168 Bulletin American Museum of Natural History. [Vol. XXVII, 



of the evolution of a "basin-shaped" molar out of a degenerate tritubercular 

 type have been cited (Osborn, 1897, and 1907, pp. 103-105). 



The writer at one time accepted this view but further study has not 

 tended to support the argument. First, the supposedly analogous cases 

 (loc. cit., p. 80) are shown by examination of photomicrographs ^ to resemble 

 the Multituberculate type chiefly in the possession of a more or less depressed 

 crown with cusps arranged in two anteroposterior rows. But little re- 

 semblance exists between the molars of any INIultituberculate and the molars 

 of the Rodent Perognath^is Q. c, p. 356), of Ccrcokptes, of Myrmecohius, 

 or of certain fruit bats. However the resemblances to Ameghino's Pata- 

 gonian Eocene genera Propolymastodon and Pliodolops, which may possibly 

 have been derived from trituberculate forms, are much more striking. 



Secondly the INIultituberculate molars and premolars retain no direct 

 evidence of derivation from a tritubercular type, but on the contrary suggest 

 a totally independent mode of origin, as shown by comparison of the denti- 

 tion of the jSIultitulierculates Ptilodus and Ctenacodon with the primitive 

 tritubercidar dentition of the Basal Eocene Oxyclsenidse. There is no 

 similarity whatever in the molars; the upper premolars of Ctenacodon have 

 it is true three cusps, ])ut the main single cusp is on the external side of the 

 tooth i. e. the reverse of the condition in Trituberculates. 



Thirdly, the INIultituberculate molar, appearing in the Upper Triassic 

 and following its own lines of evolution until the close of the Basal Eocene, 

 so far as known is a far older type than the tritubercular molar, which is 

 first known in an undeveloped condition in the IVIiddle and Upper Jurassic 

 {e. g., Amphitherium) and which ran through its principle evolution in the 

 Tertiarv. 



The conclusion indicated is that the theory of trituberculy (which is 

 most successful in explaining the dental evolution of the more typical mam- 

 mals) is unnecessarily impaired by endeavoring to make it apply to the 

 Multituberculata. 



There is somewhat better, but still very insufficient evidence for deriving 

 the INIultituberculate type of molar from a generalized triconodont type. 

 In both these supposedly INIarsupial orders, the molar cusps are at least 

 arranged in an antero-posterior line and the jaw motion was originally 

 chiefly vertical (see below). In the relatively primitive INIultituberculate 

 Ctenacodon the second molar has four cusps in a fore and aft line and an 

 external cingulum that might be a development of the one in Triconodon. 

 In both groups the w^earing svirface is chiefly on the inner sides of the upper 

 and on the outer sides of the lower molars. However these resemblances 

 may not imply close relationship. 



1 See Elliot, D. G., Land and Sea Mamm. N. and Mid. Amer. etc., Field Columbian 

 Museum, Chicago, 1904, pp. 163, 171. 



