FORMER LAND CONNECTIONS. 129 



Joeen found back uppermost with all the bones in position, as 

 though the animals had died where they were found and were 

 .covered up rapidly either by mud or dust. 



Large quadrupeds first make their appearance in the south- 

 western corner of the Karroo, and shortly afterwards, during Lower 

 Beaufort times we find many strange forms such as Pareiasauras, 

 To pinocephalus, Titanosuchus, and Dicynodon over a wider area 

 in the Cape, but not in the north or north-east. Furthermore this 

 •early fauna possesses well-marked affinities with the corresponding 

 Permian Keptilia and Amphibia of North America, a resemblance 

 that vanishes later. Broom believes that, while the two assem- 

 blages sprang from a common stock, the American offshoot remained 

 nearer the centre of their evolution. 



Migration of the Vertebrates. 



The route taken by this early Karroo fauna was probably via 

 the Northern Argentine and Southern Brazil, and, although these 

 forms have not yet been discovered in those regions, this can be 

 explained by a known stratigraphical break in the succession there, 

 during which no sediments were laid down, and consequently no 

 fossil remains preserved. 



Suggestively, too, these precursors of the Middle Karroo 

 fauna have left no representatives elsewhere in Gondwanaland, but 

 a little later, near the close of the Middle Permian, we find their 

 descendants not only in the Orange Free State and Natal, but even 

 in Madagascar and Tndia, and, though failing to reach Australia, 

 they succeeded in penetrating into Northern Russia, where on the 

 banks of the Dwina their remains were discovered by Amalitzky 

 associated with Glossopteris. Isolated on what seems to have 

 formed a peninsula in the Permian seas, the fauna became some- 

 what specialised. Since the vertebrates of this age included herbi- 

 vorous as well as carnivorous forms, for example the Pareiasauria, 

 one may be permitted to speculate that with the spreading of the 

 Glossopteris Flora — a more or less xerophytic vegetation, be it 

 noted — into this northern area, the animals followed in search of 

 pasture. 



During the Permian and Triassic, a period certainly covering 

 some millions of years, vertebrate evolution was particularly active 

 in South Africa, and among the Reptilia the advances were gener- 

 ally, but not always, towards the acquirement of mammalian 

 characters. The Middle Triassic was marked by the rise of the 

 highest forms of the Theriodontia, of which one South African 

 genus has been discovered in Brazil and an allied one in India. 



During the Upper Triassic, however, Gondwanaland, as 

 already stated, certainly became linked to Europe, and, just as this 

 union is reflected in the composition of the flora, so here again we 

 find an influx of northern types of Reptilia, Amphibia, and fresh- 

 water fishes, chief among which are the Dinosauria, those strange, 

 long-necked quadrupeds that took to sitting or even walking in a 

 semi-erect position with the long tail as a support. It is curious 

 to discover that the South African genera and species — found not 



