Autoregulation of symbiotic root nodulation in soybean seedlings (Glycine max L. Merrill cv Pride 216) was studied following double inoculation of primary roots with Bradyrhizobium japonicum 110. When the second inoculation was given 10 or 17 hours after the first, the nodulation in the first-inoculated region of the root was suppressed. The effect was eliminated if B. japonicum 110 containing Tn5 insertions in the 'common' nod ABC genes was used for the second inoculation, indicating the requirement for changes in the root mediated by these bacterial genes. When the root cortex in the suppressed basal region was examined 3 days after inoculation, cell division centers were present in numbers not significantly different from the numbers in contrQl roots given a sham second inoculation; their size distribution, however, showed a failure of enlargement compared with controls.Feedback suppression of symbiotic root nodulation ('autoregulation') was discovered in 1952 by Nutman in red clover ( 16) and has since been demonstrated in several other legume plants (see 19 for review). Nutman interpreted it as an inhibition of new nodule production by already established nodules on the root, because surgical excision of established nodules or nodule meristems allowed additional nodulation (16). Split-root techniques have since been used (12) to demonstrate the systemic nature of the response in soybean plants (Glycine max L. Merrill). Grafting experiments with soybeans have shown that feedback suppression requires the normal shoot, since the response is lost when a mutant shoot which lacks the normal nodulation control ('supernodulating') is substituted for the normal one (8,17).Double inoculation experiments carried out by Pierce and Bauer (18) showed that feedback suppression was actually a plant response which could not be explained by limitation of the number of infective bacteria. Their experiments demonstrated that a second inoculation 15 h after the first produced virtually no nodules, provided that the bacterial concentration in the first inoculum was optimized for nodule yield.The present investigation was undertaken to determine the basis for the observation (22) that very early feedback regulation appears to control the nodulation not in the apical region but instead in the basal region of the root, which is initially most susceptible to infection by the bacteria (1, 2). This region is the first one inoculated in double inoculation protocols. The number of nodules which appeared in this region was greatly reduced when a large second inoculum was given 10 h after the first inoculum (22). This result was not 865 observed by Pierce and Bauer (18), whose protocol for double inoculation was different. Two questions were therefore asked: (a) Did the effect require the 'common' nod genes in B. japonicum which are responsible for nodulation? If so, it would indicate that the effect depended on specific gene expression and was not due to nonspecific phenomena incidental to the double inoculation protocol, e.g. di...