THERETRA PORCELLUS. 105 



they are, however, clear and distinct, and the intermediate surface 

 is fairly smooth, in T. porcellus all of them have little grooves passing 

 across their posterior margins, irregularly, but, in the main, in the 

 direction of round the segment, and very few of them have anything 

 at all distinctly like a hair. The spines on abdominal segments 5, 6, and 

 7 are also somewhat different. In E. elpenor they are fairly smooth, with 

 fine wrinkles starting in all directions from their bases, in T. porcellus 

 only the tip of the spine is smooth, and the ridges run a long way up 

 the spine. If one can be got at an angle that obscures its curvature, 

 it looks like a volcano, with a smooth sharp apex, but with its sides 

 scored with gullies running out into the plain below. The way 

 in which these graduate into ordinary hair-points, and in which the 

 hair-points form nuclei or centres for the wrinkling, and, finally (at 

 other points), leave the wrinkling more or less undisturbed by them, 

 but with an intermixture of actual hairs, makes a very varied and 

 complicated graduation of structure most interesting to look at, but 

 quite impossible to deal with without elaborate figures (Chapman). 

 <? (three measured). i*2ins. — i'25ins. in length, width just over 

 •3m. Cylindrical, thickest at 4th abdominal segment ; tapering 

 gradually to either end ; the pupal envelope stout, surface rather 

 dull, due to wrinklings and corrugations of skin (although these are 

 themselves smooth and shiny when seen with a lens, the effect as a 

 whole is to produce a dull surface). Colour light wainscot or 

 flesh-brown, mottled and striped with dark brown, almost black 

 (one pupa much lighter than the two others) ; the wing-cases dark, 

 nervures pale ; antennae and maxillae dark, legs pale ; eyes and 

 headpiece dorsally dark and prominent ; the greater part of dorsal 

 area dark ; cremastral spike black. The head very prominent ; 

 maxillae large, projecting in a keel-like process at top, but tapering 

 to a narrow double sheath which is continued to end of wing-cases ; 

 antennae and legs short ; the great development of maxillae has 

 apparently forced the labrum and other usually ventral head-parts 

 into a dorsal position. The prothorax small ; the mesothorax large ; 

 the metathorax small; the 1st abdominal segment short, the 2nd 

 longer; the 3rd twice the length of the 1st, and the 4th at least four 

 times the length of 1st; the 5th. nearly as long, the 6th slightly 

 smaller; the 7th still narrower, and fused to the 8th. The cre- 

 master is broad, laterally very rough, convex on dorsal surface, concave 

 on ventral ; it narrows rapidly to a sharp polished spine curved in 

 a ventral direction. The spiracles are distinct, large, forming rather 

 long and narrow slits with the skin much wrinkled around them. 

 On the dorsal and lateral areas of the 5th, 6th and 7th abdominal 

 segments, at or near the middle of segment, but passing behind 

 the spiracle and curving towards the anterior edge near the median 

 line, is a low ridge of stout sharp conical spines, slightly curved 

 towards anal end of pupa ; in the middle of dorsal area where the 

 line curves towards the anterior margin of segment, the ridge is 

 somewhat broken up and the spines, which are scattered over a band, 

 are also smaller. [This character is well-developed in the pupa of 

 E. elpenor^ but the spines are more numerous and do not torm so 

 well-marked a ridge, but rather a band, and the spiracle is placed 

 directly in their line, forcing the spines, which are much smaller 

 hereabouts, out in a curve behind the spiracle.] A band along the 



