Accumulated data from the Honolulu Heart Program cohort of men of Japanese ancestry were analysed for 19-year trends in coronary heart disease (CHD) mortality and incidence rates, case-fatality ratios, and changes in risk factor levels. From 1966 to 1984, there was an estimated average annual decline of 2.7% for CHD mortality rates for the population from which the cohort was drawn. This decline was similar to that of US white males of the same age distribution. Within the examined cohort of men who were free of prevalent CHD at entry into the study, the temporal pattern of incidence rates of definite CHD was quite similar to the pattern of CHD mortality rates. Thus, for the examined cohort, mortality rates were an accurate indication of incidence rates. There was no evidence that the decline in CHD mortality rates was due to reductions in case-fatality ratios. There were, however, significant decreases in the percentages of men who smoked cigarettes, in blood pressure level among treated hypertensives and in the intake of dietary cholestrol. These changes, and a drop in coronary artery atherosclerosis measured during the same time period, are all compatible with the concept that the decline of CHD mortality among Japanese men in Hawaii reflects primary prevention of new disease.