On a Neiv Type of Cynodont from the Stormberg. 335 



the crown, except in the case of two teeth which are evidently 

 immature. The eight teeth in front measure 17 mm., and each 

 tooth has a section measuring 4 mm. by 1'7 or TS mm. It will 

 thus be seen that the molar series consists of a series of antero-pos- 

 teriorly compressed teeth closely set together. Behind the 8 teeth 

 more or less satisfactorily preserved is a fragment of the root of a 

 ninth. The fifth tooth as preserved has the crown in perfect condi- 

 tion owing to its being immature, and the crown of the eighth is 

 also preserved except for the loss of the tip of the main cusp. 

 In structure the crown is unlike that of any tooth ever previously 

 described, so far as I am aware. There is one large sharp-pointed 

 cusp near the middle of the crown, and on its inner side a second 

 small cusp and a third still smaller. A very small fourth cusp lies 

 in front of the third. If the crown of the tooth were found detached, 

 and one did not know that it was compressed antero-posteriorly, one 

 would readily believe it to be the premolar of some primitive mammal 

 or of some Cynodont such as Galesaurus. 



The present specimen is so entirely unlike any form previously 

 discovered that it is very difficult to discuss its affinities. The first 

 question that arises is whether the animal is a mammal or a Cyno- 

 dont reptile. There is, I think, little doubt that it is one or other. 

 The teeth are single-rooted, though of course this character is of 

 little importance. But the crowns have nothing the least like them 

 in any of the known mammalian orders. Antero-posteriorly com- 

 pressed teeth are not at all rare, as in Chysochloris, Notoryctes, 

 Kurtodon, Dryolestes, and other genera, but these teeth are entirely 

 different in structure from those of the present fossil, and are merely 

 modifications of the ordinary early mammalian type. The branchincr 

 of the maxillary nerve before opening on the face is also a character 

 unknown in mammals, but met with in Cynodonts. In Cynodonts 

 we get antero-posteriorly compressed teeth in Bauria, Triracliodon, 

 and JElurosuchus, but none of these genera can be at all nearly 

 related to the present one. In Bauria the crowns of the teeth are 

 apparently smooth and rounded, and the relations of the teeth to the 

 jugal arch and the general structure of the maxilla and jugal are 

 quite unlike. Triracliodon has molars of a highly specialised type. 

 Each may be described as having a crown with three cusps arranged 

 in a transverse row and connected by a low ridge, while the inner 

 and outer large cusps are connected by a series of very small cusps 

 arranged round the anterior and posterior borders of the tooth. It 

 is possible that the three cusps in the present fossil are homologous 

 with the three transverse cusps in the molar of Triracliodon, and 



