In 2003, the U.S. Geological Survey, in cooperation with the Colorado Water Conservation Board, Upper Colorado River Endangered Fish Recovery Program, Colorado River Water Conservation District, Colorado Division of Water Resources, and Bureau of Reclamation, initiated a study to characterize streamflow losses along a reach of the Gunnison River from the town of Whitewater downstream to the Redlands Canal diversion dam. This describes the methods and results of the study that include: (1) a detailed mass-balance analysis of historical discharge records that were available for the three streamflow-gaging stations along the study reach; and (2) two sets of discharge measurements that were made at the three stations and at four additional locations. Data for these existing streamflow-gaging stations were compiled and analyzed: (1) Gunnison River near Grand Junction (Whitewater station); (2) Gunnison River below Redlands Canal diversion dam (below-Redlands-dam station); and (3) Redlands Canal near Grand Junction (Redlands-Canal station). Data for water years 1995-2003 were used for the mass-balance analysis. Four intermediate sites (M1, M2, M3, and M4) were selected for discharge measurements in addition to the existing stations. The study reach is the approximate 12-mile reach of the Gunnison River from the Whitewater station downstream to the Redlands Canal diversion dam, which is about 3 miles upstream from the confluence with the Colorado River. For the mass-balance analysis, differences between the sum of the annual cumulative daily mean discharge at the two downstream stations and the annual cumulative daily mean discharges at the upstream station ranged from about -8,700 to -69,800 acre-feet (about -.8 to -1.1 percent), indicating that the downstream discharges generally were less than the upstream discharges. Moving 3-day daily mean discharge averages also were computed for each of the three stations to smooth out some of the abrupt differences between the downstream and upstream daily mean discharges. During water years 1995-2002, differences between the downstream and upstream moving 3-day daily mean discharges ranged from about -200 to +100 cubic feet per second (ft3/s) during one-half of each year, but the differences had absolute values as large as about 500 to 1,000 ft3/s during the other one-half of the year. The differences as a percentage of the upstream discharge ranged from 0 to -10 percent within the interquartile range and were as small or large as about -60 to +50. Two sets of discharge measurements were obtained during water year 2003. For measurement set 1 (February 5-6), discharge was measured 5-8 times over a 24-hour period at sites M1-M4, where measured discharges ranged from 527 to 608 ft3/s. Discharge was measured once each day at the Whitewater and below-Redlands-dam stations to verify discharge rating shifts; the Redlands Canal was not in operation at this time, so measurements were not needed at the Redlands-Canal station. Recorded 15-minute (unit) discharges ranged from about 575 to 615 ft3/s at the Whitewater station and from about 560 to 600 ft3/s at the below-Redlands-dam station during the February 5-6 period. Because of the inherent error in discharge measurements (5 percent for measurements rated good), and because the mean discharge at the below-Redlands-dam station, about 580 ft3/s, was only about 2.5 percent smaller than the mean discharge at the Whitewater station, about 595 ft3/s, it is concluded that there was no measurable streamflow loss along the study reach during measurement set 1. For measurement set 2 (May 14-15), discharge in the Gunnison River was about 2,000 ft3/s and increasing because of high-elevation snowmelt. Five discharge measurements were made at site M2 and discharge ranged from 1,668 to 2,117 ft3/s. Measured discharges at the gaging stations were 2,730 ft3/s at the Whitewater station, 1,268 ft3/s at the below-Redlands-dam station, and 819 ft3/s at the