The efficiency of guinea pig complement (GPC) and rabbit complement (RC) was compared in a cytotoxic assay by the 51Cr release method, and with the DBA/2 leukemia L1210 as target. Alloantisera and heteroantisera, containing antibodies of defined specificity and shown to be IgG (2ME resistant), were used. When there were limiting amounts of either alloantibody or heteroantibody present, RC was clearly more efficient than GPC. In the presence of excess antibody, however, both RC and GPC behaved similarly, except that in the presence of GPC, fewer cells were lysed. This finding is related to the target cells containing a population of cells resistant to lysis with GPC, rather than to the amount of 51Cr released from each cell. When alloantisera were mixed, variable effects occurred: a) blocking or antagonistic effects occurred in all combinations in the presence of RC; with GPC, this effect was only seen in the presence of antisera to two private specificities (H-2D.4 + H-2K.31). b) Synergistic effects were noted only with GPC when there was a mixture of antibody to a private (H-2D or H-2K) specificity with that to a public specificity. In the presence of heteroantibody, either not absorbed, or absorbed to remove species-specific antibody and leave antibody to H-2.31, the ratio of the effectiveness of RC:GPC remained unaltered, indicating little or no synergism between these two antibodies and GPC. All of these findings are in substantial agreement with the concept that GPC requires the activation of two adjacent cell surface sites for cytolysis, but that RC requires only the activation of one such site.