SUMMARY The offspring of parents belonging to the original Framingham Heart Study cohort, and

spouses of these offspring, were examined beginning in 1971. Cardiovascular examinations similar to those performed on the parents in 1952-1953 were given to offspring and their spouses. The mean blood pressure (BP) appeared to be lower in the female offspring-spouses than in the cohort mothers. Comparable mul-tivariate analyses on 3588 offspring and spouses of both sexes, ages 20-49 years, and 1842 parents, ages 30-49 years, confirmed the frequently reported importance of weight, heart rate, alcohol consumption, glucose and hematocrit or hemoglobin as independent positive correlates and smoking as an inverse correlate of both systolic and diastolic BP. In the offspring-spouse population, the additional measured variables of total serum proteins and plasma triglycerides added significantly to prediction of BP in both sex groups, and a less strik-ing negative association was found for serum phosphorus. The findings for serum proteins and triglycerides were supported in a separate population of middle-aged twin men. Even with the additional measured variables, only 28-34 % of the total BP variance is explained in these populations. EPIDEMIOLOGIC APPROACHES to the under-standing of the "mosaic " causation of hypertension have usually taken the form of cross-sectional studies that investigate possible associations betwee