It follows that the leavening element, by which the southern Negro populations have been diversely modified throughout the Bantu lands, could have been drawn only from the Hamitic and Semitic peoples of the north-east. But in this connection the Semites themselves must be considered as almost _une quantite negligeable_, partly because of their relatively later arrival from Asia, and partly because, as they arrived, they became largely assimilated to the indigenous Hamitic inhabitants of Egypt, Abyssinia, and Somaliland. Belief in the presence of a Semitic people in the interior of S.E. Africa in early historic times was supported by the groups of ruins (especially those of Zimbabwe), found mainly in Southern Rhodesia, described in J. T. Bent's _Ruined Cities of Mashonaland_. Exploration in 1905 dispelled the romance hitherto connected with the "temples" and produced evidence to show that they were not earlier in date than the fourteenth or fifteenth centuries and were of native construction[228]. They probably served as distributing centres for the gold traffic carried on with the Semitic traders of the coast. For certainly in Muhammadan times Semites from Arabia formed permanent settlements along the eastern seaboard as far south as Sofala, and these intermingled more freely with the converted coast peoples (_Wa-Swahili_, from _sahel_ = "coast"), but not with the _Kafirs_, or "Unbelievers," farther south and in the interior. In our own days these Swahili half-breeds, with a limited number of full-blood Arabs[229], have penetrated beyond the Great Lakes to the Upper and Middle Congo basin, but rather as slave-hunters and destroyers than as peaceful settlers, and contracting few alliances, except perhaps amongst the Wa-Yao and Ma-Gwangara tribes of Mozambique, and the cannibal Ma-Nyemas farther inland. To this extent Semitism may be recognised as a factor in the constituent elements of the Bantu populations. Malays have also been mentioned, and some ethnologists have even brought the Fulahs of Western Sudan all the way from Malaysia. Certainly if they reached and formed settlements in Madagascar, there is no intrinsic reason why they should not have done the same on the mainland. But I have failed to find any evidence of the fact, and if they ever at any time established themselves on the east coast they have long disappeared, without leaving any clear trace of their presence either in the physical appearance, speech, usages or industries of the aborigines, such as are everywhere conspicuous in Madagascar. The small canoes with two booms and double outriggers which occur at least from Mombasa to Mozambique are of Indonesian origin, as are the fish traps that occur at Mombasa.