MANDUCA ATROPOS. 415 



which just crosses the intersegmental furrow in both species. [The 

 resting attitude of both is also exactly similar, the head being 

 retracted and the dorsal surface of the swollen thoracic segments 

 curved into a quarter of a circle, as seen in profile. This is seen in 

 the Sphinx attitude of rest, and in other positions also.] The difference 

 between the ground colour of the thoracic segments and that of the 

 other segments is very obvious in both cases. The difference is, perhaps, 

 best described by saying that the former are yellower and paler, and 

 especially, that the colour possesses a peculiar transparency which is 

 absent from the rest of the body. The thoracic segments are, 

 however, less bright and yellow than the ground colour in the region 

 of the oblique stripes. [In a figure of the brown variety of M. atropos, 

 painted by Mrs. Owen Wilson and sent to me by Professor R. Meldola, 

 the ground colour of the thoracic segments is pink while that of the 

 other segments is brown, and the contrast is therefore much greater 

 than in the common variety. Air. Stainton, in his Manual, describes the 

 anterior segments of this variety as whitish, and the rest of the body as 

 brownish-olive. I expect that this is more correct than the impression 

 conveyed by the figure, although the specimen which Mrs. Wilson has 

 painted may have been an unusual variety.*] The black thoracic 

 spiracle of M. atropos resembles the ochreous anterior spiracle of 

 »S. ligustri and differs from that of Smerinthus ocellatus, &c, in being 

 unconcealed when the larva is at rest in the Sphinx attitude. The 

 relations of the oblique stripes and borders with those of Sphinx ligushi 

 are extremely interesting. [In Stainton's Manual, the larva of 

 M. atropos is described " with 7 oblique lateral violet stripes " and that 

 of S. ligustri 1 '' with 7 oblique lateral white streaks bordered above with 

 lilac." Certainly this correlation of the markings of the two larvae 

 seems sufficiently obvious, and is the one which is also given in other 

 descriptive works. Recent investigations, however, have shown the 

 actual relations existing between the oblique stripes and their coloured 

 borders (by a comparison of the ontogenies of Smerinthus and Sphinx 

 larvae).] A careful examination of the position of the violet bands of the 

 larva of M. atropos has resulted in the proof that these markings do not 

 correspond with the stripes of S. ligustri, but with the coloured borders 

 only — the latter markings, in fact, have persisted, while the oblique 

 stripes have become inconspicuous. This correspondence is at once 

 seen on comparing the relative positions of the caudal horn and the 

 spiracles with the oblique markings of M. atropos and S. ligustri 'respec- 

 tively. [The fact is recognised by Weismann, on p. 322 of the English 

 translation of his work, where he speaks of the blue edges of the stripes 

 of M. atropos, and compares them with those of S. ligustri. j The stripes 

 are also present in M. atropos, but, being only of a lighter yellow than 

 the rest of the ground colour, they are easily passed over. They are 

 narrow in the lower half of their length, becoming broad above, so 

 that they occupy very nearly the whole interval between the coloured 

 borders in the dorsal region. This broadening takes place as each 

 stripe crosses the furrow which separates the two segments in which 

 its course lies. In .S. ligustri the ground colour is much brighter in 



* Chaumette describes (Zoologist, ix., p. 3242) the dark form as having " the 

 thoracic segments of a pinkish-white colour . . . and the abdominal segments 

 yellowish-brown." McLaren also notes (op. cit., p. 6788) the dark variety as having 

 " the thoracic segments white inclining to pink, the abdominal segments olive- 

 brown." The pinkish tint is, we believe, not at all unusual. 



