Indomethacin 0, 15, 25 and 35 mg t.i.d. was given in randomized order for two weeks to eight outpatients with rheumatoid arthritis in a double-blind study. At the end of each treatment period the clinical response was evaluated by subjective and objective methods and the plasma indomethacin concentration was measured by GLC-mass fragmentography. Compared with non-treatment periods, indomethacin had a statistically significant therapeutic effect as judged by global assessment, duration of morning stiffness, use of escape analgesia, articular index and pain score, but there was no relation between the clinical effect and the size of the dose or the plasma concentration of the drug. Technetium uptake over the affected joints did not change during indomethacin therapy, which might reflect a lack of effect on the local activity of the disease. Lower doses than those currently prescribed are probably sufficient to alleviate symptoms in this disease.