The Bionomics of Sovfli African Insects. 475 



to pursue such comparisons further. So far as chrysipiyus 

 is concerned, I know of no facts which support the hypo- 

 thesis of the, environmental production of the forms, and 

 many which are inconsistent with it. 



The only alternative hypothesis which presents itself is 

 that of the operation of natural selection in determining 

 the very different distribution of the various forms of 

 cTirysiijpus in the different parts of its range. And in 

 attempting to solve this difficult problem I have been 

 guided by the same principles which enabled me to 

 suggest a meaning for the two widely-different seasonal 

 phases oi Precis, viz. the relation between insects and their 

 enemies, the value of warning colours under certain con- 

 ditions, their weakness and danger under other conditions. 

 I believe that the condition of desert areas corresponds to 

 that of the dry season, only differing in that they are more 

 rigid, so that cryptic colouring is still more imperative. I 

 therefore suggest that the Idn.ejii form is a development in 

 a procryptic direction in areas where the struggle is so 

 severe that even this most unpalatable and widely- 

 mimicked species must jrat off some of its aposematic 

 appearance, viz. the conspicuous black-and-white apex of 

 the fore-wing. 



There is also a peculiar faintly greenish-orange shade in 

 the area of tlie apex of the fore-wing under-side beyond the 

 sub-apical white bar of chrysippus which is wanting from 

 the corresponding part of Idugii, the difference tending to 

 bring about a further uniformity in the ground-colour of 

 the under-side of the latter. 



Furthermore, many specimens of Idugii have a ground- 

 colour quite different from that of even light individuals 

 of the type-form, gaining a distinct sand colour. This is 

 all the more striking in Africa, where the type-form 

 commonly develops a dark rich fulvous ground-colour 

 very different from the paler Oriental type. 



This interpretation is based on the assumption that 

 klugii has developed from chrysvppns and not chrysippus 

 from Idugii. and no escape from this assumption seems 

 possible. The main lines of argument ai-e these. Island 

 individuals, which are so generally ancestral, are cli.Tysipp)tts 

 and only very rarely Idugii, except near the metropolis of 

 the latter form in Somaliland. Perfect and imperfect 

 mimics, Batesian and Miillerian, are very large in number, 

 especially in Africa. Probably not one of them mimics 



