METAMORPHOSIS IN LEPIDOPTERA. 25 



the silk with the cremastral hook-pad. He adds, however, that "the 

 rectal ligament plays a most important part, and in some species a 

 more important part than the membrane itself." Osborne criticises 

 (Ent. Mo. Mar/., xvi., p. 150) Eiley's view at length, so far as relates 

 to the influence of the rectal ligament, and, it must be owned, with 

 considerable success. He writes : — " In the chrysalis of A</lais nrticae, 

 the last three or four abdominal segments are wedge-shaped, the thin 

 ends of the wedges lying together on the venter near the knobs to 

 which the ligament is attached. The terminal or anal surface of the 

 last segment has, lying on it, the ridges which terminate in these 

 knobs and the cremaster (with the hooks), making an obtuse angle 

 with the ridges, and forming, with them, a sort of bent lever (viewed 

 sideways, not unlike the open jaws of a serpent, having on its nose the 

 hooks for attachment to the silk). The extension and contraction of 

 the abdominal segments, then, cause the cremaster to move through 

 the arc of a circle, whose centre is at the point of suspension of the 

 chrysalis from the ligament, which ligament itself is highly elastic. 

 With this mechanism it is not difficult to conceive how the tail of the 

 chrysalis may work its way out of the shrivelled larval skin, stretching 

 the elastic ligament as far as necessary, and then be thrust up around 

 that shrunk-up packet of old skin so as to reach the silk." 



Chapman makes (Ent. Record, x., pp. 185 etseq.) further important 

 notes on the moulting of Sphinx ligustri, Smerinthus ocellatus, and Phalera 

 bucephala. As an example of the pupation of an ordinary underground, 

 cocoon-making larva, that of the last-named species is interesting. 

 The larva pupates at from six to ten days after going doAvn, and when 

 the pupa within has so far freed itself as to occupy only the ten front 

 segments of the larval skin, the latter splits down each side, just above 

 the ventral prolegs, a short slit in each segment, but usually continued 

 over at least two segments ; immediately after, the skin splits in the 

 normal dorsal situation, followed by the division of the larval head 

 into the two lateral and clypeal portions, whilst a tracheal tube is 

 drawn out of the meso-metathoracic (as well as the pro-mesothoracic) 

 incision. The first spiracle is so deeply placed between the pro- and 

 mesothorax at the earliest moment that it is uncovered by the receding 

 skin, as to be seen with difficulty. The margins of this spiracle at the 

 surface are not apparently tinted or hardened, the coloured portions of 

 the newly formed pupa, besides minute hair-points, being the anal 

 spines and the dorsal margins of the incision between the 9th and 10th 

 abdominal segments, which have curiously opposed dentated borders, 

 whilst there is also a trace of colour at the sites of the ventral prolegs. 

 The wing- and appendage-cases are very short, and take about twenty 

 minutes to pass from the margin of the 3rd abdominal segment to 

 their permanent position at the margin of the 4th, but it takes some 

 hours for the solid portion of the pupa (head to 4th abdominal 

 segment) to grow from one-third of the total length to three-fifths of 

 that of the mature pupa, and this occurs, not only by its own increase 

 in length, but by the shrinking of the last six abdominal segments, 

 the total length of the pupa being remarkably constant throughout all 

 these changes. 



In another specimen of /'. bucephala, Chapman observed that, three 

 minutes after the vermicular movement of the larva was first noticed, 

 the tracheae were seen being drawn out of the 7th and 8th abdominal 



