Techniques for estimating magnitude and frequency of flood-peak discharges, based on channel geometry parameters, are presented as a method for evaluating Ohio floods in geographic areas deficient in flow data. One hundred and sixty gaging stations were selected on streams with alluvial channels unaltered by man. Exposed bedrock is absent from the channels in the vicinity of the surveyed cross sections. Peak stream discharges at the selected sites were related to channel geometry parameters. Active channel width is the only statistically significant parameter. Channel geometry parameters which were tested and found to be statistically insignificant are active channel average depth, bankfull width, and bankfull average depth. The average standard errors of estimate for the resulting statewide equations, with recurrence intervals of 2, 5, 10 25, 50 and 100 years, range from 42 to 55 percent.