348 Transactions of the Soutli African Philosophical Society. 



remainder of the South African Flora, which shows pretty close 

 connections with the Tropical African Flora. The question of the 

 origin of these floral elements resolves itself, therefore, into the 

 question of the origin of the Flora of Tropical Africa which 

 cannot be dealt with here very thoroughly, although it may be 

 advisable to refer to a few points connected with it. 



Thus Stapf (1904, p. 409) says with reference to GraminecB : " The 

 connection of the tropical and sub-tropical elements amongst the 

 grasses of South Africa with the palseotropic Flora of Tropical 

 Africa is so clear that no further comment is necessary. The whole 

 sub-tropical grass region of South Africa is nothing but a continuation 

 of the whole floral region in the north. Eelations to other tropical 

 countries exist almost exclusively by way of Tropical Africa. 

 There are only three exceptions which are too isolated to be used as 

 the basis of an argument for the former direct relations to extra 

 African tropical countries." 



Similar relations can be traced in very numerous other orders. 

 Engler (1906, p. 34) has summarised the relations of the Flora of 

 the Northern Transvaal, Natal and Ehodesia to the Flora of Tropical 

 Africa as follows: "If we now ask to which of the known floral 

 regions of Africa these parts of the Transvaal or Ehodesia can be 

 associated with, the answer is not very difficult. The investigation 

 of British and German East Africa, as well as Northern Ehodesia, 

 has more and more shown that these regions belong together, as 

 also Angola, Benguela, and a large part of German South-East 

 Africa, just as it was clear long ago that the Flora of Natal is 

 related to the Flora of the Mozambique and Zanzibar coast regions. 

 There are small differences between the various subdivisions, if we 

 work carefully through the genera which are distributed through the 

 whole of the East African and South African "steppe" provinces. 

 We then find that in several contiguous sub-provinces certain groups 

 of species predominate ; on the other hand, some species (especially 

 of Cojiihretum) are distributed through several sub-provinces. From 

 the enumeration of the plants observed by me, it resulted that not a 

 few are identical with plants collected by Welwitsch in Benguela. 

 Eelations to Natal are smaller, which may be explained through 

 the fact that here we have to do with an inland Flora in which 

 the Cape and Madagascar floral elements still more pass into the 

 background than in Natal." 



One naturally looks to the Flora of the higher mountains of 

 Tropical Africa for an extension of the South African types and for 

 an indication of the channels in which those types have moved. 

 This interesting question has been referred to in various publications 



