The 1995 Food Security Act required producers with highly erodible land to develop erosion control plans to be eligible for farm program benefits. Conservation tillage was a part of many of those plans; however, mechanical incorporation of herbicides to reduce runoff losses can be a problem while trying to maintain crop residue. Soil, water, and herbicide losses (with water and sediment) were measured from continuous com runoff plots (1.7 x 12.0 m) in 1993 and 1994 under natural rainfall conditions. Four tillage/herbicide application treatments were studied: notill/herbicide broadcast sprayed (NT); fall chisel plow-spring disk/herbicide broadcast sprayed after disking (DS); fall chisel plow-spring disk/herbicide broadcast sprayed before disking (SD); and fall chisel plow-spring "mulch master" /herbicide applied with John Deere's Mulch Master (MM). Residue measurements after tillage and planting generally showed NT with the greatest percent residue cover, MM second, and SD together with DS the least. By storm event, NT generally had the least erosion and the lowest runoff volumes. For the herbicides studied, atrazine, metolachlor, and cyanazine, concentrations in sediment and runoff water were generally in the order NT>DS>MM>SD. Lack of incorporation and/or application to more crop residue with NT was believed responsible for the higher concentrations with that system. Total losses for all three herbicides each year were generally less than 2% of that applied. Because the herbicides used are not strongly adsorbed and have similar adsorption coefficients, over 95% of the runoff loss in each case was associated with runoff water. Depending primarily on runoff volumes, which in turn were dependent on the storm and the time of year, relative losses for notill were variable, sometimes being the greatest, sometimes the least; however, for the other three treatments, losses were usually in the order DS>MM>SD.