Accurate measurements of net radiation are basic to all studies of the surface energy budget. In preparation for an energy budget experiment significant differences were found between direct and component measurement of net radiation, which prompted this investigation of their cause. The instruments involved were an all-black single-dome Fritschen-type net pyrradiometer, two Eppley model 8–48 pyranometers, and an Eppley model PIR pyrgeometer. Each had recently been calibrated. The accuracy of the component instruments was considered first. Comparisons of about one hour on each of three nights between the pyrgeometer and five empirical formulas showed that the average departure over all formulas from the pyrgeometer average was −1%. Other comparisons between the pyrgeometer and an infrared thermometer viewing the surface yielded similar results. Alternate shading and unshading of the pyrgeometer looking upward during daytime resulted in a formula that was used to correct the downward longwave radiation under clear skies. The correction is dependent on wind speed, in contrast to a recent paper showing negligible dependence, but is in accord with earlier findings. Based on the manufacturer's specifications, the pyranometer calibrations were considered to be within 2% of the World Radiation Reference. Thus a series of experiments was carried out using what were believed to be reasonably accurate component measurements of net radiation and measurements from the net pyrradiometer. The results showed that the sensitivity of the latter was less in the longwave band than in the shortwave band in agreement with findings of others. Speculating on possible further dependence of sensitivity to the upward and downward streams of radiation, a method was developed to determine the magnitude of the individual net pyrmdiometer components. A reflective double-shell hemispherical cup was affixed to the upward or downward face of the net pyrradiometer such that linear regression could be applied to simultaneous measurements from the net pyrradiometer, pyranometer, pyrgeometer, and the inner cup temperature, assumed to he at air temperature, to estimate the individual components. Although a substantial difference in shartwave sensitivity was computed using this method, the result was not definitive because of the limited number and the narrow range of longwave observations. Nevertheless, the method can be employed in the field to verify uniform sensitivity of a net pyrradiometer's sensing surfaces to shortwave and longwave radiation. The method may have particular application to Fritschen-type net pyrradiometers of recently improved design after extended field use.