SMITHSONIAN INSTITUTION LIBRARIES Gift of R. S. HOUBRICK yd C DWIGHT W. TAYLOR H-Z % yy. a*. 7 ''tf. d \ £ rr* ■£■ s+'b THE TANGANYIKA PROBLEM / OMARI BIN OMARI, HEAD MAN OF THE SECOND TANGANYIKA EXPEDITION. ( From a Sketch bt the stuthor.) Frontttptece . The Tanganyika Problem cAN cACCOUNT OF THE RESEARCHES UNDERTAKEN CONCERNING THE EXISTENCE OF SMARINE cANIMALS IN CENTRAL cAFRICA By J. E, S. MOORE, F.R.G.S. Author of “To the Mountains of the Moon" WITH NUMEROUS ILLUSTRATIONS AND MAPS LONDON HURST AND BLACKETT, LIMITED 13, GREAT MARLBOROUGH STREET, W 1903 cAll rights reserved PRINTED BY KELLY’S DIRECTORIES LTD. LONDON AND KINGSTON. PREFACE. — ♦ — In completing the present account of the researches which were undertaken during the two Tanganyika expeditions, I am somewhat oppressed with the possibility that the unique opportunities and material available may not have received the justice they would have done in the hands of many naturalists who have more ex- perience than I am old enough to possess. We were actually the last zoological explorers with a great untouched field before them, and when it is remembered, that among other things, we have had to deal with nearly 200* entirely new animal types discovered during these expeditions, it will be admitted that the labour of giving an adequate account of the faunistic characters of the great African lakes has not been exactly child’s play. It should, moreover, be clearly understood that the acquisition of all this new material and information respecting Central Africa has been primarily due to the instigation and persistence of Professor Ray Lankester, and I only wish that he, instead of myself, could have found time to write the present account. On the other hand, it would perhaps have been impossible for anyone who was without a somewhat prolonged and actual acquaintance with the African interior, to attempt to deal with all the aspects the Tanganyika problem has been found to present, and I can therefore only point out to any who may feel that the present work could have been better done, that I am fully conscious of its peculiar deficiencies. Few people here, moreover, will be able to appreciate how com- pletely during the actual operations of exploring parties in practically unknown countries, whatever results may be obtained are due to the disinterested and gratuitous help which may or may not be extended to them by people, administrators and the like, who live * In this connection it may be interesting to remember that the whole of the fishes, fresh water and marine, belonging to the British Islands, only represent about 240 species. VI THE TANGANYIKA TROBLEM. upon the borders of such unknown lands. In the present case, all our results with respect to the fauna and topography of Xvassa, and much of the efficiency of the expedition further north, was due to the efforts of H.M. Commissioner, Mr. Alfred Sharpe, who allowed me to examine the lake from the gunboats, and facilitated our journey in many ways as far as the northern boundary of the British Central African Protectorate. In this way we were able to ascertain the actual depth of Lake Nyassa— 430 fathoms — a result which is sufficiently surprising, and one which, I am happy to say, has since been closely confirmed in an adjacent area by Commander Rhoades. In like manner we received much help in the selection of men and kindness generally from General Creagh, the British Resident at Aden ; from Mr. Codrington in Northern Rhodesia ; and perhaps I might also say we have been particularly indebted to Captain Bethe, and his brother officers in command of the German territory north and east of Tanganyika. In the later stages of the journey a great quantity of animals in spirit, and what remained of things in general, were hurried safely through Uganda to the coast by the help of H.M. Commissioner Mr. Jackson and Mr. Pordage at Mtebi. Turning from these details of the route to the subject of the investigations themselves, I should say at once that all the geo- graphical work, which has resulted in the first effective mapping of the region that had so long remained in dispute between Tan- ganyika and the Albert Nyanza, has been due to the energy and skill of my colleague Mr. I'ergusson, and in like manner much of the geological data with which I have had to deal in the sequel was originally collected and arranged by him. The revision of the existing geological conceptions of the African interior, which the information obtained during the Tan- ganyika Expeditions necessitates, will possibly be found the most generally interesting result that these investigations have pro- duced. It will be seen in the succeeding chapters that the whole of our current views of the geological nature of the Continent are radically unsound and incorrect. It can indeed be shown that even Suess’ conception of the graben (rift valleys) by no means always expresses the actual structure of these remarkable depressions, they are far better described under the title of PREFACE. vii “ Eurycolpic folds,” a term recently suggested to me, and which I have intentionally introduced into the present work in preference to either “ graben,” or “ rift valleys,” both of which are obviously misnomers, since the valleys in question are generally produced by folding, due to lateral pressure, and not either through rifting, or vertical collapse. Since our return, and in connection with both the first and second Tanganyika expeditions, the large number of new fishes obtained have been elaborately worked out by Mr. Boulenger of the British Museum. The descriptions contained in the present volume being his simply, while the more conspicuous genera and species have been accurately illustrated by Mr. Green. The Crustacea were examined by Messrs. Cunnington and Caiman, the Sponges by Mr. Evans, Professor Minchin and Herr Weltner in Berlin. For the description of the Mollusca, the Polyzoa and Protozoa, I myself have been responsible, but I have received unlimited help from Professor Ray Lankester himself, Mr. Edgar Smith and others at the British Museum. The anatomy of the cele- brated jelly-fish was examined some years ago by Mr. Gunther in Oxford, and in what appears in the present volume, I have consequently been able to draw largely upon his original descrip- tion, only adding some drawings of the living Medusa together with information about its life history and development, which I acquired on the spot. We have been further indebted to Mr. Prior for the complete identification of the rock specimens collected by Mr. Fergusson ; and, as 1 have explained in the text, we have been most materially helped by Mr. Huddleston in the comparison of the Tanganyika gastropods with the shell remains of the Jurassic seas. Last, but not least, I have to thank Professor G. B. Howes, who after my return from both my first and second Tanganyika expeditions, allowed me the unlimited use of the Huxley laboratory in the Royal College of Science and of his own experience and advice. The title which I have chosen for the present work, “ The Tan- ganyika Problem,” expresses the fact that there is a puzzle, a mystery, attached to Tanganyika, and the elucidation of this mystery has formed the single motive for several independent lines of enquiry described in the present volume. The Tan- THE TANGANYIKA PROBLEM. viii ganyika problem thus lends itself to the formation of a connected story, and I can point out to the general reader, who may be attracted to the puzzle as he would be to a cock-fight, that he need not be alarmed at technicalities to follow, these having been intentionally rendered conspicuous for omission in small type. Moreover to those who know Africa, I think it may be positively refreshing to find that the interior has actually got some attributes besides a continued absence of gold, and a total unsuitability for European colonization. Further, since the present volume contains the only general, illustrated account of the animals found in the great lakes, it is hoped that it may be really of use to numbers of people, who in spite of attributes, still go out to the African interior, but who up to the present time have had no means of ascertaining the nature of the sometimes extremely good fishes they are continually eating. Of the manner in which the second Tanganyika expedition was organised, and of the method of treating the Tanganyika problem which I have adopted in the present volume, it is not necessary to speak. I have already dealt with these matters in the intro- ductory chapters ; but I may point out that I have intentionally introduced as many illustrations and photographs of the localities described as possible, since it seems to me that no one can form any clear idea of the physical characters of an extensive area from maps and descriptions alone. So again, in chapter six, I have re-introduced matter relating to the formation of Park lands upon alluvial flats, which I had already published in a former work ; while lastly, owing to a misapprehension under which Sir Harry Johnston labours with respect to the relation of his own to other explorations in the Mountains of the Moon, I have had to refer to this range at greater length than I originally intended ; but I do not wish it to be understood that I undervalue Sir Harry’s work in this direction ; on the contrary, I think it distinctly interesting and valuable, since, although he did not reach a?- high an altitude as myself, in all major points he has completely confirmed the results which I had previously attained. October , 1902, Londoti. CONTENTS. — ♦ — CHAPTER I. INTRODUCTORY. PAGE Reasons for acquiring a better knowledge of the fauna of the Great African Lakes — Our first knowledge of the existence of unique animals in Tanganyika arose as a by product of Burton’s journey — The shells collected by Speke on Tanganyika were marine in aspect — But the shells obtained from the other African Lakes were not so — Extension of our knowledge of the Tanganyika shells — Possibility of the lake having been connected with the sea — Intensification of the interest in the Tanganyika problem produced by Boehm’s discovery of jelly-fishes — • Formation of the first Tanganyika expedition — As a result of this it is shown that the Tanganyika problem is larger than was supposed — Conflict between the zoological evidence and geological anticipations — Erroneous character of Murchison’s geological speculations — Actual impressions of the African in- terior— Remarkable character of the Halolimnic shells of Tanganyika — Their similarity to those of the Jurassic seas — This comparison a possible solution of the Tanganyika problem — Insufficiency of our knowledge of the Great African Lakes — Suess’ views — Possibility of a former extension of the sea into Tan- ganyika from the north — Geological and geographical interest of the country north of Tanganyika — Formation of the second Tanganyika expedition — Arrangement of matters contained in the present work ..... I CHAPTER IE ON THE NATURE AND ORIGIN OF FRESH-WATER FAUNAS AND THEIR RELATION TO THE FAUNA OF THE SEA. Animals live in fresh-water which do not inhabit the sea — Animals live also in brackish water — It is generally accepted that fresh-water organisms have arisen from marine forms and have colonized fresh-waters from the ocean — Recent additions to our knowledge of fresh-water faunas — Beaudants’ experi- ments— Views of Semper Sollas and Von Martens — Apparent impossibility of many marine organisms eyer getting into fresh-water — Inadequacy of existing views respecting the nature of fresh-water faunas — Similarity of fresh-water faunas all over the world — Distinction of primary and secondary fresh-water stocks — Origin of the second obvious, of the first not so clear — Ancestral X THE TANGANYIKA TROT LEM. PAGE characters of primary fresh water stocks — Palaeontological record of the appear- ance of the primary fresh-water stocks — This, synchronous with other geological phenomena — The universal distribution of the primary fresh-water stock inex- plicable on the assumption of migration — Want of evidence of any capacity for migration — The zoological facts seem to suggest a former difference in the sea itself — Increase of salt in fresh-water shown to be prejudical to certain animals — Effects of this as exhibited in Lake Shirwa — Artificial production of similar results — Experiments of Loeb, Ilerbst, and others, show that a change in the character of sea-water may produce variation — General evidence for a continually increasing salinity in the sea — The action of this in the production of modern fresh-water faunas — Present view of the nature of fresh-water faunas .............. 1 1 CHAPTER III. THE PHYSIOGRAPHICAL FEATURES OF THE AFRICAN INTERIOR. THE GREAT CENTRAL RANGE. Necessity of a knowledge of the topography of the Great Lake Region — Our views are changing with respect to the interior — Recent observations lend no support to the older conceptions which sprang from Murchison’s erroneous hypothesis respecting African stability — Demonstration of the existence of a Great Central Range a chief result of the Tanganyika expeditions — Characters of this range — Scott Elliot’s early appreciation of its existence — The east and west drainage slopes — The existence of deep valleys (Eurycolpic folds) parallel with the range — These, rock valleys, not formed by ice — Disturbance of old aqueous deposits, caused by the formation of the Great Central Range — Character of the Great Central Range in the region of Nyassa — Absence of true volcanoes along the ridges of the range — Wide effects of the formation of the Great Cen- tral Range — The “graben” of Suess ; Eurycolpic folds — Characters of these folds in the region of Nyassa — Characters of the folds in the region of Tan- ganyika— Disturbance of the valleys and ridges in relation to the Great Central Range ............ 31 CHAPTER IV. SUPERFICIAL GEOLOGY OF THE REGION OF THE GREAT AFRICAN LAKES. Evidence respecting the nature of the central African region before the formation of the Great Central Range — Continuation of the movements which have pro- duced the Great Central Range — Recent discover)- of vast aqueous deposits in the interior — Value of these in ascertaining the past history of the continent — Method of treating the subject in relation to the observations made during the Tanganyika expeditions — Vastness of the regions to be considered — Character of the southern section of the region of the Great African Lakes — Nature of CONTENTS. xi PAGE the Shiri Highlands — Nature of the south Nyassa region — Changes in the geological character of this region in the latitude of Mount Waller — Extent of the Mount Waller sandstones — Relation of these deposits to the other African sediments — Later deposits above them — Drummond’s beds — -African lake pleistocenes — Arrangement of these different deposits — Extent of Drum- mond’s beds — Vast extent of the old African sandstones — Demonstration of the existence of a great depression in the region of the lakes before the forma- tion of the Great Central Range — The broad features of the past history of these regions ............ 54 CHAPTER V. THE GEOLOGICAL TOPOGRAPHY OF THE REGION NORTH OF TANGANYIKA TO THE ALBERT NYANZA. Want of knowledge of the districts north of Tanganyika — Importance of these districts in the consideration of the Tanganyika problem — Wide geographical effects produced by the recent formation of the Mfumbiro Mountains — Character of the country north of Tanganyika to Kivu — Former short exten- sion of Tanganyika north — Absence of any former connection of Tanganyika with Kivu — Characters of the Rusisi channel — Characters of the Kivu basin — Peculiarity of the Kivu water — The north shore of Kivu and the Mfumbiro Mountains — The volcanoes form a dam across the lake valley — Character of these mountains — The valley of Kivu continued beyond them northwards — The fauna of Lake Kivu is totally different from that of Tanganyika, but similar to that of the Albert Edward Nyanza — Former connection of Kivu and the Albert Edward — The volcanoes have resulted in the outflow of the Kivu being turned into Tanganyika and a corresponding shrinking of the waters of the Nile — The effect of this on the districts north of Kivu — The effect of this on the districts south of Kivu — Relation of the Mountains of the Moon to the Great Central Valley of the Lakes — Structural peculiarities of these mountains They are an accentuation of the folding characteristic of the region of the Great Central Range — Misapprehensions of Sir Harry Johnston ... 76 CHAPTER VI. AFRICAN PARK LANDS, THEIR APPEARANCE ON ALLUVIAL FLATS CONSIDERED AS EVIDENCE OF RECENT PHYSICAL CHANGE. Characters of African Park lands — Artificial appearance of these districts — Their vast extent — Impermanance of Park scenery — A park is an artificial product — Relation of African parks to alluvial plains — Zoned character of the vegetation on freshly-formed alluvial plains — The relation of Euphorbias to bush patches — Gradual conversion of bush patches into forest — The production of an African park marks a phase in a gradual physical change . . . .107 TI1E TANGANYIKA PROBLEM. xii CHAPTER VII. GENERAL OUTLINE OF THE ZOOLOGY OF THE GREAT AFRICAN LAKES. The thirteen Great African Lakes, the fauna of which is more or less completely known — The lakes examined during the two Tanganyika expeditions — Other sources of information — The fauna of Lake Nyassa — Some physical characters of the Nyassa valley — Limitations of the Nyassa fauna— It is that of a great pond — The fauna of Lake Shirwa — Kela — Rukwa — Bangweolo — The fauna of Lake Mwero — The fauna of Lake Kivu — The fauna of the Albert and the Albert Edward Nyanzas — The fauna of the Victoria Nyanza — Nature of the Nyanza depression — The fauna of Lake Rudolf — The fauna of Lake Tan- ganika — Peculiarities of the fish fauna of Lake Tanganyika — The fauna of the great rivers of Central Africa — The distinctive elements of the fauna of Lake Tanganyika — Definition of the Halolimnic group . . . . . .120 CHAPTER VIII. ON SOME CURIOUS FEATURES OF THE DISTRIBUTION OF SPECIES. General poorness of the African lake faunas — The faunas of the different lakes are specifically distinct — The faunas of these different lakes do not appear to inter- colonize — The number of species in the Great African Lakes is directly pro- portional to their size — Unique climatic conditions affecting the Great African Lakes — Similarity of lacustrine faunas to island floras — On the causes which tend to the production of species characteristic of the different depressions . . . . . . . . . . . . .144 CHAPTER IX. THE FISHES OF LAKE TANGANYIKA. Descriptions of the eighty-seven species hitherto obtained • 152 CHAPTER X. THE MOLLUSCS OF LAKE TANGANYIKA. The normal and Halolimnic gastropods — Details of the structure of the Ilalo- limnic forms . . . . . . . . . . . .217 CHAPTER XI. THE AFFINITIES OF THE HALOLIMNIC GASTROPODS. Inter-relationships of the Halolimnic forms — Their primitive characters — They are not comparable to any existing fresh-water forms, but they structurally ante- cede a number of existing marine genera — The Halolimnic gastropods present the appearance of an old marine relic 266 CONTENTS. xiii CHAPTER XII. THE CRUSTACEA OF LAKE TANGANYIKA. The crabs — Platythelphusa and Limnothelphusa — The prawns PARE • 279 CHAPTER XIII. THE NEW TANGANYIKA POLYZOAN. Arachnoidia ray la>ikesteri ........... 295 CHAPTER XIV. THE TANGANYIKA MEDUSA. General structure of this organism — Its apparent life history .... 298 CHAPTER XV. THE SPONGES AND PROTOZOA OF LAKE TANGANYIKA. Spongilla moorei- — Affinities of Spongilla moorei — Spongilla tanganyikce — Affinities of Spongilla tanganyikce — Potamolepis welteneri — Livingstone’s scum . . 309 CHAPTER XVI. GENERAL CONSIDERATION OF THE NATURE OF THE HALOLIMNIC FAUNA. Recapitulation of certain matters discussed in earlier pages — The geo- graphical isolation of the Halolimnic group — Its coexistence with the normal African fresh-water fauna — The distinctive character of the Tan- ganyika fishes — The general characters of the components of the normal Tanganyika fauna — These are not structurally intermediate between those of the general African fresh-water fauna and the Halolimnic group — The Halo- limnic gastropods cannot be regarded as derivitives from any recognised fresh-water types — Similar reasoning applies to the Tanganyika jelly-fish — Impossibility of regarding the marine characters of the Halolimnic group as having been brought about through convergence — Impossibility of regarding the Halolimnic group as the relic of an old fresh-water stock — The presence of the Halolimnic animals necessitates the supposition of a former extension of the sea into Central Africa, and thereby throws light into the geological history of the continent 325 XIV THE TANGANYIKA PROBLEM. CHAPTER XVII. GENERAL CONSIDERATION OF THE NATURE OF THE HAL0LIMN1C On some peculiarities of the fish-fauna in lake Tanganyika — Relation of some of the characteristic forms to both sides of the Atlantic — The character of the fishes of Tanganyika indicate that part of the fish fauna may have originated from a western sea — Peculiarities of the Congo fauna — The Congo depression pro- bably the remains of a former extension of the sea — The age and type of the Halolimnic fauna — Comparison of the hard parts of the Halolimnic animals with fossiliferous remains — The identity between the shells of the Halolimnic gastropods and a corresponding number of forms from the Jurassic seas — Details of this comparison — General consideration of the value of this com- parison— No valid geological objection to it — The number of corresponding points renders it improbable that the correspondence is a coincidence — The Halolimnic fauna would thus appear to be the remains of that of a Jurassic sea — The characters of the other Halolimnic animals in accord with this view — Recapitulation ■ 339 LIST OF ILLUSTRATIONS. — ♦ — Head man of the Second Tanganyika Expeditions. — Frovi a Sketch by the Author ........ Frontispiece CHAPTER II. PAGE Lake Shirwa, looking west from the middle of the lake. — From a Sketch by the Author . . . . . . . .22 View of the Zambesi and Shupanga, looking south-west. — From a Sketch by the Author ...... F'acing 30 CHAPTER III. View of Mount Morumbala, from the Shiri river. — From a Sketch by the Author ........ Facing 34 View over the town of Ujiji and across Lake Tanganyika. — From a Sketch by the Author . . . . . . . 37 View across the great central eurycolpic fold, from the northern slopes of the Mfumbiro Mountains . . . . . 41 Diagrams of the geology of the Nyassa Valley . . . Facing 42 Diagrams of the Nyassa and Tanganyika Valleys . . Facing 44 The east coast of Nyassa . . 45 View of a portion of the Livingstone Range, from Nyassa . . 47 CHAPTER IV. View over the alluvial flats at the north of Lake Nyassa. — From a Sketch by the Author ...... Facing 56 View of one of the alluvial flats in the region of South Nyassa . . 59 View over the forests on the Nyassa Tanganyika plateau. — From a Sketch by the Author ........ 63 XVI LIST OF ILLUSTRATIONS. PAGE View from the top of the sandstone cliffs flanking the south-west coast of Tanganyika. — From a Sketch by the Author . . . 67 View from the southern end of Lake Tanganyika. — From a Sketch by the Author ........ Facing 68 CHAPTER V. View of the south end of Lake Tanganyika, from Kituta. — From a Sketch by the Author . . . . . . . .78 The great active cone of Kirungo cha Gongo. — From a Sketch by the Author ......... Facing 82 Diagram of Tanganyika and Kivu ..... View from the north end of Lake Kivu of the great active Facing cone of 83 Kirungo cha Gongo ....... Bird’s-eye view of the Mfumbiro Mountains and the north end of 87 Lake Kivu ........ Facing 88 View from the summit of Kirungo cha Gongo . 91 View of the steep west coast of the Albert Edward Nyanza View up the higher portion of the Mobuko Valley. — From Sketch 93 by the Author ........ 95 The snow summit of Ngomwimbi ..... 99 The northern snow ridge of Ngomwimbi 103 The broad glacier on the northern snow ridge of Ngomwimbi io5 Diagram of the region north of the Albert Edward Nyanza Facing 106 CHAPTER VI. View of the snow peak of Kanyangogwi .... Facing 108 Semi-diagram of the formation of Park-lands . hi CHAPTER IX. Protopterus aethiopicus and Polypterus congicus *54 Citharinus gibbosus ....... 156 Barbus platyrhinus ........ 157 Barbus altianalis ........ *59 Barbus tropidolepis ........ 161 Capoeta tanganicae ........ 163 Barilius tanganicae ........ 1(3 5 Synouontis granulosus . . . . : 167 LIST OF ILLUSTRATIONS. xvii PAGE Cbrysichthys brachynema . . . . . . . .169 Lates microlepis . . . . . . . . . *171 Tilapia labiata . . . . . . . . . .173 Synodontis multipunctata ...... Facing 174 Paratilapia calliura, Ectodus longianalis and Tilapia boops . 175 Paratilapia furcifer . . . . . . . . . 177 Trematocara marginatum. . . . . . . . *179 Lamprolagus modestus and Haplochilus tanganicanus . . 181 Bathybates fasciatus . . . . . . . . -183 Bathybates ferox ........ Facitig 184 Judilochromis ornatus . . . . . . . . -185 Eretmodus cyanostictus, Telmatochromis vittatus and Lamprologus fasciatus . . . . . . . . . . .187 Grammatotria lemairii . . . . . . . . .189 Paratilapia vittata and Xenotilapia sima ...... 191 Gephyrochromis moorei, Ectodus melanogenys and Tilapia grandoculis 1 93 Tilapia trematocephala, Paratilapia aurita and Paratilapia nigripinnis 195 Petrochromis polyodon . . . . . . . . -197 Lamprologus lemairii, Asprotilapia leptura and Lamprologus furcifer 199 Tilapia pleurotaenia ......... 200 Tilapia dardennii ........ Facing 200 Trematocara unimaculatum and Paratilapia leptosoma . . . 201 Lamprologus moorei ......... 203 Telmatochromis temporalis ........ 204 Tilapia microlipis and Tilapia rubropunctata . . . Facing 204 Lamprologus compressiceps, Paratilapia pfefferi and Ectodus longianilis 205 Paratilapia dwindti, Xenotilapia ornatipinnis and Barbus serrefer . 206 Tilapia microlepis and Barilius moorei ...... 208 Barilius moorei and Tropheus moorei ...... 209 Paratilapia stenosoma . . . . . . . . .211 Lamprologus elongatus, Mastacembelus taeniatus and Perissodus microlepis . . . . . . • • • - 213 Mastacembelus frenatus and Mastacembelus moorei . . . .215 CHAPTER X. Melania admirabilis, Fig. 1 . . . . . . .219 Living Typhobia horei, Fig. 2. . . . . . . .221 Semi-diagram of anatomy of Typhobia horei, Fig. 3 . , . .222 B xviii LIST OF ILLUSTRATIONS. PAGE Lingual dentition of Typhobia horei, Fig. 4 . . . .223 Semi-diagram of anatomy of Typhobia horei , Fig. 5 . . . .224 Section of otocyst of Typhobia horei , Fig. 6 . . . .225 Part of the mantle cavity of Typhobia horei , Fig. 7 . . . .226 Bathanalia howesi shell, Fig. 8 . . . . . .227 Operculum of Bathanalia , Fig. 9 . . . . .227 Lingual teeth of Bathanalia, Fig. 10 . . . . . .228 Shell of Chytra kerkii, Fig. 1 1 . . . . . . .229 Living animal of Chytra , Fig. 12 . . . . . . 229 Nerves of apporrhais pes pelicani, Fig. 13. . . . . .230 Lingual teeth of Chytra, Fig. 14 . . . . . -231 Nerves of chytra, Fig. 15 ........ 232 Operculum of Chytra, Fig. 16. . . . . . . *233 Shell of Limnotrochus , Fig. 17 . . . . . . >233 Operculum of Limnotrochus, Fig. 18a . . . . . . 233 Lingual dentition of Limnotrochus , Fig. 18 . . . . 234 Nerves of Limnotrochus , Fig. 19 . . . . . . -"235 Stomach of Limnotrochus , Fig. 20 ...... 237 Shell of Bythoceras , Fig. 21 . . . . . . . 238 Lingual dentition of Bythoceras , Fig. 22 . . . , . . 239 Nerves of Bythoceras, Fig. 23 . . . . . . . . 240 Bythoceras minor shell, Fig. 24 ... 242 Lingual dentition of Paramelania damoni, Fig. 25 . . . . 243 Lingual dentition of Paramelania cranigranuluta, Fig. 26 . . 