30 Prof. R. Broom — Permian ami Trias in South Africa. 



these beds there is a vei^ rich fauna, of which the most characteristic 

 forms are the extremel}' mammal-like Cytiodont reptiles, Cynognnthns, 

 Gomj'hognafhis, etc. Other noteworthy forms are the large Phyto- 

 saur, Enjihrosnchus ; the Guathoilont, Ilowesia, a small form allied 

 to 7Ii/perodapedon; a species of Cychtosnurtts, and a species of 

 Ceratodus. There seems little doubt that these beds are of Upper 

 Triassic age. 



Underneath these we come to a distinct but allied fauna charac- 

 terised by the abundant remains of the primitive reptile Procoloplion. 

 Other forms are Palignaiia, the oldest known true lizard, and 

 Proterosuchis, a Khynchocephalian which shows some affinity to 

 the Phytosaurs. These beds are called the Vrocolophon Beds, and 

 may be of Middle Triassic age. 



Below these are extensive beds in which land forms are very rare, 

 but in which are abundant remains of the aquatic Anomodont, Lystro- 

 sanrns, and the fish Atherstonia. These are called the Lystrosanrus 

 Beds, and they are believed to be of Lower Triassic age. 



Underneath these Lystrosaurits V>eds we come to rocks in which 

 laud forms are again numerous. Three fairly well-marked zones 

 are recognised : (1) Kidecephalus Beds, (2) Endoiktodon Beds, 

 (3) Pareiasam-ns Beds. In all of these remains of Dicynodon are 

 met with, but it is in the Endothiodou Beds that Dicynodon and 

 Ondcnodon are met with in greatest abundance. All the carnivorous 

 reptiles of this period are either Therocephalians or Dinocephalians. 

 All three zones are believed to be of Upper Permian age. 



Below the Pareiasaurns Beds a few reptiles have been found in 

 the Ecca Beds, but most of the remains are very imperfect. Meso- 

 saurus is found in the Upper Dwyka, and may be assumed to be of 

 Lower Permian age. 



While it is perhaps unwise to place too much weight on the 

 evidence, the South African faunas seem to afford a little new- 

 light on the age of the Elgin sandstones. The Stagonolepis Beds, 

 which contain Stagonolepis, Oruiihosnchns, Erpciosnchus, StenometopoH, 

 Hyperodapedon, and Telerpeton, seem to correspond to the Cynognathus 

 Beds of South Africa, which contain Erythrosuchns, Ilowesia, and 

 TJielegnatlins, forms all allied to those of Elgin. If this be so, then 

 the Stagonolepis Beds may be regarded as Upper Triassic. 



The Gordonia Beds, containing Gordonia, Geikia, and Elginia, 

 resemble most closely the Pareiasaurns Beds of South Africa with 

 Dicynodon, Oudenodon, and Pareiasaurns. The resemblance, how- 

 ever, is much closer to the Russian forms of the Upper Permian 

 of the Dwina, and it seems probable that the Elgin forms are 

 descendants of the Russian, as the Russian probably are of the 

 African. We may conclude that the Gordonia Beds are either 

 Upper Permian or Lower Triassic, more probably the former. 



(See also Abstract of paper by Prof. Broom on the Classification 

 of the Karroo^Beds of South Africa, p. 36.) 



