From 1985 to 1987, 20 experiments were carried out on shallow chalk soils, in which soil N reserves were expected to be small, to assess seasonal variations in the response of winter cereals to applied fertilizer N, and to relate these responses to measurements of soil mineral N (SMN), temperature and soil moisture deficits (SMD).

Soil mineral N measured in autumn varied from 21 kg/ha (1986) to 73 kg/ha (1985), while SMN in spring ranged from 19 kg/ha (1987) to 91 kg/ha (1985), these values were typical of soils in long-term arable rotations. Estimates of apparent net N mineralization (AM) during the growing season were small at c. 26 kg/ha and suggested large seasonal variation. The small AM is considered to be due to the shallow topsoil drying out during the growing season. Whole crop N offtake without fertilizer N was only c. 40kg/ha. Crop N offtake, grain yield without fertilizer N and AFR (apparent recovery of fertilizer N) could not be reliably predicted by regression on SMN in autumn, SMN in spring or AM. Little or none of the variation in crop yield could be accounted for by regression on accumulated temperature over winter, maximum SMD in April to July or mean temperature in April to July.

Despite optimum grain yields being only moderate at 6·59 t/ha for winter wheat and 6·78 t/ha for winter barley, response to applied fertilizer N was large, between 3·77 and 5·38 t/ha. In consequence the requirement for fertilizer N (c. 240–250 kg/ha) was also large, but differed little between seasons. This large requirement is concluded to be a result of limited fertilizer recovery and mineralization of soil N during the growing season.