150 Proceedings of the Royal Physical Society. 



be impossible to determine the nature of this difference, to 

 say whether it is due to greater vigour of the entire organism, 

 or to higher vitality in the structure or organ involved, its 

 recognition is of first-rate importance, for in no other way is 

 it possible, as far as I can see, to account for the persistence 

 of some modifications and the abrupt disappearance of others. 



By way of illustrating what I understand by natural 

 selection, I shall endeavour now to indicate how I believe 

 the recent elephants were derived from their primitive 

 Eocene ancestors. This will be perhaps best done by 

 directing attention to the modifications of the mandible. In 

 Moeritheritim the mandibular symphysis is 75 mm. in length, 

 in Palceomastodon 400 mm. in length, i.e., five times longer 

 in Palceomastodon. Part of the difference is doubtless due 

 to a difference in the size of the two animals. A more 

 accurate indication of the increase will hence be obtained by 

 contrasting the length of the symphysis with the total length 

 of the mandible. In the Moerithcrinm the mandible is five 

 times the length of the symphysis, in Palceomastodon the 

 mandible is 2J times the length of the symphysis. In the 

 case of Tetrahelodon, tlie symphysis is nearly half the length 

 of the mandible. It measures 540 mm., and is hence 7 '2 

 times longer than in Mocritherium, and 1'3 times longer than 

 in Palceomastodon. 



In the Indian elephant the symphysis, instead of measuring 

 540 mm. as in Tetrahelodon, only measures 200 mm. The 

 mandible, instead of measuring 1206 mm. as in Tetrahelodon, 

 only measures 512 mm., which makes the mandible in 

 Mephas 2| times the length of the symphysis, or relatively 

 only half as long as the mandible in Moeritherium. 



From these measurements it follows that the symphysis, 

 from being 4th the length of the mandible in Mocritherium, 

 increased until it was half the leno-th of the mandible in 

 Tetrahelodon, in which it is nearly 2J times longer than in 

 the Indian elephant. 



The question now is, How was it possible for the symphysis 

 to rapidly increase in length during the first half of the 

 Miocene period, and as rapidly diminish during the second 

 half? 



