In the late 1960s and early 1970s. several investigators observed that lymphocytes isolated directly from unimmunized, normal hosts were able to lyse certain tumor cell targets (1). Unlike conventional cytotoxic T lymphocytes (CTL),’cytotoxicity was not restricted by major histocompatibility complex (MHC) gene products. This phenomenon was designated “natural immunity” and the effectors were deemed “natural killer”(NK) cells. Initially,“NK activity” described the lysis of certain “NK sensitive” tumor cell lines by lymphoid cells from an unimmunized host. This definition, based solely on functional criteria, became problematic when investigators began examining cells generated or propagated in vitro. For example, shortterm culture of mouse bone marrow cells produced “promonocytes” that lysed YAC-1 cells, as did a cloned cell line with ultrastructural and cell surface receptors (IgE Fc receptors) typical of basophils (2, 3). To accommodate such observations, terms such as “NK-like” activity and “NK-like” cells were coined. It has remained unclear whether “NK was a function mediated by several different cell types of distinct lineages, or alternatively a homogeneous population of cells restricted to a single lineage. We believe that there is now sufficient evidence to go beyond simply a functional or morphologic definition of lymphocytes mediating NK cellmediated cytotoxicity. We propose that the term “NK activity” should be replaced by “non-MHC-restricted cytotoxicity.” This term highlights the unifying feature of such lytic activity (ie. lack of MHC-restriction) and al-lows for the possibility of several distinct in vivo (“natural”) or in vitro (“induced) effectors. This nomenclature does not suggest any preconceived notions about the relationship or lineage of the effectors. We propose that the major effectors of non-MHC-restricted cytotoxicity are lymphocytes that can be classified into two distinct categories. The predominant and most abundant effectors of non-MHC-restricted cytotoxicity constitute a population distinct from T cells in both lineage and antigen recognition. We have designated