Now though the colloidal molecules of which organisms are mainly built, are themselves highly composite; and though the physiological units compounded out of these colloidal molecules must have structures far more involved; yet it must happen with such units, as with simple units, that those which have exactly like forms will admit of arrangement into a more stable aggregate than those which have slightly-unlike forms. Among units of this order, as among units of a simpler order, imperfect similarity must entail imperfect balance in anything formed of them, and consequent diminished ability to withstand disturbing forces. Hence, given two organisms which, by diminished nutrition or increased expenditure, are being arrested in their growths--given in each an approaching equilibrium between the forces of the units and the forces of the aggregate--given, that is, such a comparatively balanced state among the units that re-arrangement of them by incident forces is no longer so easy; and it will follow that by uniting a group of units from the one organism with a group of slightly-different units from the other, the tendency towards equilibrium will be diminished, and the mixed units will be rendered more modifiable in their arrangements by the forces acting on them: they will be so far freed as to become again capable of that re-distribution which constitutes evolution. And now let us test this hypothesis by seeing what power it gives us of interpreting established inductions. § 93. The majority of plants being hermaphrodites, it has, until quite recently, been supposed that the ovules of each flower are fertilized by pollen from the anthers of the same flower. Mr. Darwin, however, has shown that the arrangements are generally such as to prevent this. Either the ovules and the pollen are not ripe simultaneously, or obstacles prevent access of the one to the other. At the same time he has shown that there exist arrangements, often of a remarkable kind, which facilitate the transfer of pollen by insects from the stamens of one flower to the pistil of another. Similarly, it has been found that among the lower animals, hermaphrodism does not usually involve the production of fertile ova by the union of sperm-cells and germ-cells developed in the same individual; but that the reproductive centres of one individual are united with those of another to produce fertile ova. Either, as in _Pyrosoma_, _Perophora_, and in many higher molluscs, the ova and spermatozoa are matured at different times; or, as in annelids, they are prevented by their relative positions from coming in contact.