The usefulness of whole-day, daytime (waking to retiring time) and overnight urine samples for assessing the relationship between blood pressure and sodium intake was examined in 301 male London civil servants, aged from 37 to 58 years old. Systolic blood pressure (SBP)/diastolic blood pressure (DBP) averaged 126/78 mmHg and the 24-h urinary excretion of sodium and potassium was 174 and 73 mmol, respectively. There was poor consistency between day- and night-time urine samples with respect to both sodium and potassium content. The urinary excretion of sodium and potassium was lower (P less than 0.001) in overnight than in daytime samples. After standardization for creatinine, the night: day ratio was 0.79 for sodium output and 0.55 for potassium excretion. Blood pressure, adjusted for age and body mass index, was significantly and positively correlated with overnight sodium excretion (SBP/DBP: slope = 0.061/0.046 mmHg/mmol) whereas the correlations with sodium excretion in daytime (0.010/0.004 mmHg/mmol) and whole-day (0.024/0.016 mmHg/mmol) urine samples were not significant. Blood pressure was significantly correlated with the sodium:potassium ratio in whole-day urine (1.941/1.968 mmHg/unit). As the agreement between daytime and overnight urine samples was low with respect to both sodium and potassium content, and due to the fact that the relationship between blood pressure and sodium in overnight samples may at least partially reflect pressure diuresis, overnight urinary sodium, even if related to sodium intake, cannot be employed to assess the association between salt in the diet and blood pressure.