Studies were undertaken to elucidate the cellular basis of skin allograft rejection mediated by long-term cultured cell lines and clones. The adoptive transfer, in vivo, of in-vitro-sensitized cells, from B6AF1 anti B10.BR or from C57BL/6 anti DBA/2 cultures, and expanded eight-fold to ten-fold for one week in lectin-free interleukin 2 (LF-IL-2) were able to mediate specific skin allograft rejection. These same cells lost the ability to mediate accelerated skin graft rejection when they were expanded more than 100-fold during three weeks of culture in LF-IL-2 even though these cultures mediated high levels of specific in vitro cytotoxicity for the appropriate allosensitizing cells. When Lyt-2+ cells were depleted using monoclonal antibodies and complement prior to in vitro sensitization and expansion in LF-IL-2, these cells lines retained the ability to mediate skin allograft rejection in vivo when expanded more than 100-fold for three culture generations in vitro. These latter lines were greatly enriched for Lyt-1+2- cells and had little or no cytolytic activity, but they retained specific in vitro proliferative responses to the sensitizing alloantigen. Several Lyt-1-2+ cloned long-term lymphoid cell lines with high levels of specific cytolytic activity against the sensitizing alloantigen were derived and none was capable of mediating the accelerated rejection of skin grafts in vivo. However, cloned lymphoid cell lines that were phenotypically Lyt-1+2- and were capable of proliferating when in contact with specific alloantigen, but were not cytolytic, were capable of mediating the accelerated rejection of skin grafts in vivo both in irradiated mice and in nude mice. These studies demonstrate that skin allograft rejection can be mediated by Lyt-1+2- cell lines with specific in vitro proliferative activity to alloantigen although Lyt-1-2+ cell lines with cytolytic but not proliferative activity to alloantigen in vitro are ineffective in mediating graft rejection in vivo. Specific proliferative activity and no cytolysis appears to be a good in vitro correlate of the in vivo activity of long-term cultured cell lines.