Herbicides transported to surface waters by agricultural runoff are partitioned between solution and solid phases. Conservation tillage that reduces upland erosion will also reduce transport of herbicides associated with the solid phase. However, transport of many herbicides occurs predominantly in solution. Conservation tillage practices may or may not reduce transport of solution-phase herbicides, as this depends on the runoff volume. Reducing herbicide application rate is another approach to minimize off-site transport. Herbicide banding can reduce herbicide application rates and costs by one-half or more. Our objective was to compare herbicide losses in runoff from different tillage practices and with band- or broadcast-applied herbicides. The herbicides alachlor [2-chloro-2′,6′-diethyl-N-(methoxymethyl)acetanilide] and cyanazine [2-[[4-chloro-6-(ethylamino)-1,3,5-triazin-2-yl]amino]-2-methylpropionitrile] were broadcast- or band-applied to plots managed in a moldboard plow, chisel plow, or ridge till system. Herbicide concentration in runoff was largest for the first runoff event occurring after application and then decreased in subsequent events proportional to the cumulative rain since the herbicide application. When herbicides were broadcast-applied, losses of alachlor and cyanazine in runoff followed the order: moldboard plow > chisel plow > ridge till. Conservation tillage systems reduced runoff loss of herbicides by reducing runoff volume and not the herbicide concentration in run-off. Herbicide banding reduced the concentration and loss of herbicides in runoff compared with the broadcast application. Herbicide losses in the water phase averaged 88 and 97% of the total loss for alachlor and cyanazine, respectively. Cyanazine was more persistent than alachlor in the soil.