A 3-year study was conducted to determine mortality rates for downstream migrant salmonids passing through spillways or turbines at McNary Dam. Mortalities were determined by releasing marked fish into the exit under study (spillway or turbine) and marked control groups immediately below the dam. Samples of released fish were recovered at stations below the dam, and survival rates were computed from the ratio of the number of experimental to control recoveries. Detailed techniques were devised to test assumptions involved in the estimating procedure. In 1955 and 1956 estimates were obtained for fingerling chinook salmon (Oncorhynchus tshawytscha) at McNary Dam. In 1957 yearling chinook were tested at Big Cliff Dam, a structure on the Santiam River with a turbine and spillway similar to those at McNary Dam. As a control measure, fingerling chinook were also tested in 1957. In all experiments the turbine was the major source of mortality. For each exit, no significant difference was found between mortalities at McNary Dam and Big Cliff Dam on fingerling chinook or between mortality rates for fingerling and yearling chinook. Estimated mortality rate of salmon passing over the spillway was 2 percent with a 95 percent confidence interval from 0 to 4 percent. Estimated mortality of salmon passing through the turbines was 11 percent with a confidence interval from 9 to 13 percent. The consistency of the results of the studies indicated that the estimates were valid representations of the mortalities experienced under the tested conditions.