An experiment was conducted to compare the quality of snow-melt runoff from fresh or composted manure that had been spread on snow. Sixteen sheet metal trays held a layer of snow and either frozen composted or fresh manure. Each tray measured 400mm x 400mm x 100mm deep and had two holes to collect runoff from the surface and bottom of the tray. Trays were mounted on racks at a slope of 20 degrees and insulated with rigid polystyrene. Total depth of compacted snow in the trays was about 90 mm (6 kg). Composted and fresh manure was spread to give nutrient loading rates in accordance with available manure management guidelines. A factorial experimental design was used with two materials, four arrangements of snow with composted or fresh manure, two collection openings, two collecting periods, and two replicates of each. Trays were placed in a temperature-controlled room to simulate a natural freeze-thaw cycle. Daily temperatures were set at 8 degrees for 10 hours and -5 degrees C for 14 hours. Runoff from the snow surface acid from the bottom of trays was collected and analyzed for soluble carbon, total nitrogen, and BOD5. The mean BOD5 level of runoff from composted manure was 45.0 mg/L compared with 1834 mg/L from unprocessed manure. The snow-material sequence, representing application season, had a significant effect on all analysis parameters from runoff of fresh manure, but no effect on runoff quality of composted manure. Although there were no significant differences between BOD5 levels of early and late runoff for both fresh and composted manure and total nitrogen for composted manure, nitrogen, and soluble carbon levels from fresh manure and soluble carbon from composted manure were significantly higher in the early runoff. The mean BOD5 level and the concentrations of soluble carbon from runoff of fresh manure collected from upper opening locations were significantly higher than those from lower ones.