Feverfew, reputed by folklore to be effective in arthritis, has in vitro properties that could be beneficial in the control of inflammatory disease. Forty one female patients with symptomatic rheumatoid arthritis received either dried chopped feverfew (70-86 mg) or placebo capsules once daily for six weeks. Allocation was random and not known by patient or observer. Variables assessed included stiffness, pain (visual analogue scale), grip strength, articular index, full blood count, erythrocyte sedimentation rate, urea, creatinine, C reactive protein, complement breakdown products (C3dg), rheumatoid factor titre, immunoglobulins (IgG, IgA, IgM), functional capacity, and patient and observer global opinions. One patient (placebo) withdrew after three days and was not included in the analysis. Treatment and placebo groups (20 patients each) were well matched at entry. No important differences between the clinical or laboratory variables of the groups were observed during the six week period. This study therefore shows no apparent benefit from oral feverfew in rheumatoid arthritis.