A strong association between HLA-DR4 and rheumatoid arthritis (RA) has been found in a number of populations. In contrast, the incidence of DR2 is decreased in patients with RA, suggesting that this specificity may confer some protection against the disease. A number of subtypes of DR2 have been defined by serology, by responses in mixed lymphocyte culture reaction, and, more recently, by restriction fragment length polymorphism. These subtypes of DR2 are in linkage disequilibrium with different subspecificities of DQw1. It is thus likely that the distribution of these subtypic DR,DQ haplotypes in DR2 positive patients with RA may be important in understanding the genetic basis of susceptibility/resistance to RA. In this paper a study of the subtypes of DR2,DQw1 haplotypes in 18 patients with RA, who required sodium aurothiomalate as a disease remitting drug, and unrelated healthy individuals is reported. Three subtypes of DR2 haplotypes, DRw15 (Dw2),DQw1.2(DQw6), DRw15(Dw12),DQw1.12(DQw6), and DRw16(Dw21),DQw1, AZH (DQw5), were analysed with a cDNA probe for the DQ beta gene. The data show that DR2 positive patients with RA carried either the DRw15(Dw2),DQw6 or DRw15(Dw12),DQw6 haplotype. No patient with RA was positive for the DRw16(Dw21),DQw5 subspecificity. In contrast, six of 29 (21%) normal healthy DR2,DQw1 positive individuals carried the DRw16(Dw21),DQw5 haplotype. These data together with earlier results on the distribution of the DR4,DQw7 haplotype in patients with RA support the hypothesis that DQB1 chain polymorphism may be important in determining susceptibility to severe RA.