FAU-NA off' TIIK AKIMCAN LAKES. 515 



wliich bears on the matter, hut the conclusions arrived at may 

 be .summarised as follows : — Since the area drained is not very 

 great, and the rainfall is relatively low, the amount of wa,ter 

 which finds its way into the lake is nearly Ijalanced by the evapor- 

 ation {cf. 190, v(,)l. ii. p. 68 et seq.). It follows that under 

 average conditions there will be little change in the surface- 

 level and no occasion for a,n outlet for the surplus water. There 

 are reasons for believing that at some former period the waters 

 of Tanganyika did not flow out at all, and were consequently 

 more saline than at jiresent. As a result of the formation of 

 great volcanic cones north of Lake Kivu, that portion of the 

 Nile drainage system which once draiiicd north into Edward 

 Nyanza became cut off, aiid its waters finally divei'ted south- 

 wards into Tanganyika (cf. 137, }). 89 et seq.). 



iSlowly raised in level by this additional suppl}', the water 

 eventually found an outlet to the west, along the course of the 

 Lukuga., a former affluent, and through a sandstone I'idge 

 denuded in part by the action of a tributary of the Congo. The 

 lake having risen abnormally before the establishment of an 

 outlet, sank at first rapidly, i-emaining at a lowej- level, subject 

 to fluctuations Avith the time of j'ear and the amount of rainfalL 

 A few consecutive seasons of unusual drought would suffice to 

 reduce the level of the lake to a point at which all outfloAv woidd 

 cease. It seems likely that under such circumstances a. barrier 

 might be formed in the bed of the effluent, due to the comluned 

 elFects of wind-blown sand and ti'opical sv^amp vegetation, and 

 vaiying in nature with the length of time taken to prodtice it. 

 If a tolerably firm dam was formed, a considerable rise in the 

 level of the lake would be necessary in order to develop enough 

 power to sweej) away the obstruction. There is reason to believe 

 that the high water level Avhich Cameron and Stanley obsei'ved, 

 and the subsequent re-establishment of an outflow, were phases in 

 a cycle of this nature. If a periodicity of this kind really exists, 

 it must of necessity be extremely variable in nature, and due to 

 irregularities in the rainfall. 



The important bearing of these inferences on the problem of 

 the fauna of Tanganyika it is hardlj^ necessaiy to emphasise. 

 If the Nile has lost a not inconsideiable drainage area, which 

 has been added to that of Tanganyika, it seems very reasonable 

 to suppose that this accessioii of water produced, for the first 

 time, the overflow of the lake. If this supposition be correct, 

 Tanganyika, must have been completely isolated and without an 

 outlet, until the formation of the great volcanic dam in geolo- 

 gically recent times. 



The lists of animals from the various lakes giAen in the body 

 of this work are, as regards extent, partly dependent on the 

 amount of biological exploration which each lake has received. 

 That is to say, a comparatively small faiuia may be merely due 

 to less complete examination. Whether the lakes considei/ed 

 here can be regarded as having been at all equally investigated, 



