242 SOUTH AND EAST AFRICA. of volcanic origin. Like all the other watercourses of this region, the Chobe has excavated its deep channel in the layer of soft calcareous tuffa formerly depo- sited on the bod of the great lacustrine basin. Daring the floods, which last from December or January to March, all inequalities of the ground disappear beneath the vast and always limpid sheet of water formed by the junction of the two streams. The annual difference between the high and low water levels varies from 20 to 24 feet. The Liba or Upper Zambese. The little river Liba, which has its source not from that of the great Lu-Lua tributary of the Kassai, is usually regarded as the true upper course of the Zambese, although both the Ku-Bango and the Chobe take their rise at a far greater distance from the Indian Ocean. One of the affluents of the Upper Liba is the Lo- Tembwa, a stream flowing from Lake Dilolo, which was discovered by Livingstone, and which presents the rare phenomenon of communicating with two distinct fluvial systems, those of the Zambese and the Congo. A great number of other " children," as the natives call the tributaries of the Liba, send their contributions to the *' mother," which soon becomes the Liambai or Zambese, that is, the " river " in a superlative sense. But the greater part of the rainfall, being precipi- tated on a too uniformly level surface, is unable to reach the main stream. It lodges in stagnant pools scattered over the reed- grown plains, which from a dis- tance resemble a boundless prairie with here and there a few wooded islets rising above the tall, waving grasses. Amongst the pci-ennial watercourses of this region, all infested throughout the year by numerous hippopotami, the most important is the Lua-Ena, whose basin stretches fur to the west. Some sixty miles from the point where the Zam- ^bese begins to become navigable, the Lua-Ena mingles its blackish waters with the yellowish current of an affluent which Livingstone regarded as the true main stream, but which is inferior to it bpth in the length of its course and in volume. This is the Kabombo tributary, first explored by Capello and Ivens. Below the confluence of the two rivers the mainstream is swollen by the waters of the Lua-Ngo Nbungo, which rises not far from the sources of the Kwa-Ndo, and traverses the extensive Lobule plains — grassy fens or waterless steppes accord- ing to the season. Beyond this junction the aspect of the land still remains unchanged, the united stream flowing directly southwards over a plain standing at a dead level, where the flood waters expand in vast shallow lagoons during the rainy season. On the surface are borne along great masses of tangled vegetation swept down by the current. With the return of dry weather the waters subside, and this apparently boundless sea assumes the aspect of a regular channel winding between steep banks of alluvial sc il intermingled with sands and many-coloured clays, where the wasp-eater and kingfisher have their nests. The river thus flows rapidly but at a uniform speed for a distance of over 180 miles, after which, beyond some wooded islands, it changes the direction of its course, trending round to the south-east. Here the stream winds between rocky
