24 BRITISH LEPIDOPTERA. 



ment is not a merely mechanical one (as if it were only slung by the 

 insertion of these projecting knobs in the pouches of the ligament), for 

 the old skin, when only held by the ligament, may be drawn towards 

 the head of the chrysalis, and still the adhesion is good. It is a curious 

 circumstance that this adhesion only takes place at pupation, and even 

 towards the end of it, and if the skin is drawn down earlier it will peel 

 off without any attachment to the horns, which will not, perhaps, have 

 been yet fully developed. What the exact nature of this attachment 

 is I am unable to say. The examination of this region in earlier 

 stages seems to yield the result that the black knob is the earlier 

 developed, and to hint at the hypothesis that the subsequent growth of 

 the white knob may include or nip in a portion of the lining coat of 

 the larval skin in such a way that it is afterwards held fast. . . . 

 I have, however, seen the ligament, after being drawn out, at first 

 cover the external white knobs and form a deflexed fold on each side, 

 and, then, suddenly slip off these white knobs with the disappearance of 

 the deflexed folds or their absorption into the main body of the liga- 

 ment which is still held fast and strained in the grooves between the 

 knobs." 



Chapman was the first observer to point out {Ent. 2Io. Mag., xv., 

 p. 78) that capillary attraction and atmospheric pressure, acting on 

 the damp surfaces of the inside of the larval skin and the outside of 

 the pupa whilst in contact, must have considerable influence on the 

 adherence exhibited between the two surfaces during the time that the 

 larval skin is being shed. Riley recognises this, in what he calls "the 

 natural adhesiveness of the moist, mucous, and membranous corium," 

 as a factor in the final exuviation of the larval skin and attachment of 

 the pupa, and characterises the other structures involved as : (1) The 

 "tracheal ligament," or the shed tracheae from the last pair of 

 spiracles. (2) The " rectal ligament," or shed intestinal canal. (3) 

 The " Osborne," or retaining membrane (membrana retinens), which 

 is but a stretched part of the membranous corium that accumulates 

 around the rectum and the anal prolegs, and that is intimately con- 

 nected with the rectal ligament. These three are essentially connected 

 with the larva and are cast off with its skin. The other structures involved, 

 but connected with the pupa, are : (1) The " cremaster," which is the 

 homologue of the anal plate of the larva, and surmounted at its apex, 

 and sometimes along the ventral ridges, by the " cremastral hook-pad," 

 thickly studded with minute but stout hooks. (2) The " sustainers " 

 (sustentores), two projections which homologise with the soles (plants;) 

 of the anal prolegs, and which take on various forms, but are always 

 directed forward so as easily to catch hold of the retaining membrane. 

 (3) The " sustentor ridges," usually connected with the sustainers, 

 embracing them on the outside and extending backwards to the inside 

 of the ventral cremastral ridges. These sustentor ridges are homo- 

 logous with the limb of the anal prolegs, and the exposed edge with 

 the posterior border of the said limb. (4) The " rectal piece," consist- 

 ing of a piece more or less well-marked, and elevated especially around 

 the closed rectum. Riley asserts that it is principally by the leverage 

 obtained by the hooking of the sustainers in the retaining membrane, 

 Avhich acts as a swinging fulcrum, that the pupa is prevented from 

 falling after the cremaster is withdrawn from the larval skin, and that 

 it is also principally by this same process that it is enabled to reach 



