1910.] Fou}uJafio)).'< of the "Wedge Theory." • 187 



(2) The upper molars in Diademodon present an internal border rising 

 to a middle point, suggestive of the protocone. 



(3) The lower molars of Diademodon have a transverse ridge analogous 

 to a protoloj)hid, and if the external tip of this ridge were produced dorsally 

 it would, in mastication, occupy i)recisely the relations of a protoconid in a 

 tritubercular molar. 



(4) The well defined ])rotolophid mentioned above is a ridge running 

 directly inward from the protoconid to the site of the metaconid. These 

 cusps are connected by an incomplete ridge in typical tuberculo-scctorial 

 molars (Fig. 12, B, C). 



It is not suggested that the Diademodon molars fulfill all the conditions 

 of the ancestral tritubercidar type, which was probably derived from a form 

 in which the protocone and protoconid were both produced into a high i)oint. 

 The comparison with Diademodon is made merely in order to show that 

 when attention is directed less exclusively to the origin of cusps and more to 

 contours of molar crowns and to actual mechanical relations of upper and 

 lower teeth, it becomes unnecessary to resort to the hypothesis of the migra- 

 tion of cusps, in order to account for the spatial relations of the parts of the 

 tritubercular molar; because these relations may have become established 

 long before the appearance of any cusps except the protocone and pro- 

 toconid, and may even antedate the accpiisition of the diagnostic mammalian 

 characters. 



The molar pattern of Protodonts and Triconodonts does not constitute a 

 valid objection to this view, i. e., it is not necessary to suppose that all the 

 pre-tritubercular molar types that ever existed are represented in the known 

 Protodonts, Triconodonts and Multituberculates; because even as far back 

 as the Cynodonts a wide adaptive radiation of molar types had already taken 

 place. 



The protolophid of Dryolestes (Fig. 12, B) aj^pears to be an important 

 structure which has been neglected by the Theory of Trituberculy in the 

 earlier stages of molar evolution, '^lliis was formed first by the upgrowth 

 of the metaconid from the basal cingulum, at a point directly internal to 

 the protoconid and secondly by the junction of this metaconid with a 

 descending ridge of the protoconid. As the protoconid became more per- 

 fectly wedge-shaped in cross section, its posterior edge, running into the 

 small metaconid below, became sharper and sheared past the anterior edge 

 of the upper molar. As the paracone grew up, the front view of the upper 

 tooth would present a concavity, facing downward, between the paracone 

 and the ])rotocone. Similarly the back view of the lower trigonid would 

 present a concavity, facing upward, between the protoconid and the meta- 

 conid. The anterior part of the paracone and the metacone, together with 



