Page:The Periplus of the Erythræan Sea.djvu/108

 This condition is corroborated by the known Arab infusion in the negro peoples on the whole coast, which is of far earlier origin than the Mohammedan colonization.

Who were the natives and what was their language, as mentioned in the Periplus? Rev. J. Torrend, S.J., in a paper read before the Rhodesia Scientific Association, included in its Proceedings (V, 2, Buluwayo, 1905), analyzes the languages of the coast and finds a striking similarity between the speech of the Tana River, which empties below the island of Lamu about 2° 40′S., and that of the lower Zambesi (18°–19°S.). He gives a long comparative list of words in these so-called Pokomo and Cizimba tongues, evidentally identical. He quotes Dr. Krapf and other German philologists as saying that the Pokomo is the aboriginal language of the coast, and that the modern Swahili is derived from it; and he himself believes that the Cizimba is even more primitive, and that it gives the key to most of the modern dialects of the southern coast. Father Torrend, full of the Sofala–Ophir theory, argues that the language was brought from the Tana River to the Zambesi, not by land because the modern tribes are of peaceful disposition, but rather by sea, and particularly by sea-traders, assuming such to have come from Arabia. The assumption is certainly far-fetched, as it is hardly likely that any traffic, however busy, would have brought this negro language and transplanted it 1500 miles down the coast to a different tribe. The suggestion is rather that this branch of the Bantu race migrated southward within historical times, through the African rift-valley, and that the modern tribes of the lower Zambesi, said to be speaking to-day the most primitive language, are their descendants, while those who remained on the Tana have had their speech modified more notably by later contact with the outside world.

The name Cizimba, borne by the modern dialect, suggests the Agisymba of the Roman geographers; which was known to them through the report of an adventurous youth, Julius Maternus, who marched for four months southward from the Garamantes (Fezzan), and brought back word of a region abounding in rhinoceros, inhabited by negroes and bearing that name (Ptolemy, I, 8, 5). It seems not an unreasonable assumption that he did rich the head-waters of the Nile and found somewhere in that great rift-vally the ancestors of this Bantu tribe which later migrated southward and formed, among other confederations, the so-called Monomotapa of the mediaeval geographers.

This rift-valley of East Africa is a striking feature of its topography, and must have had a great bearing on its early trade. A good