BACKGROUND Phototherapy units should regularly use hand-held ultraviolet (UV) meters to assess the output of treatment lamps, and these meters should be accurately calibrated. Several medical physics departments in the U.K. can calibrate UV meters traceable to national standards, but there is concern that there may be disagreement among departments. In particular, there may be difficulty in calibration for narrow-band UVB phototherapy lamps (TL-01). OBJECTIVE To ascertain the level of agreement in UV meter calibration at expert centres in the U.K., and to survey methodology at these centres, consider sources of errors and to make recommendations on calibration methods. METHODS The same UV meter with two detectors (for UVA and UVB) was calibrated by seven medical physics departments. A questionnaire on methods was also distributed and measured spectral outputs from each centre were examined. RESULTS The calibration factors for the meter varied by +/- 18% for the UVA detector and by +/- 60% for the UVB detector (2 standard deviations). Six centres performed calibration using a spectroradiometer and one centre used a reference meter method. The spectra of lamps used for calibration were similar. For the spectroradiometric methods there were some differences in methodology and instrumentation that may account for the differences in calibration factors. CONCLUSIONS UV meter calibration in the U.K. shows unacceptable variability, particularly for TL-01 lamps. An accuracy of around of +/- 10% would be clinically acceptable and should be technically achievable.