Relying more on biological N-2 fixation has been suggested as a way to meet one of the major challenges of agricultural sustainability. A 15N study was conducted to compare the fate of applied legume and fertilizer N in a long-term cropping systems experiment. Nitrogen-15-labeled red clover (Trifolium pratense L.) and (NH-4)-2SO-4 were applied to microplots within the low-input and conventional cropping systems or the Farming Systems Trial at the Rodale Institute Research Center in Pennsylvania. The 15N was applied to soil and traced into corn (Zea mays L.) in 1987 and 1988. Residual 15N was also traced into second-year spring barley (Hordeum vulgare L.). Legume and fertilizer 15N remaining in soil was measured and loss of N was calculated by difference. More fertilizer than legume N was recovered by crops (40 vs. 17% of input), more legume than fertilizer N was retained in soil (47 vs. 17% of input), and similar amounts of N from both sources were lost from the cropping systems (39% of input) over the 2-yr period. More fertilizer than legume N was lost during the year of application (38 vs. 18% of input), but more legume than fertilizer N was lost the year after application (17 vs. 4% of input). Residual fertilizer and legume 15N was distributed similarly among soil fractions. Soil microbial biomass was larger in the legume-based system. A larger, but not necessarily more active, soil microbial biomass was probably responsible for the greater soil N supplying capacity in the legume-based compared with fertilizer-based system.