Rheumatoid arthritis (RA) is likely the result of a concerted action of several inherited and noninherited factors. Although there is a high suspicion that environmental factors are important, proof is missing. Most information has been collected on genetic risk factors. The inheritance pattern for RA is complex, and there is good evidence that HLA as well as non-HLA genes are involved. Almost all racial-ethnic groups share the association of RA with the HLA-DRB1-encoded sequence motif QKRAA or QRRAA. However, the completeness of the association varies significantly in different ethnic cohorts, as can be expected in a multigene model. The sequence motif translates into a pocket in the antigen-binding site of the HLA-DR molecule. The "rheumatoid pocket" accommodates peptide side chains and has distinct binding characteristics. Epidemiologic evidence points toward a role for non-HLA genes. Candidate genes, such as transporter in antigen processing (TAP) genes are currently explored. Major advances in defining and understanding the contribution of inherited and noninherited factors in RA may come from abandoning the concept of RA as a single entity and accepting a heterogeneity model for RA. Distributions of HLA-DR genes indicate that several subsets of RA patients exist. Seronegative (prognostically good) and seropositive (prognostically worse) patients can be distinguished by the arginine versus lysine substitution at position 71 of the HLA-DRB1 gene. A different dimension of disease, rheumatoid organ disease, appears to be reached in patients with two HLA-DRB1*0401 alleles. Identification of distinct RA subsets may allow us to stratify patients into categories that differ with respect to etiology, disease course, clinical pattern, and treatment response.