Four different hydrochlorothiazide formulations were prepared, and cumulative urinary hydrochlorothiazide excretion was determined in a crossover study using six volunteers. A comparison of in vivo results showed that one formulation (Forumulation D) was significantly different from the others at 2, 3, 4, 5, 8, and 14 hr. A dissolution study was conducted on each formulation using the flask, USP basket, and magnetic basket methods at agitation speeds of 50, 100, and 150 rpm. Formulation D was significantly different from other formulations when determined using the USP basket method at 150 rpm and a sampling time of 10 min; the USP basket method at 100 rpm and a sampling time of 100 min; the flask method at 100 rpm and sampling times of 30, 40, 60, and 120 min; and the flask method at 150 rpm and sampling times of 30 and 40 min. Significant in vitro and in vivo correlations were found using a regression analysis and F test. With a correlation coefficient and 95% confidence intervals, it was established that the USP basket method at 150 rpm was the best predictor of urinary hydrochlorothiazide excretion among the dissolution methods tested.