Abstract To develop a precise assay for the T-cell progenitors of cytotoxic lymphocytes (CL-progenitors), lymphoid cells were cultured under optimal conditions in Marbrook vessels with mitomycin-treated allogeneic stimulator cells, and the total level of CL produced 5 days later estimated by a modified 51Cr release assay. Conditions were adjusted so an arithmetically linear cell dose response relationship was obtained. Three aspects of the cell dose response curve required attention. (1) At low responding cell inputs a macrophage-like cell became limiting (despite the presence of allogeneic macrophages in the stimulating cell population), leading to a lag in the response. This limitation was overcome by adding a low level of irradiated syngeneic macrophages, or by using irradiated syngeneic spleen ‘filler’ cells. (2) The slope of the resultant linear dose response region could be reduced if desired by changing from cellophane dialysis membranes to 0.1 μ pore size nuclepore membranes, suggesting a stimulatory role for some higher molecular weight soluble factor produced in the cultures. (3) At higher responding cell inputs a marked and extensive plateu was obtained. CL developing early in the response appeared to be destroying the allogeneic stimulator cells causing the response to be self-limiting. This problem was overcome by using a responding cell concentration lower than commonly employed. Assays using mixed leukocyte cultures in the lag or plateau regions could give misleading values for CL-progenitor activity. It is suggested that some examples of apparent synergism in CL generation may have resulted from these effects, rather than T-cell helper T-cell progenitor interactions.