60 



University of California Publications. 



[Geology 



large. This group is considerably advanced beyond the stage of 

 development of the nearest forms in the White River. 



The four remaining canid genera are generally referred to 

 the simocyonine division of the Canidae, having its typical rep- 

 resentative in the European Simocyon. They are all, as far as 

 known, short-headed forms with heavy jaws and a more or less 

 reduced dentition. While a reasonable doubt exists as to whether 

 these genera are really closely related to the Simocyonines, they 

 are all of a more specialized stage of development than the White 

 River dogs. In some cases there seems to be good reason for be- 

 lieving that they are the more highly specialized descendants of 

 White River forms. 



The Loup Fork Canidae are characterized by the presence of 

 the peculiar Aelurodon group; by several more or less bear-like 

 genera of the amphicyonine type, viz.: Amphicyon, Dime yon, 

 Ischyrocyon, and Cynarctus; and by a few imperfectly known 

 species closely resembling Canis. These groups all differ from 

 the John Day dogs and, excepting possibly the case of Canis, 

 there is apparently little direct connection to be traced. Even 

 Canis is not to be definitely connected with any of the John Day 

 genera. Mesocyon may not be far from its ancestral line, but 

 stands rather as the representative of a type than as the ancestor 

 itself. 



One of the earliest forms which seems to lead toward the 

 somewhat specialized Aelurodons is Tephrocyon of the Masc'all. 

 Tephrocyon may possibly be derived from a Mesocyon-like type 

 of the John Day. The amphicyonine canids of Loup Fork repre- 

 sent a branch of the family not known in the John Day and 

 White River. A satisfactory statement of the relative stages of 

 evolution reached in these divisions is difficult, as they are so 

 different, but the degree of advance exhibited by the Loup Fork 

 fauna is on the whole greater. In the Canis type an additional 

 factor is introduced, as the Loup Fork species referred to Canis 

 appear to be closer to a recent genus than to any John Day form. 



In comparison with the Deep River beds, generally consid- 

 ered as older than the Loup Fork and near the age of the upper 

 portion of the John Day, there is but little material from which 

 to draw conclusions. The genus Cynodesmus is near Mesocyon 



