NEW SOUTH AFRICAN XHEROCEPHALIAN REPTILES. 343 



18. On some new Theroceplialian Reptiles from the Karroo 

 Beds of South Africa. By R. Broom, M.D., D.Sc, 

 F.E.S., C.M.Z.S. 



[Received April 13, 1920 : Eead April 27, 1920.] 



(Text-figm-es 1-9.) 



IsDEX. Page 



Ictidosttchus longicsps, sp. iiov 3-13 



Alopecoj>sis atavus, gen. et sp. iiov 347 



MoscJiorJi inns Tcltcliingi, gen. et sp. iiov 351 



Wha itsia jplati/ceps Haiightou 354 



Thoiigli more species and genera of Therocephalians are known 

 than of either of the other two large suborders of the Carnivorous 

 Therapsida — the Goi-gonopsia and the Cjaiodontia, — the structure 

 of the skull and skeleton is much better known in these later 

 subordei-s. This is largely due to the fact that the Thero- 

 cephalians are mainly found in the Lower Beaufort zones, where, 

 owdng to the petrological conditions, very fine complete spe- 

 cimens are rare. The Gorgonopsiaiis which are found mostly 

 in the Gistece'plialus zone, and the Oynodonts which mainly occur 

 in the Upper Triassic zones, are usually represented by well- 

 pi'eserved skulls with a matrix which can be fairly easily removed. 

 Recently I have been fortunate in discovering one or two very 

 interestiug specimens of late Therocephalians formed in the Upper 

 Endothiodon and Cestecephahis zones, which add considerably to 

 our knowledge of the Theroceplialian skull. 



TCTIDOSUCHUS LOXGICEPS, Sp. nOV. 



In 1900 I discovei'ed near Pearston the remains of a very 

 interesting type of Theroceplialian which I described under the 

 name Ictidosuchus 2^'>''i'^ncKVUs . Until recently no other specimen 

 of either this genus or species has ever been discovered, and as 

 the type skull is very imperfect, a good specimen of this genus has 

 long been one of our principal desiderata. In December 1917, 

 I found at Bruintjeshooyte, between Somerset East and Pearston, 

 a good skull, which may be referred to the genus Ictidosuchus, 

 though a distinct sjjecies from /. primcevus. 



The skull is nearly complete, lacking only the incisor portion, 

 the occipital condyle, the arches of the left side and the quadrate 

 region of the right. The palate is much weathered on the left 

 side, but nearly perfect ou the right. 



As preserved, the skull measures in greatest length 135 ram., 

 and when complete was probably 152 mm. long. Tiie greatest 

 breadth is about 84 mm. From the front of the orbit to the 

 base of the canine is 52 mm., and from the front of the orbit to 



