The histories of exposure to sun through occupational, recreational and vacation activities of 595 patients with newly incident cutaneous melanoma excluding lentigo maligna and acral lentiginous melanoma, were compared to those of comparison subjects drawn randomly from the same population and matched for age, sex and province of residence in Western Canada. Significant increases in risk were seen with increasing amount of sun exposure through outdoor activities associated with recreation and vacations; activities likely to involve more intense sun exposure were associated with greater increases in risk. While a moderate amount of occupational exposure was associated with increased risk, greater occupational exposure resulted in no further increase; in men a decrease in risk was seen. These findings were independent of the effects of hair and skin colour, freckles, ethnic origin and socio-economic status. The results suggest that short-term exposure to unusually intense sunlight increases the risk of melanoma, while long-term constant exposure has no effect or may decrease risk. No simple relationship was seen between melanoma risk and total sunlight exposure. This study introduces new methods of assessing different types of sun exposure from retrospective data.