With regard to man of course the barrier has not been so effective. Evidently the Aluyi people of long ago crossed the Upper Zambezi and penetrate to the Lower Kwando or Chobe, and across this stream to the vicinity of Lake Ngami, though not much farther west; and in return, immigrants of Bechuana stock entered Barotseland from Lake Ngami in the last century. But there seems to have been a not easily explained halt in human migrations between Western Zambezia and South-West and South Africa. It would be comparatively easy to account for, if one could presume that the present sterile influence of the Kalahari desert prevailed two thousand or even one thousand years ago, and earlier still. But on the contrary from the many indications we can collect and from the legends of the natives in the regions south of the Zambezi, we are inclined to believe that the present arid nature of the lands on the western side of South Africa is quite a modern feature which has gone on rapidly increasing. So rapidly that even I who made a journey in 1882 to the Kunene river from Mossamedes can remember flowing streams (in the dry season) and forests, where the rivers and rivulets are now dry and the woodland is dead. There seem, in addition, to have taken place in comparatively recent times, changes in the disposition of water supplies, which have diverted much useful moisture from inner South Africa into the Atlantic Ocean. Not such a very long time ago the Kunene river, which brings an abundant supply of fresh water from the great knot of highlands in Southern Angola, bore this water into Lake Etosha in Ovampoland, and beyond that towards the course of the Okavango or the great vanished lake of which Ngami is the surviving vestige: Ngami, which even in Livingstone’s day—the end of the ’forties, some seventy-two years ago—was twice its present size. Into the same basin flowed the mighty Okavango (known much farther north as the Kubango), the Kwando and the upper Zambezi, then the outlet of another vast lake in the Barotse country. Quite possibly these two shallow lakes of huge size sent their overflow of water through the Guay river into the bed of the Zambezi-Kafue, and so out into the Indian Ocean after all; for these waters of Central Africa were completely hemmed in southwards by the plateaux of Damaraland, Bechuanaland and Matebeleland. But in those days the ultimate source of the Zambezi would have lain in the Benguela highlands near the Atlantic Ocean, and not in the Luvale country near to the upper Kasai. * * * * *