generations always the slowest, the smallest, or the least milk-giving individuals respectively wherewith to continue the race, is there any limit to the retrogression which would result in "assigned" time? We certainly should not bring about the reappearance, for instance, of the remotely ancestral horse, the five-toed Eohippus, for the form of the modern horse, like that of other animals, has resulted not only from evolution, but also from retrogression, the disappearance of the four lateral toes being examples mainly of the latter, while the great size of the middle toe is an example of the former; but it cannot be doubted that we should cause, in a time comparatively short as compared to that which elapsed during the evolution, such extreme retrogression as would result in an animal quite incapable of existence.
The improvements in our cultivated plants are the result of centuries of stringent selection; and here again it cannot be doubted, that if the process by which they were evolved were reversed, were we to propagate only from the most inferior plants, that the rate of retrogression would be much more rapid than was the rate of evolution.
It is to be noted, however, that some cultivated plants exhibit comparatively extreme evolution, and in fruit or flower or other particular greatly surpass the wild individuals of the species, e.g. peach, apple, pear, rose, and therefore furnish apparent exceptions to the law that rapid evolution is soon checked by an increasing tendency towards retrogression. In reality they afford the strongest proof of it. I think I am right in saying, that in every instance such plants have been propagated principally by cuttings and not by seed; i.e. they are not descendants in any true sense of their immediate predecessors, but detached portions of them. Their evolution appears to have been effected as
