440 BRITISH BUTTERFLIES. 



tessellation. Laterally, there are also longer hairs, but not so long as 

 on the back ; all the hairs have a slight curve, and, though some are 

 spiculated all round, the majority are so only along the convex 

 margin ; the spicule are short and blunt, more descriptively they are 

 rather notches than spicules. In the neighbourhood of the spiracles 

 are some leniicles, which show remarkably how closely related lenticles 

 are to hairs, as the margins of the lenticular rings are produced into 

 six or eight points, closely resembling the side- spines of the hair-bases. 

 The membrane in their lumina is finely dotted. The width of a 

 lenticle is about O'lmm., more or less, according to whether spicules 

 be measured in or no. Each of the two pads of the prolegs has seven 

 or eight pale ochreous hooks, which vary much in size, some being 

 twice as large as others, but they are not definitely in two rows. Head 

 smooth, black. In this third (penultimate) instar, the hairs and 

 lenticles are, at least, as abundant as in the second; 40 to 45 hairs 

 may be counted between the dorsal set and the spiracle on the 

 abdominal segments. There is still a double dorsal crest of hairs, of 

 which the longest is about 0*3mm., much the same length as in the 

 second instar, and which may be regarded as the seta of i, but, as 

 several other hairs run it closely, the group, rather than any one hair, 

 more probably represents tubercle i. Three of the lateral hairs are 

 nearly as long. In this instar all, but especially the smaller, hairs 

 show very strongly the peculiar character that was present in some 

 degree in the second instar, and that seems to be quite wanting in the 

 last, viz., what I have called the hacked scimitar form ; the bases still 

 have the stellate form, the trunk and branches, though hardly longer than 

 in the second instar, are decidedly thicker (pi. xxvi., fig. 1). The branches 

 are nearly (here and there quite) as long as the hairs, but contrast by 

 being smooth, cylindrical, and rounded at the ends ; the scimitar 

 hairs, curved so as soon to be nearly parallel with the skin, instead of 

 at right angles to its surface, begin with a narrow neck, where much 

 of the curvature is, then broaden out, sometimes to twice the width, 

 and then narrow to a sharp point ; the most striking peculiarity, 

 however, is that the spicules are large and bold along the convex 

 margin, making it something like a saw, and quite (or all but) wanting 

 elsewhere, and the hair consequently looks (or perhaps is) somewhat 

 flat and thin laterally, looking, in fact, just like a scimitar with the 

 edge hacked (a la FalstafT). These hairs are very numerous, and there 

 are intermediates between them and the more ordinary form, which are 

 straighter, more club-like, and with more regularly distributed spicules. 

 These hairs vary in length from 0*1 mm. to 0-1 5mm. The longest 

 hairs, however, even those of O'Smni., have something of the scimitar 

 character. The lenticles are nearly as abundant as in the previous 

 skin ; they often have three or four stellate processes nearly their own 

 width in length, and are never quite without them. The honey-gland 

 of the 7th abdominal segment is a transverse mark about 0*3mni. long, 

 without hairs, etc., but with a special line of lenticles along its posterior 

 border ; the hairs close by are short, some very short, down to O'OSmm. 

 and 0*05inm., and, in two or three cases, are hair-bases with rounded 

 tops, but no hairs, difficult to class as hairs or lenticles. The bundle 

 of fine, spiculated hairs, terminating the eversible glands of the 8th 

 abdominal segment, may be seen at the site of this gland in a cast 

 skin. The hairs of the under- surface are longer, straighter, and 



