Current scientific evidence indicates that stratospheric ozone has declined worldwide over the past 20 years. The trend estimates are markedly dependent on the geographical location and are highly seasonal. Winter trends are much more negative than those for summer and autumn. Projections based on current assumptions of chlorine release suggest that this decline will continue into the next century. On the basis of the decrease in ozone over the mid-latitudes, an increase in biologically effective ultraviolet radiation (UVR) of 4%-9% is expected, depending on the season and geographical location. However, the UVR penetration to the Earth's surface is greatly affected by clouds, aerosols and tropospheric ozone, and current increases, if any, have not been as large as this. Direct evidence for the induction of non-melanoma skin cancer (NMSC) due to UVR has been derived from animal experiments in mice and rats. Numerous epidemiological data confirm that this relationship also holds for human skin. The increase in NMSC incidence in the past two decades is not likely to be due to the decrease in ozone, given the long latency (two to three decades) associated with UVR effects on skin. A knowledge of the action spectrum for NMSC development suggests that a 1% depletion in stratospheric ozone may be expected to increase NMSC, at equilibrium, by about 2.0% The evidence on the role of UVR exposure in the development of malignant melanoma (MM) is less certain. It has been estimated that a 1% reduction in ozone may cause an increase in MM of 0.6%.