The same purpose for which the cocoons above described serve, is answered in the case of numerous _Dipterous_ insects, by a humble and less artificial contrivance--the skin, namely, of the larva; which, as was before observed[550], is never cast, but, when the insect is about to enter into the pupa state, assumes a different form and colour; becomes of a thicker and more rigid texture; and defends the included pupa, which is separate from it, till its exclusion. In this case the mouth of the larva is constantly different from that of the perfect insect, or at least has not with it those relations as to number and kind of organs, which have been observed in the mouth of other larvæ compared with the insects that they produce. The animal, immediately after it is clothed with this skin, if it is opened, exhibits only a soft gelatinous pulp, in the surface of which the exterior organs of the adult insect cannot yet be detected. Nature requires more time for their elaboration, or at least for the appearance of their outline, and to consolidate them. This pulp first takes an oblong form (_Boule allongée_ Reaum.), and afterwards that of the insect it is destined to give birth to[551]. The skin of the larva also serves for a cocoon to the pupæ of male _Cocci_[552]. The grub of the genus _Anthrenus_, so destructive to our cabinets of natural objects[553], when it assumes the pupa does not quit its skin, but only splits it open longitudinally on the back, and when it becomes an imago makes its exit through the orifice[554]. Some Lepidopterous larvæ even (_Alucita pentadactyla_, _Callimorpha rosea_, &c.) assume the pupa state within their last skin[555].

When a larva has finished its cocoon,--which with some species, that proceed so earnestly as though they had not a moment to lose, is the work of a few hours, of others about two or three days,--after a certain interval it casts its last skin, which is usually suffered to remain in the cocoon (but which one moth, _Geometra lacertinaria_, ejects through an opening purposely left in its bottom), and the pupa makes its appearance[556]. This interval is exceedingly various. Most larvæ assume the pupa state within a few days after they have formed their cocoons; but some not for several weeks, or even months. The caterpillar of _Bombyx cæruleocephala_, according to Rösel, lies three weeks in the cocoon before this change is effected; those of many _Pupivora_ and _Diplolepariæ_ Latr., according to Reaumur, six months[557]; that of _Phalæna urticata_ nine months[558]; and that of _Cimbex lutea_, according to De Geer, sometimes eighteen months[559]. Brahm observes, that such larvæ of the double-brooded moth, _Hepialus Testudo_, as form their cocoons in autumn, do not become pupæ until the following spring; while those which form them in summer undergo this change in a few days[560]. From this fact it might be conjectured, that the degree of heat prevailing at the time the insect incloses itself determines the period of the pupa's appearance; but this supposition seems contradicted by what Reaumur observed of a brood of the larvæ of _Phalæna urticata_, just mentioned, which, though they formed themselves cocoons in September, did not become pupæ till the June following[561]. I am unable, therefore, to assign any plausible cause for these extraordinary variations. The difficulty of comprehending how animals before so voracious can live so long without food may be partly surmounted, by adverting to the circumstance of its having attained its full growth, and laid up a store of nutriment for the development of the perfect insect. It is consequently no more wonderful that it should not have need of any further supply without casting off its upper integument, than that it should not eat after having done so and become a pupa.

An Introduction to Entomology: Vol. III (of 4) — Part 118