Record high near-surface ozone concentrations at two elevated sites (Chaumont, 1140 m asl, and Rigi, 1030 m asl) in Switzerland were observed simultaneously with extremely low total ozone during a fair weather period in mid-February 1993. An analysis of ozone, temperature, humidity, and wind profiles suggests that the surface ozone peaks were most possibly generated within the region in a layer between about 1000 and 1500 m asl. Mean diurnal cycles of ozone concentration during the period shows a strong increase from late morning to late afternoon at Chaumont and at the same time a decrease at the high alpine site Jungfraujoch (3580 m asl). The different diurnal ozone cycles can both be explained photochemically by taking into account the large difference in NOx concentrations (about two orders of magnitude) between the sites. Photochemical processes are also indicated by the diurnal cycles of NO2 and NO concentration. As a strong photochemical activity is not expected in mid-February at 47°N, we hypothesize that the extremely low total ozone played a role. Total ozone controls the amount of UV-B radiation reaching the troposphere and thus influences photochemical processes. Using a radiation model, we calculated an increase in ozone photolysis at Chaumont and Jungfraujoch of 73% and 83%, respectively, on the day with the lowest total ozone (243 DU) compared to average February conditions (335 DU). It is suggested that total ozone changes have the potential to stimulate photochemistry sufficiently to produce the observed surface ozone peaks at Chaumont and Rigi of 61 and 64 ppbv, respectively. A fog layer just below Chaumont during these days probably also influenced photochemistry, but on a smaller spatial scale. Our empirical results on the influence of changing UV radiation on tropospheric photochemistry are in close agreement with model studies of other groups. Although this case study represents unique conditions, a distinct anticorrelation between near-surface ozone at Chaumont and total ozone also appears in other years (1992-1997) when selecting fair weather days in mid-February. However, other influences cannot be excluded. The selected days provide evidence of a significant photochemical source of ozone in the mid-latitude lower troposphere in late winter.