Current-meter measurements of high accuracy will be required for calibration of an acoustic flow-metering system proposed for installation in the Sacramento River at Chipps Island in California. This report presents an analysis of the problem of making continuous accurate current-meter measurements in this channel where the flow regime is changing constantly in response to tidal action. Gaging-system requirements are delineated, and a brief description is given of the several applicable techniques that have been developed by others. None of these techniques provides the accuracies required for the flowmeter calibration. A new system is described--one which has been assembled and tested in prototype and which will provide the matrix of data needed for accurate continuous current-meter measurements. Analysis of a large quantity of data on the velocity distribution in the channel of the Sacramento River at Chipps Island shows that adequate definition of the velocity can be made during the dominant flow periods--that is, at times other than slack-water periods--by use of current meters suspended at elevations 0.2 and 0.8 of the depth below the water surface. However, additional velocity surveys will be necessary to determine whether or not small systematic corrections need be applied during periods of rapidly changing flow. In the proposed system all gaged parameters, including velocities, depths, position in the stream, and related times, are monitored continuously as a boat moves across the river on the selected cross section. Data are recorded photographically and transferred later onto punchcards for computer processing. Computer programs have been written to permit computation of instantaneous discharges at any selected time interval throughout the period of the current meter measurement program. It is anticipated that current-meter traverses will be made at intervals of about one-half hour over periods of several days. Capability of performance for protracted periods was, consequently, one of the important elements in system design. Analysis of error sources in the proposed system indicates that errors in individual computed discharges can be kept smaller than 1.5 percent if the expected precision in all measured parameters is maintained.