Chlorsulfuron was applied conventionally in the fall and over snow-covered winter wheat in an aqueous suspension with graphite powder and urea ammonium nitrate in February and March to determine downslope herbicide movement in runoff. Plots were established in fields with slopes of 8 to 14%. Snow depth during winter application ranged from 20 to 60 cm. Soil samples collected in the spring 0.6, 0.9, 1.5, 2.4, 3.7, 5.2, 7, and 9.1 m downslope from each treated plot were analyzed using a lentil bioassay technique to determine the presence of herbicide. Chlorsulfuron residues were detected as far as 3.7 m down-slope from plots. Less downslope herbicide movement occurred when chlorsulfuron was applied with graphite over snow than when applied as a conventional fall treatment. If herbicide is applied over snow without graphite, significant amounts may be lost as a result of erosion.