INTRODUCTION. xix 



but, in consequence of the increased size of the posterior molars, the premolars, and 

 eventually even the anterior molar, are shed as the posterior molars come into position 

 from behind. This animal attained the size of a moderately large Indian Elephant, 

 and except for the inflexibility of the mandible-supported trunk must have been very 

 similar in appearance. In the later Miocene the mandibular symphysis shortened, 

 leaving the trunk mobile and unsupported, as we now know it ; at the same time traces 

 of its original elongated condition are retained in the occurrence of deciduous lower 

 incisors in some species oi Mastodon and in the peculiar sharp process of the symphysis 

 in the Elephants. In the teeth there is a gradual increase in the number and height 

 of the ridges composing the molars, together with the great increase in the size of the 

 individual teeth. The milk-teeth are never replaced by premolars, but are early 

 displaced by the advance of the molars into position from before backwards ; the 

 culmination of this line of evolution is met in the latest Elephants, in which in old age 

 the only cheek-teeth in position are the very large third upper and lower molars. 

 The above is, of course, only a general summary of the succession of forms which 

 lead up to the modern Elephants, and no doubt there have been many checks and 

 side-branches leading only to extinction in the course of the vast period that has 

 elapsed since Moeritlierium existed. 



The position of Banjthermm in relation to the other Ungulates is quite uncertain. 

 It cannot be related to Arsinoitherium, since not only are the teeth widely difi"erent 

 in type, but the anterior part of the dentition, in the mandible at least, has undergone 

 great modification, the anterior lower premolars, canines, and posterior incisors having 

 been lost and the anterior incisors modified to form large procumbent tusks. In this 

 respect it approaches the primitive Proboscidea, but the fore limb so far as known 

 diff'ers widely from the Proboscidean type : this is especially noticeable in the distal 

 articulation of the radius, which rather resembles that seen in Uintatherium than that 

 found in the Elephants, and seems to show that the carpus difl'ered considerably from 

 that of the latter group. Another possibility must be referred to, namely, that 

 some kind of relationship with the Pyrotheriidae of Patagonia may exist ; in both the 

 molars are bilophodont, the anterior lower incisors tusk-like and procumbent and 

 separated by a long diastema from the premolars, the other incisors, the canine, and 

 anterior incisors being lost ; in the carpus also the scaphoid seems to have been lari^e 

 in both groups. The probability of such a relationship will be discussed below. 



The Artiodactyla, which are represented in the Upper Eocene by members of the 

 Anthracotheriidse and Suidee, may possibly be regarded as a foreign element in this 



