36 MR. B. M. S. WATSON Ox\ THE 



London, Broom in South Africa was describing two very com- 

 plete Gorgonopsid skulls, one associated with the anterior part 

 of a skeleton. Of these skulls Broom gave a good description, 

 bringing out the whole structure of the face and parietal region, 

 but not giving us so satisfactory an account of the palate and 

 occiput. Broom, sometimes in conjunction with Haughton, sub- 

 sequently added many new and often strange forms to the Gor- 

 gonopsida? — on the whole, emphasizing rather their resemblance 

 to the Deinocephalia and Dicynodontia and even Pelycosauria 

 than those which they show to tlie " Cynodontia." In 1914 I 

 was able to show that known Gorgonopsids could be arranged 

 as a morphological series giving a gi-adual passage in the 

 structures of the occiput, and of the basicranial and otic regions 

 between Dimetrodon a Pelycosaur and Diademodon a "Cynodont."' 

 In the same paper I described the palate of the " Cynodont '' 

 Bauria, showing that it differed much from the Cynognathids 

 and resembled the non-Gorgonopsid Theriodonts with a primi- 

 tive palate more than the Gorgonopsids. In consequence, purely 

 as a temporary measure, I revived the order Thei'iodontia and 

 divided it into four sub-orders — the Tlierocephalia, the Gor- 

 gonopsia, the Bauridfe, and the Cynodontia. Since that paper 

 was written, Haughton has published descriptions of certain new 

 forms and made important new additions to our knowledge of 

 the brain-case of the earlier Theriodonts. In his most I'ecent 

 paper he uses provisionally my 1914 classification, emphasizing 

 its insufficiency. 



In revising a paper on the relative ages of the Palaeozoic and 

 Triassic reptile- bearing I'ocks, which has occupied me at intervals 

 for some years, I was forced to deal with the joroblems piesented 

 by the fauna of the coppei-- bearing Permian sandstones of the 

 Orenburg district of the Urals. One of the most noteworthy 

 forms from this 'locality is lihopcdodon, an animal whose skull, 

 which alone is certainly known, presents many resemblances to 

 the Gorgonopsids. The necessity of discussing the systematic 

 position of this form led me to an examination of a.l] tlie 

 Theriodont material available, with the results which are set out 

 below. 



It is convenient to begin with a description of the material at 

 my disposal, then to discuss the morphological results which 

 arise from it, and, finally, consider the evolution of the group 

 and the relationship of Deuterosaurus to it. 



Arctops wiLLiSTONi Watson, Proc. Zool. Soc. 1914, p. 1026. 



Type and only known material : a skull lacking the anterior 

 part of the snout, the quadrates and quadrate rami of the 

 pterygoids, otherwise complete and practically undistorted. 

 From Howse Post, near Fort Beaufort, S. Africa, not improbably 

 Endothiodon zone. 



I described and figured the occiput and basicranial region in 

 the original description. 



