Fifty recurrent stone formers were included in a double-blind randomized study (median 3 years) performed in a Norwegian general practice to compare twice daily administration of 25 mg hydrochlorothiazide versus placebo. The number of patients with new stones was significantly higher in the placebo group than in the thiazide group (p = 0.05, one-tailed test). If a new stone was formed, thiazide, but not placebo, had the effect of prolonging the stone-free interval (p less than or equal to 0.01). The probability of not forming a new stone during the treatment period was 45% for the placebo group and 75% for the thiazide group. The thiazide effect seemed to be independent of urinary calcium, but was less beneficial in patients with hyperuricosuria. The placebo group also showed a substantial decrease in the expected number of new stones (p less than or equal to 0.01), emphasizing the importance of an adequate control group.