Spatially variable N application may improve N use efficiency, grain yield, and net returns of winter wheat (Triticum aestivum L.) in fields exhibiting wide ranges of soil characteristics. The objectives of this study were to (i) determine the variability in optimal economic grain yield and in the amount of N required to produce a unit of grain at optimum yield, the unit N requirement (UNR), among landscape positions; (ii) evaluate landscape position as a criterion for dividing fields into units of equal productivity; and (iii) assess the economic benefits of spatially variable N fertilizer application. Replicated N rate (0 to 125 lb N/acre) experiments were established on footslope, south-backslope (S-backslope), shoulder, and north-backslope (N-backslope) positions of the hillslope profile at two eastern Washington farms in 1990 and 1991. Yield potential among these landscape positions varied by up to 63% and UNRs varied by up to 70%. Landscape position, however, was not an adequate criterion for dividing fields into equal productivity units. There was little economic benefit from variable N applications in a hypothetical case analysis if N recommendations were calculated assuming a constant UNR; however, if experimentally determined UNRs were used, variable applications increased net returns by up to $14.80. The degree of economic benefit depended on the levels of misapplication that occurred when a single N rate was applied to a field and the yield responses to N which determined the results of misapplication. Spatially variable N fertilizer management will require accurate estimates of yield potential, UNR, mineralization and available soil N to be economically viable.