244 Shell of Paramelania damoni, Fig. 27 . . . . . 242 Shell of Tanganyicia rufofilosa. Fig. 2S . . . . . . 246 Semi-diagram of Tanganyicia rufofilosa, Fig. 29 . . -247 Lingual dentition of Tanganyicia rufofilosa. Fig. 30 . . . . 248 Nerves of Tanganyicia rufofilosa, Fig. 31 . . . . . 248 Semi-diagram of Tanganyicia rufofilosa , Fig. 32 ... 249 Shell of Nassopsis fiana, Fig. 33 . . . . . . .250 Semi-diagram of Nassopsis nana, Fig. 34 . . . . *. 25 1 Animal of Nassopsis, Fig. 35 . . . . . . . -252 Lingual dentition of Nassopsis nana, Fig. 36 . . . . . 253 Nervous system of Nassopsis, Fig. 37 . . . . . *254 Operculum of Nassopsis, Fig. 38 . . . . . 255 Shell of Spekia zonata, Fig. 39 . . . . . . .256 Lingual dentition of Spekia zonata, Fig. 40 . . . -257 LIST OF ILLUSTRATIONS. xix PAGE Nervous system of Spekia zonata, Fig. 41 ..... 258 Semi-diagram of Spekia, Fig. 42 . . . . . . .259 Semi-diagram of Spekia, Fig. ^3 ...... 260 Shells of Neothauma, Fig. 44 . . . . . . .261 Nervous system of Neothauma, Fig. 45 . . . . . . 263 Nervous system of Vivipara, Fig. 46 . . . . . .264 CHAPTER XL Diagram of molluscs . . . . . . . . -273 Diagram of molluscs . . . . . . . . - 275 CHAPTER XII Limnothelphusa maculata, Fig. 1 . . . . . .281 Platythelphusa armata. Fig. 2 . . . . . . .285 Limnocaridina tanganyikice, Fig. 3 . . . . . . .287 Palaemon moorei , Fig. 4. . . . . . . .293 CHAPTER XIII. Arachnoidiae ray lankesteri ........ 296 CHAPTER XIV. Living Tanganyika medusa ........ 299 „ „ „ , another 301 Semi-diagram of medusa buds ........ 303 Young bud 305 „ , rather older ......... 306 „ , still older ......... 307 CHAPTER XV. Spotigella moorei . . . . . . . . . .311 Skeleton of Spongella moorei . . . . . . . .313 Spongella tanganyikce . . . . . . . . .319 Potamolepis veltneri . . . . . . . . . 323 CHAPTER XVII. Pyrgulifera humerosa . . . . . . . . 343 Paramelania and pur purina . . . . . . • 345 Nassopsis and purpurina. . . . . . . . *34 7 XX LIST OF ILLUSTRATIONS. PAGE Amberleya and Bathanalia , 348 . . 348 Limnotrochus and Jurassic shells . . 349 Chytra and 0 nust us ....... . 35° Spekia and Niridomus ....... . 35i Melania admirabilis ....... * * 353 MAPS. Map of new positions on Lake Tanganyika Facing 8 Map of the route of the Tanganyika Expeditions >> 10 Geological Map of the Nyassa region . . . . >) 48 Geological Map of the Tanganyika region » 52 Map of the distribution of aqueous deposits in part of the African interior ......... Ft icing 75 Geological Map of the districts north of Tanganyika to the Albert Nyanza ......... Facing 80 Map of the districts explored north of Tanganyika during the Second Tanganyika Expedition by M. Fergusson . . . End of Index BIBLIOGRAPHY OF THE TANGANYIKA EXPEDITIONS. Boulenger, G. A. Report on the collection of fishes made by Mr. J. E. S. Moore in Lake Tanganyika during his expedition of 1895 and ’96. With an appendix by J. E. S. Moore. “Trans. Zoo. Soc.”, Vol. XV., 1898. Boulenger, G. A. Report on the collection of fishes made by Mr. J. E. S. Moore in Lakes Tanganyika and Kivu during his second expedition, 1899-1900. “Trans. Zoo. Soc.”, Vol. XVI., 1901. Beddard, F. On two new species of earthworms from Tanganyika. “ Proc. Zoo. Soc.”, 1901, ii., p. 190. Beddard, F. On two new earthworms of the family Megascolicidae. “Anns, and Mag. Nat. Hist.”, Ser. 7, Vol. IX., June, 1902. Calman, C. Report upon the prawns collected during the first Tanganyika expedition. “Pro. Zoo. Soc.”, 1899. CUNNINGTON, J. On the fresh-water crab Limnothelphusa from Lake Tanganyika. “Pro. Zoo. Soc.”, 1899. XXI 1 BIBLIOGRAPHY OF TIIE Digby, Miss L. On the anatomy of Chytra and Limnotrochus. “ Journ. Linn. Soc. Zool.”, Vol. 28. Evens, R. On two new sponges from Lake Tanganyika. “Quart. Journ. Micro. Sci.”, 1899. Moore, J. E. S. On the zootogical evidence for the connection of Lake Tan- ganyika with the sea. “ Proc. R. Soc.”, LXII., 1899. Moore. J. E. S. The molluscs of the great African lakes, I., distribution. “Quart. Journ. Micro. Sci.” 1898. Moore, J. E. S. The molluscs of . the great African lakes, II., the anatomy of Typhobia. “Quart. Journ. Micro. Sci.”, 1898. Moore, J. E. S. The molluscs of the great African lakes, III., Tanganyicia and the genus Spekia. “ Quart. Journ. Micro. Sci.” Vol. 42. Moore, J. E. S. The molluscs of the great African lakes, IV., Nassopsis and Bythoceras. “Quart. Journ. Micro. Sci.”, Vol. 42. Moore, J. E. S. Tanganyika and the countries north of it. “ Geogr. Journ.”, 190T. Moore, J. E. S. On the hypothesis that Lake Tanganyika represents an old Jurassic sea. “Quart. Journ. Micro. Sci.”, Vol. 41. TANGANYIKA EXPEDITIONS. xxiii Moore, J. E. S. On the zoological results of the Tanganyika expedition. “Proc. Zool. Soc.”, 1897. Moore, J. E. S. Further researches concerning the molluscs of the great African lakes. “ Proc. Zool. Soc.”, Vol. II., No. xxxi., 1901. Moore, J. E. S. The fresh-water fauna of Lake Tanganyika. “ Nature.” 1897. Moore, J. E. S. First ascent of one of the snow ridges of the Mountains of the Moon. “Alpine Journal,” 1S92. Moore, J. E. S. To the Mountains of the Moon (Hurst & Blackett, London), 1891. The Tanganyika Problem. — — CHAPTER I. INTRODUCTORY. The desirability of obtaining a fuller knowledge of the nature of the aquatic faunae of the great African Lakes arose when it was first ascertained that in Tanganyika there are animals which have not the appearance of those we have grown accustomed to regard as of almost invariable and universal occurrence in the fresh-waters of the globe. Our knowledge of this singular circumstance originated, curiously enough, as a by-product of Burton’s celebrated journey to the lake. For although it will be generally remembered that Tanganyika was discovered by Sir Richard Burton, it may not be so generally remembered that his companion Speke picked up some shells on its shore, and that these eventually found their way into the British Museum. When examined, Speke’s shells proved to be quite un- like any fresh-water forms with which naturalists were acquainted, and it was at once recognised that in their general appearance they were curiously marine. As time went on, other great lakes in the African interior were visited by many Europeans, but no shells were ever I THE TANGANYIKA PROBLEM. brought back from these lakes at all resembling the peculiar Tanganyika forms. On the other hand, the missionaries acquired further samples of shells from Tanganyika itself, and among these collections there were forms at once so strikingly different from those of any other known fresh- water lake, and so curiously marine in aspect, that when describing them Mr. Smith* drew attention to the possibility that they might eventually turn out to be relics of some former sea. But whatever interest and curiosity may thus have been raised respecting the nature of these Tanganyika molluscs was suddenly and com- pletely eclipsed by the announcement of the further discovery of jelly-fishes in the lake by the German traveller, Dr. Bohm. Their existence in Tanganyika was subsequently confirmed by Von Wissmann, and it can, in fact, be said that it was only after these announcements that what I have termed the Tanganyika problem, as such, fairly took wing. The intensification of the general interest in the fauna of the African lakes which this discovery of jelly-fishes naturally produced, is not however far to seek, for if we except the star-fishes and sea- urchins there is hardly any invertebrate type more typically marine, more characteristic of the ocean, than a jelly-fish. Like herrings, the presence of jelly-fishes in fresh-water is indicative of the past or present connection of such water with the sea. Bohm’s discovery thus in fact seemed to show, that either in present or past times, organisms like jelly-fishes could get from the sea into the lake. Eventually, through the co-operation of Mr. Frederick Moir and Mr. Swann, some of these medusae were sent to England, and on examination were found to be quite unlike any * E. A. Smith. — “On a collection of Shells from Lakes Tanganyika and Nyassa.” Proceedings of the Zoological Society of London, 1881, page 276. THE TANGANYIKA PROBLEM. 3 known forms and probably of an ancient type. Nothing more, however, could be said until some naturalist had visited the spot, and this impossibility of getting any further with the Tanganyika problem became the reason of the first Tanganyika expedition. The exploration was organ- ised by Professor Ray Lankester, who, with the help of others interested in this matter, obtained from the Royal Society the necessary grants in aid of the exploration, and I finally set out in the autumn of 1896. As a result of the journey we found that the original pro- blem presented by the jelly-fish had in no wise been solved. It had, in fact, grown enormously bigger and more difficult, for it was found that in Nyassa and Shirwa there were no jelly-fishes nor anything except purely fresh-water forms ; while in Tanganyika there were not only jelly-fishes but a whole series of molluscs, crabs, prawns, sponges and smaller things, none of which appeared in any of the other lakes I then knew, and all of which were distinctly marine in type. Further than this, however, I* found that none of these strange marine-looking animals were to be compared directly with any living marine forms, yet, in their structure, some of them certainly seemed to antecede a number of marine types in the evolutionary series, and in consequence they appeared to hail from the marine fauna of a departed age. The most definite result of the first Tanganyika expedition, therefore, appeared to be that the sea had at some former time been connected with the lake, but when or how remained a mystery. In discussing the inferences which could thus be directly drawn from our observations respect- ing the marine nature of the Tanganyika fauna before the Royal Society in 1898, I found, however, that I had unwit- * “ On the zoological evidence for the connection of Lake Tanganyika with the sea.” Proceedings of the Royal Society, vol. 62. 4 THE TANGANYIKA PROBLEM. tingly run amuck amongst some cherished geological ideas. It appeared, that in 1852 Sir Roderick Murchison, collecting such material and geological facts as were then available, had arrived at the conclusion that the interior of Africa had never been beneath the sea, and supposed, as he then said, this view to be “ confirmed by the absence south of the equator of all those volcanic activities which we are accustomed to associate with oscillations of terra Jirma In the light of our newer zoological evidence, the first part of this statement would appear consequently to be wrong, on account of the anatomical characters of the Tanganyika fauna, which relegate a portion of this fauna to a marine stock, and show that this part of Africa has been at some time connected with the sea, in order that such marine animals could get into it. While the second part of the statement, that is, the assumed evidence respecting the permanence of the African land-mass drawn from the then apparent absence of volcanic activities south of the equator, is now, also, entirely dis- proved, intense volcanic activity having been found to have occurred, and to be occurring, throughout all these regions, f In a number of subsequent papers dealing with different portions of the same problem, I therefore reiterated all that I had previously said with respect to Tanganyika having been connected with the sea. For the existence of the medusae * Journal of the Royal Geographical Society , vol. xxiv. 1864, pp. clxxv. — clxxviii. t This is shown (1) by the discovery of the active volcanoes north of Kivu, (2) by the discovery of the recent cones and active geysers all round the north and east of the Albert Edward Nyanza, (3) by the discovery of the lava fields on the west of Tanganyika, (4) by the discovery of the existence of groups of extinct volcanic cones north of Nyassa, (5) by the presence of lava flows as far south as Shirwa, and finally (6) by the demon- stration of the existence and continuance of those very oscillations of terra Jirma , which in equatorial Africa were said not to exist, on a scale only rivalled elsewhere in the region of the southern Andes. THE TANGANYIKA PROBLEM. 5 and the marine gastropods in that lake, cannot be waived in favour of the older geological speculations, which have only negative appearances to support them, and as a matter of fact the newer geological observations are not opposed, as Sir Roderick Murchison insisted, to those oscillations of terra firma which are required in order that such a state of things should have been brought about.* Moreover, the impression to which I have alluded above that the marine animals in Tanganyika must be very old, eventually bore further fruit. I remembered having been struck while still on the shores of the lake with the fact that some of the shells there were curiously similar to other shells, either living or extinct, which I had seen elsewhere, and after searching amongst the conchological representatives of the different geological eras, I found that this peculiar character, this distinctive facies , as the geologists express it, presented by the Tanganyika shells, was again presented by the fossil remains in the beds of the old Jurassic seas, that is in the marine deposits of a little later date than the English coal. The correspondence between the shells now living in Tan- ganyika and these, their long extinct marine Jurassic coun- terparts, is most extraordinarily complete; and perhaps the most remarkable feature about the comparison is that the shell of every one of the numerous marine molluscs of Tanganyika compares in what is practically a specific sense with its individual prototype in the remains of the old Jurassic seas.f This being so, it will be seen that the existence of such a correspondence between what I have termed the halolimnic * See “Tanganyika and the Countries North of It,” J. E. S. Moore, Journal of the Royal Geographical Society, January, 1901. t J. E. S. Moore, “On the Hypothesis that Lake Tanganyika represents an old Jurassic Sea.” — Quarterly Journal of Microscopical Science, vol. xli. 6 THE TANGANYIKA PROBLEM. gastropods and those of the Jurassic seas presents us at once with a possible solution of the whole Tanganyika mystery. The strange animals, the jelly-fishes, the molluscs, the sponges, etc., which appear in Tanganyika, and apparently nowhere else in Africa now, may be regarded as the relic of a time when the lake basin was in connection with an ancient sea, and consequently filled with the representatives of its ancient fauna. Moreover, the date of the lake’s connection with the sea, which this view of the nature and origin of the halolimnic fauna necessitates, is so remote, that it can easily be made to fit in with our revised notions of the past history of the continent ; and I may point out that there is nothing unnatural or strange in the isolated per- sistence of even specific forms which have elsewhere be- come extinct or greatly changed, for such ancient forms have persisted repeatedly all over the world, like the ganoids Polypterus and Amici , the scorpions and the brachiopods. But although in this manner, and as a result of the first Tanganyika expedition, we reached a tenable hypothesis respecting the nature and origin of the jelly-fishes and other marine organisms in Lake Tanganyika, it was very obvious that much remained to be done. We did not know, for example, whether there were marine organisms in any of the other great lakes, Kivu, the Albert Edward, the Albert Nyanza, the Victoria Nyanza or Lake Rudolf ; neither did we know anything of the geology of Lake Tanganyika nor of the districts north of it, as far as the Albert Nyanza. But about the same time, Suess had put forward some most interesting views concerning the nature of these very regions. He had shown that Tanganyika lies near the south end of the more westerly, and greater, of two vast series of valleys which run from the south, through Central Africa, like a couple of converging horse troughs, until they THE TANGANYIKA PROBLEM. 7 unite together in the region of the Upper Nile. Thence they pass to the Red Sea near Berbera, and continue as the valley of this sea itself as far as the Gulf of Acabah, and even through the Dead Sea to the valleys of the Jordan. In the western arm of these great valleys and north of Tan- ganyika there lie the lakes Kivu, the Albert Edward and the Albert Nyanzas. After returning from the first Tan- ganyika expedition, none of these more northerly lakes were zoologically known, nor had the remaining great African lakes, the Victoria Nyanza and the chain of lakes in association with Rudolf, been sufficiently minutely examined to show whether the halolimnic fauna existed in them or not. When, therefore, we took into account the fact that both the great series of valleys united and extended as far as the Red Sea, it appeared on the face of it, at least possible that the halolimnic fauna, or something equivalent to it, would be found in the lakes north of Tanganyika, whenever these lakes came to be zoologically explored, and that the western series of valleys might itself turn out to be the channel along which the sea had reached the lake. Moreover, the districts north of Tanganyika through Lake Kivu, as far as the Albert Nyanza, and including the Mountains of the Moon, were an almost complete terra incognita from a geo- graphical point of view. A large stretch of this country was uncharted, and thus a further journey offered exceptional opportunities for geographical work in the hands of a com- petent surveyor. It was this aspect of future exploration in these regions which gave them a distinctly geographical interest, and enabled the proposal of a second Tanganyika expedition to be supported by the promise of a very liberal grant from the Royal Geographical Society. Such an expedition was thus to be recommended on zoological, on geological, and on geographical grounds, and these three 8 THE TANGANYIKA PROBLEM. wants with respect to our knowledge of the African interior, eventually became the definitive motive for the formation of the second Tanganyika expedition. In the organisation of this new venture, Professor Ray Lankester took the initiative once more, and the expedition was finally despatched under the auspices of a committee of scientific men in England, which was formed by Professor Ray Lankester, and consisted of himself, Sir John Kirk, Sir William Thiselton-Dyer, Dr. Sclater, and Mr. Boulenger, and I had the honour to be invited to take command of the exploration in the spring of 1899. We have in this way briefly sketched the manner in which our knowledge that there was a Tanganyika problem first arose, and we have also indicated the aspect which this problem presented after the first Tanganyika expedi- tion had returned ; but, beyond this point, it would be inconvenient to treat the matter from a similar historical point of view. It will suffice to say, that on the second Tanganyika expedition an immense amount of notes and material were acquired, filling up the obvious lacunae which had presented themselves in any attempt to study the problem up to 1899. In what follows, I have dealt with the problem afresh, and with all the evidence that is now available. In the first place, I have found it necessary to consider the nature of fresh-water faunas in general, and to discuss the probabilities with respect to the origin of certain constant peculiarities which these faunas are found to present. In the next I have dealt at some length with the physical geography and geology of Central Africa, especially in relation to the fact that all the observations which have been accumulated point persistently to the conclusion that the structure and history of this portion of the continent are not what have hitherto been supposed. All the physical Hut fc TTLacJcptt L. THE TANGANYIKA PROBLEM. 9 features of Equatorial Africa being, as a matter of fact, clearly discernible as subordinate expressions of a still continuing effort on the part of the earth to produce a great mountain chain. So far, indeed, from these regions being remarkable for their stability, it is a fact that the interior of Africa is at the present time only rivalled in instability by certain districts of South America, and in the past by those records of terrestrial disturbances which we have in relation to the Alps and other mountain chains. These matters having been discussed, a review has been made of the nature of the fresh-water fauna which is found in all the great African lakes about which anything is, as yet, definitely known ; and in this way it has been shown that throughout Equatorial Africa, as in other great continents, there is a normal fresh-water fauna which has nothing peculiar about it, and is certainly not more dis- tinctive of Africa than is that of North America distinctive of the New World. Subsequently the fauna of Lake Tanganyika has been examined in detail, and it has been shown that this lake, like all the other great lakes of Central Africa, contains the ordinary fresh-water fauna of the continent, but that in Tanganyika, and in Tanganyika alone, there are a number of organisms possessing definitely marine and somewhat archaic characters. Along with these the “ halolimnic ” members of the Tanganyika fauna, as I have called them, there are others, such as the prawns, sponges and protozoa, which, although, not like the previous types, unique in being found in Tanganyika for the first time as fresh-water forms, are, notwithstanding, probably portions of the same group, for they are peculiar to Tanganyika, and are not characteristic of the general fresh-water fauna of the African continent. Subsequently, in dealing specially with the fishes of Tan- io THE TANGANYIKA PROBLEM. ganyika, and in the concluding summaries, attention has been expressly drawn to the fact, that it appears probable that the African ganoids, and certain other portions of the Tanganyika fish-fauna are in reality the now more or less scattered piscine portion of the halolimnic fauna. And in this connection I have emphasised the very remarkable fact that, notwithstanding the opportunities which always exist for fishes to migrate throughout the fresh waters of a continent, actually about half the species of Cichilidte belonging to the Old World are restricted to the confines of Lake Tanganyika even down to the present day. In the succeeding chapters the Tanganyika problem has been considered in the light of all the evidence which is now available. It has been shown in the first place that the halolimnic fauna cannot now be regarded in any other light than as something wholly distinct in origin from the general fresh-water fauna of Africa, and that it is also equally impossible to regard it as anything but the relic of some ancient sea. In conclusion, the possible mode of origin of this marine fauna has been considered together with the value and the significance of the remarkable correspondence which subsists between the shells of the halolimnic gastro- pods and the remains of those found in the deposits of the old Jurassic seas. [ To face page 10. CHAPTER II. ON THE NATURE AND ORIGIN OF FRESH-WATER FAUNAS AND THEIR RELATION TO THE FAUNA OF THE SEA. It is a fact that there are forms of animals inhabiting fresh-water which do not inhabit the sea, and which die if placed in the salt-water environment of the ocean ; while, conversely, there are animals which habitually live in the sea, and die if they are subjected to the action of water which is fresh. This is a matter of common knowledge, and when we speak of fresh-water and marine faunas we mean to describe in general the animals which can live only in one or other of these media. But it is also a fact that there are animals which can live as well in the sea as in fresh water, such as salmon ; while, again, there are others which can live in the brackish mixtures of salt and fresh water occurring at the mouths of rivers. From a variety of reasons, some palaeontological, some based on the results of comparative anatomy, and which, although they maybe rather indefinite, are probably weighty, it is generally conceived by naturalists that all the animals now inhabiting the fresh-waters of the earth originally arose in the sea. Further, it has often been held that these same fresh-water types have arisen during the past through the successful efforts of different marine organisms to colonise fresh-waters from the ocean. And, following the same kind of reasoning, it has often been sug- THE TANGANYIKA PROBLEM. I 2 gested that forms like some of the Trochoid molluscs, some jelly-fish, and many prawns which live in the brackish waters between inland lakes and the ocean constitute in themselves the modern instances of the transmutation of marine organisms into those which live in fresh-water. According to this view of the matter, fresh-water faunas are heterogeneous assemblages of organisms the forbears of which have at different times successfully colonised the fresh-waters of the earth from the ocean ; and the curiously universal distribution which is characteristic of some ot the constituents of these faunas, is regarded as having been brought about by the natural facilities which exist for the distribution and dissemination over the land of such organisms and their germs. Within the last few years, however, much definite information has been added to our knowledge of the nature of fresh-water faunas, and of the composition of those which occur in different parts of the earth. It has been shown, by the experimental researches of Beudant and others, that certain marine organisms can actually be acclimatised to fresh-water if the change is carried out with sufficient slowness, but at the same time it has become apparent that one of the most striking features which all fresh-water faunas present is the smallness of the number of the types composing them when compared to the fauna of the sea. From this it would appear that very many marine groups have never made any attempt to establish themselves in lakes and rivers, and this indication has led to the investigations of Semper* and Sollas,f who have shown that there are other obstacles besides the necessity of adaptation to new conditions of salinity which tend to prevent marine animals from * Semper, “Animal Life,” 1 88 1 , p. 149. t Sollas, W. J., Trans. Roy. Soc. Dublin , vol. iii., series ii., p. 87. THE TANGANYIKA PROBLEM. 