Colour-change in fishes, Amphibians, Reptiles, and Cephalopods, depends upon much more complex adaptations. In their case a reflex-mechanism is present which conducts the light-stimulus affecting the eye to the brain, and there excites certain nerves of the skin; these in their turn cause the movable cells of the skin which condition the colouring to change and rearrange themselves in the manner necessary to bring about the harmonization of colour. On this depends the colour-change of the famous chamæleon, and also the scarcely less striking case of the tree-frog, which is light green when it sits on trees, but dark brown when it is kept in the dark. All these are secondary reactions of the organism in which the external stimulus is, so to speak, made use of to liberate adaptive variations, either permanently or transitorily. In the caterpillars colour-changes are permanent, that is, it is only the young caterpillar which takes on the colour of its surroundings; later it does not change, even when it is exposed to different light, or intentionally placed upon a food-plant of a different colour. In fishes, frogs, and cuttlefishes, on the contrary, the reaction of the colour-cells to light only lasts a little longer than the light-stimulus, and it changes with it. The purposiveness of this difference of reaction is obvious. We cannot say to what degree the direct influence of external conditions is effectively operative on the germ-plasm, or how far, by persistently repeated slight changes, the determinants and the parts of the body determined by them may be made to vary in the course of generations; that is to say, how large a part this direct influence of climate and food may play in the transmutation of species. We can give no answer from experience, because there is an entire lack of perfectly satisfactory and clear experiments; we only know in a few cases how great the variations are which can be brought about in the body during the individual life by means of any of these factors. In most cases it is uncertain whether actually hereditary effects play any part, that is, whether the germ-plasm itself is affected. But if we wish to be theoretically clear as to how far direct climatic effects may go, we may say this, that they may operate as long as they cause no disturbance in the life of the species concerned, for at the moment that such a direct effect begins to be prejudicial to the species personal selection will step in, and, by preferring the individuals which react least strongly to the climatic stimulus, will inhibit the variation. If in any case this should be physically impossible, the species would die out in the climate in question. That a species of plant or animal has climatic limits indicates that individuals which go beyond these are exposed to influences which make life impossible and which natural selection is unable to neutralize. We are here brought face to face with one of the limits to the scope of natural selection.