SPHINGIDES. 371 



and there is usually also a well-marked depression on the 8th 

 abdominal immediately behind the scar. Poulton thinks that this 

 must be due to the bending downwards of the horn, which becomes 

 quite horizontal before the larval skin is thrown off, so that the 

 posterior edge of its base and the continuous adjacent larval cuticle 

 are depressed, and leave a permanent impress upon the yielding 

 surface of the pupa. In Mimas tiliae the area of the scar is smooth, 

 although the rest of the skin is rough and corrugated. Poulton 

 figures some interesting instances of well-developed scars (Morph. 

 of Lep. Pupa, pi. xx., figs. 2, 4, 9, 10, 13, 14), and further notes 

 that the peculiar rough plate upon the dorsal surface of the anal 

 flap of the larva of M. tiliae is represented by the extremely 

 rough dorsal surface of the terminal spine of the pupa, and this 

 is a valuable aid to the identification of these two structures 

 (loc. cit., p. 192). Sharp notes (Insects, ii., p. 381) that "the 

 pupa is remarkable from the deep cleft that exists to admit air 

 to the first spiracle, and for a deep depression on each 

 side of the anterior part of abdominal segments 5 — 7." Considerable 

 change takes place in the dimensions of the newly-formed 

 Sphingid pupa, a great difference occurring in a few hours. 

 This has been worked out in considerable detail for Sphinx ligustri 

 (Ent. Rec, x., pp. 185 et sea.), and it has been found that a pupa, 

 measuring immediately on moulting 1*87 ins. in length, made up 

 of (1) '63m. from anterior extremity to end of wings, and (2) i'24in. 

 to posterior extremity, had become respectively in two hours I'lin. 

 and *78in., so that the change represented (1) '63 : i'i, and (2) 

 1*24 : -78. The most strikingly characteristic feature in the Sphingid 

 pupa, however, is the proboscis, and we have, within the superfamily, 

 two markedly opposite lines of evolution from an ancestor with a 

 functional tongue of moderate length : (1) In the direction of atrophy 

 as seen in the Amorphids. (2) In an excessive lengthening, as seen in 

 the Sesiids (Macroglossids) and Sphingids (sens, strict.). Chapman 

 writes : " One may study the evolution of the pupa from this stand- 

 point very usefully, and, so far as the excessive length of the proboscis 

 in some species has to be accommodated in the pupa, the keeling 

 in the Eumorphids is the initial stage of the prominent curved 

 pupal tongue horn of the Sphinges, in which the proboscis, near 

 its base, projects forwards, gradually forming a loop so that the 

 imaginal proboscis passes down the front of the horn (short or long, 

 curved, coiled or straight), bends under at its extremity and returns 

 along the posterior aspect of the horn and then continues the ordinary 

 course, between and to the extremity of the wings. In Sesia (Macro- 

 glossum) the fulness hardly amounts to a keel; in most Eumorphids there 

 is a very definite keel, and in Sphinx there is almost always a horn. 

 I have no material to enable, me to say how far the transition is 

 gradual or abrupt. In Manduca, we have a recession from a 

 condition probably similar to Eumorpha ; there is some basal fulness 

 and the pupal proboscis extends to the end of the wings, but the 

 imaginal proboscis falls short of the pupal in length. The change 

 in the line of the glazed eye is very curious. It probably originates 

 in the highly-developed proboscis, involving greater development 

 of the front of the head, and so pushing it backwards, and so far 

 as the eyes are concerned, involving a rotation backwards on a 



