MAESUPIALIA. 169 



which is the fourth premolar of Plagiaulax. We must therefore regard 

 Hypsiprymnus as the descendant of a type from which the Plagiaulacidce 

 were also derived, in which some of the premolars, as far as the third only, 

 were trenchant, and in which the fourth premolar possessed the tubercular 

 character of the three true molars. Such a type would belong to Jurassic 

 and perhaps even to Triassic times, and might well have continued to the 

 Eocene. I call it provisionally by the name Tritomodon. The lines of de- 

 scent will appear as follows: 



Tritouiodoa (theoretical). 



/ ^ 



Ctenacodou. 



/ 



Plagiaulax. 



/ 



^ / 



Ptilodus. 



/ 



Catopsalis. 



/ Hypsipryinnus. 



/ \ 



Thylacoleo. Macropus. 



The discussion between Professor Owen on the one hand, and Messrs. 

 T'alconer, Kreflft, and Flower on the other, as to the nature of the food of 

 Thylacoleo, is known to paleontologists. From the form of the teeth alone 

 Professor Owen inferred the carnivorous natui'e of the food of this genus, 

 while his opponents inferred an herbivorous diet from the resemblance 

 between the dentition and that of the herbivorous Hypsiprymnus. As the 

 result of the discussion affects in some degree the genera Catopsalis and 

 Ptilodus, I recall it here. The comparison of Thylacoleo with Hypsiprymmis 

 is weakened by two considerations: first, the fact that the cutting tooth of 

 the former is not homologous with the cutting tooth of the latter; and sec- 

 ond, that the grinding series of the former is rudimental, and in the latter 

 it is complete. It evidently does not follow that because Hypsiprymnus is 



