Rhopaloceroiis forms in South Africa. 275 



variable are even tlie forms collected in a single locality, I 

 think that an instance such as this should make naturalists 

 pause before multiplying its nomenclature. Although I 

 am disposed to think that more careful observation will 

 show that this is not an isolated instance in the Pierina 

 group, it certainly would seem contradictory to the axiom 

 " Natura non facit saltum." 



I may remark that individuals of both forms vary much, 

 but in none of the instances collected by me do they show 

 any tendency to vary in the direction of the other form. 



From what I have noticed in some other PieridcB I am 

 inclined to think that the yellow coloration to which some 

 species are peculiarly subject depends in some way or 

 other on the dryness of the season. Specimens of P, 

 Severina and Mesentina ( $ s) collected in dry seasons in 

 moist localities or in dry districts are generally deeply 

 coloured. 



With reference to the changes in the colour of pupre 

 I believe a very wide field of research is open, and with 

 the improved modes of microscopical examination under 

 the spectroscope important revelations on the subject of 

 variation will be discovered. 



I here give the results of some very rude experiments 

 on this subject. 



Most of the specimens were reared in glass test tubes 

 exposed on coloured cards, in which they were partially 

 enveloped. They were constantly supplied with as little 

 food as possible, in order that their full exposure to the 

 colour should not be interfered with. 



Pup^ of C. Keiskamma. 



(1.) On dead leaves aAvay from light. Dark brown. 

 (2.) On stem and on vermilion cards. Pale ochreous. 

 (3.) On vermilion card. Pale bluish-green. 

 (4.) Exposed on bush in nature and on yellow-gam- 

 boge cards. Bright green. 

 (5.) On glass tumbler. Pale yellowish-green. 

 (6.) On green (cobalt and gamboge). Ochreous. 

 (7.) On cobalt-blue. Greenish-white. 



On a white surface, pupfe of Eronia Cleodora became 

 so pale coloured as to be almost translucent, the marks on 

 the ventral aspect of the abdomen being almost obliterated, 

 and the bright yellow-green colour usually predominant 

 almost invisible. Pupaj of Danais Chrysippus in a state 

 of nature are most variable in colour. 



