CLASSIFICATION OF THE THERIODONTIA. 63 



separated by a naiTow i-idge which runs foi-ward to tlie palate 

 proper. From the vertical sides of the basisphenoid horizontal 

 pi'ocesses arise ; these basipterygoid processes support the ptery- 

 goids. From their articulation with the basisphenoids the 

 pterygoids pass backward towards the quadrates, but do not 

 articulate with those bones, as they appear to terminate in free 

 points before reaching them. The pterygoids pass forward, 

 forming with the median ridge which continues the basisphenoid 

 a bar whose ventral surface is almost cylindrical, bi-okeu only by 

 the median fillet. At the hinder end of the palate the pterygoids 

 suddenly widen, forming very deep powerful flanges. This part of 

 the bone has a transverse suture with the ecto pterygoid. Medially 

 the two pterygoids meet in a visible suture which lies at the bottom 

 of a small depression. This suture soon terminates at the brim 

 of a much deeper and more sharply-marked hollovv^, which, as it 

 passes forward, widens and is converted into a deep open groove 

 forming anteriorly the whole roof of the much vaulted palate. 

 Throughout its extent this groove has well-marked, indeed often 

 vertical, sides. Anteriorly this groove is divided into two by a 

 ridge which rises from its surface. At about the level of the 

 last maxillary tooth this groove is bounded by roughened areas 

 of bone, which appear to have borne teeth. Tliese areas are 

 undoubtedly on the pterygoids and are separated by visible 

 sutures from the palatines, which lie laterally to the pterygoids 

 in front of the ectopterygoids. Further forwairl these sutures, 

 which form the inner border of the palatines, approach one 

 another and descend into the groove, so that its side-walls are in 

 front formed by the palatine. The ectopterygoids ai'e seoarated 

 from the palatines by visil)le sutures. 



There is no trace of a suture down the mid-line of the groove, 

 and its roof seems to be formed by a median bone, which 

 terminates at the sudden end of the groove and must be bounded 

 by sutures with the pterygoids along its edges ; of these presumed 

 sutures nothing can be seen in this specimen. 



The type-specimen oi Lycosaurus tiat'inus Owen seems to throw 

 light on the structure of the palate of Arctognathus ourviniola. 



It consists of a fragment of a snout, broken ofJ' through the 

 premaxillfe in front and by an oblique fracture on the left side, 

 but showing much of the right maxilla. It has been so developed 

 as to show a small strip of the surface of the right palatine and 

 shows a section of the palate on the hinder end. This species is 

 referred by Broom to a new genus Arctosuchus, and said to have 

 a dental formula, i. 5, c. 1, m. 4 or 5, representing a much more 

 primitive type of Theriodont than Arctognathus. The type- 

 specimen only shows two incisors, a canine, and a few cheek- 

 teeth, and it seems certain that Broom examined and used for 

 his description a snout of ScjjmnognathzLs 2ohaitsi which Lydekker 

 had referred to L. tigrinus. 



Except in the larger size and somewhat dijS'erent direction of 

 its canine, the type-specimen of L. tigrinus seems to agree 



