A public health view of diet, atherosclerosis, and coronary disease is based on the congruence of evidence from the clinic, the experimental laboratory, and from population studies. It acknowledges that hyperlipidemia and atherosclerosis are the result of a host-environment interaction and recognizes that individual and population differences in susceptibility are insufficiently explained. But it is also based on an extension of the host-environment concept to the idea that mass disease, such as atherosclerosis and hypertension, is very likely the action of powerful cultural factors on widespread human susceptibility. This is confirmed by the high proportion of human populations having elevated blood lipids and atherosclerosis, seen in many societies, as well as the high proportion of other populations with adult hypertension (up to 50%). A corollary of this concept of host-environmental interaction in mass diseases is that a very unfavorable environment would assure the maximal exhibition of susceptible phenotypes in the population, while a very favorable environment is likely to assure the minimal exhibition of risk factors and disease. The evidence for effect of habitual diet on mass hyperlipidemia and atherosclerosis is illustrated in the Seven Countries Study. Whole populations are hypercholesteremic in comparison with others. Indeed, there is virtually no overlap in total cholesterol values between the south Japanese and the east Finns. The population diet and lipid values are highly correlated. In contrast, a model of individual diet-cholesterol relationships indicates the great importance of intrinsic factors which determine individual cholesterol levels and perhaps dietary responses. That model is extended to populations on the assumption of approximately equal distribution of polygenic susceptibility. When applied to the population diets, Keys, Grande, and Anderson's experiments on the cholesterol effects of different diet compositions give estimated means and distributions of serum cholesterol which coincide well with the observed values in all Seven Countries areas. This coincidence suggests that diet is the main determinant of population differences in blood lipid levels. If, in turn, relative mass hyperlipidemia is a primary factor in population rates of atherosclerosis, then a diet high in saturated fat and cholesterol would appear to be a necessary and possibly sufficient factor for mass atherosclerosis. Individual cases of coronary heart disease (CHD) may develop in low-incidence countries and are often unexplained. Nevertheless, rates of CHD are extremely low in populations with mean total serum cholesterol (TC) values of 180 mg/dl or lower, extremely high in populations with mean values of 230 or higher, and quite variable in the range of 180 to 230 mean values. Data from other studies suggest the lack of a strong correlation between habitual diet and population HDL levels and population rates of CHD. These findings confirm the idea that LDL cholesterol is the primary diet-responsive variable and pathogen. Finally, no population-based study exists, only anecdotal reports in small isolated primitive cultures, giving any substantial evidence to refute the primary and necessary role of habitual diet composition in mass atherosclerosis and population differences in CHD. The implications for public health strategy are clear. The medical model of an individual patient-doctor approach to risk assessment and prophylactic therapy is insufficient when the disease is so ubiquitous and insidious, when the causes are predominantly sociocultural, and when the therapy is nontraditional. Under these circumstances the high-risk medical model of care requires the support of a broader, community-wide educational system. This is necessary for the true primary prevention of atherosclerosis, i.e., prevention of elevated risk characteristics in the first place.