IN DEFENSIVE COLORATION 355 



'When an abundant well-protected Acraea thus 

 approximates to two very different Danaine patterns, 

 it is obvious that we are not necessarily driven to 

 a Batesian interpretation of the forms of the female 

 Papilio cenea [the southern and south-eastern sub-species 

 of P. dardanus (merope)], which approximate to the 

 appearance of Amauris echeria as well as to . . . two other 

 Danaine types. . . . The enemies of chrysippus and the 

 species of Amauris are certainly not precisely the same, 

 and it may well be an advantage to a Mlillerian mimic to 

 secure that increased protection from [the experimental 

 attacks of] insect-eating enemies, which is conferred by 

 belonging to two or more groups. ... It is probably of 

 advantage to the whole group that the Danaine which 

 set the pattern should still be the dominant member of 

 the assemblage of which it is the centre. This dominance 

 is favoured by the individuals of an abundant species 

 joining two or more groups, instead of throwing the 

 whole of their number into a single one.' 1 We may add 

 to these considerations the advantages that would be 

 gained from differences in the distribution or local pre- 

 dominance of the various models, and from any overlap 

 they may exhibit in time. A Miillerian mimic belonging 

 to two or three combinations is far more likely to be 

 surrounded by models and co-mimics in any given place, 

 or at any given time, than a mimic which belongs to 

 a single one. Indeed, when we consider it, there is no 

 essential difference between a mimetic species such as 

 Hypolimnas misippus, ox Acraea encedon, presenting three 

 forms resembling respectively the three forms of the 

 single species, Limnas chrysippus, which serves as a model, 

 and the cases under discussion where models of two 

 or three very distinct species are mimicked. The 

 advantages conferred are probably nearly the same, and 

 a specific identity, or remote affinity, between a set of 

 models in any area would not in itself affect the selective 

 process by which they became mimicked by the forms 

 of a single species. 



It is thus seen that the polymorphism of a mimetic 

 1 Trans. Ent. Soc, Land., 1902, p. 469. 

 A a 2 



