That wings may be of use to insects under water is proved by the very interesting case of _Polynema natans_,[46] which uses its wings for swimming. This, however, is a rare case, and it is possible that the principal use of the wings was, primordially, to enable the mature forms to pass from pond to pond, thus securing fresh habitats and avoiding in-and-in breeding. If this were so, the development of wings would gradually have been relegated to a late period of life; and by the tendency to the inheritance of characters at corresponding ages, which Mr. Darwin has pointed out,[47] the development of wings would have thus become associated with the maturity of the insect. Thus the late acquisition of wings in the Insecta generally seems to be itself an indication of their descent from a stock which was at one period, if not originally, aquatic, and which probably resembled the present larvæ of _Chloëon_ in form, but had thoracic as well as abdominal branchiæ. Finally, from the subject of metamorphosis we pass naturally to that most remarkable phenomenon which is known as the “Alternation of Generations:” for the first systematic view of which we are indebted to my eminent friend Prof Steenstrup.[48] I have always felt it very difficult to understand why any species should have been created in this double character; nor, so far as I am aware, has any explanation of the fact yet been attempted. Nevertheless insects offer, in their metamorphoses, a phenomenon not altogether dissimilar, and give a clue to the manner in which alternation of generations may have originated. The caterpillar owes its difference from the butterfly to the undeveloped state in which it leaves the egg; but its actual form is mainly due to the influence of the conditions under which it lives. If the caterpillar, instead of changing into one butterfly, produced several, we should have an instance of alternation of generations. Until lately, however, we knew of no such case among insects; each larva produced one imago, and that not by generation, but by development. It has long been known, indeed, that there are species in which certain individuals remain always apterous, while others acquire wings. Many entomologists, however, regard these abnormal individuals as perfect, though wingless insects; and therefore I shall found no argument upon these cases, although they appear to me deserving of more attention than they have yet received. Recently, however, Prof. Wagner[49] has discovered that, among certain small gnats, the larvæ do not directly produce in all cases perfect insects, but give birth to other larvæ, which undergo metamorphoses of the usual character, and eventually become gnats. His observations have been confirmed, as regards this main fact, by other naturalists; and Grimm has met with a species of _Chironomus_ in which the pupæ lay eggs.[50]