The fertiliser nitrogen requirement of cereals grown on sandy soils

The responses to fertiliser‐N of winter wheat and winter barley grown on sandy soils were measured in 72 experiments in England from 1990 to 1994. Yield without fertiliser‐N (Y0) was c 1.1 t ha−1 greater following root crops than following cereals. Following potato crops given organic manures, Y0 was c 1.2 t ha−1 greater than following unmanured potato crops, but Y0 was no greater following sugarbeet to which organic manures had been applied. Only after the two driest winters was there sufficient variation in soil N supply in spring (SNSs) for this to show a relationship with Y0. However, Y0 increased with increasing N mineralisation during the growing season (AM) in the three years it was measured. There was no consistent effect of sowing date on Y0. Following potatoes, yield at optimum fertiliser‐N (Yopt) decreased as sowing date was delayed, but this was not so after cereals, sugarbeet or overall. There was no increase in Yopt with SNSS or AM, but Yopt decreased with increasing moisture stress (S) in June. The mean yield response to Nopt (ΔY) was c 0.4 and 0.8 t ha−1 smaller following potatoes and sugarbeet respectively than following cereals, but not consistently so as there were large interactions between site, year and previous crop. Following root crops, ΔY was c 0.6 and 1.4 t ha−1 less after sugarbeet and potatoes respectively that had been given organic manures. Without the addition of organic manures, ΔY following potatoes was similar to that following cereals. Regression on SNSS and AM accounted for 28 and 15% respectively of the variance in ΔY. The optimum economic fertiliser‐N application (Nopt) was similar, at c 140 kg ha−1, following cereals and potatoes. Following sugarbeet, cereal Nopt was only c 110 kg ha−1. The differences according to previous crop reported here are consistent with mineralisation of crop residues on sandy soils being more rapid than on other soils; the potato residues were rapidly mineralised in autumn and lost by leaching over winter. Residues from later‐harvested sugarbeet were mineralised during the growing season of the subsequent cereal crop. Fertiliser‐N requirements were, at c 110–140 kg ha−1, smaller than has been found on other soil types, and less than current recommendations for wheat. Requirements were significantly reduced in years of drought stress. No differences were found in Nopt between wheat and barley. These data do not justify the current advice to invariably reduce fertiliser‐N to cereals following