Hydrogeologic studies of the southern Great Basin differ widely in their conclusions about the origin of several groups of major springs discharging at valley level from Paleozoic carbonate rocks. Delineation of watersheds feeding these springs is complicated by interbasin movement of ground water through areally extensive and highly fractured Paleozoic carbonate rocks and by the structural complexity of, and sparse subsurface hydrogeologic data for, the carbonate aquifer. The mean, median, and standard deviation of the deuterium content of ground water at, and the number of samples collected from, four of these major discharge areas and one major recharge area (Spring Mountains–Sheep Range) are given below. The deuterium data are expressed in δ D units (permil deviations from SMOW); the number of samples is given in parentheses. Pahranagat Valley: −113, −113, 1, (9); Spring Mountains–Sheep Range: −102, −102, 3, (12); Ash Meadows: −106, −107, 3, (15); Death Valley: −109, −109, 1, (4); Muddy River: −100, −100, 1, (5). A nonparametric test (Kolmogorov–Smirnov statistic) of the data for the first three areas indicates different deuterium populations at the 0.01 level of significance. The deuterium data indicate that Ash Meadows springs are fed by a mixture of recharge derived from the Spring Mountains–Sheep Range uplands, 30 to 60 mi east of the springs, with underflow from Pahranagat Valley and vicinity, 90 mi northeast of the springs. Mass balance considerations suggest that the underflow amounts to about 35 percent of the spring discharge. Underflow to Ash Meadows from Pahrump Valley, a source area proposed in the literature, is not supported by the data. The Death Valley springs may be fed, in part, by water derived from Ash Meadows. The Muddy River springs may originate as precipitation on the Spring Mountains–Sheep Range area. The suggested sources for the Ash Meadows and Death Valley springs are in agreement with sources previously identified by hydrogeologic and hydrochemical studies of the region. The source proposed for the Muddy River springs, however, is not supported by water-budget studies. A major assumption in utilization of deuterium as a natural tracer of regional ground-water flow is that the mean deuterium content of recharge to proposed source areas remained relatively constant during the residence time of water in the aquifer of interest. Available data are insufficient to test this assumption for the southern Great Basin region.