1883.] NATURAL SCIENCES OF PHILADELPHIA. 79 



The characters of the six suborders will then be as follows : 



I. Incisor teeth growing from persistent pulps : 

 Canines also growing from less persistent pulps, agreeing with 

 external incisors in having molariform crowns ; i. Tseyiiodonta. 

 Canines rudiraental or wanting ; hallux not opposable ; 



II. Tillodonta. 

 Canines none ; hallux opposable ; iii. Daubentonioidea. 



II. Incisor teeth not growing from persistent pulps : 

 Superior true molars quadrituberculate ; hallux opposable ; 



IV. Prosimise, 

 Superior true molars quadrituberculate ; hallux not opposable ; 



V. Insectivora. 



Superior true molars trituberculate or bituberculate ; * hallux not 



opposable : Vi. Creodonta. 



While the above scheme defines the groups exactly, and, so far 

 as can now be ascertained, naturally, I do not doubt but that 

 future research among the extinct forms will add much necessary 

 information which we do not now possess. It is possible that the 

 group I called Mesodonta may yet be distinguished from the 

 Insectivora by characters 3'^et unknown. But I cannot admit any 

 affinity between this group and any form of '' Pachyderms," as 

 suggested bj' Filhol, or of Suillines, as believed by Lyddeker.^ 

 Such suppositions are in direct opposition to what we know of 

 the phylogeny of the Mammalia. These views are apparently 

 suggested by the Bunodont type of teeth found in various 

 Mesodonta, but that character gives little ground for systematic 

 determination among Eocene Mammalia, and has deceived palge- 

 ontologists from the days of Cuvier to the present time. The 

 only connecting point where there may be doubt as to the ungulate 

 or unguiculate type of a mammal is the family Periptychidse. of 

 the suborder Condylarthra. The suborder Hyracoidea may fur- 

 nish another index of convergence. 



^ The internal tubercle is wanting in the last two superior molars in 

 Hymnodon. This genus, of which the osteology remains largely unknown, 

 has been stated by Gervais to possess a brain of higher type than the 

 Creodonta. Prof. Scott, of Princeton, is, however, of the opinion that this 

 determination is erroneous, and that Hymnodon is a true Creodont in this 

 and other respects. If so, the genus will perhaps enter the Amblyctonidm. 



^ Memoirs Geological Survey India, Ser. x, 1883, p. 145. 



