_Transposition and multiplication of parts._--The cases of abnormal transposition or multiplication of organs--for instance, the development of teeth in the palate or of pollen in the edge of a petal--are to be explained by supposing that the gemmules unite with wrong cells instead of, or as well as, the right ones; “and this would follow from a slight change in their elective affinities.” Such slight changes are known to occur; for instance, certain plants “absolutely refuse to be fertilised by their own pollen, though abundantly fertile with that of any other individual of the same species.” Inasmuch as the cells of adjoining parts will often have nearly the same structure, we can understand that some slight change in elective affinity may affect a large area. Hence we can account for a crowd of horns on the head of a sheep, or many spurs on the leg of a fowl, etc. Frequently repeated parts are extremely liable to vary in number; in this case we have a large number of closely allied gemmules and of points for their union, and slight changes in elective affinity would be specially apt to occur. VARIABILITY.--Changed conditions may lead to irregularity in the number of gemmules derived from various parts of the body; deficiency in number might cause variation in any part by leaving some of the cells free to unite with allied gemmules. The direct action of surroundings, or the effect of use or disuse on a part, may cause corresponding modifications of the gemmules, and through these of the part in the succeeding generation. “A more perplexing problem can hardly be proposed,” and yet it receives an explanation on this hypothesis. Such causes must, as a rule, act during many generations before the modification reappears in the offspring. This may be due to the unaltered gemmules derived from earlier generations, and their gradual replacement by the increasing number of altered gemmules. Variation in plants is much more frequent in sexually produced than it is in asexually produced individuals. This may be due to the absence in the latter of that great cause of variability, changes in the reproductive organs under altered conditions. Furthermore, the former alone pass through the earlier phases of development, when structure is most plastic and yields most readily to the causes inducing variability. The stability of hybrids and of many varieties when propagated by buds, as compared with their reversion to the parent form when propagated by seed, remains inexplicable. Hence variability is explained as due (1) to the irregularity in number of gemmules, to their transpositions, and redevelopment when dormant; and (2) to their actual modification and the gradual replacement by them of unaltered gemmules.