r AFRICAN MAMMALS 159 



The rodents are scarce, and referable to the European Theridomyidae, 

 a group intermediate between the Hystricomorph rodents and the primi- 

 tive rodents of the early Eocene. This group is known only from Europe, 

 and was probably evolved in the Palaearctic ; the Fayum genera are im- 

 perfectly known, but compare with those of the late Eocene and early 

 Oligocene in France and Germany. 



The most remarkable feature about the immigrant mammals is their 

 limitation. The Perissodactyla, the most abundant mammals in the 

 northern Eocene, are entirely absent. This can hardly be a matter of 

 accident. Some barrier there must have been, whether of range or 

 habitat, which prevented any of the Lophiodonts, Hyracotheres, and 

 Palaeotheres from accompanying the Hyaenodonts and Ancodons with 

 which they are found in Europe. The al)sence of Fissipede, or true 

 Carnivora; of Anoplotheres, Xiphodonts, Dichobunids, Caenotheres, and 

 other families of Artiodactyls, and of other rarer mammals of the Euro- 

 pean Upper Eocene may be in some cases a matter of accident ; in others 

 significant. 



AuTOCHTiiONic Group 



The most remarkable of the autochthonic group is the Arsinoitherium, 

 a gigantic quadruped with elephantine limbs and a pair of great horns at 

 the front of the skull. It is the sole representative of its order, and, 

 except for distant and disputed affinities to the Hyracoids, has no known 

 relatives. It is analogous to the Uintotherium of ^orth America, but 

 most authorities concur in attributing the resemblances to adaptive anal- 

 ogy, and conclude that the Arsinoitherium evolved out of some primitive 

 Hyracoid stock in Ethiopia during the Eocene, while the Uintatheres 

 were evolving in North America out of the primitive Taligrade stock. 



The supposed ancestors of the Proboscidea, Moeritherium and Palceo- 

 mastodon, are of especial interest. There is no doubt about the position 

 of PaJceomasfodon. It is Proboscidean beyond question; decidedly more 

 primitive than any known Holarctic Proboscidea, as well as an entire 

 epoch earlier in point of time, and, so far as appears, is in the direct line 

 of descent. The position of Moeritherium is not so certain. It is found 

 both in the Upper Eocene and Lower Oligocene horizons, and is generally 

 regarded as an earlier stage of specialization of the Proboscidea, ancestral 

 to Palceomastodon, but surviving along with it. Osborn, however, regards 

 it as open to serious question whether Moeritherium can be regarded as at 

 all directly ancestral to Palceomastodon. 



The conclusion has been generally drawn that the Proboscidea evolved 

 in the Ethiopian region. This conclu^^inn, however justified by the evi- 



