The contribution of biologically fixed dinitrogen (N-2) to the nitrous oxide (N2O) production in grasslands is unknown. To assess the contribution of recently fixed N-2 as a source of N2O and the transfer of fixed N from clover to companion grass, mixtures of white clover and perennial ryegrass were incubated for 14 days in a growth cabinet with a N-15(2)-enriched atmosphere (0.4 atom% excess). Immediately after labelling, half of the grass-clover pots were sampled for N-2 fixation determination, whereas the remaining half were examined for emission of N-15 labelled N2O for another 8 days using a static chamber method. Biological N-2 fixation measured in grass-clover shoots and roots as well as in soil constituted 342, 38 and 67 mg N m(-2) d(-1) at 16, 26 and 36 weeks after emergence, respectively. The drop in N-2 fixation was most likely due to a severe aphid attack on the clover component. Transfer of recently fixed N from clover to companion grass was detected at 26 and 36 weeks after emergence and amounted to 0.7 +/- 0.1 mg N m(-2) d(-1), which represented 1.7 +/- 0.3% of the N accumulated in grass shoots during the labelling period. Total N2O emission was 91, 416 and 259 mu g N2O-N m(-2) d(-1) at 16, 26 and 36 weeks after emergence, respectively. Only 3.2 +/- 0.5 ppm of the recently fixed N-2 was emitted as N2O on a daily basis, which accounted for 2.1 +/- 0.5% of the total N2O-N emission. Thus, recently fixed N released via easily degradable clover residues appears to be a minor source of N2O.