Forty-eight patients with rheumatoid arthritis refractory to other treatments were studied in a placebo controlled trial of methotrexate (MTX) in 2 institutions. Once weekly for 6 weeks, the patients were injected with placebo (Group 1), MTX 10 mg (Group 2), or MTX 25 mg (Group 3). Then, for the next 6 weeks, Group 1 received MTX, either 10 or 25 mg/wk, and Groups 2 and 3 continued their same dose. Adverse reactions necessitated change from 25 mg to 10 mg in some patients, but no major side effects of MTX were noted. At 6 weeks, the effect of the 2 MTX doses did not differ significantly but patients on MTX had fared significantly better (p less than 0.005 - less than 0.001) than those given placebo. At 12 weeks, all indices showed significant improvement in Group 1 and maintenance or enhancement of the improvement in Groups 2 and 3. We conclude that weekly low dose MTX therapy is efficacious for refractory rheumatoid arthritis.