A creel survey was conducted from 1 July through 15 August 1994 on the Kenai River downstream of the Soldotna Bridge to estimate recreational angler effort, catch, harvest, and snag of sockeye salmon Oncorhynchus nerka. The creel survey area was divided into two strata: the Soldotna Bridge to the sonar counters (Stratum A), and the sonar counters to the Warren Ames Bridge (Stratum B). Recreational anglers exerted an estimated 53,844 angler-hours to harvest an estimated 11,624 sockeye salmon in Stratum A, and an estimated 63,204 angler hours to harvest an estimated 11,773 sockeye salmon in Stratum B. Most fish caught were retained. Total catch exceeded the harvest for both strata by only 5%-8%. The number of fish snagged was estimated to be 5,582 in Stratum A and 8,709 in Stratum B. The total inriver return (sonar estimate plus harvest estimate for Stratum A) was estimated to be 1,015,070 sockeye salmon. A second survey (a fishery survey) was conducted on the Kenai River from the Warren Ames Bridge to Kenai Lake from 15 July to 15 August 1994. The river was divided into three strata for the fishery survey: the downriver section was from the Warren Ames Bridge to the Soldotna Bridge (Strata A and B of the creel survey combined), the midriver section was the Soldotna Bridge to Skilak Lake, and the upriver section was Skilak Lake to Kenai Lake. Residency of anglers contacted in this survey was 17% Kenai Borough, 37% Alaska, 42% U.S., and 4% other. Most anglers started their fishing day between 0800 and 1200 hours. Anglers exiting the fishery from the upriver section tended to have longer fishing days than anglers exiting the downriver and midriver sections. Sixty-two percent of anglers harvested no fish, 14% harvested one fish, 7% harvested two fish, and 17% harvested three or more fish. Anglers had better success harvesting sockeye salmon in the midriver section than the upriver or downriver sections: over half of anglers in the midriver section harvested at least one fish while only about a third of the anglers in the upriver and downriver sections harvested at least one fish. In 1994, a three-fish bag limit reduced harvest by 17% but a bag limit of two or one would have reduced harvest by 23% or 53%, respectively. Angler success showed a positive relationship with the sonar counts. When fish passage exceeded the median sonar count (20,973) angler success was highest; below the median count angler success was lowest and any alteration of the bag limit would have had little effect on the harvest.