Stream-restoration projects using natural stream designs typically are based on channel configurations that can accommodate a wide range of streamflow and sediment-transport conditions without excessive erosion or deposition. Bankfull discharge is an index of streamflow considered to be closely related to channel shape, size, and slope (channel morphology). Because of the need for more information about the relation between channel morphology and bankfull discharge, the U.S. Geological Survey (USGS), in cooperation with the Montana Department of Transportation and the U.S. Department of Agriculture-Lolo National Forest, conducted a study to collect channel-morphology and bankfull-discharge data at gaged sites and use these data to improve current (2004) methods of estimation of bankfull discharge and various design-peak discharges at ungaged sites. This report presents channel-morphology characteristics, bankfull discharge, and various design-peak discharges for 41 sites in western Montana. Channel shape, size, and slope and bankfull discharge were determined at 41 active or discontinued USGS streamflow-gaging sites in western Montana. The recurrence interval for the bankfull discharge for this study ranged from 1.0 to 4.4 years with a median value of 1.5 years. The relations between channel-morphology characteristics and various design-peak discharges were examined using regression analysis. The analyses showed that the only characteristics that were significant for all peak discharges were either bankfull width or bankfull cross-sectional area. Bankfull discharge at ungaged sites in most of the study area can be estimated by application of a multiplier after determining the 2-year peak discharge at the ungaged site. The multiplier, which is the ratio of bankfull discharge to the 2-year peak discharge determined at the 41 sites, ranged from 0.21 to 3.7 with a median value of 0.84. Regression relations between bankfull discharge and drainage area and between bankfull width and drainage area were examined for three ranges of mean annual precipitation. The results of the regression analyses indicated that both drainage area and mean annual precipitation were significantly related (p values less than 0.05) to bankfull discharge.