Artificial subsurface drainage in cropland creates pathways for nutrient movement into surface water; quantification of the relative impacts of common and theoretically improved management systems on these nutrient losses remains incomplete. This study was conducted to assess diverse management effects on long-term patterns (1998-2006) of NO3-, NH4+, and PO43- loads (L). We monitored water flow and nutrient concentrations at subsurface drains in lysimeter plots planted to continuous corn (Zea mays L.) (CC), both phases of corn-soybean [Glycine max (L.) Merr.] rotations (corn, CS; soybean, SC), and restored prairie grass (PG). Corn plots were fertilized with preplant or sidedress urea-NH4NO3 (UAN) or liquid swine manure injected in the fall (FM) or spring (SM). Restored PG reduced LNO3- eightfold compared with fields receiving UAN (2.5 vs. 19.9 kg N ha(-1) yr(-1); P < 0.001), yet varying UAN application rates and timings did not affect LNO3- across all CCUANs and CSUANs. The LNO3- from CCFM (33.3 kg N ha(-1) yr(-1)) were substantially higher than for all other cropped fields including CCSM (average 19.8 kg N ha(-1) yr(-1), P < 0.001). With respect to NH4+ and PO43-, only manured soils recorded high but episodic losses in certain years. Compared with the average of all other treatments, CCSM increased LNH4+ in the spring of 1999 (217 vs. 680 g N ha(-1) yr(-1)), while CCFM raised LPO43- in the winter of 2005 (23 vs. 441 g P ha(-1) yr(-1)). Our results demonstrate that fall manuring increased nutrient losses in subsurface-drained cropland, and hence this practice should be redesigned for improvement or discouraged.