374 BRITISH BUTTERFLIES. 



say whether some of the smallest are not hair- bases that have lost 

 their hairs by accident, but, in moat of them, it is possible to note the 

 dotted diaphragm that marks the lenticles. The lines or meshes of 

 network are fine but very distinct ; rosettes are scarce, and consist of 

 minute definite rings, no wider than the lines of the network, with a 

 surrounding area poorly chitinised, about as wide as a large lenticle 

 and with radiated and dotted structure. The cover of the prothoracic 

 spiracle is (Mniin. long, of the usual columnar structure, the tops of 

 the columns are, however, of various polygonal forms. The line 

 marking off the wing from the mesothorax is more than usually 

 definite, a change taking place in the intensity of the network, the 

 darker thoracic portion ending in a more than usually straight line of 

 the netting, reaching from just outside the spiracle-cover to the notch 

 above the metathorax ; there are 17 or 18 hairs chiefly against the 

 dorsal line, a very few lenticles, and a very few rosettes ; a variation 

 in the netting suggesting wing-spines occurs at the base of the wings, 

 viz., a forward portion of stronger darker netting, and, behind this, a patch 

 where it is evanescent, and, over this area, an extremely minute dotting 

 of the tissue fills the meshes (the pupal skin apart from net- work, etc.) 

 much more plainly than usual. The metathorax has 3 to 6 hairs, 

 7 or 8 large lenticles, and 15 to 20 small ones ; the angle forming 

 the portion of the hindwing is marked off by a band in which the net- 

 work is hardly visible. The 1st abdominal segment has no hairs, very 

 few lenticles and very weak netting. The 2nd has 15 or 16 dorsal 

 hairs and about a dozen near each spiracle, where is also a flight of 

 100 or so of various-sized lenticles. The following segments are much 

 the same, the hairs rather more numerous round the spiracles, and, on 

 the 6th abdominal, with some balloon- or clubbed-hairs. Rosettes 

 are rare except along a region a little way below the spiracles. There is a 

 special structure to be observed some little way above the spiracles, an 

 area from network lines, which, as they gradually appear round it, radiate 

 from it. Then, close above this, and apparently belonging to it, is a bit 

 of tissue attached to a point, at first glance looking just like a hair, but, 

 really, it is something drawn out on moulting (like the tracheal linings 

 are), as noted in other Lycaenid pupae. The whole of this structure 

 suggests that it follows on (i.e., is identical with) the "upholstered" 

 hollow that the larva possesses just here. On the dorsum of the 

 7th abdominal is a marked transverse line of scar, structureless, i.e., with- 

 out network and surroundings, evidently the trace of the honey-gland. 

 [In some pupae of the allied Agriadet coridon, this is a very conspicuous 

 feature to the naked eye.] In the cremastral region, fine skin-granu- 

 lation is conspicuous; there is no network, and a number of points, 

 not unlike small lenticles, are probably really obsolete cremastral hooks, 

 which are homologous with rosettes (not with hairs or lenticles); 

 dorsally on the 9th abdominal segment are a few genuine lenticles ; 

 cremastrally are about a score of ordinary hairs. The head-piece has 

 about 30 hairs; the eye [inside (or rather outside) glazed eye] 3 or 4. 

 The two mandibles meet for about 0-2mm. beneath the small triangular 

 labrum; the chitinised portion of labium is very minute, and diamond- 

 shaped (exposed portion probably about 0*04 mm.). The maxillae 

 proceed for about 4mm., and then are covered by the antenna' ; they 

 proceed for about 3mm. beneath antenna 1 , and for 0*5mm. in the 

 special pocket between the 4th and 5th abdominal segments. In the 



