Using a standardized indirect cell-mediated lyrnpholysis (CML) technique, cytotoxic lymphocytes (CTLs) were raised between unrelated, fully HLA-A, B (and C) typed individuals, differing only for one identified HLA-A, B antigenic determinant. Ten CTLs involved HLA-A2+ stimulators and HLA-A2-, A28-, A9-responders. Five CTLs involved HLA-Bw35+ stimulators and HLA-B5-, Bw35-, B15-, B17-, BI8-, Bw21-responders. These CTLs were tested for cell-mediated cytotoxicity against two different panels (20 and 9 cells respectively) of HLA-A28+ and HLA-B5+ PHA-lymphoblasts. The same panel cells were used as cold competitors for cytotoxicity against the specific 51Cr labelled targets.The following observations were made: (1) most panel cells are neither lysed nor able to block; (2) significant lysis is, however, seen in the panels positive for serologically cross-reactive antigens, although lysis is generally lower than comparable lysis by anti-A28 or anti-B5 CTLs; (3) all target cells lysed in the A28 panel are able to inhibit specific lysis while 2 cells in the B5 panel are lysed but do not inhibit (Cytotoxicity-Positive-Inhibition-Negative = CYPIN); (4) a group of panel cells cannot be lysed but are able to inhibit (Cytotoxicity-Negative-lnhibition-Positive = CYNIP).These observations indicate that cross-reactive groups identified by serology may also be identified by celiology using direct cytotoxicity and/or cold target inhibition. The cross-reactivity defined by CTLs, as in serology, is inconstant and often incomplete. Whether the immunogenetic background of cross-reactivity as defined by antibody or CTL is identical is unknown but probable.The occurrence of CYNIP underlines that direct cytotoxicity and cold target inhibition are not directly comparable and indicates that different epitopes may be involved, although the discrepancy may be brought about by the experimental conditions used.The CYPIN phenomenon indicates the existence of immunodominant epitopes although non-T cell-mediated lysis cannot be excluded.