ADAPTIVE COLORATION 227 



visible only, never extending to internal characters or to such 

 as do not affect the external appearance." 



These rules relate chiefly to the Batesian form of mimicry 

 and need to be altered to apply to the Miillerian kind. 



The first criterion given by Wallace is evidently an essential 

 one and it is sustained by the facts. It is also true that mimic 

 and model occur usually at the same time of year; Marshall 

 found many new instances of this in South Africa. In some 

 cases of mimicry, strange to say, the precise model is unknown. 

 Thus some Nymphalidse diverge from their relatives to mimic 

 the Euplceinse, though no particular model has been found. 

 In such instances, as Scudder suggests, the prototype may 

 exist without having been found; may have become extinct; 

 or the species may have arrived at a general resemblance to 

 another group without having as yet acquired a likeness to 

 any particular species of the group, the general likeness mean- 

 while being profitable. 



The second condition named by Wallace is correct for 

 Batesian but not for Miillerian mimicry. 



The fulfilment of the third condition is requisite for the 

 success of Batesian mimicry. Bates noted that none of the 

 pier'id mimics were so abundant as their heliconiid models. 

 If they were, their protection would be less; and should the 

 mimic exceed its model in numbers, the former would be more 

 subject to attack than the latter. Sometimes, indeed, as 

 Miiller found, the mimic actually is more common than the 

 model ; in which event, the consequent extra destruction of the 

 mimic would at least theoretically reduce its numbers back 

 to the point of protection. 



In Miillerian mimicry, however, the inevitable variation in 

 abundance of two or more converging and protected species is 

 far less disastrous; though when two species, equally distaste- 

 ful, are involved, the rarer of the two has the advantage, as 

 Fritz Miiller has shown. His lucid explanation is essentially 

 as follows : 



Suppose that the birds of a region have to destroy 1,200 



