Absorption experiments and subsequent retesting on human peripheral lymphocytes were performed with the congenic anti-H-2f mouse serum known to give strong cytotoxic reactions with human cells which correlate with the presence of HL-A-A2 antigen [mouse serum ASP 223, donor strain B10.M (H-2f), recipients (B10 x A.SW)F1 (H-2b/H-2s) hybrids; antibodies present in serum anti-H-2.9, 8,37]. Depending on the dose of A2-negative cells used for absorption, the serum can be rendered operationally monospecific to HL-A-A2. Absorption experiments with lymph node cells of different mouse strains showed that H-2f,p,w7 haplotypes share the ability to completely absorb the antihuman activity with about 15 .times. 106 cells. Consequently, the antihuman activity is not attributable to the presence of the anti-H-2.9 private specificity in the serum. H-2d,k,r haplotypes have weak absorbing capacity, because 500 .times. 106 lymph node cells are needed to absorb 70% of the antihuman cytotoxic activity. All strains that were able to absorb antihuman cytotoxic activity from ASP 223 shared the specificity H-2.8. Thus, the H-2.37 specificity seems to be responsible for strong and the H-2.8 specificity for weak absorbing capacity. The absorbing capacity of the H-2d haplotype was localized in the H-2.K end. Mouse lymph node cells have a 10- to 20-fold higher absorbing capacity for the antihuman activity than thymus cells. H-2.K end public specificities predominantly expressed on lymph node cells are responsible for the generation of antihuman (anti-HL-A) cytotoxic activity in anti-H-2 sera.