METAMORPHOSIS IN LEPIDOPTERA. 21 



They consist, so far as can be seen, in the development of a few 

 Avhite lines under the first and second, and dorsally on the terminal, 

 abdominal segments, as well as a few in front of the second segment. 

 These indicate places where slight folds have occurred in the effete 

 skin, and probably indicate that the operations of the past few hours 

 have been more important than appears, by making the larval skin 

 everywhere free from the pupal skin beneath, breaking down any 

 remaining cellular attachments, and leaving nothing between the old 

 and new skin but a little fluid. By this time, the head of the larva is 

 bent so as to be opposite the 6th abdominal segment, and not quite 

 touching it, so that only a little straightening has taken place. More 

 active (apparently) movements now occur, strictly vermicular in 

 character, commencing at the anal segment which contracts, and pass- 

 ing forward from segment to segment. After five or ten minutes the 

 larva is L-shaped, i.e., the thoracic segments are now horizontal, and 

 the head far removed from the front of the abdominal segments. It 

 is now obvious that the meso- and metathorax are enlarged, a change 

 that has already progressed during the first stage ; the first thoracic 

 is small throughout. Additional slight folds of effete skin occur across 

 the dorsum of the abdominal segments. The third stage, if not of 

 greater activity, at least of very much more rapid progress, is entered 

 upon. The skin of the abdominal dorsum slides backwards, and 

 accumulates at the anal extremity, ventrally, and the suspensory 

 processes of the pupa, which are very large and conspicuous in this 

 species, are easily seen through the delicate skin to hook into a fold of 

 the skin of the last segment, or rather into the fold formed by the 

 incision with the next segment, and the skin of the ventral surface 

 accumulates in front of this. It is also tolerably evident, though the 

 folds of larval skin are now thick enough to rather obscure it, that the 

 cremastral process is thrown backwards at each vermicular movement, 

 acting from the suspensory processes as centres of rotation, so as to 

 push the successive folds of dorsal skin, as formed, between it and the 

 suspensory processes. The succession, or rather concurrence, of events 

 is now very rapid ; one notes the white points formed at the spiracles by 

 the dragging out of tracheal linings, quickly extending into a con- 

 tinuous line, and one notes especially such a lining drawn out between 

 the meso- and metathoracic segments, where the imago has, but the 

 larva has not, a spiracle. Suddenly one sees that the larval skin is 

 ruptured down the centre of the thorax, and soon that the head is split 

 in half, with the clypeus in a separate piece, which follows the mouth- 

 parts. Quickly the skin passes towards the tail. The cremaster has 

 pushed all the dorsal skin below it and appears through the dorsal slit, 

 and, at the second push through, it seizes the pad of silk, whilst the 

 movements of the pupa then push the larval skin off the suspensory 

 processes, and, all weight being taken off it, and receiving some upward 

 thrusts, the crochets of the anal prolegs are also loosened from the pad 

 of silk and the skin is free and falls. This pupa, chiefly perhaps from 

 its large size, but also from the colouring of the suspensory processes, 

 renders the part the latter play in the casting of the larval skin very 

 easy to observe. . . . The casting of a tracheal lining from the 

 second thoracic spiracle at pupation, I do not remember to have 

 observed before, but inferred that such occurred, since I had seen it 

 many years ago in numerous larval moultings — Antheraea yamamai, 



