Eight thousand and six Japanese men born in the years 1900–1919 participated in the baseline examination of the Honolulu Heart Study. Of the survivors, 94.6 per cent participated in a second examination two years later. All but 23 of the second exam refusers were contacted about their health. These procedures identified 101 men who had developed coronary heart disease (CHD) for the first time during this two-year interval. This incidence was associated with elevated levels of serum cholesterol, blood pressure, relative weight, uric acid, and cigarette smoking. Considered by itself, non-fasting serum triglyceride was related to CHD incidence, but in the context of other (non-lipid) characteristics this apparent association was not significant. Cigarette smoking was related to CHD other than angina pectoris (AP) but not to AP alone, while the reverse was true for glucose intolerance. In this cohort, the association between relative weight and CHD was as strong as that between serum cholesterol and CHD.