Essential features of the epidemiology and photobiology of cutaneous malignant melanoma (CMM) in Norway were studied in comparison with data from countries at lower latitudes. Arguments for and against a relationship between ultraviolet radiation (UV) from sun and sun beds are discussed. Our data indicate that UV is a carcinogen for CMM and that intermittent exposures are notably melanomagenic. This hypothesis was supported both by latitude gradients, by time trends and by changing patterns of tumor density on different body localizations. However, even though UV radiation generates CMM, it may also have a protective action and/or an action that improves prognosis. The same may be true for a number of internal cancers. There appears to be no, or even an inverse latitude gradient for CMM arising on non-UV exposed body localizations (uveal melanoma). Furthermore, CMM prognosis was gradually improved over all years of increasing incidence (up to 1990), but during the last 10 to 15 years, incidence rates decreased and prognosis was not further improved. While CMM incidence rates are twice as high in South Norway as in North Norway, the ratios of death rates to incidence rates are higher in the North, where the annual UV fluences are lower. Death- and incidence rates in Australia and New Zealand fully support this. Comparisons of skin cancer data from Norway and Australia/New Zealand indicate that squamous cell carcinoma and basal cell carcinoma are mainly related to annual solar UVB fluences, while UVA fluences play a larger role for CMM.