XXXV.] ENCEOACHMENT OF THE LAKE. 461 



The other strata showing close to the coal, which was lying 

 on granite, were limestone and red sandstone, marble and slaty 

 rocks, some patches of soft-looking gray chalk, and a reddish 

 soil like that of the Wealden area, with lumps of stone looking 

 like Kentish rag. 



All the faces of the cliffs were so much torn and seamed by 

 torrents and rains, that it is almost imjDossible, from merely a 

 passing glance, to give a reliable description. 



Jnst beyond Ras Makanyazi a sharp line seemed to divide 

 the granite overlaid with sandstone from limestone cliffs, and 

 shortly afterward the cliffs came to an end, and the mountains 

 trended back a long way from the coast, the intervening coun- 

 try being formed of low and rounded hills and level plains. 



The lake here is rapidly encroaching on the shores, and the 

 contour is constantly changing. IS^ear the mouth of the Mu- 

 samwira — the drain of the Likwa Lagoon — where large vil- 

 lages stood a year or two ago, sand-banks only are now seen, 

 and these are hourly decreasing in size. 



After passing the Musamwira, the hills again approached the 

 lake; but I observed a few inlets which might be utilized as 

 boat harbors. At Ras Kamatete the hills run back again some- 

 thing in the same manner as near the island of Kabogo, form- 

 ing a deep bay with low-lying level land around it. The south- 

 ern horn of this bay is Ras Mpimbwe, a promontory consisting 

 of enormous blocks of granite piled on each other in the wild- 

 est confusion. 



The land is composed of a hght-red sandstone, though it is 

 hardly stone at all, with large masses of granite and harder 

 sandstone imbedded: .the water washes away the soft sand- 

 stone, and leaves the harder rocks either in piles or half -sunken 

 reefs. 



I believe that exactly the same process is going on here 

 which in earlier ages formed the hills and mountains we came 

 across between Liowa's and Ugogo, the rocky hills of TJn- 

 yanyembe, and deposited the rocks in Ugogo about Usekhe 

 and elsewhere. The whole country was apparently at one time 

 an enormous lake, with a soft sandstone bottom overlying gran- 

 ite ; and as it contracted, either through a general elevation of 

 the bottom or from some other cause, the surf on the shores 



