The availability of soybean mutants with altered symbiotic properties allowed an investigation of the shoot or root control of the relevant phenotype. By means of grafts between these mutants and wild-type plants (cultivar Bragg and Williams), we demonstrated that supernodulation as well as hypernodulation (nitrate tolerance in nodulation and lack of autoregulation) is shoot controlled in two mutants (nts382 and nts ll6) belonging most likely to two separate complementation groups. The supernodulation phenotype was expressed on roots of the parent cultivar Bragg as well as the roots of cultivar Williams. Likewise it was shown that non-nodulation (resistance to Bradyrhizobium) is root controlled in mutant nod49. The shoot control of nodule initiation is epistatically suppressed by the non-nodulation, root-expressed mutation. These findings suggest that different plant organs can influence the expression of the nodulation phenotype.The development of N-fixing nodules on legume roots upon invasion of Rhizobium (or Bradyrhizobium) bacteria is subject to regulation by factors both external and internal to the plant host. In particular, the extent of nodulation is restricted by a process termed autoregulation, in which the formation ofnodules on one part of the root systemically inhibits subsequent nodule formation in other root regions (3,15). Nodulation is also severely restricted by the presence of nitrate in the soil (7). Our laboratory has recently isolated several soybean mutants with altered symbiotic features, including some tolerant to nitrate (nts3) which also supernodulate (3, 4), and others which do not form any nodules (nod-) (1, 8).Clearly nodulation is subject to control by plant factors; the sites of this control are unknown, although some experimentation has implicated shoot-root interactions (1 1-13). By means of grafts between these mutant and wild-type plants, we Table I were inoculated with B. japonicum strain USDA110 (approximately 107 bacteria per plant) and were cultured in 25 cm pots filled with vermiculite:sand mixture (1:2 ratio). Glasshouse temperatures were held between 14 and 30°C and incandescent 100 W bulbs extended the photoperiod to 16 h near summer conditions). The pots received 1.2 L of nutrient solution as described by Herridge (10) three times a week (Table I