Grafting was known and practised far back in the ages. Virgil was well acquainted with the process, and Theophrastus compared it with propagation by cuttings. The scion differs from a cutting, however, in having no roots of its own: it is parasitic upon, or rather is in symbiosis with the stock, the root and tissues of which intervene between it and the soil. Consequently the selective absorption, size and number of vessels, and innumerable other physiological and anatomical peculiarities of the stock determine what and how much shall go up into the scion, while the latter supplies the former with organic materials and rules what and how much food, enzymes, and other secretions, etc., it shall receive to build up its substance. Surely, then, if such factors as the nature of the soil, the water and mineral supplies, the illumination, and the various climatic factors of altitude can cause variations on a plant direct, these and other factors are still more likely to be effective on stock and scion, and each must affect the other. Nevertheless opinions have differed much as to whether any important effect is to be seen, and on no point more than on whether the scion can affect the stock, in spite of such examples as _Cytisus Adami_, _Garreya_ on _Aucuba_, Sunflower on Jerusalem Artichoke, etc. Recent results, especially of experiments with herbaceous plants, show that not only can the stock affect the scion (and _vice versa_) directly, but the effect of the changes may be invisible on the grafted plant and only show itself in the progeny raised from the seed of the grafted plant. In other words, variation occurs in grafts either _directly_, as the results of the effects of the environment on the graft, or owing to the interaction of scion and stock, showing as changes in general nutrition in the tissues concerned, etc., owing to special reactions of the protoplasm of the uniting cells one on the other, and of the results of the further protoplasmic secretions, sortings, and so forth, on the cells developed as descendants of these in the further growth of the graft: or _indirectly_, in that some of these changes so alter the nature of the special protoplasm put aside for reproductive purposes, that the resulting embryo in the seed transmits the effects, and they show as variations in the seedling. If these results are confirmed they should meet all objections that have been urged against the transmission of acquired characters.