13 colonising the fresh-waters of the land through the medium of the rivers which flow from them into the ocean. Semper has dwelt upon the hardness of the conditions of varying temperature, etc., which unquestionably obtain in rivers and lakes ; while Sollas has pointed out the equally unques- tionable fact that a very large number of marine forms are precluded from making any attempt to colonise the rivers by the fact that they begin their existence as free-swimming larvae, and that it is physically impossible for such larvae either to force themselves up a stream, or to maintain them- selves under the conditions which would surround them in a river. As a matter of fact, when the old question of the nature and origin of fresh-water faunas is re-examined in the newer light which these and still more recent investiga- tions can be made to throw upon the subject, it becomes more and more apparent that none of the existing concep- tions which have been entertained are fully capable of explaining the origin of such faunas, since it can be shown that no existing explanation of the peculiar composition of these faunas can be brought into accordance with our knowledge of the facts. The reasons for this statement will begin to appear if we consider the composition of the invertebrate section of the fauna in any two widely sepa- rated land masses. The fauna of a Central African lake, for example, compared with, let us say, the fauna found in the fresh-waters of the island of Celebes. To make matters still more simple we will consider only the mollusca of these two areas we have selected. If then we arrange the molluscan constituents of the Victoria Nyanza in a tabular form as on page 14 beside those ot the molluscan section of the fresh-water fauna of the island ol Celebes, it will be seen that the fresh-water mollusca of these two widely separated districts present, in the first place, a 14 THE TANGANYIKA PROBLEM. number of genera which are common to them both, and also a number of genera which are peculiar to each. In the second place, we find that the genera which are common to both fresh- water areas are those which have been recorded in innumerable places throughout the world. We should find TABLE I. Molluscan fauna of the Victoria Nyanza. Molluscan fauna of the fresh-waters of Celebes. Planorbis. Planorbis. Limnaea. Limnaea. Isidora. Isidora. Physopsis. Aucy his. Ancy lus. Protaneylus. Miratesta. Ampullaria. Ampullaria. Lanistes. Vivipara. Vivipara. Cleopatra. Bit hy nia. (?) Melania. Melania. Unio. Spatha. Mutela. Aetheria. Corbieula. Batissa. Comparison of the genera constituting the molluscan fauna of the Victoria Nyanza in Central Africa and of the fresh-waters of the Island of Celebes. them in American fresh-waters, in the lakes of Polynesia, and even in the pools and puddles of Japan. They are, in fact, almost if not quite universal in their distribution throughout the more permanent fresh- waters of the globe. We have, then, in the two particular examples of fresh- water faunas that have been chosen, two distinct types of animal THE TANGANYIKA PROBLEM. I5 life — (i) a universal series of types, and (2) another com- posed of organisms which only occasionally appear in fresh- water, and are related to a particular district or to a lake. If now we examine the character of the animals that form the types of the first series, we are confronted with this very curious and remarkable feature characteristic of them all. They have no counterparts or close allies in the ocean at the present day; and on anatomical examination they are found to stand in the relationship of ancestors to numerous well-known marine forms. If, on the other hand, we consider the forms which are peculiar to the two lakes in question, and more especially if we consider the forms similarly peculiar to a number of other lakes and distinct fresh-water areas, we find that these forms have almost invariably, like the prawns in the Lago di Garda, in Italy, close allies in the existing marine fauna, and generally exhibit closest affinities with some of the organisms which inhabit the sea nearest to the particular fresh-water which may be studied. We find, in fact, usually in the fresh-waters of the globe a fauna which is composed of an ancient and almost universal stock to which it appears that there may be added in any particular fresh- water a more or less large and variable stock of animals which have migrated into this area at one time or other from the nearest seas. The origin and nature of this latter series is obvious and plain, but the origin and relationships of the first — i.e., the ancient and universal series — is not so clear. It will, in the first place, therefore, be well to distinguish this universal series by a special name, and to speak ot it as the primary fresh-water fauna , and, dealing with the second series in the same way, to call it the secondary fresh-water series. That the types representing the groups consti- tuting the primary fresh- water fauna of the world should generally exhibit ancestral characters in their organisation is i6 THE TANGANYIKA PROBLEM. a very striking fact, and it is all the more interesting when we reflect that the characters possessed by the types of the primary fresh-water series are those which there is very good reason for believing were once possessed by some of the members of the normal oceanic fauna very long ago. Thus among the mollusca which we have just relegated to the primary fresh-water series we find that the Prosobranch genus Vivipara is with reason regarded by students of com- parative anatomy as possessing those particular structural characteristics which must have marked the transition of the old marine diatocardiate forms into the later Taenioglossa. Melania also is a form which has an extremely simple Ceri- thoid organisation, connecting up the Ceritho-Litterinoids with the Strombus and Natica groups. Limnea , which is found in almost every fresh water throughout the globe, is a form belonging to a series of gastropods which probably anteceded the modern marine Opisthobranchs. We are confronted with similar facts respecting the structural relationships of all the other molluscs belong- ing to the primary fresh-water series. It is, however, not only with respect to the mollusca that we find these peculiarities among the constituents of the primary fresh- water series : the same phenomena are encountered in re- gard to the Crustacea, and markedly among the fresh-water fishes such as the Australian, African, and American ganoids and the like. From these considerations it would appear to be suggested that the primary fresh-water stock belongs to an ancient type of fauna which there is reason to believe was once widespread in the sea, but which for some reason wholly unapparent on the face of things has latterly become restricted to the fresh waters of the globe. If, however, we examine the palaeontological evidence which exists respecting the first appearance of the types THE TANGANYIKA PROBLEM. i7 characterising the primary fresh-water series, we find that these types, such as Vivipara, Planorbis , Limnea , and Melania among the molluscs, fresh-water prawns and crabs among the Crustacea, and the immediate forerunners of the now widely dispersed groups of fresh-water fishes, such as the ganoids, arose as such about the same time. The emergence of the primary fresh-water series being as a matter of fact synchronous with a strange phenomenon already well recognised by geologists ; it occurs just at the time when a curious break is manifest in the forms characteristic of the secondary and tertiary deposits, and from this we might infer, or at least think it probable, that they owe their differentiation to the same cause which produced during this period an extraordinary multiplication of new types and the extinction of old ones. From the manner in which the primary fresh-water series appears in the geological record, it in fact seems to be suggested that there came into play some cause which was efficient to kill out a disproportionately large number of ancient marine types, and at the same time, both to produce a dispropor- tionate array of new marine forms, and to dissociate from these the representatives of the primary fresh-water series as universal inhabitants of the waters of the land ; that is to say, the facts of morphology and the facts of palaeontology, when taken together, seem to suggest that there has occurred something like a change in the character of the sea itself, which has affected the animals contained in it in such a manner that a large number of its old forms were definitely killed out, while others were driven into the fresh waters of the globe, and at the same time a very large number of entirely new marine types was produced. In attempting to form any clear conception of the nature and the origin of the curiously similar fresh- water faunas 2 1 8 THE TANGANYIKA PROBLEM. which are now found all over the world, it would, from these considerations, appear possible that we shall have to reckon with, and to define, some factor which has not hitherto been recognised, and that this is so will, I think, become quite clear if we examine the conception of the fresh- water fauna problem which is detailed in the luminous paper by Professor Sollas to which I have already referred. We find that, after pointing out the unquestionable obstacles which stand in the way of any attempt on the part of many marine organisms to colonise the fresh-waters of the land, the author inclines to the conclusion that the existing fresh- water faunas of the earth must have originated in pans, which were at one time part of the sea, and were subsequently cut off from it, the present fresh-water forms being survivors of this process, just as, if the Baltic were to be cut off now and become fresh, some of its marine organisms might persist as fresh-water types. The present curious universal distribution of the primary fresh-water fauna over the land’s surface being assumed by Sollas to have been brought about afterwards, and to be due mainly to a capacity which fresh-water organisms possess for migration. At the time Sollas wrote, the extraordinary dispersion of many fresh-water animals throughout the world, such as P/anorbis, Limnaea , Melania , Bithynia , and Vivipara, was not fully apparent, and still less were the facts relating to the actual capacity for migration which these and other fresh-water animals really possess. I shall refer to this matter again subsequently, but for present purposes, it may be stated, that the more recent observations of numerous investigators do not show that there is really any evidence whatever for a capacity on the part of fresh- water organisms, such as the above molluscs, to spread them- selves, as they are spread, into the different isolated sheets THE TANGANYIKA PROBLEM. J9 of water which occur on a great continental land mass, and throughout all the great continental land masses of the world. If we examine the great fresh-water basins and river systems of Africa, for example, the first fact which confronts us is the localisation of groups of varieties of the universal types. Thus, confining our attention, merely for the sake of simplicity, to the molluscs of Lake Nyassa, we find only one species — namely, Melania tnber- culata — existing in this lake which occurs in any of the other great lake basins.* In Tanganyika, excluding its peculiar halolimnic group, all the species of molluscs in it are peculiar to the district : they do not even occur in the adjacent lakes of Mwero, Rukwa, or Bangweolo, such types being either unrepresented in these lakes or represented by different species. So again in Kivu there is not even a Vivipara , and the only series of great African lakes that possesses a large number of identical species is that constituted by Albert Edward, the Albert, and the Victoria Nyanzas, which communicate with the Nile now, and were almost unquestionably once more or less connected together. These facts — and there are other similar ones connected with the distribution of species in the lakes and river systems of North America — seem to show in an unmistakable manner that many typical fresh- water forms have little if any capacity to migrate from their original centres of distribution, and, consequently, that the peculiar specific varieties which they present in these centres are due either to their original character or to natural selection, having operated since their origin, in a generally similar manner, in all these particular spots. In like manner, the distribution of the Characinid fishes in the American and African fresh waters is quite inexplicable * Shirwa is a lake in the same basin as Nyassa, and was at one time connected with it. 20 THE TANGANYIKA PROBLEM. on any supposition of their having originated as a relic fauna in some one arm of the sea which became cut off from the ocean, and ultimately fresh ; for there is no evidence that there has been any connection between the re- mote state land masses which these fishes now inhabit, since the origin of types now common to them both. It is the same with the Cichlidae and many other forms of fish. The idea which Sollas advocates, that the fresh-water faunas of the world are relic faunas emanating from the sea, may be true, but that these faunas have intercommunicated one with the other, and thus produced the universal distribu- tion of the primary fresh-water types, does not appear to be supported by any capacity for dispersion exhibited by their present constituents. We seem, in fact, from the evidence which is now available, to be driven either to the supposition of such a complete parallelism in the evolution of forms of different origin in different areas, that it makes generic distinction impossible, and conse- quently upsets our fundamental conceptions in a manner for which there is no warrant ; or we must find some other cause for the universal distribution of numerous fresh- water types. It has already been emphasised that the members of the universal fresh-water series possess the characters of forms which there is every reason to believe were once widely spread in the sea, and, this being so, it also becomes clear that if we can discover some cause which would have made their migration from the sea to fresh-waters a necessity, we shall have found an explanation of some of the most remarkable features of the fresh-water faunas of to-day ; for through such a cause all the members of the primary fresh-water series might have been produced about the same time, and would consequently present all over THE TANGANYIKA PROBLEM. 21 the world the similar morphological characteristics which they actually possess. This similarity would not, how- ever, have been produced as a result of intercommunication between fresh-water centres, but as a result of their all having arisen from the general sea fauna of a particular geological age. The question before us, then, is whether there is to be found in the nature of things any cause which can be regarded as sufficient to produce such an effect It is a fact that increase or decrease of the amount of saline matter in water does act prejudicially, on many, if not on all the organisms living in it, and this can be shown in a variety of ways, perhaps best by the study of a lake which, like Shirwa, was originally fresh, but which through losing its outlet has now become intensely salt. Shirwa is a large, oval body of water about 50 miles long, and enclosed on all sides, first by lacustrine plains and then by lofty granitoid hills. At the north there is an old channel which repre- sents a former connection with the Lujenda River, but this, probably through rising of the ground, has now become closed, and the waters of the lake have not flowed out at any rate for many centuries. The lake is fed by numerous rivers which flow into it from the great mountains of Mlanj i and Zomba, and the slight amount of salt which the rain washes off the land has accumulated in the evaporating-pan formed by the lake, until it is now a thick strong brine. One of the principal rivers which flows into Shirwa is the Mulmgoosi, and where it enters the lake, there is still an area of almost fresh water, bounded without by the brine wastes of the lake. In this fresh-water oasis, but not in the upper course of the stream, I found some Viviparas living, of the same species unicolor which inhabits Lake Nyassa. There was also the Nyassa Limnaea and Melania tuber - 22 THE TANGANYIKA TROT LEM. culata , as well as a few Cichlid fishes, among which there were several Nyassa species. Beyond this oasis, in the open body of the lake, and among its numerous islands, none of these animals occurred, and the only forms of life present were some curious algae, some catfish, and a small Planorbis , which lived upon the reeds and not in the lake at all. All round the marshy shores of Lake Shirwa, how- ever, there are extensive plains which were at one time unquestionably portions of the floor of the lake, and em- bedded in these there are found countless millions of Vivi- paras and Limneas ; in fact, the remains of nearly all the molluscs found now living in Nyassa. From these facts three things are obvious: (i) that Shirwa at one time flowed out and was fresh ; (2) that it was once peopled by the Nyassa fauna ; and (3) that it is now uninhabited by that fauna, except in the oases of fresh-water such as that which I have just described. At some time the lake ceased to have an outlet, and its somewhat profuse fauna died out as a result of the stagnation of its waters and the acquisition of their peculiar briny characters. The increase of the salinity of the water of Lake Shirwa must, however, have taken place with extreme slowness, much more slowly, in fact, than we could ever bring about a similar change of environment experimentally. Nevertheless, we find that the salt has eventually killed out the whole of the typical fresh- water molluscan fauna of Nyassa, and most of the fish. The consideration of the history of Lake Shirwa shows us, then, that a typical fresh-water fauna like that of Nyassa does not readily acclimatise itself to an increase of salt, such as that which has taken place in the lake, and this is so notwithstanding that the increase may have taken place with extreme slowness. I do not know the percentage or the exact composition of the saline materials which the THE TANGANYIKA PROBLEM. 23 water of Lake Shirwa contains, but, following the same line of investigation, I distributed while on Tanganyika five hundred young prawns, in equal quantities of water, among twelve vessels. Leaving two unaltered as a control to the rest, sait was added so that the solution in the remaining ten was increased daily in the proportion given in the table on p. 24 in which the course of the experiments can be seen ; and from which it will be obvious that in all cases the prawns died when the quantity of salt in solution in the vessel reached '8 per cent., no matter what the rate at which the increase had been made ; and some time before this point, although the prawns were plentifully supplied with food, their development was seriously interfered with, the salt producing a progressively stunting action proportionate to its increase. The effect of introducing salt into the environment of fresh- water organisms such as prawns leads thus to exactly the same result as that obtained with extreme slowness by nature in the case of Lake Shirwa. In both, but more especially in the former case, we have direct evidence that fresh-water organisms can stand a certain amount of salinity, but beyond a certain point, which is probably fixed within very narrow limits for any particular type, they are sud- denly killed off. When anywhere near this point has been reached, the salt acts as a poison, and, just as is now known to be the case with the action of a large number of poisons, not much effect is produced, until what may be termed the critical dose has been reached. But beyond this, wide phy- siological disturbances, or death, immediately occur. There is, however, another aspect of change in the environment of organisms like that produced by increased salinity which, in the present connection, it is very important to consider. It has been experimentally shown by a number of investi- 24 THE TANGANYIKA PROBLEM. TABLE II. Vessel in which water remained fresh. Vessels in which salt was added in the proportions indicated in the columns. 14 15 No. of days. II. III. IV. V. 1 •007 015 023 ■03 2 •014 030 •046 06 3 •021 ■045- 069 ■09 4 •028 •060 092 •12 5 Nothing added. 6 | ■O48 l •152 •2 7 O c .O98 ■2 •302 '4 8 0 a, 0 •198 '4 •602 •8 9 5 Nothing added. 10 S-4 •248 •5 752 1 1 1 •298 •6 •902 f2 12 v •345 7 1-052 13 c Nothing added. l6 1 •395 17 « e/3 •471 18 ■5 •495 19 rf t/1 548 20 00 1 ON to 21 ■645 22 •649 23 •748 24 •798 Table showing the effect of increasing the amount of common salt in the water in which Prawns were living. None left. THE TANGANYIKA PROBLEM. 25 gators that long before a poisonous close is reached the addition of chemical substances to water in which animals are normally living acts as a stimulus towards variation either in themselves or in their offspring. Thus Vernon* found that the addition of small quantities of urea and uric acid, ammonium chloride, and nitrates had an appreciable effect on the larvae of sea-urchins which were reared in such solutions. Herbstf showed that lithium salts produced very abnormal effects upon the growth of larvae. LoebJ also showed that change of salinity affected the growth of tubu- larians. It would thus appear that there is some reason to believe that a gradual increase in the salts normally contained in the water of the sea would produce effects analogous to those which unquestionably have taken place in the animals living in it during past times. Such an increase would, it appears, ultimately kill off some animals, weaken other stocks, cause what we may call rampant variation, and force a number of forms into the fresh-waters of the globe. From what has been observed there is also reason to believe that, had such an increase of the salinity of the sea occurred, it would as a matter of fact begin to affect large numbers of diverse organisms about the same time, or, in other words, the same percentage of salts would begin to tell on a large number of diverse types. If, then, the sea has become increasingly salt, this simple cause would be in itself efficient to account for a number of the most perplex- ing features which the universal fresh-water fauna of to-day presents. The study of the general nature of fresh-water faunae * Phil. Trans. Roy. Soc. Vol. i86b., p. 577 — 1895 ; also Naples. Mitthal, Vol. xiii., p. 341 — 1898. f American Journ. Phis. Vol. iii., p. 385. | Archiv. Fur Entwic. Mechanik. Vol. xi., p. 617 — 1901. 26 THE TANGANYIKA PROBLEM. suggests that some change in the sea must have taken place. In consequence of this it is incumbent on us to inquire whether there are other grounds for believing that the sea has become and is becoming steadily more salt. Turning to this matter we find that the subject has already been considered, but in general from other aspects than that which immediately concerns us at the present time. Thus, within the last few years we find both Jolly and others considering the increase of salinity of the ocean, and taking its existence for granted as a means in an attempt to ascertain the age of the earth. That the saline matters in the sea have changed in composition and in- creased in amount seems, indeed, almost indisputable from a variety of considerations. Thus Jolly points out that numerous analyses of the rocks of different ages show a marked contrast in the amount of common salt which they contain, the older the rocks the more the salt, and, further, that the percentage lost in the more recent deposits is matched bv the amount of salt dissolved in the water of j the present sea. Jolly also comes to the conclusion, and apparently with good reason, that the amount of salt which has been again lost by the sea, in the formation of rock- salt deposits, is to be considered as infinitesimal and negligible. Indeed, the same results would be forced upon us, if we were to apply the experience we have gained in the study of fresh and salt water lakes to the study of the sea itself. We find that where lakes have an outflow they remain fresh, because the salt which is brought into them is always being carried away by this outflow, but in lakes which have no outflow the modicum of salt which the rivers bring down remains in the water of the lake, so that it becomes more and more saturated until finally the THE TANGANYIKA PROBLEM. 27 salt may crystallise out on the reeds and beaches in glitter- ing white precipitates, as I have seen it repeatedly round Shirwa and the other salt-pans of the African interior. A lake having no outflow bears the same relation to its drainage area that the ocean does to the whole land sur- face of the earth, the volume of water in the sea remains constant, but vast quantities of this water are for ever evaporating into the atmosphere, and are then flung broad- cast over the surface of the land, from whence they eventually drain back in the form of rivers, but each in possession of a certain quantity of salt which they add to that already in solution in the sea. The sea is nowhere nearly saturated with salt, and the above reasoning would appear to lead to the conclusion that if we had a sample of the ocean which was sufficiently old we should find that it contained less salt than does the present-day sea-water ; whether, however, we can agree with Hunt that the water sometimes enclosed in ancient rocks may be looked upon as fossil sea-water, and consequently that the ancient oceans were richer in magnesium and potassium than the present-day sea, is a matter into which we need not at present go ; all that it is important to realise being, that it seems to be demonstrable from different sources that the sea has changed and is changing in composition with a gradual increase of common salt in solution, and that the indication of a physical change in the sea which we obtained from the purely zoological considerations, discussed in the preceding pages, seems to be in unison with various other lines of evidence. There thus appears to be reason to suppose that there has been a slow increase of common salt in solution in the sea, and if this is so, it would also appear that the general course of marine evolution may have been profoundly in- 28 THE TANGANYIKA TROT LEM. fluenced by this slowly progressive physical change. That indeed such evolution has to a large extent been an organic reflection of this change ; the change in the sea having acted as a stimulus towards variation, and also as an eliminating agent. Consequently we may have to view the primary features of distribution and character of both the fresh-water and marine faunas of the present day, as not so much due to the operation of the struggle for existence among warring animal types, as to the moulding and eliminating effect of a progressive physical change. That physical changes of a similar character are of much greater import in biology than has hitherto been supposed I have personally not the least doubt, and that they are capable of producing changes of the above kind we shall see again in the present work, for in Chapter VII. I have had occasion to advert to the effect of progressive desicca- tion upon the flora of a country, as evidence of geological change, and to show that the whole floral aspect of wide areas may be completely altered, not however as a result of the struggle for existence between different plants, but through the operation of a progressive physical change which is forcibly moulding the flora of such districts into an expression of itself. Summarising what has now been said it would appear in the first place that the sea is in exactly the same condition as an immense closed lake, and that so far as we can judge it is for ever acquiring more salt in solution than it had before, and its waters will go on getting more and more salt until they are at any rate much nearer saturation than they are at present. What effect this change will have in the future upon the animals which live in it, it is idle to speculate ; but it is important to realise the following facts : — THE TANGANYIKA PROBLEM. 