Eighty-eight untreated subjects (mean age 58.6 +/- 1.1 years; 73 males and 15 females) with diastolic blood pressure between 90 and 100 mmHg were recruited to the study. Subjects were seen fortnightly and after four pre-diet visits were randomised into a normal sodium intake group (44 subjects receiving less than 80 mmol sodium daily in the diet plus 80 mmol of sodium supplement as eight slow release sodium chloride tablets daily) or a low sodium intake group (44 subjects receiving less than 80 mmol sodium daily in the diet plus eight slow release sodium chloride placebo tablets daily). Eight weeks later, subjects crossed over to the alternate tablets while continuing with the reduced sodium diet for another period of 8 weeks. The difference in urine sodium between the low sodium phase and the normal sodium phase was 67 +/- 4 mmol/day independent of the order in which the treatments were given; the corresponding difference in urine potassium excretion was 1.2 +/- 1.4 mmol/day and was not significant. The differences in systolic and diastolic blood pressures between the low sodium and normal sodium phases for all 88 subjects were 3.6 +/- 0.7 mmHg (95% confidence intervals 2.2-5.0) and 2.1 +/- 0.4 mmHg (95% confidence intervals 1.3-2.9) respectively, and were independent of the order in which treatments were given (p less than 0.005). On the other hand, comparison of the blood pressures of the two cohorts of subjects as parallel groups during the first test phase revealed that the falls in pressure were greater by 5.3 +/- 1.4 (95% confidence intervals 2.7-8.1) mmHg (systolic) and 3.4 +/- 0.8 (95% confidence intervals 1.8-5.0) mmHg (diastolic) in subjects with low sodium intake compared with those with normal sodium intake. These differences in blood pressure reduction obtained by analysis of the crossover study and of the parallel group study were not significant, with clear overlap of the 95% confidence intervals. Moderate dietary salt restriction causes reductions in blood pressure and around 3.5-5.5 mmHg (systolic) and 2-3.5 mmHg (diastolic) of these falls can be attributed specifically to reduction of sodium chloride intake.