The 2-yr incidence of diabetes (131 cases) among 8369 Israeli men is related to area of birth and 15 other variables using a multiple logistic risk function. The variables found to be most strongly related to diabetes incidence are age, wt ht ratio and the presence of peripheral vascular disease or intermittent claudication. Also related to diabetes incidence were: cholesterol, education (negative relationship), being born in Eastern Europe (negative relationship) and being born in Africa. The model predicts markedly increased risk of diabetes incidence for persons with increased wt ht ratio and presence of peripheral vascular disease or intermittent claudication. The failure to find a relationship between the presence of ischemic heart disease and diabetes incidence suggests that the well known association between these two diseases lie only in the direction of increased risk of IHD for diabetics. None of the food variables studied was related to diabetes incidence. Other variables found unrelated to diabetes incidence in our study were diastolic blood pressure, uric acid, non-specific T-wave and an anxiety index.