182 Bulletin American Museum of Natural History. [Vol. XX^'II, 



tuberculata. It is not necessary to rehearse any of the opposing arguments 

 which have been discussed in Osborn's 'Evohition of the MammaHan 

 Molar Teeth' (pp. 7, S, 33, 227); but it nevertheless seems advisable to 

 develop here some new criticisms against various old views, to endeavor to 

 reconcile conflicting evidence and to strive for a new synthetic view of the 

 evolution of the molar teeth in the Theria. 



The ^'Premolar" Analogy Thcorij. 



One great line of argument against the theory of Trituberculy is known 

 as the "Premolar Analogy Theory" (Osborn, 1907, pp. 215-219). 



The . transformation of the premolars into molariform types among the 

 Tertiary mammals has been described by Scott (1892, pp. 405-444) . Huxley, 

 and later Wortman (1902, pp. 93-98, 1903, pp. 365-368), concluded that 

 the premolars indicate in their transformation the precise steps l)y which the 

 molars had attained the tritubercular type. 



^Yortman (1902, p. 94) figures two genera of Eocene Creodonts, Dissacus 

 and Mesonyx, the first being perhaps a collateral ancestor of the other, in 

 which there is every appearance of a progressive transformation of the 

 molars after the same manner as the premolars. He concludes (/. c, p. 97) 

 that it is "so inherently improbable that in the matter of cusp development 

 the premolars have had one history and the molars another, that the evidence 

 would recjuire to be of the most direct and positive kind even to place such a 

 proposition on the ground of reasonable probability." 



Several considerations however appear to indicate that this argument 

 should not be accepted without reservation. First, "inherent improbabili- 

 ties" have at times become demonstrated facts in science. The principle 

 of convergent and parallel evolution between entire organisms is now so well 

 known that the "inherent improbability" of convergence or parallelism 

 between closely related structures, such as premolars and molars may not be 

 taken for granted. 



Secondly, the sharp differentiation of premolars from molars is usual if 

 not universal not only in Basal Eocene and Jurassic Therians (Osborn, 

 1904, p. 322) but was present in some forms as far back as the Upper Triassic 

 (Diademodon, Dromatherium, Microconodon). In the most primitive 

 Polyprotodont Marsupials and in the Jurassic Trituberculata p.|, being 

 simple conical teeth, are wholly unlike m. \. In the Basal and Lower Eocene 

 Creodonts, Insectivores, "Primates" (Indrodon), Ganodonts, Condylarths, 

 Amblypods, etc., p^ is at first a bicuspid tooth and later becomes more 

 molariform. 



Thirdly while the molarization of the premolars is well established, the 



