Amounts of spring nitrogen (N) fertilizer (0–240 kg/ha), combined with three timing treatments (single, divided early or divided late), were tested at 14 sites in England and Wales between 1984 and 1988 to determine the optimum fertilizer N requirement for winter oats. The trials were superimposed on commercial crops of the cultivars Pennal (9 sites) or Peniarth (5 sites). Optimum amounts of N ranged from nil to 202 kg/ha (mean 119) and optimum yields varied between 5·8 and 9·9 t/ha (mean 7·3). Much (c. 60%) of the inter-site variation in N optimum was explained by differences in soil N supply, as indicated by N offtake in the grain at nil applied N. Mean yield differences between single and early (+0·08 t/ha) or late (−0·04 t/ha) divided dressings were slight, although significant (P<0·05) but inconsistent yield effects were obtained from early N at two sites and late N at three sites.

Lodging occurred at 11 of the 12 sites where lodging scores were recorded and always increased significantly (P<0·05) with applied N. The amount of crop lodging at N optimum was, on an area basis, <50% at nine of the sites. The overall extent of site lodging was also influenced by soil N fertility and hence inversely related to N optimum. However, multiple regression, using site lodging as well as soil N supply, only accounted for slightly more (65%) of the variation in N optimum, which suggests that lodging was not a major limiting factor. Lodging was unexpectedly less from early N (mean 43%), but more from late N (53%) divided dressings, compared with a single N dressing (49%). Early N reduced lodging significantly (P<0·05) at four sites, although the actual reduction was only large at one site where early N also increased yield significantly (+0·57 t/ha).

Grain N concentrations increased significantly (P<0·05) with applied N, on average by 0·12% per 40 kg/ha N increment. Timing effects on grain N concentration were very small, with mean values of 1·94, 1·91 and 1·96%N respectively from single, early and late divided dressings. Apparent recovery in grain of fertilizer N at the optimum amount ranged from 13 to 57% (mean 37), with better N recovery at the more yield-responsive sites. Changes in mean grain weight due to the amount and timing of fertilizer N were small, with an average reduction of 0·6 mg/grain per 40 kg/ha N applied. The adverse effects of N fertilizer on grain quality were slight and unlikely to have commercial significance. The agronomic implications of these results on the N fertilization of winter oats are discussed.