ON THE MOULT TO PUPA IN PTER0PH0RU8. 83 



the anterior group of hooks, and one cannot help a suspicion that 

 it reaches further and actually carries them ; in the mature pupa 

 or in the empty case it is not easy to be satisfied that this is not 

 the fact. At the period of the moult it was abundantly plain 

 that the anterior hooks belonged to segment eighth abdominal. 



I obtained some larvae of Pterophorus galactodactijlus, and was 

 fortunate enough to observe three individuals moulting to pupa. 

 This is a species that pupates beneath a leaf, and therefore often 

 inverted. The moult took place in all three examples about 

 1 p.m., after some two days' quiescence. The larva spins a 

 slight silken carpet, on which it rests, and which extends about 

 one-twentieth of an inch all round the larva beyond it, and is no 

 thicker at the critical position under the anal segments than 

 elsewhere. 



When the moult takes place the larva holds on by the anal 

 prolegs only ; the props of the ventral legs stand out stiffly, but 

 the hooks now take no hold of the silk ; how this happens I did 

 not ascertain. In many Pierids the body hangs arched away 

 from the silken pad, preventing the ventral prolegs from touch- 

 ing it ; but I do not know how this is managed in the Pierids, 

 Papilionids, and Lycfenids, where the prolegs touch the pad, aa 

 they certainly do in P. galactodactylus. The anal prolegs hold 

 well, so that they must be managed differently from the ventral 

 ones. It is less difficult to understand how all the prolegs take 

 no hold, as in pupse in cocoons, &c. 



The moulting is done rather rapidly — within, that is, about 

 fifteen minutes from the first efforts noticed. The rhythmic 

 movements, beginning at the last segments, at first push the 

 abdominal segments, about the second to sixth, forwards within 

 the larva skin, as evidenced by the tracheal threads very plainly 

 seen through the transparent skin being withdrawn from the 

 pupal spiracles. When these have passed forwards about one 

 segment, or rather when the larval skin has passed backwards 

 so far, all the abdominal trachefe are withdrawn more or less, 

 the first not quite a full segment's length. The thoracic seg- 

 ments are now crowded forwards, and the stretching of the 

 larva skin raises the front part of the larva from the surface in 

 a curve. Shortly it is seen that the anal segments of the pupa 

 still occupy those of the larva, and the skin, as it is pushed 

 backwards, gathers in a roll on the eighth abdominal segment. 

 The skin splits dorsally down the head and thoracic segments, 

 but not quite to the hind margin of the third one. The uncoil- 

 ing of the antennae from within the larval head is easily seen. 

 The crest of hairs on the pupal wing enables it to be seen 

 occupying its own segment in the larva, and rapidly expanding 

 as it assumes its pupal direction and position. When once the 

 skin splits it slips back rather rapidly, and one has to be alert 

 to see what is happening. It continues to form a roll round the 



