A large cross sectional study, the Scottish Heart Health Study, of 10,359 men and women from 22 districts of Scotland was undertaken to try to explain the geographical variation of coronary heart disease mortality. Analysis by district showed that of the classic risk factors only cigarette smoking was strongly associated with heart disease mortality among both men and women. Mean diastolic blood pressure was weakly associated with rates among men and high density lipoprotein cholesterol showed a strong negative association among women. Total cholesterol showed a weak negative association with heart disease mortality, but, because the serum concentrations of cholesterol were uniformly high in all districts, a strong association with mortality would not be expected. In both men and women many dietary factors showed moderate or strong associations with mortality from coronary heart disease in a district--of these a low consumption of vitamin C was most notable. Other factors associated with heart disease included alcohol consumption and serum triglycerides among men, and obesity, physical activity, and serum triglycerides among women. Many factors associated with heart disease showed strong intercorrelations. Clustering of risk factors (including smoking, alcohol, and diet among men, and smoking, diet, and obesity among women) was associated with much of the regional variation in heart disease mortality in Scotland.