A water-quality standard for temperature is critical for the protection of threatened and endangered salmonids, which need cold, clean water to sustain life. The Idaho Department of Environmental Quality has established temperature standards to protect salmonids, yet little is known about the normal range of temperatures of most Idaho streams. A single temperature standard for all streams does not take into account the natural temperature variation of streams or the existence of naturally warm waters. To address these issues and to help the Idaho Department of Environmental Quality revise the existing State temperature standards for aquatic life, temperature data from more than 200 streams and rivers in the salmon and Clearwater River Basins were collected. From these data, a statistical model was developed for estimating stream temperatures on the basis of subbasin and site characteristics and climatic factors. Stream temperatures were monitored hourly for approximately 58 days during July, August, and September 2000 at relatively undisturbed sites in subbasins in the Salmon and Clearwater River Basins in central Idaho. The monitored subbasins vary widely in size, elevation, drainage area, vegetation cover, and other characteristics. The resulting data were analyzed for statistical correlations with subbasin and site characteristics to establish the most important factors affecting stream temperature. Maximum daily average stream temperatures were strongly correlated with elevation and total upstream drainage area; weaker correlations were noted with stream depth and width and aver-age subbasin slope. Stream temperatures also were correlated with certain types of vegetation cover, but these variables were not significant in the final model. The model takes into account seasonal temperature fluctuations, site elevation, total drainage area, average subbasin slope, and the deviation of daily average air temperature from a 30-year normal daily average air temperature. The goodness-of-fit of the model varies with day of the year. Overall, temperatures can be estimated with 95-percent confidence to within approximately plus or minus 4 degrees Celsius. The model performed well when tested on independent stream-temperature data previously collected by the U.S. Geological Survey and other agencies. Although the model provides insight into the natural temperature potential of a wide variety of streams and rivers in the Salmon and Clearwater River Basins, it has limitations. It is based on data collected in only one summer, during which temperatures were higher and streamflows were lower than normal. The effects of changes in streamflow on the effectiveness of the model are not known. Because the model is based on data from minimally disturbed or undisturbed streams, it should not be applied to streams known to be significantly affected by human activities such as disturbance of the streambed, diversion and return of water by irrigation ditches, and removal of riparian vegetation. Finally, because the model is based on data from streams in the Salmon and Clearwater River Basins and reflects climatological and landscape characteristics of those basins, it should not be applied to streams outside this region.