ON THE COERELATION OF PATTERN IN RHOPALOCEKA. 183 



is of much interest, for it is a feature I have not found in any other 

 butterfly that I have examined. 



[n considering the whole question, however, it must be borne in 

 mind that as there are generic resemblances and specific resemblances 

 so there are both generic and specific differences, and they do not 

 necessarily pass along the same line. Investigation has taught us that 

 in the lluralidii', in its broadest sense, the tegumen is of dominant 

 generic value, that the fedoeagus and the harpagines (clasps), so far as 

 their general form goes, are a'so of generic value, but in both of the 

 latter mutation occurs which is purely specific, rarely with the 

 ffidoeagus but regularly with the harpagines, whilst so far as specific 

 divergences are generally concerned, the clasps are the most sensitive, 

 and it is in these organs that we find the smaller or larger differences 

 that are observable between species and species. 1 have referred to 

 two genera that are evidently in a period of marked mutation, Heodes 

 and Hiiralin. In both cases it was seen that the clasps were altering con- 

 siderably in difterent sections of the genera, and that this had been 

 possibly concurrent with alteration of colour and pattern, but that at 

 present, though colour was already difl'erent, yet the structure of the 

 imagines, in theii' form, their neuration, and other characters had 

 practically changed but little. The male armature, however, showed 

 definitely that mutation was in progress, and that both genera are 

 evidently in the process of splitting up. For the present, however, 

 there seems nothing tangeable, except the colour, whereby it would be 

 possible to divide them, and colour is too unstable a feature on which 

 to break up an otherwise thoroughly homogeneous family. I 

 have said that the altering of the cdasps may have been concurrent 

 with colour; as a matter of fact, I believe that colour is much more 

 sensitive to mutation than structure, and that any mutability in the 

 structure follows, rather than is followed by, mutation in colour and 

 .pattern. 



My only really thoroughly exhaustive study has been on the 

 Riiralidde, but i<-om a very considerable number of dissections made 

 by myself in other families, the same result occurs more or less in 

 most of the Rhopalocera, and I am now led to believe that pattern is 

 very generally correlated with structure. 



EXPL.^NATION OF PlATES XIV. -XX. 



All the figures are magnified x30, except those of the genera Acraea and 

 AiiKinris, which are x 7. 



Plate XIV. 



Fig. 1. Pleheius arijus. Fig. 5. Glaucopsyche. cyllarus. 



,, 2. Celustriiia urgiolus. ' ,, 6. Lymena arion. 



,, 3. Cdastrina puspa. ,, 7. Scolitantides orion. 



4. Eoeres' arqiudes. 



Plate XV. 



Fig. 8. Callophnjs ruhi. Fig. 10. Strijmon titiis. 



,, 9. Satsuvui fiivaldszkiji. ,, 11. Strymon v-albiim. 



Plate XVI. 



Fig. 12. N^olycaeiia tevgstroeini. Fig. 14. Thestor fedtschenkoi. 



,, 13. 'lliestor hallux, (not mentioned ,, 15. Laeosopis roloris. 

 in text). 