29 I. That there is progressing this change with respect to salinity. II. That it can be shown experimentally that such a change after it reaches a certain degree is both prejudicial to some organisms and causes wild variation in others. III. That these tendencies are competent to bring about the separation of marine from fresh-water faunas, the latter arising all over the world from the general marine fauna at about the same time. IV. That the facts of the morphology and distribution of the existing fresh-water faunas independently point to some such cause as having operated towards their production in the past. Accepting these various indications respecting the manner in which fresh-water faunas have become differentiated, it would appear that the present-day fresh-water faunas are to be regarded as chiefly composed of the remains of a once widely distributed and ancient sea-fauna, the ancestors of the surviving components of which were forced out of the ocean into the fresh-waters of the globe owing to a change in the character of the sea itself. This change ap- pears to have become sufficiently strongly marked to have produced an appreciable differentiation at a period roughly corresponding to the commencement of the formation of the secondary rocks. In this manner it would appear that assemblages of similar organisms have of necessity taken to fresh-water all over the world about the same time, just as one could imagine the modern universally distributed organisms of the marine littoral, such as the Trochoid molluscs and the shore crabs, being simultaneously every- where driven into the fresh-waters of the land if the sea became too hot to hold them. 3° THE TANGANYIKA PROBLEM. Once established in the fresh-waters of the globe, it is probable that the fishes and the more active of the inver- tebrates migrated slowly from point to point, but as we have seen, there is much evidence to show that in the great fresh-water areas of the different continents the invertebrate components of the primary fresh-water series have migrated very little from the point at which they first originated. In many fresh-water areas however, as in the Lago di Garda in Italy, there are often to be found organisms which have not the ancient attributes of the constituents of the primary fresh-water series, and these organisms, like the Lago di Garda prawns, are to be regarded as the com- paratively rare and conspicuous examples of voluntary colonisation from the sea. Still further, in a few places, as in the Caspian Sea, and, as we shall see later, in the case of Lake Tanganyika, there are to be found whole batches of animals which have become independently detached from the sea and which bear no proximate relationship either to the modern marine fauna, or to the primary fresh- water types. Thus we find in the fresh-waters of to-day animals which can naturally be arranged under three dis- tinct heads. We have in the first place the primary fresh- water series , in the second the sporadic and voluntary colonists of fresh-water which we may call the secondary fresh-water series , and thirdly the relics of entire marine faunas which in a few places are found to persist and which, like those of the Caspian and Tanganyika, we may call the halolimnic series of the world. VIEW OF THE ZAMBESI AND SHUPANGA LOOKING SOUTH-WEST. 31 CHAPTER III. ON THE EXISTENCE OF A GREAT CENTRAL MOUNTAIN CHAIN IN AFRICA. In entering upon an enquiry concerning the nature of the fauna of the great lakes of Central Africa in general, and that of Tanganyika in particular, it is, if not absolutely necessary, at any rate advisable, that a conception of the structure of the continent, as concise and accurate as existing observations will permit, should be formed first. On this account I have prefaced the chapters dealing with the zoology of the Great Lakes by some in which the structure and the geology of the vast regions in which the lakes lie has been described as fully as our yet scanty information will allow. In the present instance this course is doubly to be desired : our views respecting the continent are becoming changed, and some very definite theoretical anticipations which have been put forward concerning the gross geological features of the African continent appear, in the light of the most recent exploration, to have hardly even a shadow of support. These older geological anticipations are, more- over, diametrically opposed to some of the most salient zoological facts relating to the fauna of Lake Tanganyika, and consequently any fresh information about the geo- 32 THE TANGANYIKA PROBLEM. logical structure of the African interior, resulting from recent exploration, is particularly germane to the enquiry with which this work is concerned. Most, if not all, the existing views of the nature and past history of the African land-mass have been built upon the old conception that the physical characters of the con- tinent, the distribution of land and water upon its surface, the shape and the configuration of the equatorial region as a whole, have been almost unique in their stability and long permanence. That, in fact, the great mass of Africa, lay in waving hill, and plain, and ravine, much as it does now, even under the paleozoic sun ; or, at any rate, since the suns which rose over this portion of the earth, when the new red sandstone was being deposited, finally set. This conception of the past history of Africa was first definitely brought forward, now many years ago, by Sir Roderick Murchison* in a Presidential address to the Royal Geographical' Society, and it has ever since formed the chord upon which those who have concerned themselves with the geology of the continent have harped. It was, for example, resuscitated by Dr. Gregoryf while discussing Suess’ conception of the nature of the great Central African “ graben ” (rift valleys), and from this single instance, taken at random, it will be sufficiently obvious that, although we have until lately known almost nothing of the nature of the African interior, this ignorance, as usual, has in no way deterred investigators from speculating a good deal. What we require at the present time, and for the particular investi- gations upon which we are entering, is a review of the most recently ascertained facts touching the geology * Journal Royal Geographical Society, Vol. xxii. (1852) ; Ibid. Vol. xxviii. (1858). f The Great Rift Valley (1896). THE TANGANYIKA PROBLEM. 33 of the region of the great African lakes ; but what we do not want is that this should proceed in the. strongly-coloured light of antecedent speculation. We must consider the matter anew, and with as few pre-conceptions as possible, and see whether the observations will them- selves take colour and start legitimately into the shape of more generalised appreciations as we go. It has been confirmed, especially as a result of the second Tanganyika expedition, that we can, in fact must, regard all Africa as constructed in relation to an elevated ridge, which runs from the mountains of Abyssinia and those flanking the Red Sea in the north, to the continuation of these same ridges in the shape of the Drakensberg in the extreme south. In some places, as for example in the region of Tanganyika and the Albert Edward Nyanza, the ridge may be broad, constituting the crest of the interior plateaux, and resembles the Ural Mountains in that it forms the gently culminating line of two long- slopes from the east and from the west. So again although the great fold may be ten to twelve thousand feet in height, like the Urals it does not appear to the eye as an actual mountain chain from either side. In most places, on the other hand, as between Nyassa and Tanganyika, the culminating heights are narrow and rise abruptly above the surrounding country in the form of bold mountain chains. Everywhere along this line the axes of the chain run approximately north and south, and the crests of the higher summits are as lofty as those of the Rocky Mountains in the United States. Indeed, were Africa in any other latitude, there would be an immense range of ice and snow running from the extreme north to south ; while the lonely summits of the Ruwenzori Mountains, of Kenia and Kilima-Njaro, although right under the equator, are 34 THE TANGANYIKA PROBLEM. ice-capped as it is. It is only comparatively recently that the existence of these interior heights has become known, and it is only quite recently that the conception of a long axial range in Africa, bearing the same relation to that continent that the Andes do to South America, has begun to be appreciated. It is indeed to Mr. Scott Elliot* that we owe the first clear apprehension of this fact, and it is such a very important fact in all questions of African physiography, that I shall attempt in this, as in a former work,f to emphasise the appreciation of it by giving the mountain chains a special name and by speaking of them as “ The Great Central African Range.” The main river systems of Africa flow from the eastern and the western slopes of this range, and the greater rivers, partly owing to its position partly to climatic conditions respecting rainfall, flow mainly from the western slopes. The immense Congo, with all its gigantic upper tributaries, rises wholly on the west, while even the Nile, with its main components, the Albert Nile, the Sobat and the Blue Nile, drains mainly from the western slopes. So also, even the Zambesi derives most of its water from the west of the range, although it eventually cuts through the chain to an opening in the Indian Ocean. As is the case in the region of most great mountain ranges, there are to be found in Central Africa besides the chains of lofty heights, long, deep hollows which run parallel with, and between the ridges of the range itself, and these great depressions, which are now occupied by the lakes and the rivers, are often rock valleys, like the valley of Geneva. They have not, however, been formed by ice, and are not to be viewed as wholly the products of denudation operating in the past, any more A Naturalist in Mid- Africa ( 1896). t To the Mountains of the Moon. VIEW OF MORUMBALA MOUNTAIN FROM THE UPPER SHIRE RIVER. THE TANGANYIKA PROBLEM. 35 than the mountains and the great elevated plateaux of the hio-h central ridges are to be viewed as such. Both mountains and depressions are the products of igneous forces similar to those which have raised the Alps and the Caucasus in Europe ; and as is the case with the Alps, so also in Africa we find old aqueous deposits of all sorts, piled up at high angles on the flanks of the great central core. This is particularly well seen in the region of Nyassa about Mount Waller, and at the north end of the lake ; all over the Tanganyika districts and beyond them to the north in the region of the Mountains of the Moon. It will be convenient in the first place to consider the surroundings of Lake Nyassa. We find that the lake lies in a deep depression, which has the form of a vast fold in the earth’s surface, and such in reality it appears to be. The depression in which the lake lies runs also along the very top of this portion of the continent, its opposite edges here representing the highest crests of the Great Central Range. In certain places there are old sandstone deposits stretching across the present site of the depression, and these deposits have been broken along its course, in such a manner that, beyond its eastern edges, they slope away towards the Indian Ocean, while on the west they trend in a similar manner towards the opposite sea coast. The existence of these deposits shows in the first place that the trough now occupied by the lake was formed subsequently to their deposition. It is evident indeed that throughout the whole length of the Great Central Range there has been much of this local elevation and depression ; for in almost every district there may be seen faulting, tilt- ing and smashing of the old lake deposits, and other strata, which overlay these regions before such movements took place. The Central African Range appears in this region 3* 36 THE TANGANYIKA PROBLEM. to be an expression of one of those linear series of gigantic earth movements which have formed the Andes and the Rocky Mountains in America, the Alps and the Caucasus in Europe. Further, as is the case with most great mountain chains of this sort, the lofty heights that have been pro- duced along the axes of the folds, are in general not volcanic cones, or volcanoes in the ordinary sense of that term, but, at the same time, as is the case with the Alps and the Caucasus, the earth movements, the folding and the crump- ling of the globe’s crust, which appears to have brought these chains into existence, has also originated true volcanic activity in their vicinity, and thus matching the existence of the Auvergnes and the cones of the Puy de Dome in relation to the Alps and the Pyrenees, Etna, Vesuvius, and Stromboli, in relation to the Appenines ; so we find the Mfunbiro Mountains, Kilima-Njaro, Kenia, and innumerable smaller volcanoes, some active, some extinct, but always in their position more or less closely dogging the course of the Great African Range. The processes of elevation which find their maximum expression in the Alps and the Pyrenees have affected wide areas, especially to the north of these heights, and so also in Africa we find that the great tendency towards elevation along an axis running north and south through the continent has, in the same way, affected to a less degree an enormous area of the earth’s surface east and west of the range. This is evidenced by the rising of the coast line which can be seen to have occurred at innumerable places, both on the east and the west of the continent. It is very apparent for example at Mombasa, Zanzibar, Bagamoyo, Delagoa Bay, Shupanga, and near the eastern slopes of Mount Morambala, at all of which points there are to be found marine deposits of different ages now elevated above the sea level ; and THE TANGANYIKA PROBLEM. 39 though the upthrust has been far greater along the Great Central Range in the interior of the continent, it is obvious that these same earth movements have affected in a less degree an area which is at any rate as wide as the continent itself. We are, in fact, here in Africa encounter- ing again phenomena similar to those noticed by Darwin as having gone forward in the gradual elevation of that part of the earth’s surface which finds its maximum of expression along the crests of the Southern Andes. The effect of crinkling such as this, greatest along an axial line running approximately north and south in Africa, and becoming less and less as we pass from this longitudinal axis, east and west, would if it had gone on evenly and uninterruptedly have tended to raise the continent into a great hog’s back ; and south of the equator, this is as a matter of fact the form which it actually possesses, the earth’s surface has been bent up into a great arch ; but as we become better acquainted with the phenomena, it is also apparent that the changes which have produced this effect have not operated altogether evenly ; and owing to this, during their progress thev have given rise to subsidiary effects : to all sorts of parallel foldings and crinklings of the surface, which in some of their ex- pressions are extremely interesting. By far the most striking of these subordinate geological changes which the gradual upraising of the African interior has produced, are a series of great chasms, the “ Graben” of Suess, the long, deep valleys to which I have just alluded and which are found to run among, and parallel with, the ridges of the Great Central Range. These vast fold-like depressions in the surface of the earth have been noticed now by a large number of explorers. By Stanley, Stuhlmann, Cassati, Gbtzen, Teleki, and many others, and the accumulated information 40 THE TANGANYIKA PROBLEM. respecting the structural similarity of many of the grooves which wander through the African interior from south to north was luminously summarised by Professor Suess,# who showed for the first time that whatever their origin, such chasms are a related series of phenomena, and that the earth movements which have given rise to them are by no means confined to the African interior, but have brought about similar phenomena in the production of the vast walled chasm in which the Red Sea lies fifteen hundred miles away to the north. Similar chasms are indeed pro- longed into the Gulf of Akabah and the Dead Sea even as far as the depression of the Jordan itself. More detailed information respecting the nature of the lesser eastern de- pression which lies between Keniaand the New Uganda rail- way was given by Professor Gregoryt as the result of a short journey to Kenia from Mombasa, but the clear conception of these valleys as a phenomenon which is subordinate to the formation of a great internal line of elevation perhaps naturally escaped both Suess and his followers at the time. They regarded the “graben" (rift valleys) as the primary geological feature of the African interior, whereas they are far more properly viewed as the interesting by- products of the folding of a spherical surface during a modern attempt on the part of the earth to raise a grand mountain chain. Let us, however, examine the character of the great African ridge, including these valleys, in some- what more detail than has hitherto been done. If we con- struct a section of the Great Central Range in the region of Southern Nyassa, through Kota Kota for example (see dia- gram No. i ), we have the following phenomena. Beginning at a point between the lake and the east coast, there are Die Briick des Ost-Afrika. t loc. fi/. THE TANGANYIKA PROBLEM. 4i encountered a succession of rising granitoid ridges and elevations, which finally culminate in the huge mountains forming the east coast of Lake Nyassa, these mountains being in fact the last and highest of a succession of granitoid ridges separated by valleys of varying View across the great Central Eurycolpic fold from the Northern slopes of the Mfumbiro Mountains. depth and extent. The trend of the ridges is from north to south, and the valley in which Nyassa lies is the broadest, and the deepest, which we have encountered during our supposed journey from the east. Crossing the valley of the lake towards Kota Kota, we find that it sinks to, and below the level of, the sea, and then we pass up again over the alluvial 42 THE TANGANYIKA PROBLEM. flats surrounding the Arab settlement, to the moun- tains beyond the west coast of the lake. Having scaled these, we find that we have also arrived on the western slope of the continent. The gradual rise in this region is thus seen to culminate in a high ridge with two cusps and an extremely deep narrow valley in between them. The trend of the granitoid rock on both sides of these cusps is directed towards them, and they are themselves evidently the loci of vertical upthrust. Accompanying these phenomena in Nyassa there are to be found all the subsidiary phenomena which usually accompany such movements of the earth ; that is, we have everywhere abundant evidence of local elevation and depression of adjacent areas relatively to one another. Thus, if we take a section rather further to the north through a conspicuous promontory known as Mount Waller (see diagram No. 2), we find that the crests on opposite sides of the lake are not here composed of granitoid material, but of stratified sandstones and conglomerates in layers. These strata on the east of Nyassa are tilted up slightly towards the lake, until they break away immediately above its shore in an imposing series of scarps. Crossing the lake from east to west, its floor is here found to be shallower and formed of sheets of nearly horizontal sandstones, similar in all respects to those composing the eastern scarps. Along the east coast at this point there is a great line of faulting, and a vast cliff face has been upraised in the east above the beds which form the floor of the lake. On the west coast of the lake, we find the same series of phenomena repeated like a reflection on the other side. South of Deep Bay and near Mount Waller on the west there are sandstone ridges protruding above the water with a slight dip to the west (see diagram CONTINENT TOWARDS THE WE. ST COAST KQ SWAMPS KOTA : MAIN I .OPE OF CONTINENT TOWARDS THE WEST COAST MOl scale X 5. \To face page 42. THE TANGANYIKA PROBLEM. 43 No. 3), and immediately behind these there are faulted up into the air the imposing cliffs of Mount Waller itself. At the foot of Mount Waller, the igneous base on which the two thousand and odd feet of sandstones and conglomerate rest, was exposed above the water line when I visited the region in 1 896, and near this point to the north there were other sand- stone ridges which like those to the south had been faulted to a less extent than the Mount Waller mass, and bore on their upper surfaces thick stratified beds of modern chalky lake deposit. In these beds there were, moreover, fossilised shells similar in all respects to those now living in Nyassa and showing incontestably that the upthrust along the line of the west coast of the lake had been going on and had caused fresh faulting of the stratified material near the lake shore, at any rate since Nyassa contained its modern fresh water fauna. But besides thus giving evidence of modern (post pleistocene) activity and continual rising of the cusps of the great central African ridge, the phenomena I have just described near Mount Waller in Nyassa are intensely instructive in another sense. They show that where hori- zontal strata overlie a region where folding of this sort is going on, a typical faulted trough with vertical sides (a so-called rift valley) can be produced by the rising of the sides quite as well as by the falling in of a central strip of land. Lateral compression of the earth’s surface, as by the shrinkage of the globe, would actually pro- duce these effects, just as we can bend a piece of paper up into two folds with a valley in the middle ; and in this case we see also that as the lateral pressure increases the central valley in the paper tends to deepen at the same time that the two ridges are rising. All the phenomena encountered in the Nyassa region are at once perfectly intelligible if we suppose that folding of the earth’s surface 44 THE TANGANYIKA PROBLEM. of this nature has gone on ; and the only further demon- stration which is required to make the tale complete are some observations which will show that the floor of the valley of Nyassa has become, or is becoming, actually thrust down. Such evidence is readily to be obtained at the north end of the lake. The water of Nyassa has fallen many feet within no great number of years, and in consequence, wide plains of old lake deposit and alluvium covering what was once its door have become dry land near Karonga and further north. At the end of 1895, when I first visited the lake, the water of Nyassa was shown by observations made by the officers of the gunboats to be lower on the whole than it had been for many years ; but notwithstanding this, I was shown to my intense surprise a number of old trees not far from Karonga, which were standing about two hundred yards in the water of the lake, and the trunks of which were then submerged some five feet. Moreover, the old natives of the district, whom Dr. Cross interrogated on this subject for me, assured us that they could remember a time when it was possible to walk out to these same trees, which were then not near the water at all, and at whatever rate the change in level has taken place, it is quite clear that it can only have been produced by local subsidence of the ground, i.e.j by the sinking of the floor of the Nyassa Valley during the life of these particular trees. The valley of Nyassa, like that of Tanganyika, is continued beyond the lake to the north (see map, opposite p. 48), and although its magnitude becomes lessened and partially filled up, there is no doubt that the individuality of the depression can be traced as far north as the valley of Lake Rukwa, which lies to the east of the Tanganyika depression, but finally runs into and crosses it obliquely from east to west. It is in the floor of the Nyassa depression, some way to the SEA LEVEL THE TANGANYIKA PROBLEM. 45 north of the present lake, that in this region we first en- counter true volcanic action in the shape of a number of extinct volcanic cones, with circular craters, and deep blue crater-lakes, phenomena which here repeat those to be found in the great Tanganyika depression far to the north, and again in the floor of the Red Sea itself. Portion of the west coast of Lake Nyassa, near Nkata Bay. The double cuspecl character of the ridge of the great central chain is thus seen to be continued as far as the southern extremity of Tanganyika, but at this point a new depression of gigantic proportions begins. If we take a section through the region of Lake Rukwa, and through the southern extremity of Lake Tanganyika from east to west (see diagram No. 4), the following physical 4 6 THE TANGANYIKA PROBLEM. phenomena are apparent : — The country rises towards the eastern scarps which flank the trough-like de- pression of Lake Rukwa, where a line of faulting occurs. At the base of these eastern cliffs, which were found by Mr. Wallace to be about 400 feet in height, there is a flat, dusty plain of modern lake deposit, and then another series of scarps facing those on the east and marking another parallel line of faults. Above these western scarps of the Rukwa depression the land rises gradually ; it is composed in places of sandstones and conglomerates, which, further to the west, are pierced by the intrusive granitoid material of the mountains flanking the east coast of Tanganyika, and these in turn overlook to the west the great depression of this lake. On the western slopes of these heights sandstones and quartzites are again found piled up at varying angles upon the intrusive core, and sloping towards Lake Tanganyika in a succession of flat-topped, forest -clad terraces, between which there are faults represented in Diagram iv. at F.F.F. These sandstone slopes finally dip under the water of the south- eastern corner of Tanganyika at a fairly high angle. Having reached this point on the shore of Lake Tanganyika, we find before us to the west a deep depression, soundings showing a depth, in places, of 180 fathoms, and away over the lake, at a distance of 18 miles, there is the rocky promontory of Kituta and some islands. This pro- montory and these islands have the following structure: — The eastern face of the cape is a precipitous cliff of red sandstone and quartzite, rising to a height of 600 to 800 feet ; on its western side the ridge slopes much more gradually for two or three miles, until it dips under the channel which separates the island of Kinyamkolo from the mainland. Like that of the Kituta promontory, the 48 THE TANGANYIKA PROBLEM. eastern faces of these islands are steep, precipitous cliffs, com- posed of somewhat metamorphosed sandstones, which have been worn out at their bases into caves by the great ocean- like surf of the lake. On the western side these islands slope less steeply into the lake, and from their crests can be seen, over an expanse of deep blue water, 25 miles away, the gigantic scarps of the main western coast- line of Tanganyika. Soundings in the open water, between the islands of Kinyamkolo and the great western scarps, showed great depths, 200 fathoms and upwards being encountered ; but not far from the abrupt west coast, some distance north of Mbeti, there is a submerged plat- form, which rises slightly, like the islands, from west to east, and terminates in that direction in a submerged cliff, its relation to the main western scarps being shown in Diagram facing p. 44. Passing still westward over the western scarps, we find them to be composed of massive quartzites, sandstones, conglomerate and shales, 2,000 feet and more being exposed along the main western coast-line of the lake, and, finally, having ascended the magnificent red and yellow precipices which these exposures form, we reach a table-land, but one which has everywhere a slight dip to the west, and we are now, as a matter of fact, on the long main western slope of the whole continental mass. Through the contemplation of the above facts relating to the structure of the Great Central Range in the region of the south end of Lake Tanganyika, it will have become apparent that the phenomena presented by the crest of the ridge are here more complex than in the region of Nyassa, although it will, at the same time, have also become obvious that these phenomena are precisely similar in kind. The complexity in this region is due, in fact, simply to the existence of a THE TANGANYIKA PROBLEM. 