1. An association between Na and blood pressure (BP) was studied in the Cardiovascular Diseases and Alimentary Comparison (CARDIAC) Study, an international cross-sectional epidemiological study, involving 55 populations from 25 nations. 2. The partial regression coefficient of BP against 24 h urinary Na was calculated in a multiple linear regression model, by adjusting other confounding factors. This demonstrated different distribution in each of several groups stratified according to their ethnicity and geographical location. 3. The relationship of urinary K to BP regression against urinary Na excretion was investigated to clarify its role in the association of Na with BP. 4. Urinary Na/K ratio showed a significant inverse linear relationship with the regression coefficient of BP on urinary Na excretion among seven Japanese populations. 5. The effect of K was to increase urinary Na excretion, as the ratio of Na/K might modify the association between Na and BP.