Annual sampling in locks with rotenone has been a principal means of assessing fish populations in the commercially navigable portions of the Ohio River. Despite extensive use, sampling in locks with rotenone and interpretation of the data obtained have not been adequately evaluated. The purpose of our study was to determine the degree of inter- and intraseasonal variation in lock samples, estimate correction factors (CFs) for fish recovery rates, and compare lock samples to other fish collections from the navigational pools above and below the lock. Lock samples from all seasons had a greater proportion of pelagic and demersal fish than samples from the navigational pools, which contained greater proportions of littoral species. CF for non-recovery of fish were determined. Spring and summer lock collections yielded several more species and estimates of overall fish biomass were an order of magnitude higher than fall collections. Within season variation between lock samples was relatively low; however, variation in lock samples among seasons was high, equivalent to that seen among the annual samples from the 1980s. Thus, single-season sampling may not be adequate, and fall may be the least preferred season.