MANDUCA ATROPOS. 425 



disturbance ruptures it. He states that he wished to examine one of 

 these pupating larvae to see the conditions of the resting-stage, so broke 

 the earthen cocoon of one, and then put it together, as he thought, 

 very nicely ; a deformed pupa was, however, produced. Inchbald 

 says that the larvae bury several inches into the soil where they make 

 an earthen cocoon of minute granules of earth and sand with a 

 few tinv stones intermixed, and adds that there is no silken material 

 used. Morres notes that the puparia are so fragile, that, in 1897, 

 he had some 7 or 8 larvae go down in their several pots, that these 

 were placed on a table from which they had to be frequently moved 

 backwards and forwards, and that, on turning out the pupae, he 

 found them crushed and flattened out of shape, the walls of the 

 slender cells having evidently been shaken down on the newly-formed 

 pupae. In spite of the brittleness of the puparia, it has frequently 

 been found that, by allowing the earth in which they have been 

 formed to become very dry, the imagines have matured, and 

 then been unable to emerge (see Anderson, Ent., xi., p. 188). 

 Occasionally, as stated by Burrows above, the larvae pupate 

 without forming puparia. Instances of this kind are noted by 

 Reaumur ( Memoires, ii., p. 295), Anderson (Ent., xi., p. 188), Cotton 

 (Ent. Rec, xii., p. 275), &c. Hopson records (Ent. Rec, xii., p. 

 346) a larva going underground, coming up to the surface again, 

 remaining in a shrunken condition for 14 days, and then pupating, 

 the newly-formed pupa taking 3 days to assume its normal tint. 

 Pabst also asserts that pupation does not actually take place till 14- 

 16 days after the larva has constructed its cocoon. Sander says 

 (Kleine Schriften, i., p. 342) that some larvae pupate almost directly 

 after ceasing to feed, whilst others lie a long while, often 6-8 days. As 

 offering a connection between the puparium and pupal habits, Anderson 

 writes (Ent.,xi., p. 188): "Two larvae, in September, 1877, pupated on sur ~ 

 face and ten buried ; of the two pupae on surface both matured imagines 

 in November, 1877, of which one imago emerged and one died ; of the 

 others, one emerged before April and had died in the puparium owing 

 to the fine earth having got dry and hard ; two other pupae were then dead ; 

 the rest were taken from their puparia and laid on moss in greenhouse, 

 and all produced imagines between June 18th and July 5th, 1878. 



Pupa. — The newly-formed pupa* is of a bright yellow 

 colour, the posterior abdominal segments becoming pinkish- 

 red shortly afterwards, the whole skin gradually hardening and 

 deepening in colour till it reaches a logwood-brown, &c. (Graham). 

 This large pupa is of the rich deep brown colour common 

 to subterranean pupae. It has the S-curve that is associated with 

 antero-posterior flattening in the higher Sphingids (Amphonyx,&LC.) and 

 in the Sesiines. The curving here is chiefly caused by a hollow, 

 towards the end of the 1st legs in front, and to the projection of the 

 maxillary bases, all that is left of a once probably existent proboscis- 

 horn, the apex of the head being in front of the axial line, the 

 dorsal line sloping forwards at the prothorax. There is very 

 little forward curve in the abdominal segments. The hollow in front 

 of the legs is very reminiscent of the similar recession in Amphonyx, 



* Schroter says (Naturforscher, xxi., 66) that the pupa is at first wholly white, 

 that the abdomen first acquires its brown-red colour, the upper part becoming first yellow 

 and then finally brown, the full colouring only being acquired after about 24 hours. 



