FIG. 225.—Portion of a longitudinal section through a pupal wing about eight days before emergence: _s_, formative scale-cell; upper _s_, a scale. ] More recently Schaeffer has stated that the scales and also the hairs are evaginations of greatly enlarged hypodermis cells, and still more complete evidence has been afforded by A. G. Mayer (1896). In the wings of Lepidoptera, about three weeks before the imago emerges, certain of the hypodermis cells, which occur at regular intervals, begin to increase in size and to project slightly above the level of the hypodermis; these are Semper’s “formative cells,” and are destined to secrete the scales. They increase in length, and appear as in Fig. 223. In the next stage observed, the projections are much longer (Fig. 224). The hypodermis is now thrown up into a regular series of ridges, which run across the wing. Each ridge, says Mayer, corresponds in position with a row of formative cells, and each furrow with the interval between two adjacent rows. The scales always project from the tops of these ridges. The ground or basal membrane has not participated in this folding, and the deep processes of the hypodermis (_prc_) that once extended to this membrane have largely disappeared. Figure 225 represents a more advanced stage almost eight days before the emergence of the imago. The scales are originally filled with protoplasm, which gradually withdraws, leaving behind it little chitinous bars or pillars which serve to bind together the upper and lower surfaces of the scales, and finally the scales become “merely little flattened hollow sacs containing only air.” As Mayer shows (Figs. 226, 227), from the study of scales examined four days before emergence of the butterfly (Danais), “the striations upon the upper surface of the scale are due to a series of parallel longitudinal ridges,” while the under side is usually smooth. The mode of insertion is seen in Fig. 227. The narrow cylindrical pedicel of the scale is merely, according to Semper, inserted into a minute close-fitting socket, which perforates the wing-membrane, and not into a tube, as Landois supposed. Spuler describes a sort of double sac structure or follicle (_Schuppenbalg)_ which receives the hollow pedicel of the scale. This was originally (1860) observed by F. J. Carl Mayer, but more fully examined by Spuler (Fig. 228) though not detected by A. G. Mayer. [Illustration: FIG. 226.—Portion of a cross-section through the pupal wing of _Danais plexippus_, about six days before emergence: _sg_, scale; _cta.al_, wing-membrane; _cl.frm_, formative cell of the scale; _mbr.pr_, ground-membrane; _fbr.h′drm_, hypodermal fibres of pupal wings. _A_, portion of a longitudinal section through the pupal wing, eight or nine days before emergence; _prc_, processes of young hypodermis scales.—This and Figs. 223–225 after Mayer. ]