REPTILES FROM SOUTH AFRICA. 1079 



a badly weathered skull in the British Museum, which was made 

 the type of L. pardalis. Unfortunately, the genus Lycosaurus is 

 at present in considerable confusion owing partly to the bad 

 condition of the type, and partly to the fact that two other species, 

 which are not nearly related have been added to the genus. 



In Lycosaurus pardalis, Owen determined the dental formula 



.3 1 5 



to be i. 3, c. ], m. 5. Lydekker not only believes L. tigrinus Ow.to be 

 correctly referred to Lycosaurus, but thinks it is really indistinguish- 

 able from L. pardalis. In my opinion the two do not even belong 

 to the same family. The dental formula of Lycosaiorus jKtrdalis 



.53 4 



I determine as ir, or, mr. The minute canine in front of the 

 large one is a character very frequently met with in the early 

 Therocephalians of the Pareiasaurus zone. The deep square 

 symphysis with the incisor teeth carried well to the front, and 

 the small size of the premaxillary portion are all eai'ly characters. 

 Owen gives the locality of the type as " Sneewberg mountain- 

 range," but this is evidently a mistake, and Lydekker in his 

 Catalogue merely states that the specimen is from the " Karoo 

 System of the Cape Colony." It. is highly probable that the 

 specimen came from the Gouph. 



Lycosaurus tigrimos is a much later type of Therocephalian. 

 . 5 j. 4 or 5 



Its dental formula is t.~, c. ,m. ^ . As it belongs to a different 

 genus, I would suggest the new generic name Arctosicchus to 

 contain A. tigrinus Owen. 



Lycosaurus curvimola belongs to a still higher type of Thero- 

 cephalian, which is not nearly allied either to Lycosaurus or to 



4 1 5 



JElurosaurus. Its dental formula is i.~, c.~, mr. I would 

 suggest for it the new generic name Arctognaihus to contain 

 A. GU,rvimola Owen. 



Removing then these other genera and reserving Lycosaurus 

 for the type-species, we get a form resembling in many characters 

 Eriphostoma, but diifering in being larger by a half, in having 5 

 incisors as against 4, and 2 canines instead of 1 as in Eriphostoma. 

 Ictidosaurus agrees with Lycosaurus as regards the incisor and 

 canine formula, but differs in having a large number of molars 

 and a numbei- of other cranial chai-acters. The dental formulae 

 of the three genera are as follows : — 



Lctidosaurus: ^. LAA JlJ ; c. '^; ,,, 12 3 4 5 6 7 8 



Lycosaurus : i. LAAAJ.; c. — ; m. LAA_.^'. 



Eriphostoma: ?.LAAJ; c. '_? ; m. ■ ■ ■ ■ . 



The Geological Horizons of the Beaufort West Specimens. 



Some years ago I endeavoured to subdivide the Beaufort series 

 into palseontological zones. At that time it was only possible to 

 do so on very broad lines. The area is so extensive, and except 

 on the upper part of the series there are no lithological characters 



73* 



