The Bionomics of South African Insects. 477 



prevailed in Africa, because so many of its Airican mimics 

 retain this shade rather than that now borne by their 

 model, and because chrysijypns itself often exhibits the 

 paler tint, especially to tlie north and east * of the African 

 continent (see pp. 47], 472.). 



I quote below extracts from several of Mr. Marshall's 

 letters dealing with the hypothesis which I have here set 

 forth, and also referring in other ways to this interesting 

 butterfly and its forms. 



''Malvern, May 14, 1897. — I was interested to hear 

 that the L. chrysipinis, var. alcijJjJoidcs, I sent resembles 

 the West Coast specimens. Personally I have never come 

 across that variety commonly, and those I have taken have 

 mostly been very slightly marked with white, but Mr. G. H. 

 Burn, who has collected for some years in the Tugela Valley, 

 near Weenen, says they are not unconnnon there, but appar- 

 ently are most prevalent in the early winter (it is a very 

 hot dry district), and they are frequently marked quite as 

 strongly as the one I sent you." 



"Salishicry, Jan. 24, 1900. — Referring to your very inter- 

 esting remarks onZ. chrysipij'us,! was much struck by your 

 theory with regard to liitgii. The only point, however, on 

 which I do not feel satisfied is whether we can consider the 

 colouring of this variety to be really protective. Of course 

 I have never seen it in life, but I have seen many of the 

 inaria form of Hypolimnas, and I must confess that the 

 insect is very far from being inconspicuous; and even apart 

 from colour it must be remembered that slowness of flight 

 is a very important factor in rendering an insect conspicu- 

 ous (compare our large and powerfully-armed but soberly- 

 coloured wasps of the genus Belcnogaster). Again, it would 

 seem hardly consistent to regard the colouring of Mtcgii as 

 protective if we rank that of the smaller Acrteas, which 

 inhabit some at least of the same areas, as among the 

 warning colours. It would therefore strengthen the 

 theory if some other use could be assigned to the Jdugii 

 coloration, though none occurs to me at the moment. Of 

 course the correlation argument might be brought 

 forward ; but while thoroughly appreciating Meldola's 

 masterly defence of this principle, 1 must admit that I 

 have a distinct distrust in its use in such cases, as it 

 means virtually begging the question. I cannot alto- 

 gether gather from your remarks what are your reasons 

 * See F. A. Dixey in Proc. Zool. Soc, 1898, p. 373, note. 