49 succession of ridges, or to a succession of cusps in trans- verse section, and, when closely compared with the cross section of Nyassa, the structure of the Great Central Range in the region of the south end of Tanganyika will be found to have the following peculiar features : — There are, first, two lines of up-push, one on either side of Lake Rukwa, then a more or less flat space, then a greater upraised ridge flanking the eastern shores of Lake Tanganyika ; next there is the depression of Tanganyika itself, and then another up- raised ridge. In Nyassa the phenomena would be produced, as we have seen, by a double fold in the earth’s surface, like that which can be made in a sheet of paper, while in the region of the south end of Tanganyika the whole of the phenomena are at once intelligible, if we suppose that four wrinkles, four such folds, have been upraised. These are not fanciful comparisons, but appear to me to represent, in reality, exactly what has actually taken place. There is, in the physical phenomena of all this Central African region, an extraordinary simplicity, and a primitive boldness which produces on the mind an indelible impression that we are here dealing with a unique example of the initial stages in the shaping of a continental mass, and not, as in most other cases which confront the geologist, with the confused and denuded relics of activities which have long since become extinct. From the preceding examination of portions of the great central “ graben,” it will be observed that the conception which we thus gain of the nature of these valleys, as phenomena incidental to mountain building, is not by any means the same as the original conception entertained by Suess, and more recently repeated by others, all these authors having regarded the valleys as the result of vertical falling in of land from the surface 4 5° THE TANGANYIKA PROBLEM. of high and ancient plateaux. Each so-called rift, how- ever, seems far more correctly conceived as a plait-valley, or what I shall call in future a “ eurycolpic fold,” the word, from the Greek evpvKo\7roEN* [ To face page 89. THE TANGANYIKA PROBLEM. 89 Chapter VII., totally different from that of Tanganyika. It is, in fact, the fauna of a great fresh-water pond, and in the plains which run southward under the modern volcanic debris of the Mfumbiro Mountains, and to the north, actually dip under the water of the Albert Edward Nyanza, there were encountered river-cuttings, in which were exposed old lake-beds, underlying the superficial drift and gravel of the valley’s floor. In these beds- we found fossil-shells, all of which are identical with those now found living in Lake Kivu, and embedded in the magnesium incrustations of its shore. Further north, the Albert Edward was found to contain the same fresh-water shells alive which inhabit Kivu, and are found fossilised in the beds extending under the more modern volcanic dam between the lakes. There is now no connection whatever between Lake Kivu, which stands at an altitude of 4,841 feet to the south of the Mfumbiro Mountains, and the Albert Edward Nyanza, which lies 2,000 feet lower, away to the north of them, the watershed, formed by the modern volcanic cones, rising everywhere to something over 7,000 feet, between the lakes ; but we have in the above facts in- contestable evidence that the waters of Lake Kivu and those of Lake Albert Edward Nyanza were, at some no very remote time, in connection. The formation of the modern volcanic dam has simply resulted in the banking up of the water of Lake Kivu, until it finally flowed to the south over the high gneissic ridges, which originally sepa- rated the Kivu and Tanganyika valleys. The effect of this change in the position of the central watersheds of the African continent has unquestionably had an immense effect, both in the regions far to the north and far to the south of it. By the formation of the volcanic dam, the MFUMBIRO MOUNTAINS r CONGO WATERSHED WESTERN SCARPS OF CENTRALuGRABEN C REAT K I RU N GA- CH A-MOTO KIRUNGA-CHA-CONGO 11,350 FT. KARISI MBI 13,000? i SABI I N EASTERN SCARPS OF GREAT CENTRAL"CRAB,EN" ✓ 'I Bird’s-eye view of the Mfumbiro Mountains and the North End of Lake Kivu. [ To face page 89. 9° THE TANGANYIKA PROBLEM. water from the whole drainage area of Kivu has been entirely cut oft from the Albert Edward Nyanza and the Nile. And we see visible traces of this in the existence of old water-marks and beaches running round the shores of the Albert Edward and Albert Nyanzas, 50 feet above their present level. Everywhere to the north along the course of the Nile there are similar physical evidences and traditions of the disappearance of old lakes, and it is extremely probable that the shrinkage of the upper waters of the great river of Egypt, which is recorded in history, is possibly still going on, and is directly due to the recent changes which have taken place in the modern volcanic dam between Kivu and the Albert Edward Nyanza. Turning now to the south of the region of the volcanic dam, the effects produced by its formation have been no less conspicuous and strange. The whole drainage area of Kivu has been added to that of Tanganyika ; and it is a most remarkable fact that the outlet of Kivu, the Rusisi River, is five or six times bigger than the Luakuga, the outlet of Tanganyika itself. Were we, therefore, to cut off the Rusisi River from Tanganyika, that lake would altogether cease to overflow. The water of Tanganyika is somewhat salt. It seems to be fresher now than when Livingstone and Stanley examined it, while, as both these explorers aver, there are traditions among the Arabs that, in the recollection of living men, it was a lake which had never flowed out at all. From these considerations it would, therefore, seem quite probable that after Kivu had filled up through the formation of the volcanoes, its overplus of water flowed into Tan- ganyika for a great number of years, until the level of this lake was also raised to such a degree, that it, in like View from a portion of the summit of Ivirungo-cha-gongo and of an extinct secondary crater looking North towards the Albert Edward Nyanza. The dark streaks are the recent lava streams extending North for 40 miles over the plains. [ Prom a Photograph. THE TANGANYIKA PROBLEM. 93 manner, succeeded in forcing its way out, and cut a channel for its overflow to the west into the Congo. This view of the matter will at once explain the fact that Tanganyika has otherwise unaccountably fallen some 40, possibly a. great many more, feet within no great number of years, the overplus ol water in it having worn away the The steep West Coast of the Albert Edward Nyanza. The Western side of the Great Central eurycolpic fold. channel to the west to such an extent that it will never regain its ancient high level. The portion of the great central eurycolpic fold which ex- tends beyond the Mfumbiro mountains, and contains the Albert Edward Nyanza, is similar in character to the more southern portion of the same great valley which contains Lake Tanganyika, until we reach the northern shores of the Albert Edward Lake. Here, however, a new feature 94 THE TANGANYIKA PROBLEM. presents itself, in the shape of the massive ranges of the Mountains of the Moon (see diagram facing p. 106). These ranges, which rival the Alps in magnitude and in the sublimity of their scenery, lie along the eastern edge of the depression, and appear, in fact, to stand out into it beyond what was originally its eastern face. That these Mountains of the Moon are not older than this portion of the eurycolpic fold seems to be supported by many con- siderations. The undulating Victoria Nyanza plateau, which terminates abruptly in the eastern wall of the valley to the south of the range, and to the north of it in the abrupt east coasts of the Albert Edward and Albert Nyanzas, is generally composed of schists and gneiss ; but between the Albert and Albert Edward Nyanzas the plateau terminates abruptly upon the steep eastern slopes of the Mountains of the Moon themselves. Here, however, the layers of schist, of which the plateau is composed, instead of being broken, as they are along the course of the depression to the north and south of the range, are bent and piled up upon the steep flanks of the mountains themselves ; and it is only, as we found at a great height, 12,000 ft. and more, that the layers of steeply uptilted schist come to an end upon the eastern slopes of the massive old amphibolites, which form the central cores of the range, and appear to have been bodily thrust through them. That this ap- pearance of intrusion of the cores of the mountains through what was originally a horizontal mass of gneissic material, is a reality, appears to me to be supported by numerous facts relating to the structure of the range. Thus, although the schists come to an end at a height of about 11,000 ft. on the eastern flanks of the mountains, they reappear on the western slopes, pitching at a very high View looking up the higher portion of the Mobuko Valley with some of the central peaks of the Mountains of the Moon. [. From a Sketch by the Author. THE TANGANYIKA PROBLEM. 97 angle into the Semliki valley. If we consider a section through the Albert Edward Nyanza south of the Moun- tains of the Moon, we have the following structural features. Passing over the Victoria Nyanza plateau, from east to west, we find that this plateau in general slightly rises until the eastern edge of the central depression is reached, when it suddenly breaks away in a fault face, or a succession of fault faces, above the extensive plain which, here, forms the flat bottom of the lake. On the other side of the Albert Edward Nyanza, in the west, there is an opposite and corresponding series of fault faces, and instantly beyond their crests we find ourselves on the Congo watershed. If now, in compari- son with this, we consider a section through the great central depression somewhat farther to the north and through the Mountains of the Moon, we have the following features. Beginning again in the east, we find that the Victoria plateau ends upon the flanks of the great range, but that where it does so the layers of schist of which it is composed are abruptly tilted and piled upon the flanks of the range. Passing still farther to the west, we find a succession of huge ridges of intruded lower rock, which project through and above the uptilted schists ; and on the other side of these intruded ridges we encounter the schists again, lying on their western flanks, and sloping at a very high angle into the floor of the Semliki valley. The surface of this valley, where I could examine it [at the source of the Semliki from the Albert Edward Nyanza, and near its mouth in the Albert Nyanza], was composed of little dis- turbed layers of modern alluvium and old lake deposits, containing in both situations the fossil shells of the molluscs which now live in the Albert and Albert Edward lakes. 7 9S THE TANGANYIKA PROBLEM. Passing still farther to the west, these lacustrine plains came to an end abruptly beneath the western flanks of the Semliki valley, and over these we are instantly once more on the Congo watershed. It will thus be seen, that in the regions of the Mountains of the Moon, we are still dealing with precisely the same phenomena encountered when we considered the construction of the great eurycolpic depression in the region of Nyassa and Tanganyika. That is to say, the gross, geological characteristics of the region have been brought into existence by a folding of the earth’s surface along two parallel lines of up-push, each of which is apparently coincident with the faulted walls of the depression itself. The only modification of this process which has occurred in the region of the Mountains of the Moon being, that, for about a hundred miles on the eastern side of the Semliki river, the crinkling and uprising has been so violent that the lower rock masses underlying the Victoria Nyanza plateau have been forced through it, just as if the igneous base, upon which the sandstones of Mount Waller rest, was to be forced up and up, until it tilted the sandstones on both sides of it, and, finally, towered above these old aqueous deposits in ridges of intrusive igneous matter, 15,000 ft. to 16,000 ft. in height. I do not think there is the least doubt that this has been the real method and nature of the formation of the Ruwenzori range, and, consequently, we must view the massif as simply a more vigorous and local expression of the flanking mountain chains which appear along the sides of the eurycolpic folds in other places, as in the region of the Livingstone range on Nyassa, or the great outstanding masses which, at several points, tower over the general level of the sides of the Tanganyika basin. From these considerations it would appear that Stuhl- / * The snow summit of Ingomwimbi, one of the high peaks in the Mountains of the Moon. THE TANGANYIKA PROBLEM. IOI mann’s* view that the range was the expression of a double line of faulting, or was in reality a block mountain, is not altogether correct. Indeed, Stuhlmann did not acquire a sufficiently extensive acquaintance with the range in order to grasp its real nature, but his view was a distinct advance upon Stairs’ suggestion that the mountains were the denuded fragments of an old volcanic cone. Unquestionably, Stuhl- mann’s was a very shrewd guess at the nature of the mountains which confronted him, and, from his points of observation, perhaps the only supposition he could make. Scott Elliot appreciated the existence of the up-piled schists on both sides of the range, but his theory that the mountains are a sort of oval boss of raised material, independent of the surrounding geological features, cannot be made to fit in with the actual facts of the case. The range, as we have seen, seems in reality to be merely an accentuated expression of the same sort of up-push as that which has raised the eastern wall of the eurycolpic folds north and south of it. What has happened appearing to be that over this more actively raised region the lower materials, which underlay the sides of the depression, have been pushed through the surface in a succession of pro- truding, lenticular masses, the axes of which run north and south. The amphibolites, as I have found myself, do actually run out in this way towards the north of the range. Although I have in the above briefly outlined the most important structural features of the Mountains of the Moon, it is necessary to refer to certain other matters in connection with them at the present time. Before my visit to the range, in March, 1900, no one had reached the snow-line on these mountains, and consequently * Stuhlmann. Mit Emin I’acha, 102 THE TANGANYIKA PROBLEM. the rather interesting fact that it lies as low as 13,500 ft., had not been ascertained. It was, moreover, not known whether there were glaciers on the range, Stuhlmann had expressly doubted it ; and owing to a misapprehension under which Sir Harry Johnston persistently labours with respect to his own later explorations in the district, it is necessary, although wearisome, to repeat, that prior to our visit only three explorers had made any serious attempt upon the higher portions of the range. Stairs had ascended to 10,000 ft. on the north-western spurs ; Stuhlmann had reached 12,000 ft. on the west of the range, while Scott Elliot had reached heights of 10,000 ft. and 12,000 ft. on both the eastern and western flanks. My own exploration was made up the Mobuko valley, terminating on the northern snow ridge of Sitchwi, and occupied about three weeks. My highest point reached on the top of this ridge was 14,900 ft. After my return to Fort Jerry my colleague, Mr. Fergusson, made a separate journey to the mountains and reached a point nearer to the Mobuko glacier, 14,600 ft. This was the point subsequently reached by Sir Harry Johnston, who makes it 14,800 ft. Still later, Mr. Wilde, an officer of the Uganda Protectorate, reached a point on the same or an adjacent ridge of 14,900 ft., but, as he was using an aneroid his altitude is probably rather over estimated. As to the question of the height attained by any of the other snow ridges and peaks of the range, I came to the conclusion that there was nothing more than a thousand feet higher than the point on which I was, either to the north or to the south, and this would give an outside altitude for the highest peaks of the range of, say, [6,500 ft. Curiously enough, this is practically the same height which all the older explorers ascribed to the range — namely, Sitchvvi, the northern snow ridge of Ingomwimbi from a point about 12,500 feet. THE TANGANYIKA PROBLEM. 105 Stanley, Stuhlmann, Stairs and Scott Elliot. Sir Harry Johnston, on the other hand, who did not get as high as myself, holds 20,000 ft. as a minimum for some of the numerous peaks which he did not even attempt, but as he The broad glacier on Sitchwi, the Northern snow ridge of Ingomwimbi. says he is judging simply by his eye, his contention is necessarily without any weight for actual mountaineers.* * Those interested in this matter will find it further discussed by me in an article entitled “ First ascent of one of the snow ridges in the Mountains,” and published in the Alpine Journal for May, 1902 ; and if the statements contained in this are compared with Sir H. Johnston’s account of his later visit to the mountains, published in the journal of the Royal Geographical Society, it will probably become apparent that my appreciation of a “ literary capacity ” in Sir Harry’s method of treating matters of fact was not misplaced. io6 THE TANGANYIKA PROBLEM. A fairly reliable method is, however, afforded by a study of distant photographs of the range, now that the height of the snow-line has been ascertained. From the east the whole range may be seen rising from plains which stand at about 4,000 ft. In photographs obtained from these plains the snow, which I ascertained to begin at 13,500 ft., only occupies a third, at the outside, of the total height of any of the peaks, above the plains. The snow-line is, therefore, 9,500 ft. above the plains, and, consequently, these data indicate that 16,700 ft. is an outside height for any of the peaks. CONGO W / EASTERN SCARPS OF GREAT CENTRAL " G R. ABEN ” MOON CHAPTER VI. AFRICAN PARK-LANDS, THEIR APPEARANCE ON ALLUVIAL FLATS CONSIDERED AS EVIDENCE OF RECENT PHYSICAL CHANGE. In the last chapters I have referred to the changes which have unquestionably gone on in the African interior in the past, and I also pointed out that there was direct and incontestable evidence to show that in some places these changes are even at the present time in full swing. It may, therefore, not be out of place to refer in the present chapter to another series of phenomena, which have for a long time been most perplexing, but which, when rightly interpreted, appear to show, in an equally conclusive manner, the extraordinary impermanence of the terrestrial conditions over very wide- areas of the continent in which they occur. These observations have nothing to do with geology as it is ordinarily understood, being related to certain features of the flora of Central Africa, which, when first encountered, are utterly perplexing, and seem to indicate the past or present operations of a landscape gardener who is not there. If we were to land now on the banks of the Upper Shiri river, we should find, after pushing our way through ioS THE TANGANYIKA PROBLEM. the reeds and growths which fringe the waterside, that we had entered a country having for all practical purposes the characters of a park. It is made up of wide lawns of short grass, in which there stand clumps and isolated specimens of different sorts of trees. These trees are not, however, crowded together, but are grown just as they are in a piece of English park-land, so that they can be seen to the best advantage, with wide open spaces in between. The whole scenery in such a place is so peculiar and artificial in outward aspect, that I cannot perhaps describe it better than by saying that it bears a very remarkable and close resemblance to that area of kept ground which is known as the Arboretum in the Royal Gardens at Kew. There is no tangle, no forest, the scenery is delightful ; it is in fact so extremely beautiful, that, could it be divested of its sweltering tropical con- comitants, I am inclined to think that these natural parks, when compared with those of the landscape gardener at home, would be generally admitted to be the more pleasing of the two. Unnatural-looking park-lands of this description are very characteristic of vast areas of the African interior, and they have, in consequence, naturally attracted the attention and provoked the admiration of many explorers besides myself. Thus we find Stanley and Stuhlmann, Emin Pasha and Cassati, Joseph Thomson and Sir Richard Burton, all drawing attention to the existence and peculiar appearance of these parks. As a matter of fact, they cover very large areas of tropical Africa, both in the interior and on the coast. I myself have encountered them on the plains behind Beira, on the great alluvial plains bordering the Zambesi river, on the similar flats flanking the Lower and Upper Shiri river, all over the great plains sur- VIEW OF THE SNOW PEAK KANYANGOGWI And part of the snow ridge of Sitchwi. Between them is seen the Mobuko Glacier, one of the three discovered during the second Tanganyika Expedition. From the upper part of the Mobuko Valley, at an altitude of 13,500 ft. ( From a Sketch by the Author .) THE TANGANYIKA PROBLEM. 109 rounding Lake Shirwa, on the alluvial flats and plains bordering Lake Nyassa to the east, to the west and to the north. They appear again in many places on the high interior plateau between Nyassa and Tanganyika; they cover extensive regions of old lake deposits on the shores of Tanganyika itself; while they reappear on the plains south of the Albert Edward Nyanza. They are to be found again in patches mixed up with the true forest all along the course of the Semliki valley and on the shores of the Albert Nyanza. They are further to be found on old lake deposits and alluvial flats in some parts of Uganda, and I am informed that they are also a charac- teristic feature of many portions of the west coast of Africa, and of the hinterlands beyond it. Park-lands, districts having the peculiar characteristics of a kept park, thus cover immense areas in the African interior. They cover, as a matter of fact, thousands upon thousands of square miles, and the more closely we examine them the more curious and perplexing their existence appears. We have, indeed, only to look for an instant at a district such as that to which I have alluded on the Upper Shiri river, in order that a variety of questions shall present themselves, which are all more easily put than answered. In the first place, why have these districts assumed the characters of an artificial park ? Why are the trees isolated as if they had been grown for show ? Why is there no thick bush covering the ground, and converting the whole place into a thick jungle ? Why are there so very many different sorts of trees ? If we meet with a park in England the mere fact of its existence implies the present or past operations of a park- keeper or a landscape gardener, who was not only an agent I IO THE TANGANYIKA PROBLEM. independent of the natural environment of forest trees, but who acted in a persistent manner, clearing away the bushes and brambles off the lawns, in which he planted great trees and little trees, so that their limbs and foliage could grow luxuriantly and be seen. Moreover, in England, when such a park has once been formed by the agency of man, it is absolutely necessary that the operations of the gardener shall go on and continue, or the park will inevitably and quickly lose its artificial characters. Thorns and briars and bushes will quickly spring up upon the grass, and in a few years the park will have gone back again to what we are accustomed here to regard as a state of nature ; or, in other words, it will have become converted into a trackless waste of old and young jungle. In England or Europe a park-land is thus an artificial product, and is an impossibility, unless there is someone ready and willing to hold the natural tendencies of the vegetation in check. In tropical Africa, on the other hand, precisely the same floral arrangement is produced, but no human agency has had anything to do with it ; and the existence of these natural park-lands presents us with a ready-made and an extraordinary puzzle, which it is interesting to try to understand. In attempting to account for the appearance of park- lands, the most natural supposition to make is, of course, that of inequality of dampness or character of the soil, which is sufficient to allow some kinds of trees to grow in one place, some in another, and grass in between ; but although this view of the matter looks very nice and promising at first sight, its value is absolutely destroyed by the facts of the case. I have on several occasions, when in a park-land, set my men to trench and dig in different directions, and then examined the soil, with the (i.) Shore of the Albert Nyanza with reeds and a few young euphorbias springing up. (2.) The euphorbias form a spot of shadow in which more delicate plants survive. (3.) Euphorbias becoming buried in the bush which they originally sheltered. (4.) The euphorbias choked with bush, and the formation of typical park lands. The figures run from right to left. THE TANGANYIKA PROBLEM. 