﻿LEPIDOPTÉROLOGIE COMPAREE 295 



of the incrcasc of size, has an ordinary third stage skin very 

 much stretched and distended, so that : 



1 . The hairs, etc., though présent, are those suitable to an 

 ordinary third stage larva and are, till lookcd for, practically 

 invisible. 



2. The colouring mattcr which lies immediately beneath the 

 skai and made the larva when it leaves the gentian a dark red 

 brown, is attenuated so that the larva when 7 or 8 mm. long is 

 a pale flesh colour, and when full grown shows only some pink 

 shading. 



3. The larval skm is so stretched and attenuated that the 

 fat-bodies are very cons])icuous m a way rarely seen except in 

 some niternal-fecding larvae and that I ha\e certainly never 

 seen m any othcr butterfly larva. 



Lycaena alcon, then, dropped the third moult, and managed 

 to feed up to full growth m an unmodified third instar skin 

 That in once had a third moult is not only certain from the 

 ruie in other Lycaenids but is confirmed by my lucky observation 

 that it inay on rare occasions revert to havmg a third moult. 

 That the larva died, by no means proves that such a reversion 

 has necessarily a fatal resuit, as this larva had considérable 

 expériences of travel by post and was not improved m health 

 by my exammations and disturbances of it at this critical 

 moment. The fourth instar m this spécimen presentcd an ordi- 

 nary skm panoply and nothmg like the spécial fourth stage 

 skin armature of Z. arioji. 



It seems probable that the common ancestor had a history 

 something like that of L. aie on, but in the fourth instar. From 

 this condition, L. alcon progressed by dropping a moult, L. arion 

 by assuming a spécial fourth stage skm armature, or rather 

 perhaps, it retained a proper fourth stage armature, which looks 

 spécial owing to the dwindled size of the larva on entering the 

 fourth stacfe. 



