A cohort of more than 8,000 Japanese men living in Hawaii was studied for factors associated with blood pressure levels, with an emphasis on biological and sociocultural variables. Bivariate and multivariate analyses of more than 50 variables indicated that obesity, age, hematocrit, heart rate, forced vital capacity, serum triglyceride, serum uric acid, cigarette consumption, and family history of hypertension were independently associated with both cross-sectional levels and longitudinal changes in blood pressure. Serum glucose and alcohol consumption were associated with cross-sectional levels only. There was little evidence of association for specific dietary items, type of diet, diet changes, or use of salt. Furthermore, there was no support for the psychosocial hypotheses of elevated blood pressure from the stress of migration, acculturation, or status incongruity.