Scott.] ■^'JO [May 18, 



recapitulate their resemblances and differences. Agriochcerus resembles 

 the oreodonts in the following points : (1) The skull structure of the two 

 families is closely alike and the Uinta genus Protoreodon hardly ditfers at 

 all in this respect from Agriochcerus. (2) The very characteristic oreodont 

 features of the caniniform first lower premolar and incisiform canine are re- 

 peated in Agriochozrus. (3) The atlas and the peculiar odontoid process 

 of the axis are similar in both groups. (4) The elbow joint ia the oreo- 

 donts is, as is well known, very exceptional among ungulates and all its 

 peculiarities are repeated in somewhat exaggerated form in Agriochcerus. 

 (5) The tarsus is oreodont in almost every particular, and is curiously 

 paralleled in many details by that of the Loup Fork genus Merycocho&rus. 

 It should be remembered in this connection that in A. antiquus the tarsus 

 deviates less from the oreodont type than does that of the later and larger 

 A. gaudryi, and is in fact intermediate between the two. This is signifi- 

 cent, because in the long continued existence of the Oreodontidce from the 

 Uinta to the Loup Fork, there is relatively little change in the tarsus, 

 while each successive genus displays its own particular modification of the 

 carpus. (6) The phalanges, except the unguals, differ relatively little in 

 the two groups and are manifestly of the same type. (7) The articulations 

 of the metapodials are similar. 



The most important respects in which Agriochcerus differs from the 

 oreodonts are as follows : (1) The pattern of the molar teeth is very dis- 

 tinct, but this gap, especially as regards the lower teeth, is to some extent 

 bridged by Protoreodon of the Uinta Eocene, which shows that the two 

 kinds of molar may well have been derived from a single type. (3) The 

 distal end of the radius is creodont, rather than ungulate in character. (3) 

 In Agriochmrus the carpus differs decidedly from that of the "White River 

 oreodonts, both in the shape and ia the connections of its parts, the dis- 

 placement being in opposite directions. Here again Protoreodon tends to 

 connect the two extremes and displays a type of carpus not far removed 

 from that which may have given rise to both. (4) The position of the 

 phalanges with reference to the metapodials and to each other is quite 

 different. (5) Much the most remarkable difference between the two 

 groups is the presence of claws in Agriochcerus, while the oredonts have 

 hoofs. (6) Though the large Agriochcerus species from the uppermost 

 White River beds has no pollex, there is reason to believe that it was 

 present in the earlier and smaller A. antiquus. (7) A striking difference 

 in the knee-joint is observable between AgriocJicerus and the oreodonts, 

 which indicates that in the former the leg was straighter, while the proxi- 

 mal end of the tibia and distal end of the femur have quite the appearance 

 of the same parts in the Carnivora. (8) All the elements of the tarsus 

 are lower and broader, the astragalus has a deep pit for the internal 

 malleolus and a hallux is present. The latter has not been found in any 

 oreodont. (9) Another carnivorous feature in Agriochcerus is the shape 

 assumed by the neural spine of the axis. 



In brief, the dentition and skeleton of Agriochcerus show a large num- 



