A total of 74 healthy unrelated random individuals and 36 patients with juvenile rheumatoid arthritis (JRA) were typed for HLA-D antigens with the homozygous typing cell technique and typed for HLA-D/DR associated DP-antigens with the primed lymphocyte typing (PLT) technique. All patients and some of the controls were also HLA-DR typed with a limited battery of anti-DR sera. Selected PLT-cells, specific for the HLA-D/DR antigens D/DRw1-8 and the local specificity D"H" were used. The results of the PLT-experiments were evaluated with the Normalized Median Response (NMR) method and the further procedure of DP-antigen assignment was analyzed. The DP-antigen assignments could be done solely according the NMR-values in approximately two thirds of the individuals. In the remaining individuals, further interpretation of the experimental data had to be done for the assignment of DP-antigens. The correlation coefficients were estimated between the HLA-D assignments and (i) the individual PLT-cell NMR-values with a fixed cut-off for positive reactions and (ii) the DP-antigen assignments. These coefficients were 0.79 and 0.92, respectively. The correlations between HLA-D, -DR and DP-antigen assignments of the specificities HLA-D, -DR and DP1, 2, 3, 4, 7 and 8 were analyzed in 42 controls and 36 JRA patients. The total correlation coefficients were: (i) HLA-D/DR: r = 0.78; HLA-DR/DP: 0.77; and HLA-D/DP: 0.96. The DP-antigen assignments correlated significantly better with HLA-D than with the HLA-DR antigen assignments, which does not agree with other studies. The DP-antigen frequencies among the controls were calculated and the estimated sum of gene frequency corresponding to definable DP-antigens was 0.94 indicating that about 12% of random individuals possess as yet undefined DP-antigens.