SPHINGIN^E. 271 



of Sphinx ligustri and the green (yellow) form of the larva of Mandttca 

 atropos, and Weismann's observation {Studies in Theory of Descent, 

 transl. p. 145) that the violet bands of the latter correspond with 

 the coloured borders of the stripes of the former, and not 

 with the stripes themselves, these, in M. atropos, having become 

 inconspicuous. As a summary of the development of the oblique 

 stripes of Sphingid larvae, and their borders, Poulton gives (loc. cit., 

 1886, p. 148) the following: 



1. A hairy* larva with greenish or yellowish ground colour, the hairs springing 

 from light (white or yellow) tubercles. 



2. The hairs become inconspicuous, and the tubercles more distinct. 



3. The tubercles become especially enlarged and approximated along the 

 lines of the oblique stripes, thus forming the first indication of this system of marking-. 



4. The light colour spreads from the base of each tubercle, and the adjacent 

 areas coalesce, forming a continuous stripe. 



5. The tubercles disappear from the ground-colour along the anterior edge 

 of each stripe, thus producing a relatively dark border. 



6. The border is rendered more distinct by a deepening in the tint of the 

 ground colour. 



7. The border becomes a conspicuous feature in the oblique line system, gaining 

 a distinct and generally bright tint, a modification or replacement of the ground-colour. 



8. The original oblique stripes fade away until they are hardly recognisable, 

 while the bright borders become highly developed, and almost entirely represent 

 the whole system. 



Poulton remarks {loc. cit., p. 149) that " this history may be traced 

 up to the end of the 7th stage in the ontogeny of Sphinx ligustri, up to the 

 end of the 6th in Snierinthus ocellata, whilst the larva of Manduca atropos 

 supplies the 8th stage," and he adds that, although these stages occur 

 successively, it must not be overlooked, as Weismann had long before 

 pointed out, that certain stages of the ontogeny may be dropped, e.g., 

 the 7th may follow the 4th in some species, and so on. 



Chapman observes that among the imaginal characters of the 

 Sphingides is one that occurs elsewhere, but nowhere else, so far 

 as he has ascertained, affects the whole superfamily, or even 

 a whole subfamily. This is the curious development along the 

 posterior margins of the abdominal segments of a row (or rows) 

 of scales, modified into fusiform batons, with sharp points, usually 

 of a black colour, at any rate very rarely not darker than the 

 general abdominal scaling. In some cases, these are very solid 

 strong black structures with little evidence as to their being modified 

 scales, in others they present definite traces of striation, not dis- 

 similar from ordinary scales. When removed they leave, on the chitinous 

 plate, a row of pits, almost identical with those from which the scales 

 beside them have been removed, but of very much larger size. In 

 a few Sphinges, these batons are developed into a row of strong 

 spines, solidly and stiffly attached to the margin of the segment, 

 sharp and long enough to prick the fingers if properly applied 

 ( Pseudosphinx tetrio, &c). In others they are very numerous in three 

 or four rows but rather shorter and weaker than the scales around 

 them ( Hyloicus pinastri). They are well developed in Sesia and in the 



* " Hairy " must be understood as referring to the hairs in the 1st larval 

 instar of Amorphids, so that Poulton's 1st stage is the 1st in Amorphids, but 2nd 

 in Sphingids {sens, restr.). The shagreening proper does not even occur in the 1st 

 instar of Amorphids, for, although the hairs are present, they cannot be properly 

 termed shagreened as neither mammillae nor pigment-spots are present, so that there 

 would appear to be a stage with secondary hairs preceding Poulton's stage I, i.e., 

 with secondary hairs not springing from light-coloured mammillae (Bacot). 



