An intercomparison of zenith-sky UV-visible spectrometers was held at Camborne, UK, for 2 weeks in September 1994. Eleven instruments participated, from nine different European institutes which were involved with the Second European Stratospheric Arctic and Mid-latitude Experiment (SESAME) campaign. Four instruments were of the Systeme d'Analyse d'Observations Zénithales (SAOZ) type, while the rest were particular to the institutes involved. The results showed that the SAOZ instruments were consistent to within 3% (10 DU) for ozone and 5% for NO2. For ozone the results from these instruments agreed well with total ozone measurements by Dobson and Brewer spectrophotometers and integrated ozonesondes when the air mass factors for the SAOZ were calculated using the ozonesonde profiles. Differences of up to 10% in ozone and 30% in NO2 were found between different instruments. In some cases these differences are attributable to the different absorption cross sections used in the analysis of the spectra, but other discrepancies remain to be investigated. A prominent source of error identified in the campaign was uncertainty in the derivation of the amount of absorber in the reference spectrum, which can contribute an error of up to 3% (10 DU) in ozone and 1.5 × 1014 molecules cm-2 in NO2.