A study of the methods used to quantitate killing of tumor cells by antibody and complement has highlighted a number of problems. Using leukemia as a model we have found that the release of 51Cr from labeled tumor cells treated with antibody and complement can be an equivocal measure of cell viability. Combined with its restricted sensitivity (less than a 2 log range of cell killing) this makes this widely used assay of questionable value for detecting small numbers of viable cells, or for identifying subpopulations of complement-resistant cells. As an alternative we have developed a [125I]iododeoxyuridine (125IUdR) uptake assay, that combines the simplicity and rapidity of the 51Cr release technique with the sensitivity of a clonogenic assay. This method eliminates the problem of spontaneous isotope release, inherent in prelabeling assays, and variability from experiment to experiment can be avoided by including a viable cell standard curve within each assay. The sensitivity of the 125IUdR uptake method, which can be completed within a day, is similar to that of a 10 day methylcellulose cloning assay and was capable of detecting the presence of a minor subpopulation of complement-resistant tumor cells.