Low‐level wind profiles measured during Project Great Plains were used in the determination of z0, the roughness parameter, and β, the parameter related to profile curvature in Deacon's formula. Anemometers were geometrically spaced at five levels between 40 and 640 cm. The zo values, one for each observation period, found from the four lowest anemometers exceeded the ZQ values found from the four highest anemometers. The p values computed from the highest anemometers departed more from unity than did the p values computed from the lowest set. Since the apparent variability of p with height may be due entirely to the variability of z0, another analysis was conducted with the z0 value fixed for all levels and all hours of an observation period. Although the β variation was reduced by this technique, the departure of β from unity continued to increase with height.Since the Calder‐Deacon solutions of equations of diffusion from continuous point and infinite line sources are functions of β and z0 the effects of the height variation of β and z0 on predictions of ground‐level concentration were examined. The effects were found to be significant. When for each observation period, β was plotted against a modified form of the Richardson Number, the fitted curves did not coincide. There are indications that during inversion conditions the greatest departures from β = 1 occurred on nights characterized by the presence of a well‐defined low‐level wind maximum.