After the first description by Andre Govaerts of lymphocytes that could kill other cells (Govaerts 1960). the lytic mechanism used in that process became one of the holy grailsof cellular immunology. Over the ensuing years a number of reasonable candidate mechanisms were put forward, examined, and either disproved or put aside for lack of direct experimental evidence. The possibility that cytotoxie lymphocytes might use classically defined complement components was among those examined early on and eventually dropped for lack of evidence. However, the discovery of complement-like lesions on the membranes of red blood eells killed in antibody-dependent cell-cytotoxieity (ADCC) reactions (Dourmashkin et al. 19XU) led to renewed efforts to identify complement-like molecules associated with cell-mediated cytotoxicity (CMC) reactions in general. This search ultimately resulted in the discovery and characterization of perforin, a molecule with significant homology to the terminal (lytic) components of complement. Perforin has now been implicated as a cytotoxic effector mechanism in virtually every form of CMC (for reviews, see Henkart 1985, Podack et al. 1991. Tschopp et al. 1990. Young 1989). Perforin is stored in cytoplasmic granules, at least in highly activated effector cells, and released by a degranulation process upon triggering through the T-cell receptor. Whether perforin can be released through other pathways, particularly in non-granulated effector cells, is still an open question. On the other hand, recent evidence suggests that granule components other than perforin may be required for optimal perforin function (Hayes et al. 1989, Heusel et ai. 1994, Shi et al. 1992), particularly against nucleated target cells (as opposed to RBC). How these additional components (such as granzymes) could be provided for interaction with perforin in alternate delivery systems is unclear. Within a few years of its introduction, the perforin model encountered a num-