NO. 48.— -1897.] ON WALLACE'S THEORY. 



85 



At first sight it looks like a different species, but close 

 inspection shows that it is merely a common form 

 emphasized. 



In a Paper I had the honour to read before this Society 

 on the varieties of Testudo elegans, I remarked that I could 

 see nothing in these varieties for natural selection to select 

 from : the different individuals crawl about on the open 

 plains in the midday sun, and we cannot imagine that one 

 form of colouration or sculpture can be of more advantage 

 than another. Here I fell into an error of the same kind as 

 Lord Salisbury's ; if I read my Wallace right, these variations 

 under the present conditions are of no advantage : it is only 

 when the conditions change that natural selection comes 

 into play. 



Let us take the case of the money cowry. Supposing the 

 conditions of the Red Sea and of the Japanese seas change, 

 so that in the one only G. moneta survives and in the other 

 only C. annulus ; then we should certainly call them distinct 

 species ; or, supposing that the conditions of the Great Ocean 

 were so changed, that none of the varieties except the banded 

 ones could survive, then we should have a single constant 

 species over the whole area. 



If we imagine the conditions to keep constant, then this 

 banded species will tend to vary again ; on the contrary, if 

 the conditions continue to change in such a way that the 

 species can slowly adapt itself to such changes, we have a 

 case like that quoted by Romanes, in which transmutations 

 of a water snail of the genus Planorbis can be traced through 

 a long geological period in an ancient lake basin in Wur- 

 temburg. 



That such cases really occur in Nature there can be no 

 reasonable doubt. Our species of Munias are an excellent 

 example of such a case, and this could be multiplied by 

 hundreds ; and of course if we allow that species can arise 

 from varieties in this way, there is not the slightest difficulty 

 in granting that genera, families, orders, and classes may 



