78 MR. D. M. S. WATSON ON THE 



There is no step in the upper jaw, the curved tooth-bearing 

 edge of the premaxilla passing smoothly into that of the maxilla. 

 In 8. tigriceps the squamosals are not nearly so much spread, the 

 skull being much deeper in proportion to its width. If we may 

 trust the existing description of a not very satisfactory prepara- 

 tion, the palate of 8. tigriceps differs from that of 8. ivhaitsi in 

 the loss of the anteuior projection of the pterygoid and the great 

 extension of the palatines. On the other hand, judging from 

 the description by Haughton of its endocranial surface, the 

 brain-case of 8. tigriceps may have greatly resembled that of 

 8. whaitsi. 



Thus the C'istecephalus-zone animals referred to 8cymno- 

 gnathus do not belong to that genus, but differ from it by a 

 series of advances which will be seen to be all in the direction 

 leading to Diademodon . Certain of these species appear to 

 resemble Lycosaurus pardalis considerably, agreeing with that 

 animal in dentition, the short high snout with a rounded dorsal 

 surface, the large nostril, the vertical internarial bar, the small 

 exposure of the septomaxilla, the absence of a, step in the upper 

 jaw, the short and deep maxilla, and the small prefrontal. All 

 of them, however, seem to retain a large postfrontal bone. 



Two other remarkable forms, apparently from the lower part 

 of the Cistecephalus-zone, Scylacops capensis and Gorgognathus 

 longifrons, are of interest because they strongly recall Eudothiodon- 

 zone forms. 



Gorgognathus with its immensely long low snout somewhat 

 resembles Scymnognathus ivhaitsi, and its very broad inteiorbital 

 and intertemporal surfaces agree with Gorgonops. It is, however, 

 advanced in the following characters: — The loss of the step in 

 the jaw, the rounded snout, and especially the vertical occiput. 

 Haughton has pointed out another advanced feature in the 

 structure of the basicranial region. Scylacops is a small unusual 

 form with a low broad snout : it is advanced in the exclusion of 

 the frontal from the orbital margin, in the rather vertical occi- 

 put, in the lo.ss of the step in the upper jaw, and especially in the 

 loss of the anterior ramus of the pterygoid : it appears to retain 

 a, rather primitive Gorgonops-tike nose and has only small 

 temporal fossa;. Broom's figure of the occiput suggests that it 

 is advanced in the shallowness of the paroccipital processes. [It 

 is probable that the fragment of a Goi-gonopsid skull which I 

 described (Ann. & Mag. Nat. Hist. 1913, vol. xi. p. 65, figs. 1-4) 

 belongs, if not to Scylacops, at any rate to a closely allied form.] 



Arctognathus curvimola is a far more advanced form than any 

 so far discussed in this paper ; it presumabl}* comes from a higher 

 horizon in the Cistecephalus-zowe than Gorgognathus, etc. It 

 show T s advances in the following features : — The snout is short, 

 narrower than the orbital region, rounded over the mid-line. 

 The nostrils are very large and the septomaxillary foramen 

 small. The nasals do not overhang in front. The inteiorbital and 

 intertemporal surfaces are narrow, the orbits facing outwards, 



