In 96 subjects immunized artificially with antigen D HLA system antigens were determined. In 21 cases antibodies were not produced, in 70 good response to antigen D was observed, in 5 cases antibodies were produced with considerable delay. The method of two-step lymphocytotoxicity test of Terasaki was used. The reference panel included 90 sera with antibodies to antigens: 1, 2, 3, 11, W23, W24, W25, W26, W28, W29, W30, W31 from first sublocus, and 7, 8, 12, 13, W5, W10, W14, W15, W16, W17, W18, W21, W22, W27 from second sublocus. Seventy of the immunized volunteers responded by production of antibodies (group I). and 21 failed to respond (group II). The frequency of different HLA antigens was analysed statistically in both groups using the chi square test and comparing the results with the frequency in a random population of healthy non-immunized subjects. Statistically significant differences were found in antigens 2, 3 and W24 from sublocus HLA-A (LA) and W16 from sublocus HLA-B (Four). In the non-responding group antigens HL-A3 and W24 were more frequent than in the group producing anti-D antibodies. In the latter group antigen HL-A2 prevailed. It seems that the gene determining rapid recognition of antigen D is connected with presence of HL-A2 antigen.