To many persons it will seem impossible that such beautiful and detailed resemblances as those now described--and these are only samples of thousands that occur in all parts of the world--can have been brought about by the preservation of accidental useful variations. But this will not seem so surprising if we keep in mind the facts set forth in our earlier chapters--the rapid multiplication, the severe struggle for existence, and the constant variability of these and all other organisms. And, further, we must remember that these delicate adjustments are the result of a process which has been going on for millions of years, and that we now see the small percentage of successes among the myriads of failures. From the very first appearance of insects and their various kinds of enemies the need of protection arose, and was usually most easily met by modifications of colour. Hence, we may be sure that the earliest leaf-eating insects acquired a green colour as one of the necessities of their existence; and, as the species became modified and specialised, those feeding on particular species of plants would rapidly acquire the peculiar tints and markings best adapted to conceal them upon those plants. Then, every little variation that, once in a hundred years perhaps, led to the preservation of some larva which was thereby rather better concealed than its fellows, would form the starting-point of a further development, leading ultimately to that perfection of imitation in details which now astonishes us. The researches of Dr. Weismann illustrate this progressive adaptation. The very young larvae of several species are green or yellowish without any markings; they then, in subsequent moults, obtain certain markings, some of which are often lost again before the larva is fully grown. The early stages of those species which, like elephant hawk-moths (Chaerocampa), have the anterior segments elongated and retractile, with large eye-like spots to imitate the head of a vertebrate, are at first like those of non-retractile species, the anterior segments being as large as the rest. After the first moult they become smaller, comparatively; but it is only after the second moult that the ocelli begin to appear, and these are not fully defined till after the third moult. This progressive development of the individual--the ontogeny--gives us a clue to the ancestral development of the whole race--the phylogeny; and we are enabled to picture to ourselves the very slow and gradual steps by which the existing perfect adaptation has been brought about. In many larvae great variability still exists, and in some there are two or more distinctly-coloured forms--usually a dark and a light or a brown and a green form. The larva of the humming-bird hawk-moth (Macroglossa stellatarum) varies in this manner, and Dr. Weismann raised five varieties from a batch of eggs from one moth. It feeds on species of bedstraw (Galium verum and G. mollugo), and as the green forms are less