How, then, do we account for the wings of the insect? Here we can offer nothing more than speculation, but the speculation is not without interest. It may be laid down in principle that the flying animal begins as a leaping animal. The "flying fish" may serve to suggest an early stage in the development of wings; it is a leaping fish, its extended fins merely buoying it, like the surfaces of an aeroplane, and so prolonging its leap away from its pursuer. But the great difficulty is to imagine any part of the smooth-coated primitive insect, apart from the limbs (and the wings of the insect are not developed from legs, like those of the bird), which might have even an initial usefulness in buoying the body as it leaped. It has been suggested, therefore, that the primitive insect returned to the water, as the whale and seal did in the struggle for life of a later period. The fact that the mayfly and dragon-fly spend their youth in the water is thought to confirm this. Returning to the water, the primitive insects would develop gills, like the Crustacea. After a time the stress of life in the water drove them back to the land, and the gills became useless. But the folds or scales of the tough coat, which had covered the gills, would remain as projecting planes, and are thought to have been the rudiment from which a long period of selection evolved the huge wings of the early dragon-flies and mayflies. It is generally believed that the wingless order of insects (Aptera) have not lost, but had never developed, wings, and that the insects with only one or two pairs all descend from an ancestor with three pairs. The early date of their origin, the delicacy of their structure, and the peculiar form which their larval development has generally assumed, combine to obscure the evolution of the insect, and we must be content for the present with these general indications. The vast unexplored regions of Africa, South America, and Central Australia, may yet yield further clues, and the riddle of insect-metamorphosis may some day betray the secrets which it must hold. For the moment the Carboniferous insects interest us as a rich material for the operation of a coming natural selection. On them, as on all other Carboniferous life, a great trial is about to fall. A very small proportion of them will survive that trial, and they trill be the better organised to maintain themselves and rear their young in the new earth.