Origin of tJie Angiosioermous Flora of South Africa. 353 



for the Karroo Eegion (see Ficoidacece-SesuvidcE and Phytolaccacece- 

 Bivinece and -Stegnospermece). I might also have mentioned the 

 only 2 endemic South African genera of Portulaccace^ : Aiia- 

 cavipseros (placed by Pax between 2 American genera) and Portu- 

 lacaria (placed by Pax between a genus from Chile and another 

 from Madagascar). I might further have mentioned Pelargonium, 

 the nearest living ally of which is TropcBohcm (distributed in South 

 and Central America and Mexico). Nov7, while some of these rela- 

 tionships may perhaps be explained by a former Southern connection, 

 others only find their explanation if we assume a former land-con- 

 nection between Tropical West Africa and Tropical South America, 

 which must have lasted into Tertiary times. Such a connection, 

 besides clearing up some of the relations of the Flora of Tropical 

 Africa and Tropical America, would also explain many relations 

 between the Fauna and Flora of Madagascar and South America. 

 It would especially explain the relations of the Flora of Tropical 

 West Africa to the Flora of Tropical South America which, with all 

 diffidence, I venture to think are even more intimate than represented 

 by Engler in the passages quoted above. To mention only one of 

 many indications for the correctness of this view I would point to the 

 distribution of the Cyperace^. In dealing with the Gyperacece of 

 the West Indies, Clarke (1906) comes to the following conclusion : 

 " The most remarkable fact that has come out in this enumeration 

 is the large number of species common to West Tropical Africa and 

 the West Indies. This feature is also common to the mainland of 

 America and Africa ; a prominent point in this connection is the 

 almost equal distribution of the species of the distinct genus Mapania, 

 sect. Eumapania in Guiana and Guinea. 



As far as South Africa is concerned, it is noteworthy that the con- 

 nection with Tropical America is, amongst others, shown very plainly 

 in numerous Gamopetalce, and this may perhaps also be an indica- 

 tion that it has existed into Tertiary times. Obvious relations can be 

 traced between South African and Tropical American Hydrophyllacece, 

 LentihidariacecE, GesneriacecB, BignoniacecB, SolanacecB, &c. Amongst 

 ScROPHULARiACE^ we find 1 species of the genus Alonsoa, E. et P. 

 near Uitenhage, 7 or 8 in Tropical and sub-Tropical America ; in the 

 tribe ChelonecB we find 6 genera endemic in South Africa (mostly 

 Eastern), 1 Halleria, L., in South Africa, Abyssinia, and Madagascar, 

 and of the remaining 20 genera mentioned by Bentham and Hooker 

 15 are restricted to America ; the tribe Gerardiece is especially inter- 

 esting, as it contains so many parasites and semi-parasites. Its dis- 

 tribution seems clearly to demand a former land-connection of West 

 Africa and South America. Its members are grouped round Tropical 



