264 THE VARIATION OF ANIMALS IN NATURE 



requirement of the mimicry theory, but, even when the occur- 

 rence of such forms has been fully demonstrated, it requires 

 much additional evidence. The nature of the evidence required 

 may be broadly outlined as follows : 



(1) Detailed knowledge of the food of enemies (especially 



birds) in the areas where mimicry occurs. Our 

 knowledge must be quantitative to allow us to 

 arrive at any certain conclusions. 



(2) Detailed, quantitative knowledge of the rejection of 



models and mimics by enemies which prey exten- 

 sively on allied palatable or non-mimetic forms. 

 For the Mullerian aspect of mimicry we require 

 more knowledge of the process by which young birds 

 learn to recognise appropriate foods. 



(3) More evidence as to the possibility of convergence due 



to (a) parallel evolution, (b) exposure to similar 

 edaphic conditions. There is not much to go on at 

 present, but these possibilities appear to be insuffi- 

 ciently explored, and certain examples difficult to 

 explain by the ordinary mimicry theory may be 

 elucidated in this way. 



When we find how little our knowledge is of these important 

 questions it may seem remarkable that the theory has been so 

 widely accepted. We therefore wish to emphasise the following 

 points in its support : 



The existence of fairly numerous instances in which 

 the colour convergence has been brought about by 

 the modification of totally distinct structures. 



The geographical coincidence of model and mimic 

 where edaphic factors are very unlikely to be respon- 

 sible for the resemblance. This argument is even 

 •more important where several geographical races of 

 both species are involved. 



The considerable amount of data suggesting that the 

 supposed models are to some extent distasteful and 

 rejected, and that the mimics are liable to be mis- 

 taken for them. 



We suggest that the data are at present not sufficiently quantita- 

 tive to be very conclusive. There is a tendency to obtain part 



