The responses to fertilizer N of winter wheat and winter barley grown on sandy soils were measured in 74 experiments in England from 1991 to 1994. A series of curves was fitted to the data from each of the experiments to determine which best described the yield response to fertilizer N. The linear plus exponential model, adopted for data from a range of other soil types, with the rate parameter (r) determined for each experiment, proved to be the most satisfactory for these data. The median value of r at 0·987 was less than has usually been determined for other soil types, and was reduced by increasing moisture stress in June, suggesting that the reduced efficiency with which recovered N produces assimilate is partly a result of moisture stress. The model produced a mean Nopt of 143 kg/ha following cereals.

Soil N supply in spring (SNSs) was small at c. 45 kg/ha and commensurate with a large requirement for fertilizer N. Apparent fertilizer recovery (AFR) at 0·70 was greater than that previously measured on clay or chalk soils, suggesting that AFR was not restricted by lack of soil moisture. The larger AFR is considered to be a consequence of more rapid remobilization of fertilizer N in these sandy soils. The greater recovery of fertilizer N reduced the optimum requirement for fertilizer N. Nitrogen harvest index (NHI) was decreased from 0·76 to 0·70 as fertilizer N increased from 0 to 300 kg/ha, whereas dry matter harvest index (HI) decreased from 0·53 to 0·50. The economic optimum requirement for fertilizer N, at 143 kg/ha, was less than reported for other soil types.

The increase in grain yield from applied fertilizer N was less in these experiments than in those reported on other soil types in the UK. However, this overall observation masks two counterbalancing differences. Fertilizer N was recovered more efficiently on these soils; the recovered fertilizer N was used with similar efficiency to produce grain dry matter. The explanation proposed is that on these soils a large proportion of soil water is readily available to crops, and this was taken up rapidly in May accompanied by fertilizer N. However, shortage of water in June restricted the ability of the crop to produce assimilate from fertilizer N applications >c. 150 kg/ha.