22 Wortman — Studies of Eocene Mammalia m the 



Eutheria. Moreover, evidence of the existence of such a group 

 remains to be discovered. On the contrary, all the facts point 

 very strongly to their origin, along with the Carnassidentia, 

 from Implacental or Marsupial Metatherians. It is likewise 

 conceivable that from this same general substratum the other 

 Eutherian orders arose. 



Some day, when our knowledge of these matters is vastly 

 more extensive and accurate than at present, it will perhaps be 

 necessary to abandon this horizontal system of classification, 

 imjDlied by the use of such terms as Prototheria, Metatheria, 

 and Eutheria, and substitute for it the phyletic or linear sys- 

 tem — the only one expressive of the genetic relationship which 

 we seek to discover. It will then be determined, without 

 much doubt, that these several ordinal groups have recogniz- 

 able chains of ancestry, penetrating not only well down into 

 the Metatherian substratum, but almost to the very bottom 

 or beginnings of Mammalian existence. We shall then give 

 names to these lines of descent rather than to successive stages 

 of their development. But until this Utopia in the subject is 

 realized, we must content ourselves with cruder methods, more 

 in keeping with our ignorance. 



Stnnmary. 



Having now completed a study of all the Eocene Carnivora 

 in the Marsh collection, I append herewith a brief summary 

 of the more important discoveries, opinions, discussions, etc., 

 embodied in the foregoing paper. They are the following : 

 The general organization of the order and the relationship of 

 its more primitive members to the Metatherian Marsupials are 

 discussed ; the order is divided into three suborders, Creodonta, 

 Carnassidentia, and Pinnipedia, and their relations are con- 

 sidered ; of the Eocene Canidse, the type of the genus Yul- 

 pavus is figured, a new species added, much of its osteology 

 described, and the progressive modification of the family con- 

 sidered ; a new related genus, J^eovulpavus^ is proposed ; addi- 

 tional characters of TJintacyon^ together with its position in 

 the group, are given ; Prodaphmnus is considered to be the 

 forerunner of the Amijhicxjon series, and four main lines of 

 canine descent are pointed out ; the Yiverravidse are deiined, 

 the type species of Yiverravus is figured, and another species 

 added ; a new genus of this family, Oodectes, is proposed and 

 a large part of its osteology described ; the relations of the 

 Yiverravidae to the living Yiverridse are examined, and the 

 position is taken that the descent of the modern civets is prob- 

 ably traceable to this source ; the type species of the three 

 genera Triacodon^ Ziphacodon, and Ilarpalodon, are figured, 

 and the opinion expressed that they are not valid genera ; the 



