Seven field trials were conducted on winter barley to define relationships between rate of applied N, the amount of nitrate-N present in the soil after harvest and the ratio of soil nitrate-N to grain yield. Applying N up to the economic optimum rate (estimated from yield and N rate data from individual trials) was associated with small increases in soil nitrate-N after harvest (the mean increase was 4 kg N ha(-1)). Where the optimum N rate was exceeded, soil nitrate-N levels increased to a greater extent. In every trial, the ratio of soil nitrate-N to yield showed a minimum at a fertilizer N rate below the economic optimum. However, the value of the ratio was always lower at the optimum N rate (mean value 6.0 kg N t(-1)) than at the zero-N treatment (mean value 8.9 kg N t(-1)) and the difference between the minimum value (mean 5.6 kg N t(-1)) and that found at the optimum N rate was small. Overall, application of fertilizer N up to the economic optimum rate for practical purposes could be regarded as consistent with the objective of minimising the risk of nitrate leaching per hectare and per tonne of grain in the trials.