between the scales. I have been unable to discover hitherto any evidence suggesting an external stimulus which could cause this remarkable process of development in feathers, or indicating that the function of flight would involve such a stimulus. For the present, therefore, we must conclude that feathers are not an adaptation, and not due to somatogenic modification, but must be result of a gametogenic mutation. Feathers, having been evolved, served in the wings and tail as important organs of flight. There is reason to believe that, once present, the growth of feathers was modified greatly by the degree of stimulation applied to the papillae at roots by the movement and bending strain of the feathers. The modification of the hones and of the wing, shoulders, and sternum by the functional stimuli involved in flying are obviously adaptations, and in my opinion are only to be explained as the hereditary effects of functional stimulation, like all skeleto-muscular adaptations. The strains produced in bones by muscular contraction produce hypertrophy of the part of the bone to which the muscles are attached and thus we can understand the origin of the carina of the sternum in flying birds, and its absence in flightless forms. In bats and in pterodactyls also the sternum is produced into a carina along the median line. The reduction of the digits of the wing in birds to three, with the bones firmly united together, would follow from their use in flight and their disuse as digits, and it would seem, from the fact that the flight-feathers must have been always on the posterior edge of the wing, and that the ulna is larger than the radius, that the three digits which have persisted are the 3rd, 4th, and 5th, and not the 1st, 2nd, and 3rd as usually taught. A comparison of the hind-limbs of birds with those of bats and pterodactyls suggests strongly that the patagium flyers have arisen from arboreal or climbing animals, while the birds arose from terrestrial forms which acquired the bipedal habit, as certain reptiles have. An arboreal animal would necessarily use all four limbs, as climbing animals actually do. The wings of birds, on the other hand, would have arisen, from the endeavour to increase speed by movements of the fore-limbs. The perching birds would therefore have arisen by later adaptations after the power of flight had been evolved.