88 



ON THE CLASSIFICATION OF THE THERIODONTIA. 



natural one, the conservative main stem which leads on to the 

 Cynognathids being i'e[)re.sented by a series of Gorgonopsids of 

 which Gorgoiwps itself is one of the most primitive members. 

 From this stem side-bi'anches arise, which retain the broad 

 parietal region and other primitive features, but present either 

 an acceleiated development of certain regions or are indi- 

 vidually specialised. From still earlier members of the main 

 stem arose the groups of animals, resembling one another in 

 the precocious convei'sion of the broad intertemporal region into 

 a sagittal crest and in the acquirement of suborbital vacuities, 

 which are usually included in the Therocephalia and belong to 

 many independent stirps, each in all px-obabiJity being dependent 

 on the mam Gorgonopsid stock. It is shown that there is a. 

 series of evolutionary trends which persist throughout the whole 

 gx-oup of Anomodonts fx-om Varanosaurus to Diademodon, and 

 that the special I'apid advances which sepai-ate the Therocephalia 

 fx^om the Gox^gonopsids, iix the main, merely follow out the pi-e- 

 determined evokxtionaxy txack proper to the group. 



Thus any classification of the Thexiodontia is necessarily 

 complicated, as involved and difficult of constx^uctioxx as that of 

 the Theria themselves. Existixxg material is so incomplete that 

 any attenxpt at detailed classification, even if only into families, 

 is dangerous, in that it will load the litei-atuxe with xxxxdefined 

 gx'oups, whose chax^acteristic forxxxs may only be knowix fx-oxxx the 

 fx'oxxt exxd of the skxxll or the dentitioxx. 



The detailed descriptions of skull-structixres in this paper show 

 how unx'eliable, eveix for genexic distinction, are the characters 

 presented by the teeth of Thei-iodonts. 



Thixs, for the present, I aixx ixxcliixed to retain my former 

 division of Theriodontia into Goi'gonopsidas, Thexocephalidse, 

 Oynognathidfe, axxd Baux-omoi-pha,, fully recognising that these 

 groups — or, at any x^ate, the fix-st two — cover a xnultitude of fox'ms 

 not directly of common origin and only held together by two 

 or thx'ee striking characters. 



It remains to discuss the connectioxxs of the Theriodontia 

 with the other gx-oups of South Afx-ican Aixoxxxodonts — the 

 Deiixocephalia, Dromosauria, axxd Dicyixodoixts. 



In the copper-beax^ing sandstones and associated limestones of 

 the Ux-al Mountains, which imxnediately sxxcceed the Artinsk 

 beds and are shown by a coxxiparison of x^eptile and amphibian 

 faunas to be slightly older thaxx the Tajiinoce'phalus-T.oxi^^ ax^e 

 found thx^ee types of Anomodonts, each x^epiesented by skulls or 

 jaws: of these Detiterosaurus is cleai-ly a Deinocephaliaix of the 

 Tapiixocephaloid gi'oup recalling man}^ Soxxth Africaxi fox'ms. 



Deinosaurus [=Oliorhizodoii) is represented by jaws, ixx oxie 

 case associated with a palate whose dox'sal sixxface is well exposed. 



RhojMlodon is a naxxxe coverixxg not only several jaw-fragments 

 but also a complete skxxll, which was descxibed hj Prof. Seeley. 



Of this skull, x'emarkably beautiful lithographic drawings of the 



