In a double-blind crossover study 12 patients with essential tremor were treated with propranolol and a placebo; 8 improved with propranolol and 3 with the placebo; the degree of improvement with propranolol was greater. In a similar study with diazepam 5 of 12 improved with diazepam and 4 of 12 with the placebo; the degree of improvement was less than that achieved with propranolol. Response in 21 patients to treatment with propranolol for 2 to 4 years was excellent in 4, good in 4 and fair in 10; the condition of 1 was unchanged and that of 2, worse. Excellent response was maintained for as long as 4 years, but response tended to deteriorate with time if initially it was less than excellent. Response decreased with increasing age. No patient 60 years of age or older had an excellent response, and the four with an excellent response were under age 55, three being under age 35; all four had had their tremor less than 12 years. Patients with essential tremor should be given a 3-month trial of propranolol at 120 mg/d; if no significant response is seen the dose should be decreased, then the drug discontinued.