Most comparisons of TOMS estimates of surface UV irradiation with measured values from ground-based instruments have indicated a bias of the TOMS estimates toward larger values. A portion of this bias results from absolute uncertainties in the ground-based instruments. The comparison reported here is based on ground-based data from four sites in the UGA/EPA Brewer network. The raw data from the ground-based instruments has been corrected for (1) stray light rejection, (2) the cosine errors associated with the full sky diffuser, (3) the temperature dependence of the response of the instruments and (4) the temporal variation in the instrument response reducing the estimated errors of the absolute irradiance values of each spectral measurement to < ±7%. Comparisons of TOMS with the surface measurements are performed both at spectrally resolved wavelengths at the time of overpass and for erythemally-weighted daily-integrated doses. These comparisons are made for all days and for clear-sky days only. The comparisons are carried out using both linear regressions of scatter plots of the two sets of data and for mean differences with respect to both TOMS and the Brewer measurements. It is found that spectrally resolved comparisons suffer from inconsistencies at some of the sites that are believed to result from wavelength uncertainties in the Brewer; they are therefore of more limited use than wavelength integrated data. A comparison based on daily-integrated doses shows only a small positive TOMS bias (4%) for clear-sky days with a somewhat larger bias (8%) for data taken from all days.