MIMICRY 225 



of diverse orders, the mimetists retaliate by regarding 

 every resemblance as a mimetic one. The generalized 

 resemblance between one species of black and white 

 or yellow Pierine and another is solemnly quoted as 

 an instance of mimicry. But a vast number of 

 Pierince are white or yellow with black tips or black 

 borders to the fore-wings ; such a type of colouring is 

 almost as important a characteristic of the sub-family as 

 (e.g.) the structure of the fore-legs or the method of 

 pupation, and to assume that mimicry is the cause of 

 the resemblance is unreasonable. We know that the 

 Pierince of some islands in the Malay Archipelago are 

 characterized by a certain amount of melanism, the 

 wings being very heavily bordered with black ; we 

 know too that numbers of Celebes butterflies have the 

 fore-wings long, narrow, and pointed, and we attribute 

 these characters to similar environmental conditions 

 acting in a similar manner on butterfly protoplasm. 

 Natural selection will not affect the characters if they 

 are harmless, while if they are correlated with char- 

 acters of immense importance to the species in the 

 struggle for life, natural selection may be said to 

 have produced them. Whole-hearted supporters of 

 natural selection regard variation as indefinite and 

 infinite, and only controlled by natural selection ; but 

 I am heretic enough to believe that variation is de- 

 fined and limited and controlled only partially by 

 natural selection. I regard the lines along which 

 variation in any organism can proceed as limited in 

 number ; to use a metaphor, I look on variation as 

 an engine which can proceed only along certain rails ; 

 there may be numbers of such rails going in different 

 directions, but the engine cannot get off the rails. I 

 16 



