OF THE GENUS HYPOLIMNAS. 347 



misippus in this direction from the very first, and sufficient pro- 

 tection being thus conferred there would be no tendency towards 

 the production of other forms. In either case we must look 

 upon the selective process as chieily responsible for the result. 

 It is impossible to deny abundant powers of variation to misippus, 

 when we reraember its faithful resemblance to the special changes 

 undergone by D. chrysippus. But variation being under the 

 guidance of selection in one direction only, has produced nothing 

 in any other direction. It is easy to imagine conditions under 

 which a. bolina might become equally restricted If Euploea 

 core had the distribution of Danais chrysippus, it is probable 

 that no other mimetic variety would have been produced. Or if 

 Danais chionippe of Celebes had the range and abundance of 

 D. chrysippus, it is probable that the superior advantages attend- 

 ing the resemblance to it might cause the ultimate predominance 

 of this one out of the many mimetic forms of H. holina. 



If, then, we are right in believing that the results are deter- 

 mined by the range and abundance of the mimicked form, 

 because this, through selection, determines the number and kind 

 of the mimicking varieties, it is clear that selection rather than 

 unguided variation is the essential cause of the phenomena, 

 always assuming the necessary amount of variation for selection 

 to act upon. 



The fact that selection follows, where possible, the path of 

 least resistance as regards variation, is well seen in S. holina. 

 Not one of its many mimetic forms departs so widely from the 

 ancestral appearance as those of misippus, and for the production 

 of most of them comparatively small changes are necessary. In 

 India and Malaya, with a single exception, various dark-coloured 

 Euploeas are mimicked. The interesting exception of the 

 chianippe form proves that much greater divergence is possible, 

 and that the path provided by the easiest and most probable 

 variation is only followed when it is advantageous. When we 

 pass into Africa, we find that the place of the genus Euploea is 

 taken by the Danais genus Amauris, and dark-coloured butter- 

 flies of this specially protected genus have afforded ready models 

 for mimicry, so that here too the necessary conditions have been 

 met by less divergence than has been necessary for H. misippus. 



My thanks are due to Messrs. Godman, Salvin, and Crowley 

 for examples of various mimetic forms, and especially to Professor 

 Poulton for much kindly assistance in deducing the above con- 

 clusions. 



