Nitrogen fixation by field-grown soybean (Glycine max [L.] Merrill) was assessed by the natural '5N abundance and ureide methods. The field sites (five) and genotypes (six, plus two levels of inoculation on Bragg) were chosen to provide a range of proportions of plant N derived from nitrogen fixation (P). Genotypes K466, K468, ntslO07, and ntsIll6 and Davis were included on the basis of their reported tolerance of the suppressive effects of nitrate on nodulation and nitrogen fixation. Bragg was included as a 'nitrate-sensitive' genotype. Seeds of all genotypes were inoculated at sowing with Bradyrhizobium japonicum CB1809 (USDA136). Amounts of nitrate in the soil profile (0-1.2 meter depth) at sowing ranged from 70 (site 3) to 278 kilograms per hectare (site 5), resulting in large effects on plant nodulation, on the 615N values of nodulated plants, on the relative abundance of ureide-N in vacuum-extracted sap (VES) and stem extracts, and finally on the estimates of P. There was no relationship between amount of soil nitrate at sowing and the P15N of the plant-available soil N. Correlation matrices of the measured and calculated parameters indicated generally weak correlations between crop growth (dry matter and N) and the parameters of symbiotic activity (nodule weight, 615N, relative ureide-N); correlations were strong and highly significant between nodulation and the measures of nitrogen fixation (615N, relative ureide-N; r = 0.79-0.92). Estimates of P ranged between 0 and 68% (B15N) and between 6 and 56% (ureide) and were highly correlated (r = 0.97). Results indicated that the ureide method can be used with confidence to assess P by field-grown crops of soybean.