The composite plant formed by grafting a scion on a rootstock provides a unit of considerable scientific interest and commercial importance. The combination of two plants of different genetic constitution, thus having different inherent characters, affords new opportunities of studying these characters, and of making wider use of them.One of the first general reviews of rootstock/scion relations was that of Hatton, in 1930 (1). This included much of his own pioneer work, which led to a great expansion of interest in clonal rootstocks (2, 3). Tukey & Brase (4) included a useful review in a paper on the influence of scions and inter stocks. Hilkenbaumer (5) reviewed apple, pear, and plum rootstock results in Germany in 1953, and Hoblyn (6) gave a brief practical summary of results in England up to 1950. Tubbs (7) in a general review of work at East Mailing, illustrated the extensive range of rootstock effects with a list of 29 scion characters found to be influenced by the rootstock. Reviews on special as pects of the subject include those of Argles (8) on stock/scion incom patibility, and Graves (9) and Katyal (10) on double working; Roberts (11), writing on "Theoretical aspects of graftage," provided a comprehensive catalogue of 479 references with concise comments on the papers cited. The present paper is not intended to cover as wide a range, but will deal mainly with the part played by different components of the composite tree, incom patibility, viruses and certain other aspects, concluding with a discussion of some possible mechanisms of rootstock effect.At the outset, it must be said that the discovery of latent viruses in some clonal rootstocks and scion varieties has introduced a new factor of uncer tainty into some past rootstock trials. Varying proportions of certain clones, particularly M.lX and M.I, were found to be infected with rubbery wood virus, which has been shown to have a dwarfing effect [Posnette & Cropley (12,13); LuckwiII (14)]. Further work by Posnette & Cropley with apple mosaic (15), cherry (16), and pear (17) viruses has shown marked dwarfing effects, both with and without leaf symptoms.I t is reassuring, however, that tests of each tree in the classical Malling apple rootstock trials have shown that the original clones were evidently not infected with rubbery wood or mosaic and that the original conclusions on rootstock vigour remain valid (18).In discussing other papers it can only be assumed that the rootstock results are not clouded by virus infection; but the importance of using only virus-tested material in future is evident.