The associations among obesity, height, cardiovascular risk factors, and the incidence of clinical diabetes mellitus were investigated in the Norwegian population-based Finnmark Study of 11,654 men and women aged 35-52 years at baseline in 1977-1978. A total of 87 cases of diabetes among men and 75 cases among women were registered during 12 years of follow-up. The incidence of diabetes was 1.1 per 1,000 person-years in women and 1.2 per 1,000 person-years in men, but sex-related differences in risk factors were noted. Body mass index was the dominant risk factor in men and predicted diabetes in a dose-response relation in both sexes. However, in women, the association between body mass index and diabetes was greatly attenuated after multivariable adjustment. Serum lipid concentrations were similar in prediabetic men and women; thus, prediabetic women had a relatively more adverse metabolic risk profile as compared with nondiabetics of the same sex. In multivariable analysis, high density lipoprotein cholesterol was inversely related to diabetes in women (relative risk per 0.3 mmol/liter, 0.53; 95% confidence interval 0.41-0.70) but not in men (relative risk, 0.97; 95% confidence interval 0.78-1.19). Serum glucose was a highly significant predictor in both sexes, while height was inversely related to diabetes only in women (relative risk per 5 cm, 0.71; 95% confidence interval 0.58-0.87).