Nitrogen fixation and transfer in vetch and vetch-oats mixtures

15N-labelled organic matter and ammonium sulphate fertilizer (20 and 60 kg N ha −1) were applied as 15N sources to estimate N 2 fixed in field-grown pure vetch ( Vicia sativa L.) and vetch grown in two mixtures with oats ( Avena sativa). The decline in 15N enrichment in the soil during the seven sequential harvests and deduced from the atom % N excess values in the oats reference crop was significantly slower in the organic matter labelled plots than with the 15N-labelled ammonium sulphate plots. Total N uptake patterns of oats and vetch differed markedly for the last two or three harvests. Also, estimates of N 2 fixed (Ndfa) in vetch differed with the 15N source. The mean % Ndfa (or total Ndfa) at 106 days after emergence was 76% (95 kg N ha −1) with the ploughed-in organic matter, 45% (61 kg N ha −1) or 41% (41 kg N ha −1) when 20 or 60 kg N ha −1, respectively, of 15N-labelled ammonium sulphate was applied. Estimates of N transferred from the legume to the cereal were similarly influence by the 15N source. Little (maximum 5 kg N ha −1) or no N transfer from vetch to oats was detected, using the organic matter labelled plots. In contrast, several significant values (up to 29 kg N ha −1) of N transferred were estimated using the inorganic 15N-labelled plots. With the greater effects of fast 15N: 14N ratio decline in soil on reference crop errors, we estimate that the values of N 2 fixed and of N transferred under the more stable 15N: 14N conditions, i.e. with 15N-labelled organic incorporated into soil, should be the most accurate. Thus, vetch derived on the average, over 70% of its N from fixation, while N transfer from vetch to oats was not significant.