Cli. XXL] EXPLANATION OF MIMETIC FORMS. 383 



and was not to be distinguished except when accidentally 

 shaken out ; it is the larva stage of a species of Phasma. 

 The extraordinary perfection of these mimetic resem- 

 blances is most wonderful ; and I have heard this urged 

 as a reason for believing that they could not have been 

 produced by natural selection, as it has been argued 

 that a much less degree of resemblance would have pro- 

 tected the mimetic species. To this it may be answered, 

 that natural selection not only tends to pick out and 

 preserve the forms that have protective resemblances, 

 but to increase the perceptions of the predatory species 

 of insects and birds, so that there is a continual progres- 

 sion towards a perfectly mimetic form. This progressive 

 improvement in means of defence and of attack may be 

 illustrated in this way. Suppose a number of not very 

 swift hares and a number of slow-running dogs were 

 placed on an island where there was plenty of food for 

 the hares but none for the dogs, except the hares they 

 could catch, the slowest of the hares would be first 

 killed, and the swifter preserved. Then the slowest- 

 running dogs would suffer, and having less food than the 

 fleeter ones, would have least chance of living, and the 

 swiftest dogs would be preserved; thus the fleetness of 

 both dogs and hares would be gradually but surely 

 perfected by natural selection, until the greatest speed 

 was reached that it was possible for them to attain. I 

 have in this supposed example confined myself to the 

 question of speed alone, but in reality other means of 

 pursuit and of escape would come into play and be 

 improved. The dogs might increase in cunning, or com- 

 bine together to work in couples or in packs by the same 

 selective process ; and the hares on their part might 



