color-sense exists only in those animals which would be decidedly benefited by its possession. And for these reasons it seems improbable that insects ever developed such a faculty until the need for it arose among the beautiful flowers.
Now that we have arrived at this theoretical conclusion, let us hark back again for a while to the reactions which the color-sense, thus aroused, produced upon the flowers which gave it birth.
We may take, as a capital example of an insect-fertilized flower, an English dog-rose. Compare this mentally with the wind-fertilized blossoms, such as grasses and catkins, and it is at once obvious that the great difference between them consists in the presence of a colored corolla. No wind-fertilized plant ever has a whorl of gay petals; and though the converse is not quite true, yet almost all insect-fertilized plants are noticeable for their brilliant tints of red, white, blue, or yellow. The structures in which these pigments reside have no function whatsoever, except that of attracting the insect eye. They are produced by the plant at an enormous physiological expense; and, if their object were not to secure the visits of insects, they would be just so much dead loss to the species. Nor is it only once that these colored corollas have been developed. They occur, quite independently, in both great divisions of flowering plants, the monocotyledons and the dicotyledons. This coincidence could hardly have happened had it not been for that original tendency which we already noticed for pink, scarlet, or orange pigments to appear in the neighborhood of the floral organs. Nor is it twice only, in all probability, that flowers have acquired bright petals through insect visits, but a thousand times over. In almost every family, insect-fertilized, self-fertilized, and wind-fertilized species are found side by side, the one with brilliant petals, the others with small, green, and inconspicuous flowers.
For comparison with the dog-rose, one could not find a better type than that common little early spring blossom, the dog's mercury. It is a wind-fertilized flower, and it does not wish to be seen of insects. Now, this mercury is a very instructive example of a degenerate green flower. For, apparently, it is descended from an insect-fertilized ancestor with bright petals; but, owing to some special cause, it has taken once more to the old wasteful habit of tossing its pollen to the wandering winds. As a consequence it has lost the bright corolla, and now retains only three green and unnoticeable perianth-pieces, no doubt the representatives of its original calyx. Almost equally instructive is the case of the groundsel, though in this case the process of degradation has not gone quite so far. Groundsel is a degenerate composite, far gone on the way of self-fertilization. No class of flowers have been more highly modified to suit the visits of insects than the composites. Hundreds of their tubular bells have been crowded on to a single head, so as to make the greatest possible attractive display; and in many cases the outer blossoms of the head, as in the
