The development of a terminal-labelling microcytotoxicity assay is described in which target cells (fetal fibroblasts) were labelled with 125I-iododeoxyuridine after effector (lymphoid) cells had been incubated with them for 24 h. The time-course for the development of cell-mediated cytotoxicity was assessed following allogeneic skin grafting. 'Non-specific' cytotoxicity detracts from the sensitivity of all microcytotoxicity assays and the terminal-labelling assay using 125I is no exception. The non-specific effects can be reduced but not eliminated by the removal of adherent cells. The optimum target cell/effector cell ratio would seem to be between 1:100 and 1:250. Residual lymph node cells did not appear to incorporate enough label to affect the test results. In vivo correlates of in vitro findings are still not easy to determine.