170 BULLETIN 169, UNITED STATES NATIONAL MUSEUM 



fragments. This agrees very closely with Torrejon specimens referred 

 to Psittacotherium multi/ragum, and judged from his figures and 

 descriptions the same is true of Douglass' single tooth. The generic 

 reference is beyond doubt, and the specific reference highly probable." 

 The species is surely and the genus probably different from IPsitta- 

 cotherium lohdelli Simpson, 1929, from Bear Creek, which is definitely 

 more advanced. 



Order CARNIVORA Vicq d'Azyr, 1792 

 Suborder Creodonta Cope, 1875 



Matthew has repeatedly discussed and carefully defined this primi- 

 tive carnivore suborder, which includes all known Paleocene carni- 

 vores. The only serious criticism that has been made of liis general 

 arrangement (for instance by Wortman or Osborn) is that the Mia- 

 cidae, being structurally ancestral to the Fissipedia, shoidd be placed 

 in the latter group. This would be in accord with phylogenetic classi- 

 fication, but as Matthew protested and as most students must agree, a 

 completely phylogenetic classification is a practical impossibility. 

 Tliis case is one in which departure from it seems desirable and 

 necessary. The Miacidae have many characters allying them with 

 creodonts and cutting them off from their descendants the fissipedes 

 and furthermore if they are removed from the Creodonta that group 

 ceases to exist not only as Cope defined and conceived of it but also as 

 a natural and practical group. Alatthew's retention of the "hori- 

 zontal" unit Creodonta, including the Miacidae, seems sound and is 

 adopted here. 



In the Paleocene there are five typical groups of creodonts: Oxy- 

 claeninae, Arctocyoninae, and Triisodontinae (these three subfamihes 

 forming the Arctocyonidae), Mesonychidae, and Miacidae (Viverra- 

 vinae only in the Paleocene).''* Of these the Oxyclaeninae are far the 

 most primitive, without carnassial teeth and with decidedly generalized 

 dentition and skeleton. The Arctocyoninae are also primitive and 

 indeed intergrade with the Oxyclaeninae but are generally larger forms 

 with flat, broad, bearlike teeth. The Triisodontinae are without 

 shearing teeth but with pecuHar blunt, heavy, and extremely simple 

 teeth (probably secondarily in part). The Mesonychidae, so aberrant 

 that they have been excluded from the Carnivora (Gregory) although 

 probably belonging there (Matthew), developed a pseudotriconodont 

 and semihomodont lower dentition and are still more strildngly 



'3 The Torrejon specimens show great variation, and Cope named three species, which were, however, 

 considered synonyms by Wortman. Matthew (Pale. Mem.) believed it possible that more than one species 

 occurs there but did not redefine them separately. 



'< Other groups begin to appear before the nominal end of the Paleocene, but they seem to be Eocene fore- 

 runners, not typically Paleocene mamm als. 



