Thirty-five patients with definite or classic rheumatoid arthritis (RA) entered a prospective, double blind, randomized study of 26 weeks duration. All patients had active RA that was unresponsive to greater than or equal to 2 nonsteroidal antiinflammatory medications and/or antimalarials. Eighteen patients were randomized to receive methotrexate (MTX) 12.5 mg weekly (oral and noncycled) + placebo, and 17 patients to gold sodium thiomalate (GSTM) 50 mg IM weekly + placebo (schedule I) after initial test dose of 10 and 25 mg. Dose reductions from Schedule I to Schedule II (i.e. MTX 5.0 mg and GSTM 25 mg) were permitted at Weeks 6 and 12. The GSTM group showed statistically significant improvement at Week 26 compared with baseline status in all of the clinical efficacy variables and the MTX group in 7. There were no statistically significant differences in these outcome variables between the 2 groups at Week 26. However, this small sample size may not have detected a clinically significant difference between treatment groups (alpha = 0.05, beta = 0.20). Two of the 18 patients treated with MTX and 6 of the 17 patients treated with gold withdrew because of drug toxicity, but this difference was not statistically significant. In conclusion, GSTM and MTX are similarly efficacious in the short term treatment of RA.