The hypothesis that homoeopathic potencies are placebos was tested in a randomised, double-blind, placebo-controlled trial. The study model chosen compared the effects of a homoeopathic preparation of mixed grass pollens with placebo in 144 patients with active hayfever. The homoeopathically treated patients showed a significant reduction in patient and doctor assessed symptom scores. The significance of this response was increased when results were corrected for pollen count and the response was associated with a halving of the need for antihistamines. An initial aggravation of symptoms was noted more often in patients receiving the potency and was followed by an improvement in that group. No evidence emerged to support the idea that placebo action fully explains the clinical responses to homoeopathic drugs.