LYCAENA ARION. 339 



very dirty, circumstances that are too clearly recorded in the 

 photographs. 



PI. xxxviii. shows the true legs of the two larva x 45, again of 

 identical dimensions. 



PL xxxix. shows the 7th and following segments of the autumn 

 larva, and a portion of the skin of the same region of the full-grown 

 larva, showing the identity of the hair-bases x 100. 



PI. xl. shows two portions of the prothoracic region of the full- 

 grown larva (xlOO), to be compared with pi. xxxvi., fig. 1, and pi. 

 xxxvii., fig. 1; the hairs and bases are identical, though their aggrega- 

 tion is so different. In pi. xxxix., fig. 2, and pi. xl., fig. 2, the reticu- 

 lar structure of the skin is well shown. 



In the newly-hatched larva there is one peculiarity that separates 

 it from all Plebeiid larvae that I have examined, and from the Everid 

 and Lycaenopsid larvae, of which, however, I have only seen minimus 

 and argiolus. All these show immediately below the dorsal tubercules 

 and above the supra-spiracular, two lenticles on each abdominal 

 segment, these are, in fact, very conspicuous features of these larvae. 

 L. avion has only one lenticle in this space (apparently the upper one) 

 and it is by no means a large and conspicuous one. The second lenticle 

 is, however, occasionally present ; I have a specimen in which the 

 fourth abdominal segment has it on both sides. The other features of 

 the larva do not distinguish it from the other first stage Lycaenids 

 referred to. The minute hair in front of i and ii is present. The 

 posterior of the two hairs at iii is reduced to a point. The central 

 of the three subspiracular hairs is much the longest. The hairs and 

 lenticles on 1 , 8, 9, and 10 (abdominal) vary from their distribution on 

 those in front in much the same w T ay as in other first stage Lycaenids. 

 These and other details will be seen in pi. xxxiii. better than by 

 description. 



In the second instar (pi. xxxiv.) the hairs are more numerous, and 

 tend to be arranged across the segments in tw 7 o row 7 s. The dorsal 

 lenticles do not increase in number, but several tend to appear 

 (irregularly in different segments, and even on opposite sides) near the 

 spiracles, e.g., in pi. xxxiv., these appear near the left, and two close 

 above the right, spiracle, on the first abdominal segment. 



In the third instar the tendency for the hairs to divide into an 

 anterior and posterior set on each segment is very obvious in pi. xxxv. 

 as also the circumstances that two dorsal (apparently ii., not i.), and 

 a sub-spiracular hair are conspicuously long (0*6mm.), and that the 

 marginal hairs are well-developed. Plates xxxvi.-xl. will give a better 

 idea of the developments of hairs, and on the last instar both before 

 and after hibernation, than much description. The honey gland 

 already conspicuous in the second instar is very well developed. 



I am indebted to Mr. A. L. Eayward for a specimen of the pupa, 

 and of a larva of L. avion just about to moult to pupa, my own 

 search for these, over which I spent a week in Devonshire, having been 

 quite barren, although I turned up various stones, and dug up not a 

 few ants' nests, and made unavailing examinations in other directions. 

 The result of my examination of this material was absolutely to con- 

 firm Mr. Frohawk's conclusion, that the larva that disappears in the 

 autamn, shortly after entering its fourth instar, is still in the same 

 instar when it is discovered quite full-grown in early summer. 



