34 



University of California Publications. 



[Geology 



rior pair in Philotrox. The otic bullae in both genera are extra- 

 ordinarily large. 



Tomarctus Cope from the Loup Fork Miocene of Colorado is 

 evidently a more specialized form, both as regards the reduction 

 of the premolars and the development of M^. 



Philotrox resembles the Old World Simocyonines in the short- 

 ness of the jaw, and the rather heavy type of the premolars. It 

 differs from Simocyon in the greater development of P 2 and P.,, 

 in the much smaller size of M 2 , and in the presence of M 3 . The 

 heel of Mj is also less distinctly basin-shaped in Philotrox. From 

 Cephalogale it is distinguished by the absence of P 1 and the 

 greater development of P 2 and P 3 . The cusps of the lower car- 

 nassial are also somewhat more elevated and the metaconid seems 

 to be less prominent. In other respects there is much in common 

 with Cephalogale, particularly in M 2 and M,, and in the form of 

 the heel of M r 



Of the five genera referred to the Simocyoninae all are re- 

 moved from Philotrox by quite a distance. Cephalogale is pos- 

 sibly one of the nearest, while Simocyon is one of the farthest 

 removed. Oligobunis is evidently not closely related; Enhydro- 

 cyon and Hyaenocyon appear to be about as far removed as 

 Oligobunis, but are not well enough known to warrant a definite 

 statement. 



In the present state of our knowledge it is difficult to deter- 

 mine exactly the relative stages of evolution reached by the John 

 Day types of short-faced dogs. Philotrox does not appear to 

 represent either the most primitive or the most specialized type 

 of the four genera. As far as is known, it seems to show a some- 

 what higher development of the crushing functions in the tuber- 

 cular region of the inferior molars than we find in the other 

 forms. 



It is also difficult to make even a tentative statement as to the 

 probable origin of Philotrox. While it exhibits some resemblance 

 to Cephalogale, it is to be doubted whether its affinity in this 

 direction is really closer than with some of the members of the 

 Cynodictis group. Particularly is this true of the cusps of the 

 inferior sectorial which are hardly of the type of Cephalogale. I 



