The effects of various methods for removal of monocytes and residual erythrocytes from human peripheral blood mononuclear cells (PBMC) on T cell subsets were examined. T cell subsets were determined in a fluorescence activated cell sorter (FACS) using fluorescein-conjugated monoclonal antibodies (anti-Leu-1, anti-Leu-2a and anti-Leu-3a antibodies). Removal of monocytes by methods based on adherence or phagocytosis decreased yields of lymphocytes and caused changes in the percentage and/or the fluorescence intensity of T cell subsets. Exclusion of monocytes from the FACS analysis by setting of the scatter gates was incomplete (about 80%). Removal of residual erythrocytes after Ficoll separation by ammonium chloride treatment also changed the percentage of T cell subsets. A method using PBMC without removal of monocytes and erythrocytes was chosen as best and simplest for routine use in FACS analysis of lymphocytes. Erythrocytes could be excluded from the FACS analysis by setting the scatter gates and the percentage of T cell subsets was corrected after measurement of monocytes, identified by peroxidase staining. The reproducibility of measurements of T cell subsets made at different times was examined using PBMC obtained from the same healthy man during 12 weeks. For standardization of the assay, the peak positions of scatter and fluorescence intensity of each PBMC labeled with anti-Leu-3a were adjusted to standard values in each FACS analysis. Under these conditions, variations of other parameters of this standard PBMC were very small in 12 different assays. Using this standard PBMC, satisfactory, reproducible results were also observed on PBMC obtained from another normal subject. Therefore, this standard PBMC labeled with anti-Leu-3a was used as a standard in FACS analysis. Under these accurately standardized conditions, it was demonstrated that the peak position of fluorescence intensity of Leu-2a (suppressor/cytotoxic T) cells was significantly lower (P less than 0.01) in women than in men.