1910.] The "Premolar Analogy" Theory. 183 



presumed evolution of the true molars from a premolariform type must have 

 taken place during the INIesozoic, since the early stages of the process are 

 never observed in the Tertiary. But no Cretaceous or Jurassic forms 

 have been discovered in which the upper molars show the one essential 

 feature of the presumed j)remolariform mode of complication, namely: 

 "protocone" represented by a small basal ledge springing from the internal 

 side of the tooth. Owing to the scarcity of the upper teeth of Mesozoic 

 mammals this of course is not conclusive; but certainly in the only Jurassic 

 Trituberculates in Avhich the upper teeth are known, namely Peralestes, 

 Kurtodoii and Dryolestes, the supposed pi'otocone is the largest cusp on 

 the crown. This argument has already been advanced in outline by Osborn 

 (1904, })p. 321-323) bvit he omitted a very important link, namely a demon- 

 stration that the high internal cusj) in those genera was truly homologous 

 with the protocone of "normal" tritubercular types. 



This link has been partly supplied in the preceding discussion (pp. 174, 

 180) of the teeth of Peralestes and Amphitherium, in which it is shown that 

 the internal cusp probably fitted into the talonid and therefore functioned 

 like the protocone of Creodonts and all other typically tritubercular mammals 

 Gidley (190G, p. 96, quoted in Osborn, 1907, p. 220) has supplied evidence 

 for the same view in the case of Dryolestes. Being, however, an adherent 

 of the "premolar analogy theory" Mr. Gidley sees in the protocone merely 

 a greatly hypertrophied "heel." In Dryolestes he says "the specialization 

 has apparently been centralized in the development of the high, narrow, 

 heel-like cusp and its supporting fang on the inner side of the molar." He 

 applies also the same idea to Dicrocynodon. The outer parts of the crown 

 of both Dryolestes and Dicrocynodon he regards as homologous with the three 

 cones and two fangs of Triconodon ; but of the inner portion of the crown of 

 Dicrocynodon (/. c. fig. 207) he says "on the internal side a large secondary 

 cusp has been developed differing widely in character from that of Dryolestes. 

 This cusp is a laterally compressed cone supported by tico rudimentary 

 fangs and is joined to the outer portion of the tooth by a high, wedge-shaped 

 ridge." 



Now this large internal cusp in Dicrocynodon presumably fitted into the 

 basin-like talonid of the lower molar and therefore functioned like a normal 

 protocone and like the protocone of Dryolestes. Why, therefore, should the 

 great difference in the shape of the protocone in the two genera and the 

 division of the supporting fang in Dicrocynodon imply that the protocone 

 itself is a neomorph? 



A most valuable part of Mr. Gidley's paper is that he has emphasized 

 the fact that the protocone of the upper tooth always fits into the talonid of 

 the lower (cf. Fig. 13). Because the talonid of the premolars of many 



