We measured nitrate (NO− 3) leaching, ammonia (NH3), and nitrous oxide (N2O) emissions, denitrification, crop offtake of N, and wet N deposition on two fields at each of two sites for five seasons. The soils were loamy sand and silty clay loam. The objective was to quantify all the major N fluxes over 2 arable rotations on contrasting soil types and to assess the impact of these balances on soil N. The greatest N input was fertiliser-N, which was between 89 and 96% of the total, except when linseed was grown to which no fertiliser-N was applied. Wet N deposition was 7–15 kg ha−1 which contributed ca. 5% to total N inputs. Outputs of N were mainly as crop offtake and NO3 − leaching. Crop N offtakes were 54 to 98% of total N outputs, ranging from 45 kg ha−1 for linseed to 245 kg ha−1 for winter wheat. N offtake was usually least ( 200 kg ha−1) were usually by winter wheat crops when straw was removed. At both sites average leaching followed the expected pattern of sugarbeet < cereals < potatoes. Total annual losses of N2O never exceeded 2 kg ha−1 N, between 0.2 and 2.8% of fertiliser-N applications. Estimated losses of N2 gas tended to be greater, especially on the heavier soil, up to ca. 7 kg ha−1 in 1998/99. Fluxes of NH3 were usually small (1–3 kg ha−1 N), and nearly half of the measured fluxes were net deposition. Net N inputs of NH3 exceeded outputs in only one of the four fields monitored. Thus, the greatest potential reduction in N losses from arable fields not given organic manures is to minimise NO3 − leaching. This would have the additional benefit of reducing subsequent indirect emissions of N2O which at 2.5% of leached N, would be as great as direct soil emissions of N2O. However, it may be difficult to make significant reductions in leaching from rotations such as these, receiving only mineral fertiliser-N. Most crops in this study received no more fertiliser-N than was currently recommended at the time for optimum yield, and so reductions in N input would potentially have led to loss of income. Moreover, inputs and outputs of N were, at best, in balance and in one of the four fields monitored N outputs exceeded N inputs, which is likely to lead to a depletion of soil N. Any further decrease in soil N might have repercussions for the sustainable use of soil.