This can be clearly demonstrated in the case of flowers. That the wild pansy (_Viola tricolor_), which lives in the plains and on mountains of moderate elevation, is fertilized by bees, and the nearly allied _Viola calcarata_ of the High Alps by Lepidoptera, is readily intelligible, since bees are very abundant in the lower region, and make the fertilization of the species a certainty, while this is not so in the High Alps. There the Lepidoptera are greatly in the majority, as every one knows who has traversed the flower-decked meads of the High Alps in July, and has seen the hundreds and thousands of butterflies and moths which fly from flower to flower. Thus the viola of the High Alps has become a 'butterfly-flower' by the development of its nectaries into a long spur, accessible only to the proboscis of a moth or butterfly. The chance which led certain individuals of the ancestral species to climb the Alps must also have supplied the incentive to the production of the changes adapted to the visits of the prevalent insect. The hypothesis of a predestinating Power of Development suffers utter shipwreck in face of facts like these. We have, furthermore, an excellent touchstone for the reality of the processes of selection in the _quality_ of the variations in flowers and insects. Natural selection can only bring about those changes which are of use to the possessors themselves; we should therefore expect to find among flowers only such arrangements as are, directly or indirectly, of use to them, and, conversely, among insects only such as are useful to the insect.