582 PEor. E. B. POULToisr : natural selection 



the dominant types of common warning colours. These and 

 other reasons, such as the great number and wide geographical 

 range of species belonging to the same genus and adopting 

 a single metiiod of concealment, compel the bt-lief that examples 

 of protective resemblance are extremely ancient in the past 

 history of the species as compared witb examples of mimicry, so 

 that we can well understand how it is that in the former, when 

 the female differs it is ancestral as compared with its male, while 

 in the latter the converse relationship obtains, and the appear- 

 ance pi-esented by the male is comparatively ancestral. 



The main conclusion which emerges is that the advantageous is 

 the thing that is attained. If an ancestral appearance is advan- 

 tageous it is retained, especially in the sex that needs it most ; 

 if a new appearance is advantageous it is attained, especially by 

 the sex that needs it most. The female sex becomes conserva- 

 tive or progressive according to the needs of the species, and 

 natural selection is limited by no bounds of constitutional dif- 

 ference between the sexes as regards the preservation of the old 

 or the initiation of the new. 



(10) The Space and Time Relationsliips of the Resemhlances 

 in Question. 



A mimetic group is found in the same locality, or at least the 

 mimic (Batesian) is not found beyond the range of its model. 

 The types of Common "Warning Colours are remarkably local, 

 although probably certain members of the group (being ex 

 hypothesi all specially protected) may sometimes have a wider 

 range than others. When such a mimic as Sypolimnas misippus 

 can invade and thrive in South America and the Antilles in the 

 absence of its model (Limnas chrysippus), we probably have to do 

 with a Miilleriau rather than a Batesian association (see also 

 a paper on Mimicry in the Genus Sypolimnas by E. B. Poulton 

 in Eeport Amer. Assoc, for Adv. of Sci., Detroit Meeting, 1897, 

 where other arguments in support of this conclusion are urged). 



Looking at the examples broadly the phenomena are charac- 

 teristically local. This, although harmonizing with the other 

 suggested explanations, is quite unintelligible if the theory of 

 internal causes be adopted. Why should these results if 

 attained independently in the evolution of various forms be 

 attained in the same locality ? The number of patterns and the 

 number of forms is so vast that we must expect a certain amount 



