Nitrogen balance sheets for the Rothamsted Continuous Wheat Experiment show that N fertilizers were used very inefficiently in the past. Thus, over the period 1852-1967, the apparent recovery (in grain and straw) of the fertilizer N applied to the plot receiving 144 kg N ha -1 each year was 32%. Recently, however, the apparent recovery of fertilizer N has increased and for the two years 1979 and 1980 the mean value in this plot was 86% . Increased recovery is mainly due to new, high-yielding varieties of winter wheat that take up more fertilizer N, to better control of pests and diseases and to earlier sowing in autumn. Early sowing can increase the amount of N overwintering inside the plant instead of outside in the soil, where it is subject to losses through leaching and denitrification. Recoveries of N by spring-sown barley on the Rothamsted Continuous Barley Experiment were lower than with winter wheat, as would be expected under a cropping system in which the soil is without plant cover from August until March or April. Tentative national N balance sheets for the two major cereal crops in the U.K., wheat and barley, are set out, based on crop survey data. In these balance sheets, between one-third and one-half of the chemically combined N reaching the soil is not recovered in grain and straw. Recoveries of N on the Rothamsted Classical Experiments show that there is considerable scope for improving the recovery of N on a national scale. Research into the quantitative aspects of the nitrogen cycle in the field is needed in order to understand how this N is lost and how the losses can be diminished.