The relationship between tumour incidence rate and habitation patterns was investigated in 3289 patients registered between 1959 and 1968 as cutaneous malignant melanoma in the Swedish Cancer Registry. A linear correlation analysis between latitude and melanoma incidence rate in various regions showed a coefficient, r, was -0.74, which implies a decreasing incidence with increasing latitude. This result supports the hypothesis that ultraviolet irradiation is the predominant cause of melanoma. However, considerable deviations from the regression line were seen in some regions. Moreover, a comparison between town and country indicated melanoma incidence increased with population density, an increase not explained by overdiagnosis and thus not in agreement with the working hypothesis of UV-irradiation as the predominant cause of melanoma. Increase of foreign travel, as estimated by passport issue, may explain the overrepresentation of melanoma in some regions namely city and county of Stockholm, city of Malm and Gteborg county. As foreign travel in Sweden generally means sunshine trips, this town versus country paradox may also be explained on the basis of increased irradiation, as may the anomalies observed in some counties. A regression analysis of the epidemiological index for UV-irradiation and melanoma incidence, adjusted for foreign travel, demonstrated a close agreement with the results of the latitude gradient analysis.