It is upon this that the operation of natural selection, that is, personal selection, must depend--that the germ-plasms in which the favourable variation-direction is in the majority are selected for breeding, for it is this and nothing else that natural selection does when it selects the individuals which possess the preferred variations. The ascending process is thus considerably advanced, because the opposing determinants are more and more eliminated from the germ-plasm, till the preferred variations of _X_ are left, and among these, as ascent in the direction begun continues, the opposing variations are again set aside by germinal selection, and so on. Reducing divisions and amphimixis are thus powerful factors in furthering the transformations of the forms of life, although they are not the ultimate causes of these. Now that we have made ourselves familiar with the idea of germinal selection we shall attempt to gain clearness as to what it can do, and how far the sphere of its influence extends, and, in particular, whether it can effect lasting transformations of species without the co-operation of personal selection, and what kind of variations we may ascribe to it alone. First, I must return for a moment to the question we have already briefly discussed--whether the variation of a determinant upwards or downwards must so continue without limit. We might be inclined to think that the great constancy which many species exhibit was a plain contradiction of this, for if every minute variation of a determinant necessarily persisted without limit in the same direction, we should expect to find all the parts of the organism in a state of continual unrest, some varying upwards, some downwards, always ready to break the type of the species. Must there not be some internal self-regulation of the germ-plasm which makes it impossible that every variation which crops up can persist unlimitedly? Must there not be some kind of automatic control on the part of the germ-plasm, which is always striving to re-establish the state of equilibrium that has once been attained by the determinant system whenever it is disturbed?