XXIX 



rate in different genera, are not the same, it must accommodate itself constantly 

 to these. A very clear instance of this I have quoted in the introduction to 

 my third monograph, p. ix in the evolutionary process which causes the 

 destruction of the so-called horn of the Sphingidae larvae. This horn usually 

 takes a different form in the various genera, or shows a different stage of the process. 

 In the same way in butterflies the specific or generic difference will include a 

 difference in physical structure, although it may be too slight for us to observe, 

 and this may be such as not to admit a progress of the colour evolution in 

 the same way, but demand a special accomodation to a special physical 

 structure. On p. xm of the introduction to my first monograph I pointed 

 this out, and quoted J. F. Cunningham's opinion on the subject. In particular 

 it is the structure of the wings which is of importance in this connection ; as 

 discussed on p. xlix of the introduction to my third monograph, distinct 

 ridges or indurations may, for instance, develop on the wings in one species, 

 and not in another, and as persistent spots, may play an important part here. 



Observation shows that these evolutionary processes do not as a rule proceed 

 in the same way on the upper- and underside of the same butterfly. We 

 have discussed this also in the passage referred to, showing that it must be 

 due to a difference of structure of the two wing surfaces, which would influence 

 the development of this evolution in the manner above referred to. 



Observation further shows, that the rate of progress of colour evolution — 

 and indeed of all evolutionary processes — is very unequal ; even in the undoubtedly 

 extremely ancient process by which the fore-legs of the Rhopalocera gradually 

 disappear, the genus Taraka does not seem to be in the same stage of development 

 yet as the other Lycaenidae. It teaches us, moreover, that there are periods 

 of complete stagnation, of epistasis, which may last for a very long time, and 

 which may occur in one sex or species, while in another the process continues; 

 this is another cause of great dissimilarities. I have already discussed this 

 matter in my second introduction, (p. xvii). Finally, we have here to deal 

 with two processes of colour change, that of fading, and that of darkening; 

 processes which develop independently of each other and by no means simul- 

 taneously, and whereby the latter not frequently renders the effect of the former 

 invisible, and then again in some cases, by disappearing more or less, brings 

 the first once more to light. Moreover the phenomenon of structural colouring 

 further increases the variety which the wealth of colour presents to us; the so 

 far unexplained local influences which play such an important part in the 

 Danaidae, have, however, never been observed in the Lycaenidae. There are 

 obviously, factors enough which may account for the differences here under 

 discussion. If we still further consider, that the number of species of Lynaenidae 