113 result that I could find no difference whatever, either in its dampness or its consistency, at least, any that it was possible to correlate with the different plants that grew upon it. In the soil under a great clump of acacias there was as little moisture, and it was of the same consistency as that upon which there was nothing but a scanty covering of grass ; neither can difference of climate or rainfall be invoked. Park-lands occur in the Semliki valley where it is very wet, and also on the Albert Edward plains where it is very dry. From general observations, however, it soon becomes apparent that park-lands are by no means distributed haphazard over the surface of the country ; they never occur, for example, on hill sides or upon rocky ground. They are invariably found upon alluvial plains like those formed by rivers, or upon old lake deposits ; that is, they only occur on flats made up either of blown sand or of ground of aqueous origin ; but although they are definitely related to fiats of the above sorts, this fact at first sight perhaps makes the whole matter more perplexing- still, for such flats are by no means invariably covered with parks. Thus there are wide districts on the Semliki plains which are covered with heavy forest, and there are similar alluvial areas covered with heavy forest only along the Zambesi river, and indeed in many places elsewhere. What can it be, then, which in some places inaugurates and maintains a natural African park ? This question is a great puzzle, and the answer to it is not at all apparent on the surface of things. I obtained, however, what there is every reason to believe is a clue to the whole matter, during my visit to the Albert Nyanza. On that journey I descended from the western slopes of the Mountains of the Moon on to the plains of the great central valley, which I 8 !I4 THE TANGANYIKA PROBLEM. crossed diagonally until I reached the lake itself. On this series of marches we passed off the mountain slopes on to plains of alluvium and old lake deposit in which there are the remains of fresh-water shells similar to those which still live in the lake ; and from this, it is obvious that the lake once extended over these regions which are now covered with thick forest of different sorts. As we neared the lake we passed out of the forests, first into park-lands and then over steppes with only a very few trees, and finally on to the absolutely treeless salt wastes bordering the shores of the Albert Nyanza. We were travelling over the old lake deposits all the way, and the shelly remains became fresher and fresher until we actually reached the shores of the lake itself. Moreover, on the western shores of the lake we found old water marks and beaches which show, in as con- clusive a manner as similar terraces on the west coast of the Albert Edward Nyanza, that the lake has steadily fallen and receded to the north during a number, but an unknown number, of years. From this it will be obvious that in our marches from the Mountains of the Moon to the shores of the Albert Nyanza, we were passing over land which had been covered with water at a more and more recent date, and conversely, as we returned over the road we had come, we were passing over land which had been land for a longer and longer time owing to the gradual northern recession of the lake ; and the different age of the land was related to a different type of flora which was growing on it. This differ- ence in the character of the vegetation encountered during the journey has been represented in the figure on page 1 1 1, which is a combination in sequence of a number of drawings and photographs I obtained of the kinds of vegetation through which we passed. By the lake shore there was a belt of reeds, and beyond this, almost desert THE TANGANYIKA PROBLEM. 115 steppes over which the fierce tropical sun blazed without any protection for many hours during the day. In such places the noon-day heat is fearful, and the men on this particular occasion, as is often the case on exposed plains near the Equator, were hardly able to walk with their bare The pebble beaches of the West Coast of the Albert Edward Nyanza and old water-marks. feet on the hot ground. The surface of the earth was desiccated and sandy, but a few inches below there was an appreciable amount of moisture, due to the occasional storms which sweep over such plains and disappear almost as quickly as they form. Nothing but grass grew near the lake, and even this had evidently had a very bad time, for it was scraggy and white and bleached, and alternated with 8* 1 16 THE TANGANYIKA PROBLEM. patches of absolutely bare sandy soil. On these plains there were, however, in places, scattered over the surface of the ground, a few young euphorbia trees, the seeds of which had evidently been disseminated over the plains by the wind or birds, and as these hardy plants grew bigger on the older land further from the lake shore, 1 noticed that in the hot glare of noon their massive structures threw a patch of deep cool shadow round their feet. Farther away from the lake where the land was older and the euphorbias had con- sequently had time to grow proportionately bigger, the noon-day spot of shade had also correspondingly increased, and in the area of such shadow there were to be found varieties of plants, besides the grass, which here found protection from the fiery glare and heat, and were con- sequently able to grow. Among these plants struggling against the naturally adverse conditions of the plains under the euphorbia shadows there were thorn trees, climbing plants, and flowering shrubs, and when once these plants had in this manner got a footing on the plains they prospered like one of Germany’s protected industries and throve amazingly ; so much so indeed, that on land that was still further from the lake, and consequently still older, the thorns and bushes of various sorts were enveloping the euphorbias, which now appeared as rather choked growths in the centre of the bushy patches. Further away again from the lake there were many clumps of bushes and trees scattered in all directions over the country ; and in many of these were still to be found the dead or dying remains of the original euphorbia, to the protec- tion of which the bush patch owed its growth. The seven lean kine had here eaten up the seven fat kine, and in such districts we entered the typical scenery of an African park. THE TANGANYIKA PROBLEM. 117 Once started, the groups of trees and patches of bush which marked the graves of their former bene- factors, the euphorbias, spread gradually under the pro- tection of their own shadows, until finally the patches ran together and more or less coalesced into the ragged forest which covers the higher portions of these long- alluvial slopes. It will be obvious that we have thus, in the simple natural dispersion and growth of euphorbias over desert steppes and in their mode of growth, a completely satisfac- tory explanation of the formation of park-lands, and their relation to steppes and forest ; the process of their forma- tion being a natural sequence of events following upon the scattering-, through the agency of the wind or birds, of the seeds of a single tree. But at the same time the appearance of a park-land is seen also to be one phase in a series of changes which follow the retreat and drying up, or the change in position, of water on the face of the land. And further, it appears to be as true of the natural park as of the artificial one, that unless it is kept up, it must, in the course of time, disappear and become converted into more or less thick forest. Its appearance is simply the ex- pression of progressive physical change. There appears to be no agency, except a park-keeper which is capable of maintaining a park, any more in Africa than in England ; and perhaps the most singular, or at any rate the most interesting thing, which the foregoing observations teach us, is that the African parks are absolutely impermanent, and are, in reality, direct and incontestable evidence in them- selves of wide-spread physical changes in the lands on which they exist. But not only is the existence of a park-land evidence of recent change in any district in which it may occur, it is THE TANGANYIKA PROBLEM. 1 18 also indicative of change in one particular direction — that is, of the gradual drying up and shrinking of whatever lakes and open waters the district may possess. Hence the almost universal distribution of park-lands all over Central Africa is clearly indicative of one of two things : either the rainfall is becoming less over the whole of the equatorial regions, or the land is being gradually moved and changed and shifted in such a manner that the water on it is being drained off the interior as a whole. There is no evidence of any decrease in the rainfall, but as we have seen there is abundant evidence of continuing geological change ;■ and therefore we are driven to conclude that the existence of these extensive parks must be due to movements and shifting in the watersheds, and the general configuration of the land. Now, the only direction in which earth move- ments on a slow and extensive scale could effect this draining, is by that of a gradual raising and humping up of the interior ; and it is extremely interesting to find that the study of the features of natural parks, thus leads exactly to the same conclusion respecting the impermanence of the terrestrial conditions of the interior, that were indicated by the geological and physiographical considerations which I discussed in the last chapter. Although it will thus be seen that the history of African park-lands affords us another mode of demonstrating the geological impermanence of the African interior, and of the existence of a progressive series of physical changes which are still going on there, it should not be overlooked that their history is also not without a rather wide im- portance from a purely biological point of view. It teaches us in an analogous way to the matters connected with the formation of fresh-water faunas in general, which I described in Chapter II., that the main iloral characters of a country THE TANGANYIKA PROBLEM. n9 may not be clue, as we have hitherto been in the habit of supposing, to the struggle for existence between contending species, but entirely to the sorting influence of a progressive physical change to which the biological phenomena are related in a manner of a perpetually responding reflection. The struggle for existence betw-een contending species has had nothing to do with the chief features of the flora ol vast areas in the African interior which are now covered with natural parks. I 20 CHAPTER VII. GENERAL OUTLINE OF TIIE ZOOLOGY OF THE GREAT AFRICAN LAKES. In order that a general appreciation of the nature of the Tanganyika problem may be obtained, I have prefaced the more detailed examination of the components of the fauna of the lake, with an account, or rather with a series of enumerations, of the different animals which have hitherto been recorded in each of the other great African lakes. By this means certain features of the distribution of the fresh- water fauna of equatorial Africa are rendered self-evident, and numerous repetitions which would otherwise be necessary in the sequel can be avoided. But besides this, so much that is new has recently been added to our knowledge of the composition of the faunas which exist in the great African lakes, that it is now for the first time possible to deal with the general characteristics of the equatorial African fresh-water fauna, and the results of such a study lead not only to conclusions which are interesting, but which are also by no means without importance from the broader biological point of view. There are thirteen great African lakes, about the fauna of which it can be said that something definite is known at the present time. THE TANGANYIKA PROBLEM. I 2 I Beginning in the south these lakes are : — Shirwa, Nyassa, Kela, Bangweolo, Rukwa, Mwero, Tanganyika, Kivu, the Albert Edward Nyanza, the Albert Nyanza, the Victoria Nyanza, the chain of lakes in association with Beringo, and Lake Rudolf. The lakes actually examined during the Tanganyika expeditions were : — Shirwa, Nyassa, Kela, Tanganyika, Kivu, the Albert Edward Nyanza, the Albert Nyanza, the Victoria Nyanza, and Lake Nivasha. For our information about Lake Bangweolo we are dependent upon the observations of Livingstone, Weatherley, and the late M. Foa. Rukwa has been examined geo- graphically by Mr. Wallace, and some information respect- ing its fauna was collected by the German explorer, Dr. Fulleborn. Mwero has been examined, and the general characteristics of its fauna been ascertained through the exertions of Mr. Crawshay and H.M. Commissioner, Mr. Alfred Sharp. Beringo and the minor lakes in association with it were examined by Professor Gregory ; while for Rudolf we are dependent on the somewhat scanty observations made by Messrs. Donaldson Smith, Cavendish and Harrison. In making a general survey of the faunistic characters of these usually vast and often remotely isolated inland waters, . it will perhaps be most convenient to begin with what is known of the fauna of Lake Nyassa. The faunistic characters of this lake are to a large extent typical of those in the majority of the African lakes, and are an indi- vidualisation of what rnay be called the fresh-water fauna of Africa generally, this in turn being but a slight modification of the fresh-water fauna of the world. Nyassa lies in an immensely long and relatively narrow 122 THE TANGANYIKA PROBLEM. valley, the surface of the lake and the floor of the valley in which it is contained being far below the general level of the surrounding country. In consequence of this, whether the lake shores are flat and sandy or rocky and steep, they are always sooner or later backed up by precipitous flanking ranges which constitute the steep sides of the gigantic trough in which the lake lies. The surface of Nyassa is now 1,500 feet above the level of the sea, and it extends for about 350 miles north and south with an average breadth of 40 miles. In most parts of the valley the flanking ranges rise to heights varying between 7,000 and 10,000 feet, but at the extreme south there is a narrow valley lying between part of the Kirk range and the Mngochi moun- tains which serves to carry away the surplus waters of Nyassa over the Murchison Falls in the shape of the Shiri River which eventually opens into the Zambesi itself. Throughout the lake there is abundant evidence that its waters have fallen greatly, and also that they have fallen in a succession of drops ; for almost everywhere there are to be seen at least three old water-marks, old beach terraces in fact. The most conspicuous of these stands about 14 feet above the water level, and upon the well-marked flat top of the terrace which it makes all round the lake there are always numbers of immense baobab trees. Baobab trees are known to be extremely slow growing ; they will not grow, and they die, if they are in any way submerged, and consequently it is extremely probable that many centuries have elapsed since the water of Nyassa first sank from this old beach line, and there are at least two more clearly discernible terraces above and behind it. These general falls in the water level have in all probability been due to the successive wearing away of definite obstacles in the floor of the Murchison Falls. There are THE TANGANYIKA PROBLEM. 123 also not wanting indications that Nyassa was at one time directly connected with Lake Shirwa, and it is probable that at some more remote period both these lakes as one threw their surplus water down the valley of Lujenda River. Besides general falls in the level of Nyassa, there are, however, indications of local disturbances of level, such as those to which I have referred in Chapter III., and these changes are undoubtedly due to the geological distortion which is still going on along the course of the Great Central African Range, within the folds of which Nyassa lies. The floor of the lake valley is very deep indeed, reaching we found in one place 430 fathoms, and these results have been subsequently closely confirmed by my friend Lieu- tenant Rhoades. Over large areas, generally, in fact, the floor of the lake is as low and lower than the sea level ; and consequently we may say broadly that the lake valley is a chasm, or fold, in the earth’s surface over 300 miles long and about 50 miles broad, and 10,000 to 11,000 feet deep. Owing to its latitude the climate of Nyassa is warm and tropical, but since the lake’s surface is 1,500 feet above the sea level, the monotonous heat experienced on the equatorial African sea coasts is perceptibly tem- pered, especially at night. Observations have been made on the fauna of Nyassa by a number of explorers, by Sir John Kirk, Livingstone, Joseph Thomson, Mr. Crawshay and others, and in 1895 and 1899 extensive examinations of the lake were made by myself and my colleagues during the Tanganyika expedi- tions. The investigations undertaken during both these journeys extended over three months, and the results obtained throughout all the different regions of the lake which we examined were of such a definite and similar character that there can be no doubt we have pretty well 124 THE TANGANYIKA PROBLEM. exhausted the fresh-water fauna of Nyassa, at any rate so far as a knowledge of the different types composing it is concerned. When the fauna of Nyassa is considered in comparison with the fauna of the other great lakes of the world, it is perhaps chiefly characterised by its limitations. For it will be seen from the subjoined lists that it consists almost exclusively of fishes and mollusca. The fishes of Nyassa include representatives of seven families, and the contained genera and species which have hitherto been recorded are represented in the following table : — ClCHLID/E. I. Paratilapia robust a Gthr. 2. 9 9 afra Gthr. 3- 9 9 modesta Gthr. 4- 5 9 livingstonii Gthr. 5- 99 intermedia Gthr. 6. 9 9 dimidiata Gthr. 7- 9 9 longiceps Gthr. 8. 99 nototaenia Blgr. 9. Corematodns shiranus Blgr. 10. Petrochromis nyassa Blgr. 11. Tilapia shirana Blgr. 12. 9 9 mossambica Ptrs. I3- 99 kirkii Gthr. 14. 9 9 squamipinnis Gthr. !5- 99 rendalli Blgr. 16. 9 9 lateristriga Gthr. i7- 9 9 subocularis Gthr. 18. 99 johnstoni Gthr. 19- 99 lethrinus Gthr. 20. 9 9 tetrastigma Gthr. 21. 99 callipterus Gthr. 22. 99 williamsi Gthr. 23- ,, ait’-ata Blgr. 24. Docimodus johnstoni Blgr. 25. Cyrtocara moorii Blgr. 26. Hemitilapia oxyrhynchus Blgr. Mastacembelid.e. 27. Mastacembelus shiranus Gthr. SlLURIDiE. 28. Bagrus meridionalis Gthr. 29. Anoplopterus platyehir Gthr. 30. Synodontis zambesensis Gthr. CyPRINIDj®. 31. Labeo tnesops Gthr. 32. Barbus trimaculatus I’trs. 33. Bari litis guentheri Blgr. 34. Engraulicypris pit ignis Gthr. i 35. Pelotrophtis microlepis Gthr. j 36. ,, microcepkalus Gthr. CHARACINID/E. j 37. Alestes iniberi Ptrs. Cypri nodgnti i>.k. 38. Haplochilus johnstoni Gthr. Mormyrid.e. 39. Monnyrus discorhynchus Ptrs. 40. ,, catostoma Gthr. 41. Monnyrops zambanenje Ptrs. Curiously enough, beyond copepods, eye lops and other minute forms, there are no Crustacea in Nyassa besides THE TANGANYIKA PROBLEM. I25 the fresh- water crabs. There are no shrimps or prawns in the lake, and when we pass to the molluscan section of the fauna we find it to be constituted by the following characteristic fresh-water genera and species : — 1. Limnaea natalensis Krauss. 2. Isidora nyassana E. Sm. 3. ,, succineoides E. Sm. 4. Physopsis africana Krauss. 5. Planorbis (sp. indt.). 6. Ancylus (sp. indt.). 7. Ampullaria gradata E. Sm. 8. Lanistes olivaceus Sow. 9. ,, ovum Ptrs. 10. ,, ellipticus Marts. IX. ,, nyassanus H. Dohrn. 12. Vivipara unicolor 01. 13. Bithynia humerosa Marts. 14. ,, Stanleyi E. Sm. 15. Melania tuberculata Mull. 16. ,, simonsi E. Sm. 17. ,, nodicincta Dohrn. 18. Melania pergracilis Marts. 19. ,, polymorpha E. Sm. 20. ,, turritispira E. Sm. 21. ,, arcuatula Marts. 22. Unio mossambicensis Ptrs. 23. ,, leideri Marts. 24. ,, lechaptoisi Ancy. 25. ,, borelli Ancy. 26. ,, kirki Lea. 27. ,, nyassaensis Lea. 28. ,, hypsiprymnus Marts. 29. Spatha kirki Ancy. 30. ,, nyassaensis Lea. 31. Mutela alata Lea. 32. Corbicula radiata Phil. 33. ,, astartina Marts. There are no worms in Nyassa, and I never came across any polyzoa, nor did I, after careful searching, find any hydroids, and there appeared to be only one very incon- spicuous spongillid of doubtful affinity. Besides the above forms, there were representatives of the usual fresh-water protozoa, and nothing else. Nyassa, however, is extremely interesting from a zoological point of view. The fauna, as will have been seen, is typically that of a pond, and con- sequently it is highly instructive to ascertain how these fresh-water pond animals which stock it, and which usually live in the still waters of pools and puddles, have succeeded in colonising the deep waters of the huge lake, which sink in places to more than 400 fathoms, and over which a heavy surf is often running at all seasons of the year, a surf, in fact, which breaks in ocean-like lines of foam on 126 THE TANGANYIKA PROBLEM. the open beaches and wears out the hard masses of the rocky coast into fantastic points and bays. An examina- tion of the 800 to 900 miles of coast line which Nyassa possesses showed that the fauna, including the crocodiles and the fishes, was almost exclusively restricted to a narrow littoral zone, its molluscan section living, in fact, in groups in bays and sheltered inlets, and being hardly ever found alive along the sandy surf-swept beaches which occur in many portions of the lakes. These molluscs, moreover, did not extend into the deep water of Nyassa, and it was found, both on the first and second Tanganyika expeditions, that beyond 100 to 150 feet the lake was practically a fresh- water desert, there being encountered in its deeper water nothing but a few dead shells, the fragments of crabs’ shields and legs and other organic refuse, enclosed in fine, grey mud.* East and south of Nyassa there exists, at about the same level, Lake Shirwa. The faunistic characters of this lake have already been described in Chapter II., and conse- quently need not be fully repeated here, all that it is necessary to remember being the fact that the fauna of Shirwa unquestionably appears to have been at one time identical with that of Nyassa. This is shown, as we have seen, by the semi-fossilised remains which occur around the shores of Shirwa, but, owing to the geographical change which has separated Shirwa from Nyassa, and which has finally resulted in Shirwa having no outflow, this lake has become extremely salt, and the old Nyassan fauna has been killed out of it, except in the curious fresh- water oases which are still maintained at the mouths of the permanent rivers flowing into the lake. * For tables of the bathymetric distribution of mollusca in Nyassa see paper in Quart. Jour n. Micro. Sci., Vol. 41. THE TANGANYIKA PROBLEM. 127 Passing north from Nyassa, we find that, with the exception of the small crater pools occupying extinct cones which occur along the continuation of the lake’s valley towards Rukwa, there are encountered no fresh waters of any dimensions until we reach Kela, a small, nearly circular, mere, with practically no fauna at all in it, beyond some siluroid fishes and swarms of a minute species of Bithynici. Curiously enough, Kela is within 20 miles of Tanganyika, and yet there is not a single specific identity between the fauna characteristic of the two lakes. Far to the east of Kela lies the larger lake Rukwa ; it is a shallow sheet of water occupying the floor of a continuation of the Nyassa valley. Our most definite information respecting this lake is that obtained by Mr. Wallace,* who, in 1897, made an extensive geographical examination of the region, but who observed nothing but minute shells, probably those of Bithynici , and a number of siluroid fishes. f Far to the west, again, we have the shallow and extensive expanse of water which constitutes Lake Bangweolo. This lake is known to contain siluroid fishes, but Mr. Weatherley reports that it is “absolutely shell-less.” Somewhat further to the north lies Lake Mwero, a deep lake lying in a secondary fold similar to that of Lake Rukwa, to the east of Tanganyika. It has a much more profuse fauna than either Rukwa, Bangweolo, or Kela. Among the fishes occurring in Mwero, there have been recorded the following fourteen different types : — * Journal of the Royal Geographical Society , I S9 7- t I believe some further information respecting the fauna of Lake Rukwa has more recently been obtained by Dr. Fiilleborn, but no account of this appears yet to have been published. See Chapter IV., p. 73, of present work. THE TANGANYIKA PROBLEM. 1 2S LAKE MWERO. FISHES. 7. Chrysichthys sharpii Mgr. 8. Auchenoglanis biscutatus Geofl'r. 9. Synodontis zambesensis Ptrs. 10. Synodontis ornatipinnis Blgr. ClCHLID/E. 11. Paratilapia macrocephala Blgr. 12. Paratilapia mocruensis Blgr. 13. Tilapia natalensis M. Web. 14. Tilapia polyacanthus Blgr. In the above list the genera and species printed in italic are endemic to Lake Mwero. It is curious to note, moreover, that the Viviparas of Mwero are not absolutely unlike the Neothauma of Lake Tanganyika. In continuing: a general consideration of the faunas in the remaining great African lakes it will be convenient to pass now far to the north, omitting for the present Tanganyika altogether, and to consider the remarkable fresh water which is constituted by Lake Kivu. Kivu lies, as I have shown elsewhere, in part of the continuation of the same depression which contains Tanganyika, about 100 miles to the south of it. But beyond the fact that the outflow of Kivu finds its way into Tanganyika, the lakes appear to have no connection one with another, and their faunas are entirely distinct. The surface of Lake Kivu is 4,800 and odd feet above the level of the sea, and its water, as I have explained in Chapter V., contains a certain percentage of magnesium carbonate in place of the more usual modicum of sodium chloride which lake waters usually possess. From its great height the climate of Kivu is distinctly cool, and the fauna is poor in the extreme. We obtained the following fishes : — KIVU. 4. Paratilapia bloyeti Sauv. 1883. 5. Barbus altianalis Blgr. 1901. Mormyrid.*:. 1. Gnathonemus stanleyanus Blgr. 2. Mormyrus longirostris Peters. Characinhve. 3. Ilydrocyon lineatus Blkr. 4. Alestes macrophthalmus Gthr. 5. Alestes lemairii Blgr. SlLURID.-E. 6. Schilbe mystus L. (?) 1. Tilapia nilotica L. 1757. 2. Tilapia burtoni Gth. 1893. 3. Paratilapia vitata Blgr. 1901. THE TANGANYIKA PROBLEM. 129 There was a small variety of Planorbis , a small Bithynia and Melania tuberculata among the gastropodean molluscs, one or two species of fresh-water bivalves, closely allied to the unios found generally in the African lakes, and apparently nothing else ; the most striking feature about the fauna of Lake Kivu being the apparent absence of Viviparas. North of Kivu we have the Albert Edward and the Albert Nyanzas, the faunas of which can very well be treated together. These two lakes are distinctly interesting. They lie at the respective altitudes of 3,000 and 1,700 feet above the level of the sea ; or, in other words, they repeat the conditions with respect to altitude and climate which were encountered in the case of Tanganyika and Nyassa. The fauna of both the Nyanzas is abundant — perhaps quite as abundant as that of Nyassa — but in both lakes it is composed of only a very few forms of fishes, and half a dozen species of inollusca at the most. In many portions of both the Nyanzas the beach is composed of actually nothing else but the shells of Melania tuberculata. In the Albert, so far as is at present known, the fauna consists of the following types : — ALBERT NYANZA. FISHES. Petrochromis Andersoni (Blgr. ).* 1. Planorbis adowensis Bgt. 2. ,, apertus Marts. 3. Ampullaria stuhlmanni Marts. 4. Vivipara rubicunda Marts. 5. Cleopatra pirothi Jick. 6. Bithynia alberti E. Sm. 7. ,, walleri E. Sm. MOLLUSCA. 8. Melania tuberculata Mull. 9. ,, liricincta E. Sm. 10. Unio acuminatus H. Ad. 11. ,, bakeri H. Ad. 12. Mutela nilotica Fer. 13. Corbicula radiata Phil. 14. Sphrerium species indeterminate. * It is to be noted that most, if not all, the other species of fish found in the Albeit Edward occur in the Albert as well. Only the new forms were, however, collected. 9 130 THE TANGANYIKA PROBLEM. The fauna of the Albert Edward Nyanza is practically identical with that of the Albert Nyanza, the following types having been obtained : — ALBERT EDWARD NYANZA. LePIDOSIREXID/E. 1. Protopterus aethiopicus Heac. Cyprinid*. 2. Barbus fergussonii Blgr. 3. ,, edwardianus Blgr. FISHES. Si lurid*:. 4. Clarias lazeras. 5. ,, moorii Blgr. ClCHLID*. 6. Tilapia nilotica Blgr. 1. Planorbis sudanicus Marts. 2. Ampullaria erythrostoma Rv. 3. Vivipara unicolor Ol. 4. Cleopatra pirothi Jick. 5. ,, exarata Marts. 6. Bithynia alberti E. Sm. MOLLUSCA. | 7. Melania tuljerculata Mull. 8. Unio stuhlmanni Marts 9. ,, ngesianus Marts. 