THE DIFFERENTIATION OF SPECIES. 75 



gether and not separated by deep water show the same species, 

 like James and Jervis, or Hood and Gardner. The faunas of 

 the larger central islands are again closer related to each other 

 than are the faunas of the more isolated islands, like Tower, 

 Bindloe, Abingdon, Hood, and Charles. The Tortoise of 

 Duncan is closest to the Tortoise from Abingdon ; at the same 

 time the Tropidurus of Duncan comes nearest to that of 

 Abingdon. The prevention of intercrossing after the separation 

 of the islands, the time of separation, and the difference in the 

 conditions on the different islands, are the factors which pro- 

 duced the different races. By the subsidence theory every 

 difficulty is explained in the easiest way. 



To give further support to my opinion on the continental 

 origin of the Galapagos, I shall now consider the West Indian 

 Islands. That these islands formerly were connected with 

 each other and with Florida, Central America, and also a por- 

 tion of South America is, I believe, considered to-day an 

 established fact, notwithstanding most of these islands are 

 surrounded by a very deep sea. 1 We even have evidence that 

 this connection still partially existed at a relatively late period ; 

 for remains have been discovered of the Edentates Megalonyx 

 on Cuba, of Mastodon on the Bahamas, and of large rodents 

 and deer in caves of Anguilla. 



If the West Indian Islands are really the product of the 

 splitting up of a greater area of land, we ought to find the 

 same harmony in the geographical distribution of the fauna, as 

 we found on the Galapagos. 



The following table giving the distribution of the species of 

 different genera of birds on some of the islands, speaks for 

 itself ; we find exactly the same harmony: 2 



1 Suess, Eduard : Die Antillen, in : Das Antlitz der Erde. Wien, 1892. Vol. 

 I, pp. 698-712. 



' 2 Cory, Charles B.: Catalogue of West Indian Birds. Boston, 1892. 



