas of cretaceous Europe, America, or Africa.1

The type of shells possessed by these forms has been, however, repeated so often by so many widely separated molluscan types, such as in the Melanias, Litorinas, Pur-purinas, and the like, that it is pardonable if zoologists require something more than merely conchological characters to establish an identity among these forms.1 But this supposed homology between the shells of a living and extinct species of Gasteropod (about the anatomy of neither of which up to the present anything whatever has been known) is the one fragment of positive evidence which can be produced in favour of the relation of the Halolimnic fauna to an extinct fresh-water stock.

The hypothesis, moreover, is combated by the same objections emanating from the facts of distribution of the Halolimnic animals that were fatal to the first hypothesis, and they have here equal force. If the Halolimnic fauna of Tanganyika is the remnant of an old African fresh-water stock, it must have been present at one time in all the lakes which are as old as Tanganyika ; but we have seen that with respect to Lake Nyassa this does not appear to have been the case. It is very improbable that many of the remaining so-called rift-valley lakes are not as old as Tanganyika, yet we have seen that they do not contain the Halolimnic forms. Therefore, in order to support this second hypothesis, we shall be obliged to have recourse to hypothetical catastrophes which must be supposed to have destroyed the Halolimnic fauna in every lake but one. Hypotheses of this sort spring, however, from the carcass of a theory only after it is dead, and our second hypothesis is therefore opposed to the facts of distribution as they at present stand; its acceptance would, moreover, be revolutionary to many zoological conceptions of the present time. It would necessarily lead us to believe that deep-water crabs may be indigenous fresh-water forms ; that deep-water Gastropods and sponges were common in Cretaceous times ; that jelly-fish were once fresh-water organisms, and so on through a number of consequences, which, when the nature of the evidence supporting the original hypothesis is weighed, must seem little better than grotesque. On the other hand, all the facts of distribution and the like, as well as the superficial character of the Halolimnic animals themselves, are absolutely in accord with the third hypothesis, i, e. that the Halolimnic fauna is a relatively recent importation from the sea.

But before accepting this conclusion, as the natural teaching of the facts and observations which we have been discussing, it is absolutely necessary to be quite sure that in the nature of the country itself—that is, in the past geological history of Africa—there is nothing which renders impossible the realisation of such a theory in fact. Now, on turning to the geological aspect of the questions we have just discussed, it is apparent that there is an accumulation of negative evidence drawn from what is now known to geologists of the nature of the African interior, which, although it does not specifically favour the view of the ancient fresh-water origin of the Halolimnic forms certainly renders evident a gap in the confirmation of the theory of their marine origin.

In 1852 Sir Roderick Murchison 1 advanced the hypothesis that Africa, south of the Sahara, was a continent of great antiquity and simplicity, the greater part of which has never been changed or covered by the sea, at any rate since the age of the formation of the new red sandstone. This theory has appeared to be supported by the discoveries of Livingstone, Burton and Speke, and Speke and Grant, and it was finally re-advanced and summarised by Murchison in 1864,2 when he described this part of Africa as geographically unique “in the long conservation of ancient terrestrial conditions.” But he immediately fell into the now exploded error of assuming that “this impression is further supported by the concomitant absence throughout all the larger portion of this vast area, i.e. south of the equator, of any of those volcanic rocks which are so often associated with oscillations of the terra firma.” This latter speculation is now shown to be in no sense true, for there is abundant evidence of volcanic action and of volcanic materials all the way from Kilimanjaro and Ruwanzori in the north, to the little group of volcanic cones near the coast of Lake Nyassa. “The first part of Murchison’s theory, however, which affirms that Central Africa has never been below the level of the sea, is still in harmony with the known geological facts, for no deposits of a certainly marine origin have as yet been discovered in the interior.” The sedimentary rocks described by Burton and Speke to the west of the Victoria Nyanza have yielded no fossils to indicate the conditions under which they were formed. The triassic Ganoids and Gastropods unearthed by Drummoud1 at the north end of Lake Nyassa have been generally regarded as fresh-water forms.2 The great red sandstones and shales which stretch from the north of Nyassa far up the coasts of Tanganyika, which were examined by Joseph Thomson, and more recently by myself, have not yet been found to contain any animal forms ; the only indication which might lead to a belief that fossiliferous rocks occur in these regions being the fact that the natives of the west coast of Tanganyika are said to wear necklaces of beads which they dig out of limestone rocks, and which, if this statement is true, are probably the disarticulated segments of crinoid stems.

