62 BRITISH LEPIDOPTERA. 



of the pupa. Similarly for the other segments in the Incompletae, the 

 pupal coverings of the wings, legs, &c, separate more or less from each 

 other and retain their attachments to their own proper segments, so 

 that the empty pupa-case often affords more information as to the true 

 relation of its several parts than the living pupa does. On the other 

 hand, an empty obtect pupa shows at once that the 5th and 6th 

 abdominal segments were free in the living pupa, and no others ; but 

 an empty incomplete pupa-case leaves much doubt as to which were free 

 segments, because, at many places where no movement was allowed in 

 the pupa, movement has taken place in dehiscence, e.g., in the 

 Tortricid pupa the free abdominal segments are 4, 5, 6 (and 7 in male), 

 yet, on dehiscence, it would appear as if the wings were partially free 

 from abdominal segments 2 and 3, and, as if these segments, too, were 

 free. It may be here well to note that some obtect pupae have, in the 

 male, a more marked incision between the 7th and 8th abdominal seg- 

 ments than has the female, even looking, sometimes, as if the incision 

 admitted of movement. This is never the case, but it appears to prove 

 unquestionably an ancestry in which the 7th segment was " free " in 

 the male at a later period than in the female. 



The pupa-incompleta is remarkable for its ability to leave its 

 cocoon ; thus the pupa of Adscita statices usually emerges entirely from 

 its cocoon before the imago appears, and the pupa-case of the Eriocraniids 

 is also found at some distance from the empty cocoon. Few obtect 

 pupa? can leave their cocoons before the emergence of the moth — those 

 of Endromis versicolor and Choeroccunpa elpenor have, however, retained 

 this ancestral habit, and that of Macrothylacia rubi is said to move up 

 and down in its long cocoon. Scudder states that the pupae of the 

 Sphingid genus Macrosila emerge from the earth for the escape of the 

 imago, using for the purpose certain flanges in the spiracular region. 

 All pupae -incompletae, however, do this, and are armed with various 

 rows of hooks and spines to facilitate the process. The attachment of 

 the Pterophorid pupae makes them exceptions to this general rule, for 

 although pupae-incompletae in structure and dehiscence, they are, of 

 course, quite unable to change their position from that they have once 

 taken up. Chapman also notes that Epermenia has a pupa of incom- 

 plete structure, but does not emerge from its cocoon. Whilst discuss- 

 ing the matter of pupal movement, it may be well to call attention to 

 the fact, discovered by Chapman, that certain Gelechiad pupae can only 

 move the free segments antero-posteriorly, with "something of the 

 manner that belongs to the movement of the click-beetles." 



The pupae-mcompletae show more variety than the obtect pupae, and 

 their incomplete character suggests that they are of a lower form, a 

 suggestion comfirmed, Chapman says, by their close resemblance to 

 many Tipulid pupae, especially those of some gall-gnats, a resemblance 

 that appears to be one of relationship rather than accidental, from the 

 dipterous pupa having hindwings, although the imago has halteres, 

 and from the imago possessing scales of quite a lepidopterous character. 



With regard to the movement of which the butterfly pupa is capable, 

 we may notice that by means of the movable incisions 4-5, 5-6, 6-7 the 

 posterior segments of the Papilionid pupa can be moved in any direction, 

 whilst in the Pierids, by the development of certain dorsal tubercles at 

 the margins of the segments (possibly due to the fusion of the anterior 

 trapezoidal larval tubercles), antero-posterior movement is lost, only 



