In a double blind study 57 patients with active erosive rheumatoid arthritis without malalignment or deformities (median disease duration 13 months) were randomly treated with 50 mg gold sodium thiomalate (GSTM) or 15 mg methotrexate (MTX) intramuscularly for 6 months. In the GSTM group, 5/28 patients had to be withdrawn because of side effects; in 2/28 the dose was reduced. In the MTX group, 2/29 were withdrawn, one was lost to followup. The number of swollen joints improved by over 6 in 18/26 (MTX) and in 15/21 (GSTM). Five clinical variables and the sedimentation rate improved significantly in both groups without significant intergroup differences. The radiographs of hands, wrists and forefeet (32 joints evaluated according to Larsen) showed a radiological progression in 11/26 (MTX) and in 8/20 patients (GSTM); however, the deterioration of the mean Larsen index was not significant. While there was no significant difference in effectivity, tolerability was better in the MTX group.