The nitrogen economies of six contrasting field pea (Pisum sativum L.) genotypes were examined at three widely separated sites in south Western Australia, using the 15N natural abundance technique to asssess proportional dependence on fixed N, harvests at peak biomass to assess total N yields and harvests at crop maturity to examine partitioning of N between seed and non harvested crop residues. The budgets for one site (Wongan Hills) included N of nodulated roots which on average comprised 12% of total plant N at peak crop biomass and 25% of recoverable plant N after harvest of seed. At this site maximum potential (residual) benefits to a following crop (peak total plant biomass N - N uptake from soil and N taken off as harvested seed) varied between genotypes from 8 to 41 g N ha-1 (mean 26; n = 6). Data for the other two sites, based solely on N budgets of above-ground parts, provided evidence of substantial site- and genotype-specific differences in N balance in terms of shoot residues (i.e. + 7 to - 24 kg N ha-1 (mean - 5) at Avondale, + 40 to - 29 (mean + 3) at Mt Barker). The results collectively indicated a general relationship between peak biomass N of a crop and its potential or otherwise to effect a net input of residue N to the ecosystem. There were, however, considerable variations between genotypes and sites due to differences in proportional dependence on fixation (range across all sites and genotypes 60-91%) and crop harvest indices for N (corresponding range 53-90%). Correlation plots were constructed from the data for N2 fixed against crop dry matter yield and residual nitrogen benefit against nitrogen harvest index. Results are discussed in relation to values for N balance of field pea and other grain legumes obtained elsewhere by other investigators.