THE DESICCATION OE AFRICA 
149 
Tanganyika had one outlet on the western side to the 
Congo, the Lukuja river. I have it from Mr. Irwin (now in 
Mombasa) who was on the lake in 1890, that the old Arabs 
regularly took large dhows down the Lukuja and into the 
Congo. In my time the mouth of this river was nearly closed 
by vegetable growth, and no outflow was perceptible. 
The Chambezi, Lake Bangueulu, Lua'pula River , and Lake 
Mweru . — These can be taken in the above sequence, and 
comprise the most westerly and perhaps one of the largest 
headwaters of the Congo. The Chambezi river is of fairly 
rapid, though fluctuating, volume. Most of its upper feeders, 
such as the Mansia, Luatikila, and Lukulu are permanent 
streams, which originate on the Chambezi- Tanganyika cum 
Nyasa watershed. On the flats towards Bangueulu the stream 
is considerable and well defined. It was once supposed to 
flow into Lake Bangueulu, but is now known to flow into the 
vast swamps at the south end of the lake. From these swamps 
another river arises known as the Luapula, which evidently 
takes the Chambezi water in addition to drainage from the 
lake and swamps. The Luapula then turns sharp to the 
north, passing Bangueulu to the westwards, and flows into 
Lake Mweru. Thus a continuous flow of permanent water 
can be traced and which continues in the Luvna, which flows 
from the north end of Lake Mweru into the Congo. But a 
different story is unfolded on examination of the vast flats 
on the east, north, and west confines of Lake Bangueulu. It 
is then that one can realise the former extent of the lake. 
Although Bangueulu is marked as well defined on most maps, 
it is in reality most difficult to define on its south end, because 
immense sheets of water occupy the old permanent lake bed 
during the rains. This water is held on the surface by the 
vast and continuous deposit of kaolin clay, bog-iron, and 
limonite, which underlies the whole country between the 
Chambezi on the east and the Luapula on the west. The 
lake has undoubtedly shrunk, but without these vast clay 
flats with a gentle fall lakewards, it is conceivable that the 
lake would by now have been yet smaller in area. Of course 
much evaporation must take place on the confines of this basin, 
where the water advances and recedes every year, but the 
