the fertilized germs produced by the same parents, the physiological units derived from them respectively will bear a different numerical ratio to each other in every case. If the parents are constitutionally quite alike, the variation in the ratio between the units they severally bequeath, cannot cause unlikenesses among the offspring. But if otherwise, no two of the offspring can be alike. In every case the small initial difference in the proportions of the slightly-unlike units, will lead, during evolution, to a continual multiplication of differences. The insensible divergence at the outset will generate sensible divergences at the conclusion. Possibly some may hence infer that though, in such case, the offspring must differ somewhat from each other and from both parents, yet that in every one of them there must result a homogeneous mixture of the traits of the two parents. A little consideration shows that the reverse is inferable. If, throughout the process of development, the physiological units derived from each parent preserved the same ratio in all parts of the growing organism, each organ would show as much as every other, the influence of either parent. But no such uniform distribution is possible. It has been shown (_First Principles_, § 163), that in any aggregate of mixed units segregation must inevitably go on. Incident forces will tend ever to cause separation of the two orders of units from each other--will tend to integrate groups of the one order in one place and groups of the other order in another place. Hence there must arise not a homogeneous mean between the two parents, but a mixture of organs, some of which mainly follow the one and some the other. And this is the kind of mixture which observation shows us. Still it may be fairly objected that however the attributes of the two parents are variously mingled in their offspring, they must in all of them fall between the extremes displayed in the parents. In no characteristic could one of the young exceed both parents, were there no cause of "spontaneous variation" but the one alleged. Evidently, then, there is a cause yet unfound.