SELECTION 149 



Take, for example, the absence of colour on the 

 lower surface of the flat-fish, which is evidently a 

 case of degeneration. The stimulus of light being 

 absent for many generations, retrogression has taken 

 place, and now the old characters can with difficulty 

 be recalled. 



This is different from atavism, in which a latent 

 character takes the place of a newer one. In de- 

 generation the character dwindles, nothing replaces 

 it, and it is finally lost. So long as an organ is 

 stimulated, although perhaps the stimulus may be 

 different from the original one, the organ may 

 remain ; as the web in the foot of the Upland-goose 

 of Patagonia. But when quite unstimulated the 

 organ disappears altogether. 



The incompleteness of Natural Selection. We 

 may now proceed to examine the difficulties in the 

 way of considering natural selection as the sole 

 agent in evolution. The first difficulty is the exis- 

 tence of numerous characters, either of colour or of 

 form, which are called non-utilitarian, because they 

 are not of any use to their possessors. Darwin 

 always saw this difficulty ; but at first he thought 

 that it could be overcome. In the first edition of 

 the "Origin of Species" he attempts to explain 

 these useless characters as being principally due to 

 the inheritance of characters, which were useful to 

 ancestors of the present possessors, but partly as 

 the results of correlation of growth. That is, he 

 thought that they might be necessarily connected 

 with other characters which are of use. No doubt the 

 first explanation will account for the presence of use- 



