336 LEPIDOPTERA CHAP. 



and P. marcellus. It is uncertain how many generations there 

 may be in one year of this species, as the length of the life- 

 cycle varies greatly according to circumstances. But in West 

 Virginia all the butterflies of this species that emerge from the 

 chrysalis before the middle of April are the form marcellus ; 

 those produced between the middle of April and the end of May 

 are telamonides ; while those that appear after this are ajax. 

 P. telamonides is not, however, the offspring of marcellus, for both 

 forms emerge from pupae that have passed through the winter 

 (and are the offspring of ajax), those that emerge early being 

 niiu-cellus, those that appear later telamonides. 



In various parts of Asia and Africa the butterflies produced 

 during the wet season differ more or less markedly from those of 

 the same species produced during the dry season. These are 

 called " wet " and " dry season " forms. Their aetiology has not 

 been investigated, this discovery being comparatively recent. 



Turning to the early life we find that some larvae vary in 

 colour, and that this variation is sometimes of a definite char- 

 acter, the larva being one of two different colours green or 

 brown. In other cases the variation of the species is less 

 definitely dimorphic, a considerable range of variation being 

 exhibited by the species. In tracing the life-histories of Lepi- 

 dopterous larvae it is not rare to find species in which the larva 

 abruptly changes its form and colour in the middle of its life, 

 and so completely that no one would believe the identity of the 

 individual in the two successive conditions had it not been 

 shown by direct observation ; in these cases the change in 

 appearance is usually associated with a change in habits, the 

 larva being, perhaps, a miner in leaves in its first stages, and an 

 external feeder subsequently. In the case of the larval variation 

 we have alluded to above, it is understood that there is no 

 marked change of habits. Poulton has shown 1 that it is not 

 infrequent for some of these latter kinds of variable larvae to 

 change colour during life, and he considers that light or conditions 

 of illumination, that he speaks of as " phytoscopic," are the in- 

 ducing causes. Great difference is, however, exhibited according 

 to species, some variable species not being so amenable to these 

 influences as others are. In dimorphic forms the change was 

 observed to take place at a moult, the larva changing its skin 

 1 Tnins. cut. Soc. London, 1892, p. 293, etc. 



