tibia with large spine. Astragalus with both inferior and anterior 

 pulley-shaped surfaces. Ulna much reduced distally, behind the radius, 

 which includes almost the whole of the carpal articulation. 



This arrangement violates previous views less than any other that 

 would recognize the primary characters of the Eobasil&us. The difficulty 

 of determining the limits of the two first named orders, is partially 

 caused by the fact that the Hyracoidea present the radius of the Proboscidia 

 with the hind foot of the Perissodactyla. These animals are, however, 

 well regarded as a distinct order. Whether all the animals to be included 

 in the Proboscidia possessed a proboscis or not, is of secondary importance. 

 It is neverthele'ss highly probable that Loxolophodon and Eobasileus pos- 

 sessed one, and not unlikely that such forms that approach still nearer 

 the tapirs were not without an organ such as they possess, and which 

 Cuvier ascribed to the PalcBotheria and other allies. 



The divisions of the Proboscidia are as follows : 



No incisors, nasal bones short ; astragalus articulating 

 with navicular only : No third trochanter Proboscidia vera. 



No incisors, nasal bones elongate ; astragalus articu- 

 lating with both navicular and cuboid : No third tro- 

 chanter Dinocerata . 



Dentition complete ; i. e. incisors present ; ? nasal 

 bones. Astragalus articulating with both navicular 

 and cuboid : A rudimental third trochanter Pantodonta.. 



These suborders present a series of approaches to the Perissodactyla. 

 Thus the Dinocerata agree with the typical Proboscidia in addition to the 

 above points, in the posterior expansion of the scapula, and its apical 

 acumination ; in the very short cervical vertebrae ; in the flat carpal 

 bones ; in the absence of pit for round ligament of the femur ; in the flat- 

 tened great trochanter, contracted condyles, and fissure-like intercondylar 

 fossa of the same bone. In the longitudinal crest of the tibia separating 

 glenoid articular faces which are on a transverse line. In the short cal- 

 caneum which is wider than long, and tubercular on the inferior face. 

 In the five digits ; the acetabulum not separated by a peduncle from 

 the iliac plates, and the lack of angular production of the latter beyond 

 the sacrum. Also in the three distinguished sacral vertebrae, as con- 

 trasted with the five closely coossified ones of the Rhinocerotidce. These 

 characters are, some of them, of subordinate value only. 



The chief differences are seen in the cranium, though here also there 

 are important resemblances. Thus in Loxolopliodon the malar bone forms 

 the middle element of the zygomatic arch, sending a narrow strip only for- 

 ward to the neighborhood of the lachrymal. In Uintatherium, accord- 

 ing to Marsh, its extension towards the side of the face is rather greater, 

 but still much less than in Perissodactyla. The dentition is not far 

 removed from that of Dinotherium, and the mode of succession of the 

 teeth was in all probability similar. The premaxillaries and nasals are 



