128 BRITISH BUTTERFLIES. 



although this is by no means always the case. Massey records (Ent. 

 Bee, xx., p. 144) on the larger size of the specimens of this species 

 in the north of Britain than in the south, those from the Grange 

 district of North Lancashire, Ireland, and West Scotland, being larger 

 than those from the south of England, the Grange and Irish being in 

 turn larger than the Scotch. We believe that the largest British examples 

 that we have ever seen came from the Isle of Lewis, and have always 

 supposed this larger size to be due to the longer feeding-period, the 

 species in these districts being usually later in appearance than in the 

 south, and, to a much less extent, partially double- or even triple- 

 brooded, for it is generally the case, as noted above, in all the districts, 

 at home and abroad in which we have collected that, in doable- 

 brooded areas the second brood is usually smaller than the first, 

 and the examples of single-brooded areas at fair elevations larger than 

 either. One of the finest races, for size, that we have captured, is that 

 of Bourg St. Maurice, at the foot of the Little St. Bernard, where the 

 species occurs abundantly in July and early August, of a size that is 

 equal to the finest examples of var. clara of Irelat d or Scotland, but 

 the $ s of a lovely lilac-blue, and not bright blue as in the latter. 

 Blachier notes (in litt.) that, in the Geneva district, the species 

 averages, $ 30mm., 5 29mm., whilst the largest examples he has 

 measure, $ 34mm., $ 31mm., and the smallest, $ 21mm., $ 17*5mm., 

 the last named of the hyacinthus form ; altogether, Blachier says that 

 his collection contains a series of twelve $ s and $ s less than 23mm. 

 in expanse, the specimens coming from the Geneva district, the south 

 of France, and the Italian Simplon. The largest $ s and 2 s in the 

 British Museum coll. reach 38mm., the smallest $ 22mm., and the 

 smallest $ 20mm. The colour variation in the $ s is very considerable, 

 extending normally from a w T arm lilac-blue, through purple-blue to a 

 bright bellan/us- or escheri-like tint ( = clara, Tutt), and occasionally to a 

 brilliant glossy hylas-hlue(hi/lasoides,n. ab.) a colour that becomes almost 

 normal in the Central Asiatic races. Occasionally, a specimen of eros 

 colour, or colour approaching coridon is recorded (ab. eros, Stphs.),and 

 Barrett mentions one of dark iron-blue colour, somewhat like that of 

 Lycaena avion ( = ab. arionoides, n. ab.). Besides these, a pale form is 

 developed, as a very rare aberration, with a slaty rather than a blue tint 

 ( = ab. Uvida, Gillm.). Blachier says that, in the Geneva district, the 

 $ s are normally more or less deep violet- blue (agreeing probably with 

 our designation warm lilac- blue), very rarely quite blue, although 

 occasionally of a brilliant azure-blue suggesting that of Polyommatus 

 escheri ( = ab. clara); Reverdin also notes the colour as being sometimes 

 pure blue, but usually a little violaceous or mixed with red, specimens of 

 both tints being found on the same da}' in the same place. The typical 

 lilac-blue form very frequently weakens off in its scaling in the 

 summer races to the ab. pallida, a very pale lilac form ; another 

 typically-coloured lilac form is plentifully supplied with long white 

 silky hairs, the costa being markedly white ( — don/las. -Term.), 

 whilst the blue form with little or no lilac or purple in its tint, 

 and with well-developed long silky white hairs was named candaon 

 by Bergsthisser. Jenner-Weir notes (Ent., ix.. p. 254) a g 

 of this species, in which one of the wings, compared with the 

 other three, is of a distinctly more lilac-colour. The narrowness 

 of the marginal border in the J s of this Bpecies is such, that 

 any development beyond the normal is quickly noticeable, and 