10. Corbicula radiata Phil. 11. Sphcerium species undetermined. Passing to the east, we encounter the Victoria Nyanza, the largest of all the African fresh-waters. This great lake is swarming with living things, but the fauna, in a specific sense, is by no means very profuse. When the genera of which it is composed are considered, the fauna of this lake is very comparable to that of Nyassa, and up to the present is known to consist of the following types : — VICTORIA NYANZA. FISHES. ClCHLID*. 1. Paratilapia longiros/ris Hilg. 2. ,, cavif rons Hilg. 3. ,, retrodens Hilg. 4. ,, serranus Blgr. 5. 'I'ilapia nilotica Cuv 6. ,, nuchisquamnlata Hilg. 7. ,, sauvagii Pfeff. 8. ,, oblit/nidens Hilg. Mastacembelid*. 9. Mastacembelus , sp. Si LURID*. 10. C/arias Lay era C. & V. 11. Synodontis afrofischcri Hilg. Cyprixid*. 12. Labeo forskalii Rupp. 13. Labeo rueppellii Pfeff. ,, victorianus Blgr. 14. Barbus pagenstecheri Fisch. 15. ,, tri macula/ ns Ptrs. 16. Discognathus johnstoni Blgr. THE TANGANYIKA PROBLEM. [3> i7- 18 3* 4- 5- 6. 7- 8. 9- io. IX. 12. I3- r4- !5- 1 6. 17- 18. Fishes — continued. Characinid.e. Alestes rueppellii Gthr. Cyprinodontid.e. Fundulus tceniopygus Hilg. Mormyrid/e. 19. Mortnyrus oxyrhynchus Geoff r 20. ,, longibarbis Hilg. Lepidosirenid.e. 21. Protop ter us annectens Ow. , , cethiopicus Hack. MOLLUSCA. Limncea nyanzse Marts. ,, debaizei Bgt. Isidora trigona Marts. ,, strigosa Marts. ,, transversalis Marts. ,, forskali Ehrbg. Physopsis ovoidea Bgt. Planorbis sudanicus Marts. ,, choanomphalus Marts. ,, victorise E. Sm. Ancylus stuhlmanni Marts. Ampullaria nyanza1 E. Sm. ,, gordoni E. Sm. ,, bukobas Marts. ,, ovata 01. ,, emini Marts. Lanistes schweinfurthi Ancey. Vivipara unicolor Ol. 19. Vivipara rubicunda Marts. 20. ,, meta Marts. 21. ,, phthinotropis Marts. 22. ,, constricta Marts. 23. Cleopatra guillemei Bgt. 24. Bithynia humeroso Marts. 25. Melania tuberculata Mull. 26. SEtheria elliptica Lm. 27. Unio lourdeli Bgt. 28. ,, hauttecceuri Bgt. 29. ,, multicolor Marts. 30. Spatha trapezia Marts. 31. Mutela subdiaphana Bgt. 32. ,, bourguignati Ancey. 33. Sphrerium nyanza.1 E. Sm. 34. ,, stuhlmanni Marts. 35. Eupera parasitica Parr. POLYZOA. 1. Species indt. The Victoria Nyanza stands at an altitude of nearly 4,000 feet. It is higher than either Tanganyika or Nyassa, in most parts its waters are shallow and its shores reedy swamp fringes, but there are many island archipelagoes where the coasts are steep, and the waters which surround them deep and rough. In nature the depression in which the Victoria Nyanza lies is totally unlike those of most of the other African lakes which we have hitherto con- sidered. It is, however, very comparable to the depression occupied by Lake Bangweolo, the Victoria Nyanza being, in fact, nothing more than a gigantic rain puddle, occupying 9* 132 THE TANGANYIKA PROBLEM. a shallow, saucer-shaped depression, lying between distant eastern and western lines of mountains and valleys running north and south. Further to the east again we have another series of lakes, lying mostly along the course of the great fold which runs from the vicinity of Kilima Njaro to the Red Sea. In these lakes Rudolf, Stephani, Beringo, Nivasha, Elimantita and Nakaroo, there is little of zoological interest. But it cannot be said the faunistic characters of Rudolf are as yet fully known, still at the same time, from the observations of Mr. Donaldson Smith and Mr. Harrison, who both spent some time about the lake, it is quite clear that neither in the direction of fishes nor shells is it faunistically remarkable in any way. In Rudolf there have been recorded the following forms of fishes : — POI.YPTEKID.I-:. I. Polypterus Senegal us. Characintd.-e. 2. Hydrocyon forskalii. 3. Alestes baremose. 4. ,, nurse. 5. Citharinus geoffroyi. 6. Distichodus. RUDOLF. FISHES.* Si lurid.*:. 8. Clarias mossambicus. 9. Auchenoglanis biscutatus. 10. Synodontis schall. 11. ,, zanzibaricus. 12. ,, citernii. 13. Mochocus niloticus. Serkanid.e. 1 4. Lates niloticus. ClCHI.ID/E. Cyprin ID,®. 15. Tilapia nilotica. 7. Neobola bottegi. | 16. ,, zillii. Nothing sufficiently definite appears to be known of the invertebrate fauna of Lake Rudolf to make a list advisable. * This list of fishes from Lake Rudolf is partly taken from Vinciguerra Ann. Mus. Civ. Genova (2) xix., 1898, p. 241, the nomenclature having been kindly revised for the present work by Mr. Boulenger. THE TANGANYIKA PROBLEM. i33 From the characters of the faunas encountered in the widely separated African lakes which we have now briefly considered, it will have become clear that throughout a large number of these fresh-water lakes the fauna is the same in type although the number of genera and species composing that of any lake in particular may vary considerably. On this account I have intentionally postponed the con- sideration of the fauna of Lake Tanganyika, so that the very important fact that there is a type of fresh-water fauna which generally characterises the tropics of Africa may be fully appreciated, and in order that the exceptional fauna of Lake Tanganyika may be considered in comparison with the general characteristics of the rest. Turning now to the fauna of Tanganyika, we find that up to the present time the lake has been found to contain some 200 and odd different species of aquatic and semi- aquatic animals. There are, to begin with, the hippor potami, the crocodiles, amphibia and water tortoises common to the majority of the African lakes and rivers. Next we have to deal with nearly 100 different species of fishes, some 50 species of molluscs, four crustaceans, one gymnolaematus polyzoan, four sponges and coelenterates, and perhaps 20 recognisable protozoan types. Prior to the first Tanganyika expedition only four different species of fish were known, some 15 species of molluscs had been described from their empty shells, one species of crab had been described by Milne Edwards, and there was besides one coelenterate in the shape of the famous Tanganyika jelly-fish. With the exception of the few fishes col- lected by the officers of the Free State, the whole of our knowledge of the fauna of Tanganyika as it exists up to date, has been acquired as a result of the 134 THE TANGANYIKA PROBLEM. two Tanganyika expeditions. It will be unnecessary in the present work to consider the mammalian, reptilian, and amphibian vertebrates, as they are similar to those contained in other African lakes, and are already quite well known ; consequently, we may proceed at once to examine the remaining inhabitants of the lake. In roughly surveying the whole field, perhaps the most striking fact is that nearly half the total number of species representing the population of Tanganyika should be made up of different kinds of Teleostian fishes. A fish-fauna with ioo species or there- abouts, in a fresh-water lake, is extraordinary anywhere ; but it is doubly strange when, as in the case of Tanganyika, we find that 76 out of the 87 species which have actually been discovered and described in the lake, are endemic forms ; that is to say, they occur, so far as is at present known, nowhere else in the world. The majority of the fishes in Tanganyika are, in fact, endemic, and when we consider the profusion of species which is present, this fact is extremely interesting and suggestive in itself. Consider for a moment in comparison the fish-fauna of Lake Mwero : up to the present, only 14 species have been found in it ; so also in Nyassa we find that there have been recorded 41 species of fishes, in Rudolf only sixteen. These facts point in themselves to what is from other considerations unquestionable, i.e., that Tanganyika has been practically isolated and undisturbed for, at any rate, a considerable time, and the extraordinarily large number of endemic forms present, can only be viewed either as the result of the formation and multiplication of species through natural selection and other similar causes, in the lake itself, or as a survival in this lake of some old fauna which was rich in such types of fishes. The 87 species of Tanganyika fishes are divided among the following nine families : — o o THE TANGANYIKA PROBLEM. ‘35 Polypteridse, Lepidosirenidse, Characinidse, Cyprinidse, Siluridae, Cyprinodontidae, Serranidae, Cichlidae and Masta- cembelidae. Of these, the species occurring in Tanganyika which represent the Polypteridae, Lepidosirenidae, Characi- nidae and the Siluridae, are widely spread throughout the fresh-waters of the African continent. But all the seven species of the Cyprinidae, and the six species of Masca- cembelidae, which occur in Tanganyika, are endemic forms. The single species representing the Cyprinodontidae and the Serranidae are the same, while among the 2 1 genera and the 58 species of the Cichlidae, only one species, Tilapia Burtoni, is found outside the confines of Tangan- yika, this single species having made its way up the Russisi river into Lake Kivu. It is, however, only among the Cichilidae that endemic genera occur. The Cichlidae are found in the fresh-waters of the Old and the New World, especially in the rivers of South America, and we are further confronted by the extraordinary fact that nearly half of the Old W orld species are confined to the limits of Tanganyika itself, a circumstance which, on the face of it, would almost at once suggest that in Tanganyika we have the remains of the ancient point of origin from the sea of this particular group of fishes. Such a conception of the mode of origin of the Tanganyika fish-fauna would fit in admirably with a number of other facts relating to it. Thus, it was shown by Dr. Gunther that the fish of the African fresh waters are related to each other in such a manner as to suggest that they have spread out from a centre to which Tanganyika would correspond ; while more recently Mr. Boulenger, as the result of an examination of the fishes which I obtained during the Tanganyika expeditions, has confirmed this view, and shown that the morphological attributes of the Tanganyika 136 THE TANGANYIKA PROBLEM. Cichlids are on the whole primitive as far as that group is concerned. If, moreover, the above view of the origin of the Tanganyika fish-fauna be correct, the fact that the genera of Characinkke and Siluridaj living in Tanganyika are not endemic would be perfectly intelligible. The first of these groups is a very old one, and might easily have become a wandering fresh-water stock before Tanganyika became a land-locked depression ; while the second contains forms which are known to be capable of migration in a high degree. The same kind of reasoning thus obviously also applies to the non-endemic species of the latter group. Some such view of the case would at once explain other matters. If Tanganyika became stocked from an ancient western sea, this would account for the presence of Cichlidse in the Congo, and the presence of forms like bass only in the Congo, the Nile, and Tanganyika, as well as the existence of several species of the ancient ganoid Potypterus in these different waters. For it is a geographical fact, which should be clearly understood, that at certain times and at certain seasons it is possible to push one’s way in a boat from the upper tributaries of the Congo into those of the Nile, the watersheds in this region being extremely ill defined. Up to the present time there are known in Tanganyika the following series of forms : — FISHES. POL.YPTERID.iE. 1. Polypterus congicus Blgr. 1898. Lepidosirenid.e. 2. Protopterus sethiopicus Heck. 1851. Characinid^;. 3. Hydrocyon lineatus Blkr. 1863. 4. Alestes macrophthalmus Gthr. 1S67. 5. ,, macrolepidotus C. & V. 1849. 6. Citharinus gibbosus Blgr. 1899. Cyprinid.e. 7. Capoeta tanganiccc Blgr. 1900. 8. Barbus platyrhi>ius Blgr. 1900. 9. ,, altianalis Blgr. 1900. 10. ,, serriRr Blgr. 1900. 11. ,, tropidolepis Blgr. 1900. 12. Barilius moorii Blgr. 1900. 13. ,, tanganica Blgr. 1900. THE TANGANYIKA PROBLEM. i37 F ishes —continued. SlLURID/E. 14. Clarias robecchii Vincig. 1893. 15. ,, liocephalus Blgr. 1898. 16. Chrysichthys cranchii Leach, 1S10. 17. ,, myriodon Blgr. 1900. 18. ,, brachynema Blgr. 1900. 19. Auchenoglanisbiscutatus Geoffr. 1829. 20. Anoplopterus platychir Gthr. 1864. 21. Synodontis granulosus Blgr. 1900. 22. ,, multipunctatus Blgr. 1898. 23. Malopterurus electricus Lacep. 1801. Cyprinodontid^e. 24. Haplochilus tanganicanus Blgr. 1898. Serraniimc. 25. Lates microlepis Blgr. 1898. Cichlidje. 26. Lamprologus tetracanthus Blgr. 1899. 27. „ elongatus Blgr. 1898. 28. 3 9 tretocephahis Blgr. 1899. 29. ,, modestus Blgr. 1898. 30- „ lemairii Blgr. 1899. 3i- 99 hecqui Blgr. 1899. 32- 99 moorii Blgr. 1898. 33- 39 * brevis Blgr. 1899. 34- ,, compressiceps Blgr. 1898. 35- ,, fasciatus Blgr. 1898. 36- L fitrcifer Blgr. 1898. 37- Julidochromis ornatus Blgr. 1898. 38. Paratilapia vittata Blgr. 1901. 39- „ aurita Blgr. 1901. 40. 5 5 bloyeti Sauv. 1883. 41. 55 pfefferi Blgr. 1898. 42. 55 calliura Blgr. 1901. 43- ,, ?tiacrops Blgr. 1898. 44. ventralis Blgr. 1898. 45- ,, dewindti Blgr. 1899. 46. 5 5 fitrcifer Blgr. 1898. 47- 55 steuosoma Blgr. 1901. 48. 55 leptosoma Blgr. 1898. 49- 5 5 nigripinnis Blgr. 1901. 50. Bathybates ferox Blgr. 1898. 51. ,, fasciatus Blgr. 1901. 52. Pelmatochromis polylepis Blgr. 1901. 53. Ectodus descampsii Blgr. 1898. 54- ,, melanogenys Blgr. 1898. 55- >> longianalis Blgr. 1899. 56. Xenotilapia sima Blgr. 1899. 57- , , ornatipinnis Blgr. 1901. 58. Grammatotria lemairii Blgr. 1899. 59. Trematocara marginatum Blgr. 1899. 60. ,, unimaculatum Blgr. 190*1. 61. Telmatochromis vittatus Blgr. 1898. 62. ,, temporalis Blgr. 1898. 63. Gephyrochromis moorii Blgr. 1901. 64. ' Tropheus moorii Blgr. 1898. 65. ,, anneciens Blgr. 1900. 66. Simochromis diagramma Gthr. 1893. 67. Tilapia nilotica L. 1757. 68. ,, burtoni Gthr. 1893. 69. ,, horii Gthr. 1893. 70. ,, rubropunctata Blgr. 1899. 71. ,, dardennii Blgr. 1899. 72. ,, labiata Blgr. 1898. 73. ,, pleurotcenia Blgr. 1901. 74. ,, trematocephala Blgr. 1901. 75. ,, boops Blgr. 1901. 76. ,, grandoculis Blgr. 1899. 77. ,, microlepis Blgr. 1899. 78. Petrochromis polyodon Blgr. 1898. 79. ,, tanganiae Gthr. 1893. 80. Asprotilapia leptura Blgr. 1901. 81. Eretmodus cyanostictus Blgr. 1898. 82. Spathodus erythrodon Blgr. 1900. S3. Perissodus microlepis Blgr. 1898. 84. Xenockromis hecqui Blgr. 1899. 85. Plecodus paradoxus Blgr. 1898. Mastacembelid^e, 86. Mastacembelus frenatus Blgr. 1901. 87. ,, moorii Blgr. 1898. 88. ,, ellipsifer Blgr. 1899. 89. ,, tanganicce Gthr. 1893. 90. ,, tceniatus Blgr. 1901. 91. ,, ophid.um Gthr. 1893. The names of fishes in this list printed in italics are those which are endemic in the lake. It should also be noted that the five species of fishes obtained from Kivu are in- cluded in this list. THE TANGANYIKA PROBLEM. i >S CRUSTACEA. Brachyura. 1. Limnothelphusa maculata (Cunnington) 2. Platythelphusa armata (Milne Edwards) Atyida;. 3. Limnocaridina Tanganyika; (Caiman) Pal.'F.monid.-e. 4. I’alremon moorei (Caiman) MOLLUSC A. 1. Limnrea debaizei Bgt. 2. ,, jouberti Bgt. 3. ,, laurenti Bgt. 4. ,, natalensis Krauss. 5. Isidora conlboisi Bgt. 6. ,, randabeli Bgt. 7. Physopsis tanganyicre Marts. 8. Planorbis sudanicus Marts. 9. ,, tanganicanus Bgt. 10. ,, tanganicensis E. Sm. 11. Ampullaria bridouxi Bgt. 12. ,, ovata Ol. 13. Lanistes jouberti Bgt. 14. Vivipara sp. ? 15. Neothauma E. Sm. 16. Cleopatra guillemei Bgt. 17. Melania tuberculata Mull. 18. ,, tanganyicensis E. Sm. 19. ,, horei E. Sm. 20. Typhobia horei E. Sm. 21. Bathanalia Howesei Moore. 22. Bythocerus iridescens Moore. 23. ,, mimor Moore. 24. Chytra kirkii Moore. 25. Limnotrochus Thomsoni E. Sm. 26. Paramelania Damoni E. Sm. 27. ,, cranigranulata E. Sm. 28. Tanganyicia rufofilosa Crosse. 29. Speika zonata Woodw. 30. Nassopsis nassa E. Sm. 31. Syrnolopsis lacustris E. Sm. 32. Stanleya neritinoides Bgt. 33. Rumella callifera Bgt. 34. Unio bohrni Marts. 35- >> gerrardi Marts. 36. ,, calathus Bgt. 37. „ horei E. Sm. 38. ,, burtoni Woodw. 39. ,, rostralis Marts. 40. ,, tanganyicensis E. Sm. 41. ,, thomsoni E. Sm. 42. Spatha tanganyicensis E. Sm. 43. Mutela soleniformis Bgt. 44. Iridina spekei Woodw. 45. Burtonia tanganyicensis E. Sm. 46. Corbicula radiata Phil. 1. Arachnoidia lankesteri (Moore) 1. Lemnochnidia. 1. Spongilla tanganyikae Evans. 2. ,, moorei Evans. POLYZOA. HYDROZOA. SPONGES. 3. Potamolepis weltneri Moore. i Associated with the above there are in Tanganyika some conspicuous Protozoa, chief among which is a curious peri- dinia-like form possessing cilia. This organism forms the yellow scum on the lake which was noticed by Livingstone in his last journey. Associated with it there is a large THE TANGANYIKA PROBLEM. T39 condylostema, and besides these, representatives of the universally distributed fresh-water protozoa. This enumeration brings us to an end of our survey of the African lakes which have up to the present time been explored. I have no information about Chad, and there is at present little or nothing to be ascertained about the numerous minor lakes which are associated with the tributaries of the Congo, but it will be admitted that as a result of the investigations carried out during the Tanganyika expeditions, and through the scattered explorations of a number of other observers, we now know definitely what are the faunistic characters of the principal African fresh waters, and consequently we are for the first time in a position to deal in a general way with the meaning of the facts of distribution and the character of the African fresh-water faunas, and with the conspicuous anomalies which the fauna of Lake Tanganyika presents. But before proceeding to this inquiry it may be pointed out that, besides the actual lakes of Central Africa, there are the great rivers and the backwaters of these rivers, which, as Mr. Boulenger has shown, contain an extraordinary fish- fauna, if nothing else. From our own experiences on the Zambesi, Shiri, Rusisi, Ruchuru, and Semliki, it would appear that in the majority of these African rivers the fauna consists of little else but fish. In backwaters and the like there are often encountered the typical fresh-water insects, mollusca, and minute Crustacea which occur generally over the African interior, but in the actual course of the rivers little if anything but fish. This is, I am convinced, what will be found to be the case in the majority of African streams, both great and small, but it is perhaps to be anticipated that the Congo, its backwaters and its tributaries, will eventually prove more or less of an exception to this rule. The Congo 1 4° THE TANGANYIKA PROBLEM. is fed from Tanganyika, and numbers of the animals which, as we have seen, are peculiar to Tanganyika must be annually carried into its upper reaches, and may possibly continue to subsist in the backwaters of the river. Moreover, the Congo lies in what was probably at one time a former extension of the sea, and, as I have said, its fish-fauna is unique ; it is therefore quite probable that the backwaters of this great river will be found to be zoologically fruitful. There have, indeed, been unconfirmed statements of the occurrence of jelly-fishes, whether similar to the Tanganyika form is not known, high up in the course of the river. If we consider the constituents of the faunae ol the African lakes other than Tanganyika, described in the preceding pages, it becomes clear that the animals encountered are, at any rate in a generic sense, the same. One or other or more genera maybe omitted in the case of any particular lake, but with the solitary exception of Tanganyika there is no lake which contains numerous genera peculiar to itself, and certainly no lake which contains not only genera which are peculiar to itself, but which are not found else- where in the fresh waters of the globe. The genera which in Lake Tanganyika possess this remarkable characteristic are constituted, as we have seen, by a whole series of inver- tebrates and a number of cichlid fishes. These forms how- ever, do not replace the ordinary fresh-water fauna in the lake : they simply coexist along with it and consequently we have in Tanganyika a series of animals which are super- added to the ordinary fresh-water fauna of the continent. With the exception of Tanganyika, the ordinary fresh- water fauna of Africa has nothing novel or striking about it. In all the lakes which we have examined there is found in each case an ordinary fresh-water stock, which may or may not have become individualised by the pro- THE TANGANYIKA PROBLEM. 141 duction of specific varieties, and this rather monotonous sameness in the nature and composition of the fresh-water faunas of the great lakes of Central Africa has now been definitely shown to hold good as a rule, characteristic of them all with one solitary and conspicuous exception. The fauna of Lake Tanganyika alone contains forms, and many of them, which have not the characteristics of any usual fresh- water types, but, although this fauna thus differs from that of all the other great African lakes, it is after all only a partial exception to the rule of uniformity in type which characterises the fauna of the great African lakes in general. It was seen that Tanganyika contains all the genera of fish and invertebrates which are found in Lake Nyassa, and the character of the normal fresh- water con- stituents of the fauna of Lake Tanganyika differs no more from the fresh- water types contained in the Victoria Nyanza or Nyassa than the constituents of the faunas of these two latter lakes differ from one another. Tanganyika is ren- dered peculiar, not by the general characters of its fresh- water fauna, but simply by the additional possession of a number of forms which are peculiar to that lake. The animals forming the invertebrate section of this peculiar group have an obviously marine aspect, and on that account I have spoken of them elsewhere* as forming a halolimnic series in Lake Tanganyika — that is to say, they form a group of animals which, although living in a fresh-water lake, have at the same time the characters of animals that are typical of the sea. The Tanganyika animals which possess par excellence these characteristics, are the endemic gastropods, the gymnolaematus polyzoa, and the jelly-fishes. But besides these forms there are other invertebrates which, although not so markedly marine in character, * Proceedings of the Royal Society (he. cit.). 142 THE TANGANYIKA PROBLEM. are nevertheless highly remarkable from the fact that they are only found in Lake Tanganyika among the fresh waters of the African interior. To this category of animals belong the prawns, some crabs, the sponges, some protozoa, and a host of Cichlid fishes. Fishes of course are highly migratory forms, and, once established in fresh water, are sure to spread to a large extent through the rivers and lakes of the continental land-mass in which they happen to be ; still there is this peculiarity about the fishes which are at the present time en- countered in Tanganyika ; some of the cichlids, as Mr. Boulenger has shown, are primitive representatives of the group ; while, in Tanganyika, but not in Nyassa and the lakes to the south, there exists the ancient ganoid Polypterus, and several characinid fishes which although not restricted to Tanganyika, do nevertheless share the somewhat primitive characters which the other halolimnic animals undoubtedly possess. On account of this, I am inclined for the present to regard the African Polypteroids and at any rate the genera of Cichlidse peculiar to Tanganyika, as a piscine portion of the halolimnic group. The reasons for the incorporation of these fishes which are not wholly restricted to Tanganyika is again discussed further on, and, as thus defined, the halolimnic fauna of Tanganyika contains the following forms : — among the Teleostei fourteen genera of Cichlkke, Polypterus among the ganoids, seventeen genera of gastropods, possibly two Lainellibranchs, two crabs, two prawns, one gymnolaematus polyzoan, one medusa and some endemic protozoa. All these forms, with the exception of Polypterus, which is ob- viously an animal capable of migration, are the peculiar and characteristic feature of Lake Tanganyika, and it will be admitted that if the invertebrates of this halolimnic fauna DWIGHT W. i - THE TANGANYIKA PROBLEM. 143 have had a different origin from the general African fresh- water constituents, it is highly probable that some of the fishes which we now, from their dispersal, regard as belong- ing to the normal African fresh-water series, are in reality scattered members of this same halolimnic group. At all events it is quite clear that it is not correct to suppose that the fish-fauna of Tanganyika is at all similar to that of the other African lakes, and this strikingly distinctive character of the Tanganyika fishes constitutes the primd facie evidence for supposing that at least some of the fishes which characterise the fauna of Lake Tanganyika are of the same stock as the halolimnic invertebrates themselves. The halolimnic fauna consists then of a group of animals, the invertebrate section of which is rigidly confined to the lake, and which have no obvious relation to the normal African fresh-water invertebrates, but it is also indicated that some of the African fishes which we have hitherto regarded as normal fresh-water fishes of the African continent may belong in reality to the now scattered vertebrate section of the halolimnic group. It remains for us now then to examine in much greater detail the characters of the individual components of the halolimnic group, and thereby to attempt to acquire the information necessary in order to ascertain the actual affinities of these forms ; while finally we shall have to consider the question of the nature and the origin of the whole of the halolimnic group in Tanganyika, and this question itself forms the Tanganyika problem as it now exists. But before proceeding to discuss in detail the structure of the halolimnic forms, it is desirable to con- sider certain widespread phenomena relating to the normal fresh-water fauna of the African lakes in general. O 144 CHAPTER VIII. ON SOME CURIOUS FEATURES OF THE DISTRIBUTION OF SPECIES. In the majority of the lakes in Equatorial Africa, in all of them in fact, with the solitary exception of Lake Tangan- yika, there is a fauna which is in no way peculiar, and although the vertebrate section of this fauna is rendered distinctive of Africa by the profusion of ganoids of chari- cinidsc and cichlidae, its invertebrate sections are dis- tinctly poor. Thus it appears to be a fact from what we have seen that in most of the greater lakes of Central Africa the fresh-water fauna consists of fishes and of molluscs and practically of nothing else. The majority of the great African lakes do not at all support the generally pre- valent idea, that the fresh waters of the tropics are in posses- sion of a profuse fauna. In Mwero, in Beringo and in the Albert Edward Nyanza, or even in the Albert and the Victoria Nyanzas for that matter, there is hardly so much variety of life as there is in an ordinary American or European puddle. There is often an extraordinary profusion of some particular type as in the case of the Nyassa Viviparas, or of Melania tuberculata in the Albert Edward Nyanza, but there is no diversity or even modification among the constituents of the primary fresh-water series, which THE TANGANYIKA PROBLEM. 