Here, then, we have a distinct case of alternation of generations, as characterized by Steenstrup. Probably other cases will be discovered in which insects undeniably in the larval state will be found fertile. Nay, it seems to me possible, if not probable, that some larvæ which do not now breed may, in the course of ages, acquire the power of doing so. If this idea is correct, it shows how the remarkable phenomenon, known as alternation of generations, may have originated. Summing up, then, the preceding argument, we find among insects various modes of development; from simple growth on the one hand, to well-marked instances of the so-called alternation of generation on the other. In the wingless species of Orthoptera there is little external difference, excepting in size, between the young larva and the perfect insect. The growth is gradual, and there is nothing which would, in ordinary language, be called a metamorphosis. In the majority of Orthoptera, though the presence of wings produces a marked difference between the larva and the imago, the habits are nearly the same throughout life, and consequently the action of external circumstances affects the larva in the same manner as it does the perfect insect. This is not the case with the Neuroptera. The larvæ do not live under the same conditions as the perfect insects: external forces accordingly affect them in a different manner; and we have seen that they pass through some changes which bear no reference to the form of the perfect insect: these changes, however, are for the most part very gradual. The caterpillars of Lepidoptera have even more extensive modifications to undergo; the mouth of the larva, for instance, being remarkably unlike that of the perfect insect. A change in this organ, however, could hardly take place while the insect was growing fast, and consequently feeding voraciously; nor, even if the change could be thus effected, would the mouth, in its intermediate stages, be in any way fitted for biting and chewing leaves. The same reasoning applies also to the digestive organs. Hence the caterpillar undergoes little, if any, change, except in size, and the metamorphosis is concentrated, so to say, into the last two moults. The changes then become so rapid and extensive, that the intermediate period is necessarily one of quiescence. In some exceptional cases, as in _Sitaris_ (_ante_, p. 30) we even find that, the conditions of life not being uniform throughout the larval period, the larva itself undergoes metamorphoses. Owing to the fact that the organs connected with the reproduction of the species come to maturity at a late period, larvæ are generally incapable of breeding. There are, however, some flies which have viviparous larvæ, and thus offer a typical case of alternation of generations.