Fifty patients with osteoarthrosis of the knee and hip were entered into a double-blind trial of d-2-(6'-methoxy-2-naphthyl)-propionic acid (naproxen) 500 mg daily versus indometacin 100 mg daily. Two simultaneous trials of identical design were performed, one consisting of patients with unilateral or bilateral knee joint arthrosis only, the other of patients with unilateral or bilateral hip joint disease. Patients were assigned to 4 weeks treatment on naproxen or indometacin, then crossed over to the other drug. Assessments were made at --2, 0, 2, 4, 6 and 8 weeks, including subjective grading of pain and objective measurement of joint movements, stair climbing time and walking time. In almost all parameters there was significant improvement from baseline on both drugs, the magnitude of improvement being statistically equivalent. Side-effects recorded during the naproxen treatment period were significantly fewer than during indometacin treatment.