145 is everywhere found to people the greater and the lesser Central African lakes. In the preceding examination of the components of the in- dividual faunas of the different African lakes, it will have been observed that if we exclude the obviously migrating vertebrates, the fishes, the amphibia, and the like, the remain- ing invertebrate constituents of these faunas are almost always in a specific sense, different in each of the lakes. There are, in fact, only one or two specific forms, such as Melania tuberculata , which occur generally in all the lakes — that is to say, we have in each depression a fauna which, in the species composing it, differs from the fauna in any of the other depressions. It is obvious from this fact that even when lakes are within twenty miles of each other, as in the case of Tanganyika and Kela, the invertebrate faunas of two such depressions do not readily, at any rate, communi- cate with one another. There is, in fact, very little indica- tion, if any, that the invertebrate faunas of the lakes intercommunicate or inter-colonise at all. This is parti- cularly well seen in Lake Kivu. Kivu is a great lake and must have been in much the same condition that it is now, at any rate, for centuries, yet there is not a Vivipara to be found in it, although they swarm in the lakes a hundred miles to the north and to the south. Kivu is, in fact, in direct water connection with Tanganyika, but so far as is at present known, only one small fish, Tilapia burtoni , has ever migrated from the one lake to the other. From these facts and similar ones which we encountered again and again, and which have been further verified by the observa- tions of numbers of other travellers who have been among the 'Central African lakes, it would seem that the inverte- brate components of the fauna of the different African fresh waters do not tend to migrate at all, while the fishes of 10 146 THE TANGANYIKA PROBLEM. these same fresh-water areas do not migrate with anything like the vigour that one might have been naturally led to suppose. There is, however, another characteristic of the faunas of the different great and small African lakes which is certainly interesting, and may very possibly be of wide theoretical importance. If we examine the lists of species consti- tuting the faunas of these lakes, we find that for some reason or another the number of different species present in any particular lake is directly proportional to its size. The fauna of Nyassa consists of about 75 species. The ordinary fresh-water fauna of Tanganyika of 80 species. The fauna of the Victoria Nyanza, although not completely known, contains about 60 species. The fauna of the Albert Nyanza, which is much less than either of the preceding lakes, contains only 20 forms. The Albert Edward, about the same number ; Kivu only 10 species; while the really small lakes, Kela, Eliinantita, Beringo, Nivasha, Nakaroo, etc., contain only about half-a-dozen species each. It is thus obvious that from some cause or other the number of specific forms in an African lake is roughly proportional to the size of the lake itself. The physiographical features of the African lakes are unique ; nowhere else in the world have we numerous sheets of fresh-water of all sizes scattered over an area bigger than Australia, and yet all of which are under practically the same climatic conditions. In this way the faunistic pheno- mena presented by the African lakes are distinctly simpli- fied ; we have not to reckon with the wide climatic differ- ences that are forced upon the faunas of the lakes which occur in the latitudes of North America, Europe and Asia, and consequently we cannot legitimately invoke difference of conditions to account for the specific richness of Nyassa as THE TANGANYIKA PROBLEM. 147 compared with the poorness of the fauna of the Albert Nyanza. The water of the Albert Nyanza is quite as full, if not fuller, of animals than the water of Nyassa ; but the number of species which the lakes contain is, as we have seen, widely different. In this way it would appear to be a fact that the number of animals living in any particular lake may be, and probably is, related to the food supply which the lake presents, but the number of species is for some reason a function of the area of the lake and not of the food which it contains. This is a very remarkable circumstance, and has, in all probability, a wide significance ; the only analogous phenomena with which I am acquainted being the contrast which has been found to subsist between conti- nental and island florae. It has been shown by botanical enquirers that the number of species of plants which flourish upon an island is less than the number of species which flourish upon a similar continental land area, and that, roughly speaking, the number of species which constitutes an island flora is proportional to the size of the island ; the less the island, the fewer the species of plants which it sup- ports, although the island may be just as thickly covered with vegetation as a similar continental area. From these considerations it would appear probable that in the case of the different sized lakes of Africa and in the case of island floras we are dealing with analogous phenomena. For in the one case we are dealing with sheets of water of different sizes, which are, so to speak, detached from the sea, and in the other, with different sized pieces of land occurring in the ocean which are detached from the surface of the continents. In the case of island floras it has been found, however, that there is a distinction to be drawn between what have been called oceanic islands and continental islands, or in 10* 148 THE TANGANYIKA PROBLEM. other words islands which have long been remotely isolated in the midst of the ocean, and islands which are more or less narrowly-detached annexes of a continental land-mass ; the floras being different in character on these two types of islands. On comparing the floras of continental with oceanic islands, it is found that on continental islands there are more genera and fewer species as compared with the fewer genera and more species found on oceanic islands. The continental islands are dominated by migrations of certain genera from the continent ; whereas the floras of oceanic islands are less exposed to the effect of the incursion of vigorous continental forms ; and the floras of such islands are able to gradually adapt themselves to the varying conditions which the island presents in various parts, the flora of an oceanic island be- coming affected by the natural restrictions which deliminate different portions of the island from one another ; and in different areas individual varieties gradually attain specific rank. It is apparently the oceanic island which is most nearly comparable to the remotely inland lake, for it is ob- vious that when the depressions holding lakes upon a con- tinental area have become established, the migrations of new forms into these depressions from the ocean will have been rendered difficult or impossible, as the case may be. From what we saw in Chapter VII. it would appear that to a large extent the fresh-water fauna of a continent is something which in its origin is bound up with the origin of the con- tinent itself, and consequently the production of species in the fresh waters of a continent is, in a sense the measure of the age of the fresh waters to which these species may belong. From general considerations it would appear that the specific varieties of the genus Melania characteristic of the fauna of Nyassa cannot have migrated into it, for they do not exist outside the lake — they are, in fact, THE TANGANYIKA PROBLEM. 149 characteristic of the fauna of Nyassa, and consequently they would appear to have originated in the lake itself. In a great lake like Nyassa the origin of new forms peculiar to the region appears to be related to the existence of conditions there which have already been observed to be conducive to the separation of definite variations among animal forms. As Darwin pointed out, it is definitely known that the fencing- off of cattle in a park tends towards the formation of stock which differs on different sides of the fences, and in a great lake like Nyassa we find that the fauna is often actually very definitely fenced off into isolated areas by the natural condi- tions which exist in such a lake. Thus the molluscs in Nyassa live chiefly in the sheltered bays and creeks which communicate with the open body of the lake, each bay forming in itself a colony wherein a number of molluscs live practically isolated from all the other colonies which are dis- tributed round the shore. Storms and floods effect the dis- persion of members of these colonies at different times, and in this way every colonisable area of the lake becomes filled with molluscs, but in general the individual population of each colonisable area remains cut off and isolated from all the rest. The actual illustrations of these phenomena are often very interesting and instructive. Thus the Vivipara unicolor , which is generally found in Nyassa, and frequents more diversely conditioned parts of the lake than any other form, is replaced in Monkey Bay by a type nearly three times as large, and, so far as I know, this variety occurs in no other part of the lake. So again in Tanganyika, there are at least three very well marked varieties of the ally of the genus Vivipara Neothauma. One of these, with the type of shell repre- sented on page 261 ( a ), occurs exclusively at the south end of the lake, swarming in the broad and more or less sheltered *5° THE TANGANYIKA T ROB LEM. reaches into which the southern end of Tanganyika expands. In the narrow, surf-swept, and turbulent portion of the lake, which stretches between the north of Cameron Bay and Tembwi, Neothauma is only found in the little bays and sheltered places occurring along both shores, and here the character of the form changes, the double-keeled shell of the former variety being replaced by the elongated type shown on page 261 ( b\ Northward the lake terminates again in more or less sheltered expanses, like the Gulf of Ubuari, the deep bays near Ujiji, and the extreme northern extremity of Tanganyika. In these the form of the genus again changes, the two more southern varieties being generally replaced by the curious rounded form represented on page 261 ( c ). The same phenomenon is again remarked in the distribution of the beautiful Typhobia shells of Tangan- yika. In the north we have a type which differs from that in the south. So again with regard to the distribution of the Tanganyika genus Paramelania and its allies ; the shell of Paramelania damoni is represented on page 242, and occurs more or less abundantly throughout the lake ; but in certain areas it is to a greater or less degree replaced by several definite varieties ; one of these, represented on page 238, is so peculiar and well marked as to constitute a new generic type. This form, Bythoceras , occurs only in the southern half of the lake, and is practically exclu- sively related to Cameron Gulf and the bays which lie opposite the gulf on the east coast. Another variety which is much more closely related to Paramelania damoni is the form Paramelania crassigranulata , and this variety is also restricted to certain areas on Tanganyika, being encountered about Karema on the east coast of the lake. From these observations it would appear to be indicated that the cause of the production of a greater number THE TANGANYIKA PROBLEM. 151 of species in a great lake as compared to a small one is really due to the existence of a greater number of more or less definitely isolated aquatic areas in the greater lake as com- pared to the less. The greater lake corresponds to a park in which the cattle have been fenced off into a greater number of pens, and thereby a greater number of varieties have been produced, while the smaller lake corresponds to one of these areas, and contains only a single variety. The phenomena presented by lacustrine faunas would thus seem to be very similar to the phenomena presented by island floras, for, as we have seen, the number of species en- countered in the different African lakes vary directly with the size of the lake, but it should at the same time be clearly apprehended that the specific varieties characteristic of individual great African lakes are usually only so many structural changes which have been rung by time and cir- cumstance upon the types of fresh-water organisms which are universally distributed throughout the earth. These varieties which appear in the greater African lakes have nothing to do with the halolimnic forms peculiar to Tangan- yika, although these forms may and have been, as we have just seen, affected in a similar way. It is, moreover, not always the same forms which tend to vary in the different lakes. In Nyassa there are numerous endemic varieties of Melania , Ampul-aria , and Vivipara ; in the Victoria Nyanza numerous varieties of the Melania , to which there are however added noticeable modifications of normal fresh- water Lamellibranchs, as in the case of AEtheria. *5* CHAPTER IX. THE FISHES OF LAKE TANGANYIKA. As will have been seen in the preceding chapters, the fishes that have been recently discovered in Lake Tan- ganyika consist of 87 different species, of which 74 are new to science. Of these latter, four were originally collected by Captain Hore, 21 by the Lemaire Expedition and the officials of the Congo Free State, the remaining 49 being collected by myself and my colleagues during the first and second Tanganyika Expeditions. For the elaborate and careful working out of this somewhat unique series we have been entirely indebted to Mr. Boulenger, of the British Museum, who published two fully illustrated reports on them in the transactions of the Zoological Society.* In the present work I have consequently merely incorporated Mr. Boulenger’s original descriptions accompanied by reproductions of Mr. Green’s careful and accurate drawings of the most conspicuous genera and species. In what follows the fishes have been arranged in their natural order, and only the descriptions of the few forms found in Tanganyika, which like Polypterus were already well known, being omitted. * For references to these and other papers containing descriptions of the fishes collected during the Tanganyika expeditions by Mr. Boulenger, see bibliography of the expeditions at the end of the volume. Polypterus congicus. See p. 154. 154 THE TANGANYIKA PROBLEM. POLYPTERID/E. 1. P. Congicus. — Blgr. 1898. (Fig., p. 153, lower). LEPIDOSIRENIELE. 2. Protopterus /ETHiOPicus. — Heck 1851. (Fig., p. 153, upper). CHAR AC I N IILE. 3. Hydrocyon lineatus. — Blkr. 1863. 4. Alestes macrophthalmus. — Gthr. 1867. When this species was first recorded from Lake Tanganyika, it was only known from the Gaboon. It has since been found in Lake Mweru and in the Congo. 5. Alestes macrolepidotus. — C. and V. 1849. Kalambo. 6. ClTHARINUS GIBBOSUS. — Blgr. 1 899. (Fig., p. 155.) Kalambo. The largest specimen measures 500 millim. CYPRINIDzE. 7. Capoeta tanganic/E. — Blgr. 1900. (Fig., p. 163.) Depth of body 3I to 4 times in total length, length of head 5 times. Snout broad and rounded, as long as or slightly longer than the eye, the diameter of which is 3.} times in the length of the head and nearly twice in the interocular width ; the width of the mouth equals § that of the head ; a minute barbel, hidden under the lip at the angle of the mouth. Dorsal III. 9; third ray very strong, ossified, smooth ; the fin, which is equally distant from the eye and from the caudal, has the free edge notched, and its greatest depth equals the length of the head. Anal III. 5 ; the longest ray measures \ the length of the head. Pectoral acutely pointed, as long as the head, not reaching the ventral, which is inserted under the first rays of the dorsal. Caudal forked. Caudal peduncle twice as long as deep. Scales 68-70 — 14, 9 or 10 between the lateral line and the root of the ventral. 14—15 Olive above, each scale darker at the base, silvery white beneath ; fins greyish. Total length, 320 millim. Described from three specimens from the north end of Lake Tanganyika. The discovery of a species of this genus in Lake Tanganyika is particularly interesting from the fact that only one was known from Africa, viz., the Abysin- nian C. dilloni , C. and V. ; this is distinguished by the absence of barbels and the greater size of the scales (30 to 32 in the lateral line). In the presence of a pair of barbels and the small size of the scales, C. langanica belongs to the typical section of the genus, inhabiting south-western Asia ; but it has the enlarged dorsal ray neither feeble, as in C. fundulus , Pall, and allied species, nor serrated, as in C. truita, Heck. Cithasinus gibbosus. Seep. 154. J56 THE TANGANYIKA PROBLEM. 8. Barbus platyrhinus. — Blgr. 1900. (Fig., p. 157.) Depth of body 3J times in total length, length of head 4 times. Snout broad and rounded, twice as long as the diameter of the eye, which is contained 5.I times in the length of the head and 2 \ times in the interocular width ; mouth small, with two pairs of subequal barbels, the length of which equals the diameter of the eye. Dorsal III. 8 ; third ray not enlarged, not serrated ; the fin, which is equally distant from the eye and from the caudal, has the free edge convex. Anal III. 5 ; the longest ray not quite § the length of the head. Pectoral a little shorter than the head, not reaching the ventral, which is inserted below the middle of the dorsal. Caudal forked. Caudal peduncle 1$ as long as deep. Scales 40 — — — - 3^ between 5i the lateral line and the root of the ventral. Olive-brown above the lateral line, golden yellow beneath. Total length, 390 millim. Described from a single specimen from south of Usambura. This species appears to be more nearly related to B. capensis , Smith, from which it differs in the much shorter and broader snout. 9. Barbus altianalis. — Blgr. 1900. (Fig., p. 159.) Depth of body equal to or slightly greater than the length of the head, which is contained 4 to 4$ times in the total length. Snout moderately broad, rounded, scarcely projecting beyond the lower jaw, ii to l§ times as long as the diameter of the eye, which is contained 5 to 5.4 times in the length of the head and twice to twice and one-fourth in the interocular width ; mouth small, with two pairs of subequal barbels, the length of which equals or exceeds a little the diameter of the eye. Dorsal III. -IV. 9; third or fourth ray very strong, ossified, not serrated; the fin, which is equally distant from the occiput and from the caudal, has the free edge notched, and its greatest depth is but slightly less than the length of the head. Anal III. 5; the longest ray measures about J the length of the head; the fin, when folded, reaches nearly the root of the caudal. Pectoral a little shorter than the head, not reaching the ventral, the first ray of which corresponds to the origin of the dorsal. Caudal forked. Caudal peduncle nearly twice as long as deep. Scales 34 35 — 3 between the lateral line and the root of the ventral. Olive Si- brown, very dark above. Total length, 450 millim. Described from two specimens from Lake Kivu and one from the source of the Rusisi River. B. altianalis is extremely closely related to B. mariquensis. Smith. It differs only in the somewhat broader snout, the stronger third dorsal ray, and the somewhat longer caudal peduncle. c n s-< c3 Ph m d PQ 58 THE TANGANYIKA PROBLEM. io. Barbus serrifer. — Blgr. 1900. (Fig., p. 207, lower.) Depth of body 3 to 3^ times in total length, length of head 4 to 4' times. Snout rounded, not projecting beyond the lower jaw, as long as or a little longer than the diameter of the eye, which is contained 4 to 4', times in the length of the head and ij to ii times in the interocular width; mouth small, with two pairs of barbels, the posterior of which are the longer, and measure twice the diameter of the eye. Dorsal III. 7 ; third ray very strong, ossified, serrated behind ; the fin, which is equally distant from the eye and from the caudal, is not notched, and the largest ray is a little shorter than the head. Anal III. 5 ; the longest ray jj the length of the head. Pectoral £ to J the length of the head, reaching, or nearly reaching, the base of the ventral, the last ray of which falls under the first of the dorsal. Caudal forked. Caudal peduncle ii to 1 4| as long as deep. Scales 28-30 4- ■ -L1 3 between the lateral line and the root of the ventral. Olive-brown 5* ’ above, silvery white beneath ; a greyish stripe along each side of the body above the lateral line ; a small blackish spot at the base of the caudal. Total length, 120 millim. Described from three specimens from the north end of Lake Tanganyika. Allied to B. kessleri, Stdr. Distinguished by the smaller eye, the longer barbels, the more numerous scales, and the presence of only 7 branched dorsal rays. ix. Barbus tropidolepis. — Blgr. 1900. (Fig., p. 161.) Depth of body 3 times in total length, length of head 4 to 4^ times. Snout broad and rounded, strongly projecting beyond the mouth, to twice as long as the diameter of the eye, which is contained 4.J to 5.4 times in the length of the head and 2 to 2b times in the interocular width ; mouth small, inferior ; a minute barbel almost entirely concealed in the angle of the lips. Dorsal III. 9; third ray very strong, ossified, not serrated, its length at least § that of the head ; the fin, which is equally distant from the occiput and the root of the caudal, is notched. Anal II. 5 ; the longest ray about £ the length of the head. Pectoral about f the length of the head, not reaching the base of the ventral, the first ray of which falls under the origin of the dorsal. Caudal forked. Caudal peduncle iA to i£ as long gi as deep. Scales 44-46 -rp 5 between the lateral line and the root of the ventral ; o.j in breeding specimens, the scales, those at least which are above the lateral line on the caudal portion of the body, bear a median swelling or obtuse keel, these keels forming together very regular longitudinal lines. Olive above, silvery white beneath. 12. Barilius moorii. — Blgr. 1900. (Figs., p. 208, 209, lower.) Depth of body equal to length of head, 4 times in total length. Head a little over twice as long as broad, with slightly curved upper profile ; snout pointed, not extending beyond the lower jaw, as long as or a little longer than the diameter of vd LO d. C/3 PQ i6o THE TANGANYIKA PROBLEM. the eye, which is contained 4 to 4^ times in the length of the head ; interorbital width a little greater than the diameter of the eye ; mouth extending to below the anterior third or the centre of the eye ; no barbels ; the naked space between the pneopercle and the suborbitals less than half the width of the latter. Gill-rakers very short, almost rudimentary, Son lower part of anterior arch. Dorsal III. 9, originating at equal distance from the anterior border of the eye and the root of the caudal ; its border is not notched, and its depth equals about 3 the length of the head. Anal III. 13-14, originating under the middle of the dorsal ; its anterior rays a little longer than those of the dorsal and much longer than the posterior rays, forming a rounded lobe. Pectoral pointed, shorter than the head, not reaching the ventral, which extends to the origin of the anal. Caudal forked. Caudal peduncle twice as long as deep. Scales 56-60 , 3 between the lateral line and the root of the ventral. Silver)', brownish on the back ; more or less distinct dark vertical bars on the side of the body, about 10 in number ; dorsal blackish at the end. Total length, 115 millim. Several specimens from the north end of Lake Tanganyika. 13. Barilius tanganica:. — Blgr. 1900. (Fig., p. 165.) Depth of body equal to length of head, 4j times in total length. Head a little over twice as long as broad, with straight declivous upper profile ; snout very pointed, not extending below the lower jaw, once and a half the diameter of the eye, which is contained 53 times in the length of the head ; interorbital width once and a half the diameter of the eye ; mouth extending to below the posterior border of the eye ; no barbels ; the naked space between the praeopercle and the subor- bitals about 3 the width of the latter. Gill-rakers short, 10 on lower part of anterior arch. Dorsal III. 10, originating at equal distance from the occiput and the root of the caudal, the posterior third of its base above the anal ; its anterior rays are longest, measuring a little more than half the length of the head. Anal III. 17, strongly notched, with rounded anterior lobe, the longest rays of which measure J the length of head, whilst the posterior rays measure barely \ . Pectoral pointed, | length of head, not reaching the ventral, which extends to the vent. Caudal forked. Caudal peduncle a little over twice as long as deep. Scales 82 — , 4 between the lateral line and the root of the ventral. Silvery, olive on back; 16 or 17 blackish vertical bars on each side of the body, equally distant from the median dorsal line and from the lateral line. Total length, 260 millim. Described from a single specimen from north end of Lake Tanganyika. SILURIFLE. 14. Clarias robecchii. — Vincig. 1893. 162 THE TANGANYIKA PROBLEM. 15. Clarias liocephalus. — Blgr. Vomerine teeth in a narrow band, without posterior process. Depth of body, 5.$ times in total length ; length of head, 5} times. Head smooth, covered with soft skin, slightly longer than broad ; occipital process very short, angular ; diameter of eye, 3 times in length of snout, 6 times in interorbital width ; maxillary barbel as long as the head, nasal barbel a little shorter ; inner mandibular barbel 2 length of head. Dorsal 70. Anal 50. Caudal free. Pectoral £ length of head, not extending to the vertical of origin of dorsal fin. Uniform blackish brown. Total length, 80 millim. Described from a single specimen from Kinyamkolo. 16. Chrysichthys cranchi 1. — Leach, 1810. 17. Chrysichthys myriodon. — Blgr. 1900. Depth of body, 4.^ times in total length ; length of head, 3J to 3.$ times. Head broad and much depressed, J longer than broad, rough on the vertex and occiput ; snout broadly rounded, scarcely projecting beyond the lower jaw, J length of head and twice the diameter of the eye, which is contained 6 times in the length of head and 2 to 2h times in the interocular width ; the occipital process, which is rough like the occiput, in contact with the interspinous shield ; nasal barbel, } or 1 the diameter of the eye ; maxillary barbel a little more than half the length of the head; inner mandibular barbel, ^ length of the head, outer a little less than half. Vomero-pterygoid teeth very fine and closely set, as are also the premaxillary and mandibular teeth, forming a broad, horseshoe-shaped, uninterrupted band ; its width in the middle a little less than that of the premaxillary band, but increasing at the sides, where it much exceeds the latter. Dorsal I. 6 ; spine, rugose, not serrated, nearly half the length of the head and ^ to J the length of the longest soft rays. Adipose dorsal, a little longer than deep, its base § that of the rayed dorsal and i the distance which separates it from the latter. Anal IV. 8-9. Pectoral spine a little longer and stronger than the dorsal, feebly striated, and bearing on its inner edge about twenty strong retrorse serrae. Ventral not reaching the anal. Caudal deeply notched, with obtusely-pointed lobes, the longest rays measuring double the length of the median. Caudal peduncle ijj to twice as long as deep. Olive-brown above, white beneath. Total length, 470 millim. This description is taken from three large specimens, one received from Albertville by the Congo Museum, through Captain Hecq, one from Tembwi, and one from Kinyamkolo, obtained during the second Tanganyika expedition. Compared with specimens of C. cranchii of similar size, C. myriodon differs by its smaller and more numerous teeth, the greater posterior width of the vomero-pterygoid band, the larger eye and the more strongly serrated pectoral spine. It has also a higher numl»er of vertebrae (20-27). Z 2 Some young specimens, measuring up to 130 millim., collected at Kibwesi, in VO