Auranofin (triethylphosphine gold) was administered to 21 patients with juvenile rheumatoid arthritis during an open-ended, open-label, noncontrolled trial designed to establish safety and preliminary efficacy. Initial dosage was 0.1 mg/kg/day; incremental increases to 0.2 mg/kg/day were allowed. Aspirin (80 mg/kg/day), tolmetin (20 to 40 mg/kg/day), and naproxen (400 to 600 mg/m2/day) were allowed as rapidly acting anti-inflammatory agents. All patients attained measurable plasma concentrations of gold during the study. Clinically significant improvement (greater than 25%) occurred in more than half the patients with regard to the number and severity of joints with swelling, pain on motion, and tenderness. The number of joints with active arthritis decreased by at least 25% in nine of the 19 patients. Group mean changes between the initial and final visit indicated improvement in all articular disease indices measured. Eleven of 16 patients with an elevated erythrocyte sedimentation rate showed decreases of at least 25%. The group given the higher dosage had a greater proportion of responders with decreases in erythrocyte sedimentation rate. Four of six patients whose sera contained rheumatoid factor showed decreases in its titer. Discontinuation of auranofin was necessary in two patients because of headaches and because of hematuria and anemia associated with a severe flare of polyarticular disease, respectively. The results from this trial are sufficiently encouraging to merit a double-blind trial of auranofin in children with juvenile rheumatoid arthritis.