Spectral measurements of the solar UV irradiance at the earth’s surface depend critically on the characteristics of the spectroradiometers and the procedures used in their calibration. Two UV spectroradiometers were subjected to a series of laboratory tests to investigate which factors were most significant in limiting the reliability of the absolute irradiance measurements. Three independent standards of spectral irradiance were scanned by both instruments under a range of bench conditions. The results were consistent to within about 3%, most of the uncertainty being due to scattered light in the laboratory. An UV laser was used to determine the slit function of each spectrometer and the influence of internal stray light. Significant departures from the ideal cosine and azimuth responses were measured by a xenon lamp. Both spectroradiometers were kept indoors throughout the experiments. The relevance of these laboratory results is discussed with respect to the task of measuring solar UV radiation in the field.