Low-dose, weekly methotrexate is now standard second-line therapy for rheumatoid arthritis. Methotrexate is the single agent to which new agents or combinations of agents have been compared in efficacy studies, and methotrexate is included in effective multiple-drug regimens as well (see other articles elsewhere in this issue). Despite a general expectation that methotrexate-induced toxicity would be common and severe, most patients tolerate the drug quite well. Because methotrexate was empirically reintroduced for the treatment of rheumatoid arthritis (RA) without any specific understanding of its mechanism of action, no follow-up agents with less toxicity, greater efficacy, or both have been introduced to replace methotrexate. Recent developments in a number of laboratories suggest several viable molecular mechanisms of action for methotrexate. This article reviews the hypotheses concerning the mechanism of action of methotrexate, analyzes the evidence supporting the proposed mechanisms of action, and discusses the implications of these various hypotheses for the design of new therapies for inflammatory diseases.