MANDUCA ATROPOS. 417 



segments (ist to 8th, both inclusive). Over an area rigidly limited 

 by the inferior edge of the lower part of an oblique border in each 

 segment, the shagreen-dots form the centres of relatively large 

 circular patches of purple. These patches are larger upon the borders 

 than elsewhere, and are especially large upon the superior posterior 

 part of each border. The patches are especially small and scarce 

 posterior to the upper part of the sixth border on the 7th abdominal 

 segment, while upon the 8th abdominal segment there are only a very 

 few faint patches, all of which are placed on the upper part of the 

 seventh border, which traverses this segment. The purple patches 

 have evidently spread from the bases of the dots, which are still 

 visible in the centre of each as a small light spot (which, however, is 

 not raised above the general surface of the larva). These dots have 

 entirely lost the hairs in this last stage, but their former presence is 

 probably indicated by a minute central scar-like point. The patches 

 are also seen to be connected with dots because of their arrangement 

 upon the eight annuli, into which each abdominal segment is divided 

 (except the first, which is only divided into six or seven annuli). The 

 occurrence ot the dots upon the coloured borders is quite exceptional, 

 but then, in this larva, the former become the centres of patches of the 

 same colour as the latter, although it is noteworthy that the patches 

 are always purple, while the upper parts of the stripes become blue. 

 It may be that the patches point to the origin of the borders upon 

 which they persist, but, if so, such a development of this marking is 

 most unusual, for, in other forms, the borders arise — first by the absence 

 of dots, secondly by the darkening of the ground colour, and finally 

 by its replacement by a different tint. On the other parts of the larva, the 

 shagreen-dots are more normal when they are present. Thus, upon the 

 underside, there are extremely minute hairs borne by exceedingly 

 small traces of dots, exactly as in the larva of S. ligustri in the last 

 stage. Again, the caudal horn is well known to be rough, and this 

 condition is caused by the persistence of true shagreen-dots upon it, 

 many of which bear simple hairs. Comparing the larva with that of 

 S. ligustri in the same stage, M. atropos exhibits phyletically older 

 features in the persistence of normal shagreen-dots upon the horn, 

 and of (much modified) dots upon the dorsal surface, and in the 

 persistence of the "eighth stripe" (border). On the other hand, it is 

 phyletically younger in the extreme development of the coloured bor- 

 ders, and, probably, in the fading away of the oblique stripes (Poulton, 

 Trans. Ent. Soc. London, 1886, pp. 143-148). 



Development of larval markings. — First instar : Light 

 yellow-green, with a long straight black horn, half as long as the larva 

 itself. Second instar* : The striped pattern of the later stages becomes 

 visible, the oblique stripes whitish. Third instar: Yellow or yellow- 

 green with blue markings ; the long horn now yellow at base. Fourth 

 instar: Darker yellow ; with darker oblique stripes ; the head yellow- 

 green bordered with black ; the horn yellow, roughly granulated, 

 somewhat curved downwards ; the spiracles black. Fifth instar: This 



* Mathew notes {Entom., xxxi., p. 116) that, on November 10th, 1897, 

 at Marmance, on the coast of Syria, a sharp frost having occurred the previous nignt, 

 he found a fullgrown larva of M. atropos and also a small one, only a few days old, 

 the latter bright green with pale greenish-yellow stripes and a long slender and almost 

 straight horn, 



D I 



