The efficacy of propranolol and metoprolol in the treatment of essential tremor was compared in a double blind crossover placebo-controlled study in 16 patients. Each treatment was given for a period of 4 weeks at two different dosage regimens (150 and 300 mg daily for metoprolol, 120 and 240 mg daily for propranolol). Each dosage regimen lasted for 2 weeks. Tremor assessment was carried out by accelerometry, clinical evaluation, patient's self-rating and a battery of performance tests. At the lower dosage, propranolol was found to be superior to placebo on the basis of performance tests and patient's self-assessment. At the higher dosage, propranolol was superior to placebo on all methods of assessment. By contrast, the tremorolytic effect of metoprolol was not significantly different from that of placebo, irrespective of the dosage or of the method of assessment used. Propranolol (120 mg daily) was better than metoprolol (150 mg daily) on the basis of clinical evaluation and patient's self-assessment. Propranolol (240 mg daily) was superior to metoprolol (300 mg daily) on the basis of patient's self-assessment. Both drugs antagonised standing tachycardia to a similar extent. These results indicate that the effectiveness of metoprolol, previously demonstrated in a single-dose study in the same patients, is not fully maintained during prolonged administration. In the absence of specific contraindications, propranolol represents a better choice in the treatment of patients with essential tremor.