During 2002, Colorado experienced the State's worst drought since 1977. In 2003, the U.S. Geological Survey entered into cooperative agreement with the Colorado Department of Public Health and Environment to evaluate the general effects of drought on the water quality of streams in Colorado during summer 2002 by analyzing a water-quality data set obtained during summer 2002 in cooperation with a variety of State and local governments. Water samples were collected at 148 stream sites in Colorado and were measured or analyzed for field properties, major ions, nutrients, organic carbon, bacteria, and dissolved and total recoverable metals. Mean annual streamflow was analyzed at 134 sites in Colorado, and mean summer (July-September) streamflow for 2002 was determined for 146 sites for water years 1978-2002. Mean annual streamflow for 2002 had an average percentile of 29.4 and mean summer streamflow for 2002 had an average percentile of 7.6 relative to 1978-2002. These results indicate that streamflow in Colorado was substantially less than median streamflow for the period and that the effect of drought on streamflow was greater during summer 2002 than during water year 2002 (October 1, 2001, through September 30, 2002). Few measured constituent concentrations or values were elevated or depressed on a widespread basis during summer 2002. Specific conductance was elevated (in the upper quartile relative to historical data) in five of the seven basins that had sufficient data for characterization, indicating that specific conductance likely was affected by drought in those basins. Chloride concentrations were elevated in three of five basins with sufficient data and indicate that chloride concentration generally was affected by drought in those basins. Sulfate concentration was elevated in four of six basins with sufficient data. The widespread elevation of specific conductance and concentrations of chloride and sulfate indicates that salinity generally was affected by drought in Colorado streams during July-September 2002, likely because streamflow at most sites was dominated by base flow of ground water, which usually has substantially greater salinity compared to runoff from precipitation. Total-recoverable iron and manganese concentrations were depressed (in the lower quartile of historical data) in the Arkansas River Basin, which likely was due to reduced land-surface washoff of sediment containing oxyhydroxides of these metals. Of the 246 water samples collected at 148 sites during the summer of 2002, constituents in 115 exceeded Colorado water-quality standards. Constituents that exceeded water-quality standards were pH (all 9.0 standard unit exceedances; 9 samples), chloride (1 sample), sulfate (9 samples), dissolved ammonia (10 samples), dissolved nitrite nitrogen (3 samples), E. coli (Escherichia coli) bacteria (34 samples, 20 in Arkansas River Basin), fecal-coliform bacteria (18 samples, all in Arkansas River Basin), dissolved copper (1 sample), dissolved iron (3 samples), total-recoverable iron (3 samples), dissolved manganese (13 samples), dissolved selenium (10 samples), and dissolved zinc (1 sample). Of these 115 exceedances, historical data were sufficient to conclude that 21 probably were affected by drought, that 39 probably were not affected by drought, and that 55 were of indeterminate nature. Specific conductance indicates that the San Juan River Basin (average percentile 95.2) experienced the greatest effects of drought on water quality during summer 2002 compared to other basins in Colorado, followed by the Upper Colorado (90.0) and Dolores River (85.7) Basins. The South Platte River Basin (70.9) experienced the least effect of drought, and the Yampa and White River Basin group (73.7) had the second smallest effect. The Gunnison River (82.1) and Arkansas River (81.2) Basins had intermediate drought effects. The Rio Grande had insufficient data to rank the relative effect of drought on salinity.