CH. Ill] MIGRATIONS OF REPTILES. 147 



another river system. We can only understand the min- 

 gling of the faunas of two river systems by wide-spreading 

 floods or by alterations in mountain ranges. Mountain 

 ranges are necessarily modern, geologically speaking, 

 owing to the potency of the effects of subaerial denudation. 

 Different relations existed in river systems before their 

 upheaval; and it is probably to facts of this kind and 

 possibly also to whirlwinds that we must trust in reflecting 

 upon the migrations of the Amphibia, which cannot be 

 much carried about by birds, as they are absent from 

 oceanic islands 1 , even so close to the mainland as is 

 Fernando Noronha. 



Dispersal of Reptiles. 



It will be pointed out later that as a rule the only 

 terrestrial vertebrates of purely oceanic islands, if there 

 are any at all, are Reptiles. This necessarily argues that 

 Reptiles possess some power of crossing the sea which is 

 denied to other vertebrates. There are a certain number 

 of facts which favour this view. While there are com- 

 paratively few mammals which voluntarily take to the 

 sea, there are not a few reptiles which either live almost 

 habitually in sea water without any corresponding modi- 

 fication of structure, or which occasionally are met with 

 in salt water. 



1 There is however Bufo dialophus of the Sandwich Islands ; it may 

 be that the strings of eggs produced by the toad are more portable by 

 birds than the masses of eggs of the frog. 



10—2 



