156 THE BOKKEVELD FAUNA 



several possible explanations. The faunas compared 

 were of littoral habit and may have been effectually se- 

 parated by a deep ocean, and there may also have been 

 climatic reasons for the diversity. Other evidence for 

 the existence of an Atlantic continent is seen in the dis- 

 tribution of late Mesozoic and early Cainozoic fossils, 

 terrestrial as well as marine ; and also from supposed 

 remnants of continental rocks thrown from the volcanoes 

 of the Atlantic Islands, and from the serpentines of St. 

 Paul's Eocks, but it has not been proved that the land 

 was continuous at any one time, nor is the evidence 

 from the rocks independent of theoretical considerations 

 as to the nature of the sub-oceanic crust. 



The recent discovery in the Sahara of Devonian fos- 

 sils said to be identical with, or closely related to, Cape 

 and American forms l (Homalonotus herscheli, Leptocoelia 

 ftabellites, Conularia africana, Actinopteria aff. boydi) is not 

 in agreement with the hypothesis of a continent which 

 prevented communication between the southern and 

 northern seas. At present, therefore, the question must 

 be regarded as an open one. 



The country occupied by the Bokkeveld beds north of 

 the Langebergen and in the Worcester and Kobertson 

 Divisions south of that range is characterised by strongly 

 marked escarpments and valleys, so that from the top 

 of a prominent hill in a suitable position the lie of the 

 rocks can be made out over a very wide area. The most 

 accessible of such hills are the Brand Vlei Mountain 

 near Worcester, Gydo Berg north of Ceres, the high hill 



1 Haug, E., Comptes Rendns, cxli., 1905, p. 970. 



