Betaxolol is a cardioselective beta-blocker, which has a bioavailability of 90% and a T 1/2 of 20 h. A four group, cross-over double-blind trial was conducted to select between betaxolol 20 mg and 40 mg for long term trials. 60 patients were allocated randomly to one of the sequences placebo-20 mg, 20 mg-placebo, placebo-40 mg and 40 mg-placebo, each treatment lasting for 2 weeks. Groups were homogenous for baseline diastolic blood pressure (DBP), age and male/female ratio, and were slightly unbalanced for weight. A two-way ANOVA (3 treatments, 2 sequences) showed no treatment-sequence interaction nor sequence effect. The mean reduction in DBP was 14.2 +/- 1.8 mm Hg following 20 mg and 18.0 +/- 1.8 following 40 mg betaxolol, and 4.0 +/- 1.2 mm Hg during placebo (p less than 0.001). Age, weight, baseline DBP and duration of hypertension did not influence the treatment effect. The 95% confidence intervals of the reduction in DBP were 10.4 - 17.9 for 20 mg and 14.3 - 21.6 mm Hg for betaxolol 40 mg. Aiming at a mean reduction to 90 mm Hg, betaxolol 20 mg would appear to be adequate in similar patient populations.