Marine, Jurassic, triassic, and probably carboniferous deposits have been found along the coast at many points from Mombasa to the Cape, but these have never been shown to extend any distance inland, and they seem to have no connection with the great sedimentary deposits of the interior, such as those north of Lake Nyassa, which underlie Drummond’s fresh-water triassic (?) beds. There is thus at present no geological evidence of the sea, or of even an arm of the sea, ever having been in the region of Tanganyika within reasonable geological times.

There has, however, been steadily accumulating a mass of observations relating to the formation of the so-called rift valleys, the general tenor of which has been to reveal a great instability of the regions in which Tanganyika lies ; an instability which has been quite sufficient to conceivably account for any amount of upheavals and depressions which may have been requisite for the marine contamination of that lake. The valleys in which the north of Nyassa, Lakes Tanganyika, Albert Edward, Albert, Baringo, Rudolph, and about twenty-seven minor lakes lie, are really part of a connected series of depressions formed by faults which run approximately north and south through an immense distance, and can be traced as far as Berbera on the Red Sea, thence north along the Red Sea shore itself, the coasts of which are to a great extent of similar formation, and they terminate finally in or about the Dead Sea, and the valleys of the tributaries of the Jordan. All the country traversed by this immense series of faults from the north of Nyassa in the south, to the ancient sites of Sodom and Gomorrah in the north, is filled with native traditions of catastrophes, of floods, of earthquakes, of volcanic outbursts, and the like ; and the geological investigations of Gregory 2 and others have shown that much of the above faulting and volcanic activity must have occurred, geologically speaking, in quite recent times. The existence of these singular rift-valley faults has divided the African lakes into two distinct series ; one series of lakes being always, like Nyassa, Tanganyika, and Rudolph, long, narrow, and deep ; the other, like the Victoria Nyanza, Bangweolo, and Shirwa, broad, shallow, and round.

We have, therefore, evidence of great geological instability in the very regions in which the Halolimnic animals now live ; but, like the palæontological record, it affords no insight as to how or when, if Tanganyika ever was connected with the sea, this connection could really have been made; but it is a singular fact that the one lake in which the Halolimnic animals now live is that which lies at the bottom of the biggest and most conspicuous inland rift.3

From all this it will be seen that, unless we are to assume that the Halolimnic group came into Tanganyika from the sea in very ancient times indeed, and that they are far older than their characters in any way appear to warrant,1 we are without any direct evidence from geology that the sea, or even an arm of the sea, has ever been in the Tanganyika region of the interior. So far as positive evidence goes, geology is absolutely silent upon this subject, it offers no evidence of any sort; and the theory of marine contamination—if it occurred, let us say, during a later period than Jurassic times—is thus diametrically opposed to a geological theory of the nature of the African interior which is at present accepted by many competent authorities.

The only way in which the nature and origin of the Halolimnic group can be really satisfactorily determined is, therefore, through a minute knowledge of the morphology of the individual members of the group themselves, and the best types belonging to the Halolimnic group for this kind of work are the Gastropods, because these organisms, unlike the lake Medusae, can be more or less directly compared with all sorts of analogous organisms, ancient and modern, fresh-water and salt. If it can be shown from the study of their morphology that the Halolimnic Gastropods in Tanganyika are really morphologically most closely related to the fresh-water Gastropods at present known, then the theory of the ancient fresh-water origin of the Halolimnic group is probably true. If, on the other hand, it turns out that the Halolimnic Gastropods are really most closely related to typically marine genera, then there will be little doubt that the Halolimnic group originated in the lake through marine contamination, and geological conceptions will have to make room for the fact of the interior of Africa having been connected with the sea as best they can.

