FURTHER NOTES ON LIBYTHEA CELTIS. 193 



In the 2nd instar the hairs are much more numerous, the primary 

 that were present in the first skin are still to be distinguished as longer 

 and larger (about 0"16mm. long)]; between dorsum and spiracle are, 

 in addition to the three primary, 8 to 10 shorter hairs (aljout 0-08mm. 

 long, except one just above spiracle very minute, about 0-015mm. 

 long), all these have the same structure. The prolegs have 17 or 18 

 hooks, of which 4 or 5 are spread on the outer margin, the rest (in two 

 sizes) packed along the inner (pi. xiii., fig. 3). My specimens of this 

 instar are not clear enough to apportion the hairs to the several 

 subsegments. 



In the 3rd instar the hairs are more numerous, 15 or 16 secondary 

 between dorsum and spiracle (pi. xiv, fig. 4). This photograph 

 shows the 4 subsegments very distinctly and how the hairs are 

 arranged in a central line across each subsegment. The division of 

 each segment into four subsegments is distinct throughout all the larvl 

 stages. 



In the 3rd instar the prolegs have about 19 hooks alternately larger 

 and smaller, they are of a full horse-shoe form, i.e., a complete 

 circle, but for a small gap at inner margin on the posterior sets, the 

 first pair (on 3rd abdominal segment), however, show a complete circle 

 with about 5 smaller hooks only along front margin. In the 4th 

 instar the hooks are sparse and small on one margin, but a complete 

 circle would be a correct description of most specimens. 



In the 4th instar (pi. xiv., fig. 5), the hairs are very numerous and 

 are so spread as to be irregularly placed across each subsegment, 

 perhaps about 70 or 80 to the half segment above the spiracle, varying 

 in length from 0-3mm. downwards, the prolegs have about 28 to 30 

 hooks without a very decided break in the circle except on the claspers, 

 where the outer portion is bare for about a fourth of the circle. There 

 are very numerous fine sharp skin-points, these are aggregated, with 

 partial clear spaces. The places they occupy are dark on the photo- 

 graph, looking almost like plates across the subsegments. 



In the 5th (last) instar (pi. xv., fig. 6) the hairs are very numerous, 

 especially the smaller that are tolerabljr evenly distributed, even 

 between the subsegments. These minute hairs seem to be pointed, 

 not coronate like the larger ones. The larger ones seem to be pro- 

 portionally fewer than in the preceding instar, as though some of the 

 smaller of that instar had joined the swarm of short ones in this. The 

 longest hairs are about 0-4mm. long. The patches of sharp skin-points 

 are less extensive than in the preceding skin, and show dark in the 

 photograph. The prolegs in this instar have about 50 hooks closely 

 packed in their alternating sizes on the inner side, but with only 

 about seven on the outer margin, but these so spaced as to make it 

 difficult to say there is an actual break in the circle, or where it is if 

 there is one (pi. xvi., figs. 9 and 10). 



Plate xvii. shows the larval heads from cast skins in all the instars 

 except the 4th, of which I do not appear to have preserved a present- 

 able specimen. That of the last instar is somewhat damaged. They 

 are fairly regular in their increase in size from stage to stage, the 

 ratio being apparently an addition of 50%, or as 2-3. The actual 

 widths would be in millimetres — 1st, •4mm.; 2nd, -Gmm.; 3rd, •9mm.; 

 4th, 1^35mm. ; 5th, 2mm. 



The full grown larva and pupa are beautifully shown in the 



