412 PROCEEDINGS OF THE ACADEMY OF [1892. 



alternate, so that each lower molar opposes two upper ones ; nor 

 can the talon be functional until there is some displacement of the 

 superior protocone and the tritubercular stage is reached, giving a 

 cusp which the talon can oppose. 



In Trlconodon these conditions are not fulfilled. Following 

 Fleischmann's line of reasoning, it might be shown that the posterior 

 cusp of the cat's inferior sectorial was the " metamere " and equiva- 

 lent to the talon, but such a conclusion would be entirely erroneous 

 as it has been demonstrated that in the cats' molar the talon has 

 disappeared. Furthermore, the history of premolar development 

 proves that the talon may be superadded to the primitive triangle, 

 and that the latter may be complete before the former makes its 

 appearance. At all events, no conclusion as to such a problem can 

 be safely drawn from the examination of teeth so far removed 

 from the primitive condition as are those of Dasyurus. 



The mode of origin of the trigonodont (or tritubercular) molar 

 which has been suggested by Cope and Osborn will be assumed as 

 correct in comparing the premolars with the molars, an assumption 

 which is justified by the evidence at present available. In the 

 absence of any well defined phyletic series of mammals from the 

 Triassic to the Puerco, however, the theory cannot be regarded as 

 finally established. The premolars have quite a different history, 

 and, as I have elsewhere shown (No. 11, pp. 48-9) even when these 

 teeth have become completely molariform, the elements which 

 correspond in function and position to those of the molars, are not 

 homologous with them, the key to these homologies being given by 

 the position of the protocone. 



The premolars do not display quite the same degree of constancy 

 in the order of succession of their component cusps as do the molars. 

 For this reason the fourth superior premolar will be taken as the 

 standard. The primitive form of the premolar is a simple cone, 

 implanted by a single fang, which is still preserved in several exist- 

 ing genera and which obviously corresponds to the protocone of the 

 molars. As early as the Puerco, however, we find that p * in every 

 known genus is complicated by the addition of a second cusp upon 

 the inner or lingual side of the protocone, which may be called the 

 deuterocone. This bicusjiid tooth represents a pattern from which 

 all the pre:nolar types of the higher mammalia may be simply and 

 naturally derived by the continued addition of new parts, which in 

 many groups reach the same or an even greater degree of complication 



