Whereas gerontologists usually regard ageing as an intrinsic process that increases the risk of death, investigators of coronary heart disease (CHD) tend to believe that its causes are largely environmental. If the latter view is correct and if the same is true of many other fatal diseases (as is commonly alleged), then gerontologists should not regard the age pattern of mortality in a population as an index of ageing. An attempt is made in this paper to assess the roles of intrinsic and extrinsic factors in the aetiology and pathogenesis of CHD. The analysis is carried out in terms of the author’s theory of age-dependent autoaggressive disease although the conclusions are largely independent of that theory. Two parameters of the mathematical formulation of the theory prove to be important in this context: S, the fraction of the study population that is genetically predisposed to fatal CHD, and λ, the average latent period between the end of the intrinsic stochastic process of initiation and death from CHD. From published studies it is shown that the ‘risk factors’ cigarette smoking, high relative weight and (probably) hypercholes-terolaemia associate only with λ and not with S. High levels of each of these factors associate with low values of λ. However, the ‘risk factors’ lack of exercise, hypertension and diabetes mellitus associate with both S and λ. The associations with S probably have a straightforward genetic interpretation: persons genetically predisposed to lack of exercise, hypertension and diabetes mellitus are more likely – at least in various Caucasian populations – to be genetically predisposed to CHD than persons in the general population. Our main problem, therefore, is to decide whether the negative associations between the level of a risk factor, and the average duration, λ, of the latent period, should be interpreted in ‘causal’ and/or genetic terms. Secular (temporal) trends in sex-specific and age-specific death rates from CHD in England and Wales, from 1921 to 1973, are illustrated and they give no indication of any appreciable shortening in λ – in spite of increases in some major risk factors. The average latent period remains at about 10 years for men and 20 years for women throughout the period. A wide range of other pertinent epidemiological evidence is reviewed, all of which is consistent with the genetic interpretation and much of which is inconsistent with the ‘causal’ interpretation of the associations between the classical risk factors and λ. (There is no doubt, however, that high death rates from CHD immediately following extremes of environmental temperature should be attributed to a causal mechanism.) Theory predicts that, at high levels, the classical risk factors will not be independent. Recent epidemiological studies confirm this prediction. It is improbable that the classical risk factors make any appreciable causal contribution to the pathogenesis of CHD, but even if they do, the intrinsic biological process of initiation dominates the age dependence of CHD. Similar considerations extend to other major fatal diseases. In regarding the age pattern of mortality in medically advanced countries as resulting predominantly from the phenomenon of biological ageing, gerontologists are unl