Two approaches are examined for the removal of tidal currents from acoustic Doppler current profiler (henceforth ADCP) measurements. The first uses predictions from a numerical model (Foreman and Walters, 1990) whose domain includes the measurement region. Although the numerical model is barotropic, adjustments are made to correct for baroclinicity in accordance with historical observations from stationary current meters in the region. The second approach (originally proposed by Candela et al. (1990)) performs a least squares polynomial or biharmonic spline fit to the ADCP data in order to determine the spatially varying amplitudes and phases of major tidal constituents and the magnitude of the residual flow. Both tide removal techniques are applied to ADCP data taken during a 3-day cruise off the southwest coast of Vancouver Island, and their relative effectiveness is compared. Some of the detided currents are seen to compare well with geopotential surfaces and clearly show the Juan de Fuca Eddy (Freeland and Denman, 1982). A discussion of the results and limitations of each technique is also presented. ¿American Geophysical Union 1991