The reach of the Colorado River from the Gunnison River confluence to the Utah Border, and tributaries in the Grand Valley, are on the State of Colorado 303(d) list of impaired water bodies because the concentrations of dissolved selenium in these streams exceed the State of Colorado chronic standard of 4.6 micrograms per liter at the 85th percentile level. In response to concerns raised by a local watershed initiative about the issue of selenium in the Grand Valley, the U.S. Geological Survey, in cooperation with Mesa County and the City of Grand Junction, developed a study to characterize and determine the sources of selenium and how these sources are related to changes in land use. This report describes the methods and results of a study of concentrations and loads of selenium in three tributaries to the Colorado River in the Grand Valley. The study area consists of three subbasins, Persigo Wash, Adobe Creek, and Lewis Wash, each representing transitional agricultural to residential, agricultural, and residential land-use types, respectively. These subbasins represent different land-use types and the tributaries that drain each subbasin contribute moderate to high concentrations and loads of selenium to the Colorado River. Two synoptic-sampling events were conducted in each tributary to characterize variations in water quality during the nonirrigation season. Water samples were collected for analysis of dissolved selenium, total nitrogen, and total dissolved solids (salinity). Streamflow was measured by either the tracer-dilution or standard current-meter method. In Persigo Wash selenium concentrations generally decreased or remained constant in a downstream direction whereas selenium loads increased. Effluent from the Persigo Wash wastewater treatment plant diluted selenium concentrations in Persigo Wash and increased the selenium load. The concentrations and loads of salinity and total nitrogen generally increased downstream in Persigo Wash. Concentrations and loads of selenium correlated well with concentrations and loads of total nitrogen (R2 = 0.80 and 0.83, respectively). Concentrations and loads of total nitrogen also correlated well with streamflow (R2 = 0.89 and 0.99, respectively). In Adobe Creek concentrations and loads of selenium generally increased downstream. The largest selenium loads in Adobe Creek were observed between a 1.6-mile-long reach extending approximately from the Grand Valley Canal to the Main Line Grand Valley Canal, where selenium load increased 0.72 pounds per day. This reach accounted for about 81 percent of the total selenium load at the mouth of Adobe Creek (site AC1). Results from the synoptic sampling in Adobe Creek indicated that there was very little seasonal variation in selenium concentration during the nonirrigation season. Salinity concentrations were more variable than selenium concentrations during the nonirrigation season. The concentrations and loads of salinity and total nitrogen generally increased downstream. Concentrations and loads of selenium correlated well with concentrations and loads of total nitrogen (R2 = 0.89 and 0.98, respectively). Streamflow also was related to concentrations and loads of total nitrogen; results indicated a fair correlation for concentration (R2 = 0.51) and a good correlation for load (R2 = 0.95). In Lewis Wash concentrations and loads of selenium generally increased downstream. Selenium concentrations measured in Lewis Wash were lower than those measured in Persigo Wash or Adobe Creek. Salinity concentrations were similar to those measured in Persigo Wash and Adobe Creek. Salinity concentrations were similar among sites during each synoptic-sampling event. Salinity loads in Lewis Wash were highest during the beginning of the nonirrigation season. Concentrations and loads of total nitrogen generally increased downstream. There was a fair correlation for selenium and total nitrogen concentration (R2 = 0.71). Step-trend analysis