,M i Annals of the South African Museum. 



seen between the two thin ascending plates of the ptery golds. These 

 latter, as stated above, seem to face anteriorly and form the median 

 bone which is clasped by the prevomers. Anterior to it, the longi- 

 tudinal section shows another well-ossified plate in the nasal septum: 

 and the pterygoid plate is pierced by a large opening just above the 

 internarial bar. 



There is a true sphenethmoid, very similar to that of Dicynodon 

 seen on the posterior surface of the specimen. It forms the inner 

 wall of the orbit, and is seen to articulate with the frontal, prepa- 

 rietal and parietal. The body of the bone separates the orbit from 

 the anterior prolongation of the brain. A similar ethmoidal element 

 has been described by Watson in Sci/mnognathnx whaitsi. The rib 

 along the ventral surface probably rested in the groove in the upper 

 surface of the parasphenoid which, in our specimen, is displaced to 

 one side. 



Broom described a similar structure in the type of Aelurognat/iiis 

 tiijriceps. He says the ascending plates of the pterygoid "pass up- 

 wards and clasp the front of the median sphenoid. In front of the 

 sphenoid they become ankylosed, and form a median plate which 

 extends forwards to meet the vomer" (i. e. the prevomers). "Whether 

 this large, thin, median plate is entirely made up of the fused ptery- 

 goids, or whether there is a median basi-cranial element as well, 

 cannot be made out in the specimen". 



One of the most interesting features which has arisen from a 

 re-study of the line series of types in the Museum collection is con- 

 nected with the question of the presence or absence of a median 

 vomer on the palate. In the earlier descriptions of the Gorgonopsian 

 palate, such as that of Scylacops capensis hy Broom, there was no 

 mention of a median hone on the palate between the pterygoids 

 such at occurs in some at least of the Cynodonts, e. g. Diademodon. 

 Broom considered then that the interchoanal bar was the true vomer 

 and that there were no paired prevomers. This position he dubiously 

 maintained in his C'roonian Lecture of J9i:>. pointing out that "till 

 further specimens are studied it will be impossible to settle the 

 question conclusively". 



In his 19"21 paper on the Classification of the Theriodontia Watson 

 figures and describes a median vomer forming part or most of the 

 roof of the median vaulted area of the palate in Gorgonops torvus, 

 Scymnognathus wliditxi and Arctognathus curvimola among the Gorgo- 

 nopsia; whilst, on the other hand, from my examination of material 

 for the present paper 1 have been unable to discover a single Gorgo- 



