The frequency and severity of adverse effects during the first 14 days of treatment with amlodipine (5 mg once daily), nifedipine retard (20 mg twice daily) or placebo were compared in a multicentre, three-way, cross-over study involving 97 patients with mild-to-moderate hypertension. All three groups of patients were well matched for age, sex and baseline blood pressure. Amlodipine and nifedipine retard produced highly significant and comparable reductions in blood pressure, indicating that the doses were therapeutically equivalent. The incidence of adverse effects considered to be definitely or probably related to nifedipine retard treatment (41%) was significantly higher than for placebo (16%, p less than 0.01) or amlodipine (27%, p less than 0.05). There were no significant differences in the incidence of vasodilator-related adverse effects between amlodipine and placebo. In contrast, headache, flushing and dizziness were reported more frequently by patients while on nifedipine retard than on placebo or amlodipine. The convenience of once-daily dosing, together with a lower incidence of adverse effects, with consequently fewer withdrawals from therapy, suggests that amlodipine has clinical advantages over nifedipine retard in the treatment of hypertension.