Comparison between CDC and ADCC in a human allogeneic model using the same cytotoxicity marker, i.e. 51Cr release from labelled target cells, provided a better definition of the optimal technical conditions for revealing ADCC. Testing 229 sera from human renal allograft recipients after transplantation, we found that CDC is detected only during rejection (43% of cases) whereas ADCC can be found both in those recipients tolerating their grafts and in those undergoing rejection, more frequently in the former than in the latter (39% vs. 21%, p less than 0.05). These in vivo results and other in vitro experiments support the assumption that two distinct varieties of antibody can mediate the in vitro ADCC phenomenon, some directed against the A, B and C HLA loci and having a deleterious effect on the graft, and others capable of exerting a protective effect and which could be directed against D related HLA determinants.