CELASTEINA ARGIOLUS. 451 



fcurve, getting sharper as it proceeds, passes over the highest point, and 

 reaches the attached surface by finally bending slightly forwards, the 

 extreme posterior end of the pupa (apparently the middle of the 9th 

 abdominal segment) being about 08mm. from attached surface and 

 0-2mm. beyond it. Viewed from above, the wing-spines project 

 obviously as slight rounded elevations ; the pupa is 2*4mm. wide across 

 them, l-4mm. from the front. In front of this the head and pro- 

 thorax form a little more than a semicircle, being quite rounded ; 

 thence, the sides gradually diverge (along the wings) to 5 -6mm. from 

 front to the widest part of the pupa (nearly 4mm.) at 3rd-4th abdominal 

 incision ; thence again, the lines form a continuous elliptic curve 

 round the posterior extremity. The markings are a dorsal series, a 

 series halfway thence to spiracle, and another a little nearer spiracle, 

 again just below spiracle, and another series a little more ventral. 

 There is usually a dark mark at the end of the 1st abdominal segment; 

 it belongs to the subdorsal series, but is present when these are wanting 

 in next segment or two ; there are others round the wing-spines. 

 These marks vary much in different pupae, both in size and intensity, 

 but are not, as in many Lycamids, obviously aggregations of dots. 

 The wings are more or less mottled with dark, the neuration often 

 showing well as paler lines. "Poulton's line" is not well marked, but 

 the margin beyond it slopes down to the surface of the 4th abdominal 

 segment. The prothorax is about 1mm. long down the central suture, 

 and from this to the pointed outer angle measures about 2mm. ; the 

 borders are parallel for half this distance, then they approach the 

 posterior the more ; it has also a curved hollow opposite the spiracle- 

 cover of the mesothorax ; it carries 60 or 70 hairs on either side, with 

 some lenticles ; attached to its anterior border, after dehiscence, as 

 before, is the dorsal head-piece. Either half of the latter is less than 

 lmm. transversely, and about 0*2mm. from back to front at its broadest, 

 and about 01mm. in the middle line ; it has the ordinary lines of 

 netting somewhat faintly marked, but no hairs or lenticles. The 

 mesothorax is about 3'5mm. long in the middle line (suture), about 

 2-Omm. at the base of the wings, where the metathorax, as usual, arches 

 up into it, the latter being about l-3mm. long here, and about 0'25mm. 

 in the middle line. The mesothorax has perhaps 140 to 150 hairs on 

 either side, the metathorax about 35. In each case (in the pupa ex- 

 amined), some hairs have been broken off, and, without a minute study 

 in each case, it is difficult to say whether the ring left is a hair-base 

 or a lenticle. In each case, the hairs, etc., cease where the wings 

 begin, these having only reticulations. The cover of the 1st spiracle 

 is in the wing area, sprite its own length from the ordinary surface area 

 (with hairs, etc.) ; this spiracle-cover is about 0-5mm. long and 01mm. 

 broad (from back to front) ; it is raised so as to be convex over its 

 whole surface ; the latter seems to be a pavement of polygonal cells, 

 about 2000 or so in number, where, however, a view of the margin is 

 obtained, each of these cells is seen to be the expanded top of an 

 upright pillar, a structure identical with that observed in Lampides 

 boeticus, except that here they all appear to be welded together in the 

 pupa of C. aryiolus, whilst, in that of L. boeticus, each little glassy pillar 

 with its slightly cupped, nearly flat, top, appears to be unattached to its 

 neighbours (see p. 355, pi. xxii., fig. 2). The 1st abdominal segment is 

 about 3mm. across (side to side), 0'4mm. in length at the dorsal line. In 



