Spacial smoothing by the radar beam as well as post-detection integration reduce the variability of the distribution of rainfall rate in space. It is shown that when radar data are compared with instantaneous point rainfall rate a random error and a bias are introduced by the smoothing. This could account for some of the difficulties in the hydrological use of radars. It is shown that when raingage data are smoothed in time there is an optimum smoothing time interval such that the random error and the bias are reduced to a negligible level. A method is suggested for the optimum comparison of radar and raingage data and the possibility of a determination of Z-R relationships from such comparisons is discussed.