454 BRITISH BUTTERFLIES. 



continuous in distribution with, those already referred to as forming 

 the micro-like rows. These rows of spicules have no apparent relation 

 with any larval structures. The full-grown larva of R. phlaeas has 

 spiculated hairs, but no skin-points, the skin-surface being divided 

 into a mesh of hexagonal cells by a fine network of lines. It seems 

 difficult to avoid looking for some relationship with some micro- 

 ancestor to account for them, and yet it is almost more difficult to 

 explain their survival, since they must have been useless for their 

 original functions for many ages. It is, however, no easier to 

 suggest any other origin for them, or to imagine what useful 

 functions they can now perform. To return to the cremastral area 

 and its hooks, I find it impossible to satisfy myself as to the limits 

 of the 9th and 10th abdominal segments. On the ventral line, the 7th 

 segment is clear enough, but the 8th is so contracted and fused with 

 the 9th, that even its limits are doubtful. Except on the ventral line, 

 the posterior margin of the 8th is definite enough. Within the circle 

 it encloses, to take the specimen of 0. dispar before us, and specimens 

 of II. phlaeas agree with it, we find first, in the dorsal half, an area 

 much like the rest of the pupa, with buttons, ribs, and trumpet-hairs, 

 but with a small central area smooth, except for some lines radiating 

 from its centre. This has all the appearance of a scar, not unlike that 

 of the horn in Sphingids, but whether of some injury or normal might 

 be doubtful, were it not that other specimens present a very similar 

 appearance. Turning to the ventral half of the area, we find it more 

 delicate and transparent, and divided across the middle by a suture, 

 which does not, however, reach either side. The whole of the area is 

 armed with the cremastral hooks, excepb a portion in the middle line, 

 slightly behind the suture noted, but chiefly between it and the front 

 of the segment. In the middle of this clear area are two projecting 

 points side by side, and, running forwards from between them, two fine 

 ridges with a groove between, ending in front by widening out into a 

 rounded lappet, with a surface of extremely fine spiculations. This 

 appears to be at a different level from the portion of segment that 

 seems to overlap it from either side and carries the hooks, and one might 

 suppose this to be the 9th and the hooks on the 10th, but those immedi- 

 ately behind the surface are continuous, without intervening suture. 

 Living pupa (changed June 9th, 1906) (rittilus): At first, pale greyish- 

 green, but soon develops darker markings. There persists, on the 2nd, 3rd, 

 4th, and 5th abdominal segments, a yellow-greenish series of oblique 

 marks, the "oblique" lines, which are hardly visible in the larva. 

 The pale coloration also persists beneath. The chief darker markings 

 at present (6 to 12 (?) hours after pupation) are a dark olive (thin black 

 over green) mediodorsal line the whole length of dorsum ; a number of 

 dark (black) spots, abundant on sides of mesothoracic dorsum (which 

 is pale between these and dorsal line), also abundant on metathorax, 

 but smaller and less numerous on prothorax. There is a dark shade 

 down the wing-base from prothorax to wing- spine, and thence down 

 the wing between nervures la and 1/;. The wing also has a dark shade 

 along the inner margin. The abdomen is pale on each side of the dorsal 

 line on the 1st and 2nd abdominal segments, the "oblique" line here 

 stretching inwards on the following segments 3. 4, 5, and indeed 6, 7, 8, 

 and 9; the area between the dorsal and oblique lines is olive, with a 

 faint rufous tint. Outside the oblique lines it is deep olive, with two 



