MANDUCA ATROPOS. 429 



them to the ordinary sculpture. On these segments are several 

 depressed spots. One of these is a little outside and anterior to 

 the proleg cicatrix ; fully developed, it is a raised ring with a central 

 pit. A similar smaller one may exist between it and the scar, and 

 two other very minute ones are associated with it closely in 

 a line extending outwards and backwards. Another is directly 

 in front of this at the anterior border of the segment, and 

 looks as if a group of the ordinary pits had been closely aggre- 

 gated into a depressed spot. Various other arrangements of 

 the sculpture seem to be peculiar to particular spots, but description 

 would be interminable and not by any means luminous. The scar 

 of the larval horn is like that of the prolegs, a smooth spot with 

 radiating lines. There is no hollow behind it, but, instead, are 

 usually two small tubercles, one on either side, and, occasionally, a third 

 behind them in the middle line. The anal spike, 3*5mm. long, 3mm. 

 wide and 2mm. deep, is very coarsely mammillated with rounded con- 

 volutions, and ends in a small bifid spine. Its undersurface has a 

 longitudinal groove. The anal scar is marked by a longitudinal 

 groove about 2 -5mm. long, with several fine parallel ridges. The 

 male tubercles are large and flat, quite symmetrical and well separate 

 from front margin of 9th segment, but encroach by about as much on 

 10th. On the 8th segment are obscure markings as of an obsolete paired 

 arrangement ; this is sometimes fairly distinct, in other specimens 

 hardly visible. The female area presents the ventral aspect of 

 8, 9 and 10 smoothed out, without trace of the segmental incisions, 

 but with two or three swellings marking the line of hind margin 

 of 8. The anterior two-thirds of 8 is marked by a central impressed 

 line, with deeper pore-like spots at each of its ends and usually 

 another smaller centrally. There is a good deal of variation, and either 

 of these three may be obsolete. In one specimen the posterior one is 

 advanced to the middle of the segment, and has two small mammillae 

 just like the male tubercles, but smaller ; in another the slit is confined 

 to the front margin of the segment and cannot be differentiated. 

 In one specimen the anterior margin of 9 is very distinct as a 

 fine line and comes forward and encloses the posterior pore. In 

 a small starveling specimen, 51mm. long (normal length 65mm. -70mm.), 

 the claspers are represented by large rounded bases on either side 

 of the front end of the anal scar, and there is the depression behind 

 the scar of the larval horn, so common in Sphingid pupae, but quite 

 absent in all normal pupae of M. atropos. The exaggerated larval scars are 

 due to weak development of the pupa, partly by larval characteristics 

 not being vigorously thrown off, but chiefly from want of promptness 

 and quickness in moulting the larval skin, which allows the soft 

 pupal skin to begin to harden within the larval skin as a mould, 

 instead of the larval skin being got rid of before any hardening 

 commences. Carried a little further, the same weakness results in 

 the larval skin being only partially cast, or even not cast at all. 

 Microscopic hairs occur around each spiracle to the number of 8 or 10, 

 but elsewhere are very scarce, not so numerous as if they represented 

 the ordinary tubercles only. The pits, as transparent objects, look 

 very like those fish scales that radiate from a centre, or crystallisations 

 that form a radiating arrangement ; the margins are sinuate, and to 

 the sinuations darker and lighter shades radiate from a central point 



