Careful study of the imaginal discs of the wings in a caterpillar (fig. 10) made by examining microscopically sections cut through them, shows that the epidermis is pushed in to form a little pouch (C, p) and that into this grows the actual wing-rudiment.  Consequently the whitish disk which seems to lie within the body-wall of the larva, is really a double fold of the epidermis, the outer fold forming the pouch, the inner the actual wing-bud.  Into the cavity of the latter pass branches from the air-tube system.  In its earliest stage, the wing-bud is simply an ingrowing mass of cells (fig. 10 A) which subsequently becomes an inpushed pouch (B).  Until the last stage of larval life the wing-bud remains hidden in its pouch, and no cuticle is formed over it.  When the pupal stage draws near the bud grows out of its sheath, and projecting from the general surface of the epidermis becomes covered with cuticle to be revealed, as we have seen, after the last larval moult, as the pupal wing.  Thus all through the life of the humble, crawling caterpillar, ‘it doth not yet appear what it shall be,’ but there are being prepared, hidden and unseen, the wondrous organs of flight, which in due time will equip the insect for the glorious aerial existence that awaits it.

[Illustration:  Fig. 10.  A, B, C, Sections through epidermis and cuticle, showing three stages in growth of the imaginal disc (w) of a wing in the caterpillar of a White Butterfly (Pieris). ep, epidermis; cu, cuticle; t, air-tube, whence branches pass into the developing wing.  In C, cu’ represents the new cuticle forming beneath the old one, and (p) the pouch within which the wing-disc (w) lies.  Highly magnified.  After Gonin, Bull.  Soc.  Vaud. XXX.]

As mentioned above, this hidden growth of the wing-rudiments, in butterflies, beetles, flies, bees, and the great majority of the winged insects, has been emphasised by Sharp (1899) as a character contrasting markedly with the outward and visible growth of the wing-rudiments in such insects as cockroaches, bugs, and dragon-flies.  The divergence between the two modes of development is certainly very striking, and a conceivable method of transition from the one to the other is not easy to explain.  Sharp has expressed the divergence by the terms Endopterygota, applied to all the orders of insects with hidden wing-rudiments (the ‘Metabola’ or ‘Holometabola’ of most classifications) and Exopterygota, including all those insects whose wing-rudiments are visible throughout growth (’Hemimetabola’ and ’Ametabola’).  Those curious lowly insects, belonging to the two orders of the Collembola and Thysanura, none of whose members ever develop wings at all, form a third sub-class, the Apterygota (see Classificatory Table, p. 122).