Farm fields are traditionally fertilized as one homogeneous soil unit. Most fields, however, contain two or more soils with different crop yield potentials. This study was conducted to (i) measure crop yield differences between contrasting soils within fields, and (ii) compare the economics of "farming soils, not fields," where contrasting soils in a field receive different vs. uniform rates and formulations of fertilizer. Crop yield variability studies were conducted along 1600 ft transects across several soil units in each of four fields during 1987. Grain yield, test weight, and returns over variable costs varied greatly among soil units in each field (P < 0.05). Soil fertility studies also revealed differences in grain yield, test weight, and returns among soil units in five fields during 1987 and 1988. Fertility studies indicated yields were similar for small grains when recommended fertilizer treatments were applied as soil unit treatments rather than as a field treatment. Returns were $2.06 to $5.14 greater per acre for the soil treatment than for the field treatment in three of five fields, but overall, the returns were not significantly different. A recommended fertilizer treatment was not always the optimum treatment, however. In two fields, additional returns of $21.68 to $23.51/acre resulted when optimum soil treatments were applied rather than the field treatment. The data reveal the importance of appropriate crop yield goals, accurate soil tests, and reliable fertilizer recommendations when developing a strategy for generating greater returns by farming soil, not fields. Given these caveats, our work suggests that farming soils, not fields will increase fertilizer profitability.