In arriving at the conclusions contained in the preceding paragraphs I have virtually fulfilled the object which I had before me in collecting and examining the facts concerning the distribution of the African lake faunas, before entering upon any detailed examination of the evidence which can be gathered from the study of the morphology of the Halolimnic animals themselves. We have seen that the collateral evidence afforded by the facts of distribution and the like at once clear away the likelihood of the Halolimnic group having originated at any time or in any manner, de novo, in Tanganyika, and that there is finally brought on a more or less direct issue between the supposition of an ancient marine contamination of Lake Tanganyika and the ancient fresh-water origin of the Halolimnic group. All the facts of distribution which we have examined appear to me to strongly favour the former of these hypotheses; and although we are at present ignorant of the precise manner in which the marine contamination of Lake Tanganyika may have been effected, there is no positive geological objection to the view that it has occurred, while there is the certainty of a sufficiently great geological instability throughout the very districts in which Tanganyika lies to have easily accounted for it.

1

It is quite possible that many of the old so-called fresh-water deposits are in reality marine, since the forms which became exclusively fresh water as time went On probably made their appearance in the sea first, as so many of the more recently derived fresh-water types have done,—prawns, for example.

1

The affinities of the deep-water sponge which I obtained in Tanganyika have not yet been determined, but its striking external character and its remarkable deep-water habitat have inclined me to regard it as a member of the anomalous section of the fauna which the lake presents.

1

‘ Ann. des Sci. Nat.,’ septième série, ix, x, 126, 1890.

1

On further examination it appears—(1) That Hie genus Paramelania of Tanganyika is similar to the cretaceous Pyrgulifera; (2) but that the genus Pyrgulifera, so far as some of its representatives go, is oonchologically indistinguishable from the old marine Jurassic genus Purpurina, and that the Nanopsis of Tanganyika corresponds to one section of this genus, the Paramelania to the other. (See Hudleston’s figs., Plates I and ii, and text p. 85—95, ‘ Jurassic Gasteropoda,’ Palseontographical Society, vol. xli, 1887.) It would thus appear that the marine genus Purpurina became a fresh-water form, as so often happens in Cretaceous times. We find, however, that other Halolimnic Gasteropods, Bathanalia, the socalled Lithoglyphus, and Limnotrochus, are also indistinguishable from marine Jurassic forms, which are not found in any Cretaceous formation, fresh-water or marine. Consequently the geological evidence on this matter distinctly favours the old marine origin of the Halolimnic fauna; but it places their original marine existence much further back than I had even dared to suggest. I shall discuss this most interesting line of investigation fully in a special memoir.

1

I am quite aware that this statement cuts at the roots of many geological determinations ; but I am prepared to maintain that the criticism is sound.

1

Murchison, President’s Address, ‘ Journ. Royal Geog. Soc.,’ vol. xxii, 1852.

2

‘ Jouru. Royal Geog. Soc.,’ vol. xxxix, 1864, pl. xxxvii, pp. 201—205.

1

Drummond, ‘ Tropical Africa.’

2

It appears, however, that these fossils have been by no means satisfactorily described.

1

See Suesa,’ Die Brücke des Oust Afrika.’

2

The Great Rift Valley.

3

The southern two thirds of Nyassa is not in a rift; and in contrast to Gregory I do not believe that the north of Nyassa lies in the main eastern rift, but in one which through Lake Rukwa is continuous with the western Tanganyika series. See Gregory’s ‘ Rift Valley,’ p.’ 7 ; also my paper in the * Journal of the Royal Geographical Society,’ September, 1897.

1

See foot-note on page 161.