Basic data from 67 gravel-bed river reaches in Alberta, Canada, are used to test widely adopted equations for computing average velocity in natural channels for the case of relatively high in-bank flows. Relations proposed by Cowan, Strickler, and Limerinos to estimate Manning's n are tested. The equations of the form presented by Keulegan and Lacey for computing average velocity are also evaluated. Computations based on these methods indicate that the Limerinos-Manning equation is the most acceptable approach for estimating average velocity in gravel-bed river reaches. The Lacey equation provides satisfactory results if no bed material data are available. The basic data are also used to develop best-fit relationships for computing Manning's n and the friction factor for gravel-bed river reaches. Particular emphasis is placed on an evaluation of the significance of bed material data in equations used to compute average velocity in gravel-bed river reaches.