METAMORPHOSIS IN LEPIDOPTERA.. 27 



the brown base, a pale green diaphragm, exactly like an eyelid, and 

 then rose again. It occurred once every three seconds, the opening 

 being now the shortest part of the sequence, and lasting half a second. 

 This continued till the spiracular openings became too narrow to see 

 the movements distinctly. As the observation was made on the 

 naturally suspended pupa, the descent of the diaphragm was from the 

 posterior margin of the spiracle, the final narrow slit being parallel 

 with the lines of incision, or nearly so. Chapman thinks that these 

 were valvular arrangements, associated with the pumping process 

 necessary to the expansion of the appendages. At the same time, he 

 suggests that the movements might be merely a development occurring 

 for the occlusion, more or less complete, of the spiracles during the 

 pupal state. Further observations were made by him as to the details 

 of this striking phenomenon on other newly-formed pupae — Sphinx 

 liyustri, Phalera bucephala, &c, and published Ent. Record, &c, x., 

 pp. 185 et seq. In Sphinx liyustri, the abdominal spiracles were at first 

 widely open, and the winking movement of a membranous fold from 

 the posterior wall of the trachea a short distance within was easily 

 seen. There are no dark chitinous parts in connection with these 

 valves in this species. The movements continued as long as observa- 

 tion could be made, although the period of closure became longer and 

 longer. The movements of the valves were synchronous with the 

 vermicular movements, but not with the pulsations of the dorsal 

 vessel. During actual moulting the valves were quite closed. In 

 Phalera bucephala, the abdominal spiracles (on segments 2-7) are wide 

 oval openings immediately on moulting, and it is easy to look into the 

 spiracular trachea as far as it joins the longitudinal trunk and gives 

 off a bunch of branches. At a depth from the outer opening, equal to 

 about its widest diameter, is a narrowing or second opening, marked 

 by a narrow semilunar brown chitinous piece running across the 

 anterior surface of the tube, with its free end just standing out into 

 the lumen of the tube. Opposite this, posteriorly, one sees nothing at 

 first, but, at intervals, a thin fold from the posterior surface passes 

 forwards and meets this chitinous margin, closing the tube. At first 

 the closing lasts only a second, and it remains open for five or six 

 seconds, but gradually the intervals become more equal, and before the 

 observation ceases the period of closure is rather the longer. The 

 observation ceases because, in about ten minutes, the anterior and 

 posterior lips of the external opening, or spiracle proper, become by 

 gradual approximation, very close together, so close that, knowing 

 what is going on, one can still see that it is continuing, but still so 

 little, that, without that knowledge, little or nothing could be made of 

 it, and, in from fifteen to twenty minutes, even this amount of view is 

 shut out, and the outer spiracle is quite closed. In another 24 hours 

 or so, this sinks into the bottom of a depression, such that, except in 

 a very good light, one cannot see it at all, and might fancy the spiracle 

 was simply a wide patent opening. 



We have already shown, in our account of the moulting of Phalera 

 bucephala, that it takes a considerable time for the various pupal organs 

 and structures to assume their proper proportions after having got rid 

 of the larval exuviae. Chapman has worked out, in considerable det iil, 

 the differences existing between the proportion of the various parts in 

 newly-formed and matured pupa-, e.y., a pupa of Sphinx liyustri 



