the fingers of one and how slender those of another; or it needs but to note the unlikeness of gait of passers-by, implying small unlikenesses of structure; to be convinced that the relations among the variations of co-operative parts are anything but fixed. And now, confining our attention to limbs, let us consider what must happen if, by variations taking place miscellaneously, limbs have to be partially changed from fitness for one function to fitness for another function—have to be re-adapted. That the reader may fully comprehend the argument, he must here have patience while a good many anatomical details are set down.
Let us suppose a species of quadruped of which the members have for long past periods been accustomed to locomotion over a relatively even surface, as, for instance, the "prairie dogs" of North America; and let us suppose that increase of numbers has driven part of them into a region full of obstacles to easy locomotion—covered, say, by the decaying stems of fallen trees, such as one sees in portions of primeval forest. Ability to leap must become a useful trait; and, according to the hypothesis we are considering, this ability will be produced by the selection of favorable variations. What are the variations required? A leap is effected chiefly by the bending of the hind limbs so as to make sharp angles at the joints, and then suddenly straightening them; as any one may see on watching a cat leap on to the table. The first required change, then, is increase of the large extensor muscles, by which the hind limbs are straightened. Their increases must be duly proportioned, for if those which straighten one joint become much stronger than those which straighten the other joint, the result must be collapse of the other joint when the muscles are contracted together. But let us make a large admission, and suppose these muscles to vary together; what further muscular change is next required? In a plantigrade mammal the metatarsal bones chiefly bear the reaction of the leap, though the toes may have a share. In a digitigrade mammal, however, the toes form almost exclusively the fulcrum, and if they are to bear the reaction of a higher leap, the flexor muscles which depress and bend them must be proportionately enlarged; if not, the leap will fail from want of a firm point d'appui. Tendons as well as muscles must be modified; and, among others, the many tendons which go to the digits and their phalanges. Stronger muscles and tendons imply greater strains on the joints; and unless these are strengthened, one or other dislocation will be caused by a more powerful spring. Not only the articulations themselves must be so modified as to bear greater stress, but also the numerous ligaments which hold the parts of each in place. Nor can the bodies of the bones remain unstrengthened; for if they have no more than the strengths needed for previous move-