111 untreated subjects (mean [SEM] age 58.4 [1.0] years; 93 male, 18 female) with diastolic blood pressure between 90 and 100 mmHg were seen fortnightly, and after four pre-diet visits they were randomised into a low sodium intake group (53 subjects; diet containing less than 80 mmol sodium/day plus 8 placebo tablets daily) or a normal sodium intake group (55 subjects; same dietary sodium plus 8 slow-release sodium chloride [10 mmol] tablets daily). 103 subjects completed the intervention phase of 8 weeks. Urinary sodium fell significantly in the low sodium group but not in the normal sodium group. Urinary potassium excretion did not change in either group. Mean (SEM) systolic and diastolic blood pressure fell by 6.1 (1.1) and 3.7 (0.6) mmHg, respectively, in the low sodium group, but by only 0.6 (1.0) and 0.9 (0.6) mmHg in the normal sodium group. Multivariate analysis allowing for the effects of pre-diet blood pressure, weight, and age, reduced the effect of lowering the sodium intake on the systolic pressure from 5.5 (SEM 1.5) mmHg to 4.8 (1.3) mmHg (p < 0.005) but the effect on diastolic pressure was not changed significantly.