Passage of chinook salmon Oncorhynchus tshawytscha in the Kenai River was estimated using side-looking dual-beam sonar technology from 1987 through 1994. Sockeye salmon O. nerka migrate concurrently with chinook salmon and are far more numerous. To exclude as many sockeye salmon as possible from counts of chinook salmon only those targets which exceeded minimum-range criteria and target-strength thresholds were counted. In 1995, in addition to normal procedures for estimating fish passage, the dual-beam system was replaced with a split-beam system and three ancillary experiments were conducted to evaluate the accuracy of chinook salmon passage estimates. The first study used spatial location data from the split-beam sonar to determine direction of travel for each target and estimate the proportion of fish migrating downstream by day, bank, and tide stage. In the second study, live fish of known size and species were tethered in front of the transducer to evaluate the effectiveness of the target strength threshold and determine whether there were other measurable acoustic parameters that could be used instead of or in conjunction with target strength to separate chinook salmon from other species. The third study used drift gillnets to address the question of whether there is sufficient spatial separation between chinook and sockeye salmon to distinguish between the species. Downstream-migrants comprised 12% and 5% of fish during the early and late runs respectively, with greater downstream proportions on the left bank than on the right bank. The proportion of downstream fish differed little by tide stage. The relationship between mean target strength and length of tethered fish was imprecise (r2 = 0.23, P < 0.0001) and distributions of mean target strength overlapped almost completely between species and between length classes. Tethered chinook salmon had a mean target strength of -24.8 dB (SD = 2.5), sockeye salmon averaged -25.5 dB (SD= 2.0) and coho salmon averaged -26.1 dB (SD = 1.1). Fish shorter than 650 mm (n = 55) averaged -25.9 dB (SD = 1.9), whereas fish longer than 650 mm (n = 38) averaged -24.2 dB (SD = 2.3). Most sockeye salmon tethered in front of the split-beam sonar had mean target strengths exceeding the target strength threshold of -28 dB. Target strength measurements were sensitive to fish movement, position, and transducer aim. Other acoustic parameters based on pulse duration showed potential as size and species discriminators. Substantial numbers of sockeye salmon were present in mid-river during the study period, although relative proportions of sockeye and chinook salmon could not be estimated due to net selectivity biases. We conclude that chinook salmon passage estimates include some sockeye salmon, and further ground-truthing of sonar estimates with an independent abundance estimate is recommended.