A fall in nocturnal plasma melatonin occurs in patients with cluster headache, suggesting that melatonin may play a role in the promotion of attacks. During a cluster period, we administered melatonin to 20 cluster headache patients (2 primary chronic, 18 episodic) in a double-blind placebo-controlled study of oral melatonin 10 mg (n = 10) or placebo (n = 10) for 14 days taken in a single evening dose. Headache frequency was significantly reduced (ANOVA, p < 0.03) and there were strong trends towards reduced analgesic consumption (ANOVA, p < 0.06) in the treatment group. Five of the 10 treated patients were responders whose attack frequency declined 3-5 days after treatment, and they experienced no further attacks until melatonin was discontinued. The chronic cluster patients did not respond. No patient in the placebo group responded. There were no side effects in either group. Although the response rate is low, melatonin may be suitable for cluster headache prophylaxis in some patients, particularly those who cannot tolerate other drugs.