The influence of seed and soil inoculation on bradyrhizobial migration, nodulation, and N(2) fixation was examined by using two Bradyrhizobium japonicum strains of contrasting effectiveness in N(2) fixation. Seed-inoculated strains formed fewer nodules on soybeans (mostly restricted to the tap and crown roots within 0 to 5 cm from the stem base) than did bradyrhizobia distributed throughout the soil or inoculated at specific depths. Nodulation was greater below the depths at which bradyrhizobial cells were located rather than above, even though watering was done from below to minimize passive bradyrhizobial migration with percolating water. The most profuse nodulation occurred within approximately 5 cm below the point of placement and was generally negligible below 10 cm. These and other results suggest that bradyrhizobial migration from the initial point of placement was very limited. Nevertheless, the more competitive strain, effective strain THA 7, migrated into soil to a greater extent than the ineffective strain THA 1 did. Nitrogen fixation resulting from the dual-strain inoculations differed depending on the method of inoculation. For example, the amount of N(2) fixed when both strains were slurried together onto the seed was about half that obtained from mixing the effective strain into the soil with the ineffective strain on the seed. The results indicate the importance of rhizobial distribution or movement into soil for nodulation, nodule distribution, strain competitiveness, and N(2) fixation in soil-grown legumes.