These experiments were conducted to evaluate the combined effects of intercropping and N source and the interaction of these two factors with tillage on N dynamics. Maize (Zea mays L.) was planted in two consecutive years in Costa Rica to measure changes in soil mineral N, soil C, and maize grain N in response to monoculture or polyculture and legume residues or NH 4 NO 3 fertilizer under conventional (CT) or minimum (MT) tillage. Soil mineral N levels were greater with fertilizer than with residues at the end of the first season and greater with residues at the beginning of the second season, reflecting a residual effect from the previous season. There were no treatment interactions. The dominant form of mineral N shifted from NO 3 - -N (first season) to NH 4 + -N (second season). The shift corresponded to an overall increase in soil C (from 27.6 to 30.2 g kg -1 ), which was significantly greater under monoculture and with residues (33.2 and 33.4 g kg -1 , respectively) than under polyculture and fertilizer (31.1 and 30.9 g kg -1 , respectively). A positive cropping system x N source interaction resulted in 67% greater maize grain N in the polyculture under residues than with fertilizer N. Maize response to polyculture and legume residues, together with soil N and C data, were used to propose a mechanistic explanation for how these strategies may help maintain productivity (and minimize N loss) in low-input farming systems.