PRE-JURASSIC TETRAPODS. 185 



clearly in the loss of tlie screw-shaped articular surface of the 

 head of the humerus, and the corresponding resti-iction of the 

 glenoid cavity to the scapuln and posterior coracoidal element 

 alone. 



It is separated from other Anomodonts with advanced limbs 

 by the retention of the large quadrate. The mode of articulation 

 of the stapes shows that it cannot have been derived from any 

 Pelycosaur more advanced than a Poliosaurid, 



The order Dromosauria is referred to the super-order by its 

 lower jaw and occiput. It is distinguished from all other South- 

 African types by the short face, very large orbits, slender limbs, 

 and long tail. 



Palceohatteria is only placed here provisionally ; it is certainly 

 an Anomodont, and in the structure of its temporal region very 

 strongly recalls Galepus. The absence of the posterior coracoidal 

 element may only mean that, as in Varanoops, it is unossified, 

 although present as a cartilage. 



It will possibly be found, when fuller knowledge of Bolosaurus 

 becomes available, that that type is really allied to the 

 Dromosaurs, and that the two orders should be combined. 



The order Dicynodontia includes the first known and typical 

 Anomodonts. The order is a very compact group, thoroughly 

 distinct from all others and of uncertain derivation, perhaps from 

 the Deinocephalian stock. 



The order Theriodontia includes maiiy very diverse animals, 

 still known almost wholly from skulls. Its members are held 

 together by the presence of a reduced quadrate in all of them and 

 by a well difi'erentiated carnivorous dentition. 



The suborder Gorgonopsida appears to be ancesti'al to the Cyno- 

 dontia, thfe Therocephalia to the Baiiriamoi'pha, and it is probable 

 that other lines will be distinguished. Each suborder certainly 

 contains many families, but in the absence of detailed knowledge 

 of the skull and of all knowledge of the post-cranial skeleton in 

 most forms, it seems at present useless to found families on the 

 dentition, 



I have recently discussed at length the relations of Eiinoto- 

 saurus to the Chelonia ; if rightly referred, it is so much more 

 primitive than any other Chelonian that ordinal separation seems 

 justified. 



The various animals included in the Sauropterygia differ so 

 widely amongst themselves, and the whole group is so distinct 

 from all othei'S, that its ordinal rank is unquestionable. 



Merriam s quite distinct ordei' Thalattosaui'ia is still rather 

 incompletely known, but its members certainly have only a 

 single arch and the temporal region of the skull is reminiscent of 

 the earlj' Pelycosaurs. 



The Ichtln'osaui's form a very compact gi^oup, whose origin is 

 quite imknown. 



The inclusion of the two arched reptiles in one super-order is 

 now coiumonly accepted. I havie excluded the Hhynchocephalia 



Proc. Zool. See— 1917, No. XIIT. 13 



