Epidemiologic studies have documented a rapid rise in the mortality and incidence of malignant melanoma of the skin in many countries. This paper presents an analysis of malignant melanoma time trends in data from the Norwegian Cancer Registry, 1953-1978, by subsite and sex. The purpose of this study is to quantify previous qualitative analysis of the same data, mainly to distinguish birth cohort effects from time period effects. This analysis shows that time trends in malignant melanoma in Norway can be explained economically by a common age effect and a separate cohort effect for each subsite in both sexes or alternatively, by a common cohort effect but different age effects for each sex and subsite. In neither case is a time period effect required. Mathematical modelling can never fully explain any situation exactly and it is more than likely that the increases observed have been produced by a measure of both extreme descriptions. The different cohort effects and age effects are discussed in terms of their implications for etiology.