312 THE INFLUENCE OF INANIMATE SITKEOUNDINGS. 



known to "Wallace but discovered by myself, that at a fojmer 

 period an elephant"' was found on Mindanao, the most 

 southerly of the Philippines, can scarcely be explained except by 

 the supposition that a direct connection existed between this 

 island and the Indian continent, or an indirect one by a junction 

 with the larger Malayan islands. For any transportation of 

 this species, which is very nearly allied to the dwarf variety of 

 liidian elephant, by a passage across the sea is not to be thought 

 of. Nevertheless, I believe that this hypothetical connection of 

 the islands and mainland is not sufficient by itself to explain 

 even those facts that are already known to us as to the distribu- 

 tion of Indian and Australian forms on the islands lying between 

 the two continents. Even Wallace himself falls back on a 

 number of other causes, and in order not to abandon his general 

 principle he suggests a hypothetical history of upheavals and 

 subsidences, so numerous and various in the different islands 

 that, in the total absence of all geological proof of them, we feel 

 ourselves gradually withdrawn from the terra firma of justifi- 

 able speculation and floating in the clouds. It seems to me 

 that there is a very general predilection for too readily construct- 

 ing sunken continents. Whenever any extensive resemblance 

 between the faunas of two distant countries is discovered, or 

 even imagined, a bridge of mainland is always freely brought in 

 as the only mode of accounting fortius resemblance. No doubt 

 it is the most convenient of instruments, and all the more easy 

 to work with, i.e. to use as evidence for a theory, because it is 

 absolutely impossible to prove the fallacy of the hypothesis by 

 the method of observation, the only way open to the natu- 

 ralist. 



But until the question is finally settled whether two parallel 

 series of animal development might not have proceeded inde- 

 pendently in two countries remote from each other, we can 

 never venture to regard the resemblance of two faunas as con- 

 clusive evidence of their primaeval actual connection ; nay, it 

 even seems to me that the two historical series of species of the 

 horse, recently discovered both in Europe and America, may on 

 the contrary be regarded almost as a proof that each series was 

 developed independently in the two continents and yet led to 



