238 BRITISH BUTTERFLIES. 



long hairs. Just above this and halfway (or nearly) to spiracles, skin- 

 points are rare or absent ; below it, is again a clear space followed by 

 a second (really third, the first being obsolete) flange with a good 

 many hairs. The subsegmentation of the abdominal segments is 

 proportionally 4:1:1:1 in width ; on the 1st subsegment, skin- 

 points are very numerous and irregularly placed ; on the 2nd, they 

 are in two incomplete rows; on the 3rd and 4th, they are in one 

 row (at least dorsally). The head is 2-4mm. across, the 1st thoracic 

 l'6mm., and the broadest portion of the larva is 3'7mm. (from 2nd 

 to 6th abdominal segment). The head is black, really a very deep 

 brown, with golden (very short) hairs; these arise from ordinary hair- 

 bases which are situated at the bottom of cup-shaped hollows. The 

 whole head-surface is pitted with such hollows so closely set that they 

 interfere with each other so as not to be often circular, and frequently 

 want a portion of the margin. Except that the hairs arise from the 

 largest and most perfect of these, so that they seem to be the primary 

 feature of the surface- sculpturing of the head, it would be at least as 

 easy to describe the sculpturing as consisting of a network of fine 

 sharp ridges, with rounded hollows in the interspaces, the ridges being 

 of varying heights. The cups, in fact, form the ridges by intersecting 

 each other and not by being squeezed together. It is futile to describe 

 the mouth-parts without figures; the maxillag and their palpi are well- 

 developed, and the terminal bristle of the antenna is long (0*5mm.). 

 The protboracic plate has no median suture, but has a transverse line 

 rather behind the middle ; it carries a lenticle on either side near the 

 middle and one at either end ; they are small and dark, hardly bigger 

 than the hair-bases. The meso- and metathoracic segments have four 

 subsegments, of the mesothoracic the first is evanescent in the medio- 

 dorsal line, the second projecting forwards. This and the two following 

 are of width as 2 J: 1:1. The metathorax has the 1 st subsegment complete, 

 and the width of the four are as 1:1-5:1:1. The lenticles throughout 

 are small and (except on the thoracic plate) of nearly the same colour as 

 the skin, so that it is extremely difficult to be quite sure of them. There is 

 one on each thoracic segment, rather below spiracular level. On the 

 abdominal segments there is one a little above (and in front of) the 

 spiracle, and a smaller one rather further from it and directly below. 

 There are apparently two close together on the 2nd and 3rd thoracic seg- 

 ments, at level of the upper of these. On the 2nd, 3rd, 4th, 5th, 6th, 7th 

 (not on 1st and 8th) abdominals are two dorsal lenticles, at, approximately, 

 the position of tubercle i. The skin-hairs are very numerous dorsally ; 

 on an abdominal segment there are about 100 on a square of the 

 width of the segment. The larger are less numerous, about O'SSmm. 

 long, the shortest are about a fourth of this length ; it is impossible to 

 identify any of the longest as being those of primary tubercles. They 

 are all, apparently (large and small), bifid at the extremity, one or two 

 even trifid, but this may be an aberration. Many of the smaller widen 

 out beyond the middle, and terminate in three, four, or even five points, 

 suggesting a relationship to the trumpet-hairs of Nisoniades tages larva, 

 and of a Lycaenid pupa. Some of the larger hairs are not dissimilar, 

 but two other forms are commoner. One has the shaft widened and 

 flattened, and looks as if two hairs were combined by a thin membrane 

 joining them ; these then approach each other terminally, but before 

 meeting end in separate points, with often another between them, on 



