[Illustration: Fig. 173—A, section of part of wing showing the complex androconia of _Thanaos tages_, a Skipper butterfly. The turning over of the costal margin of the wing is in this case part of the arrangement. _a_, Upper covering-scales attached to the costal portion of the under surface of the wing; _b_, edge of costal margin of the wing; _c_, costal nervure with its scales; _d_, field of the wing next the costal nervure, bearing stunted scales; _e_, the androconia proper, or male scales; _f_, posterior covering scales; _g_, lumen of the costal nervure: B, a portion of the costal area flattened out and seen from above; lettering as before: C, section of androconium on the second nervure of _Argynnis paphia_. (After Aurivillius.)] Resuming our consideration of the DEVELOPMENT of the WINGS, we may remark that the history of the changes during the pupal state is still imperfect. By the changes of relative size of the thoracic segments the hind wing is brought to lie under the anterior one (_i.e._ between it and the body), so that in the newly formed pupa the arrangement is that shown in Fig. 174. The wings are two sacs filled with material surrounding peritracheal spaces in which run tracheae. The subsequent history of the tracheae is very obscure, and contrary opinions have been expressed as to their growth and disappearance. We have alluded to the fact that in some nervures tracheae are present, while in others they are absent; so that it is quite possible that {333}the histories of the formation of the nervures and of their relation to tracheae are different in various Lepidoptera. This conclusion is rendered more probable by the statement of Comstock and Needham,[197] that in some Insects the "peritracheal spaces" that mark out the position of the future nervures are destitute of tracheae. Gonin thinks the nervures are derived from the sheaths of the peritracheal spaces, and a review of all the facts suggests that the tracheae have only a secondary relation to the nervures, and that the view that a study of the pupal tracheae may be looked on as a study of the preliminary state of the nervures is not sufficiently exact. It is, however, probable that in Lepidoptera the pupal tracheae play an important though not a primary part in the formation of the nervures; possibly this may be by setting up changes in the cells near them by means of the air they supply. Semper long ago discovered hypodermal cylinders traversed by a string (Fig. 170, B), placed near the tracheae in the pupa.[198] It appears probable that the "wing-ribs" found in the nervures (Fig. 170, A _fr_ and B) are the final state of these cylinders, but the origin and import of the cylinders are still unknown.