The aim of this study was to establish the validity of four lymphocyte isolation methods. The effects of three different erythrocyte lysing methods commonly used in the analysis of human cells, namely, lysis by ammonium chloride (AC), Becton Dickinson lysis (BDL) and the Coulter Q-Prep (CQP) preparation system were established by flow cytometry on rat lymphocyte subsets. The results were compared with those obtained with a Ficoll-Isopaque (FI) density gradient procedure adapted for use with rat cells. Lymphocyte isolation by AC or FI gradient was performed before labelling the lymphocyte subpopulations, whereas the BDL and CQP methods were performed after staining the cells in whole blood. The FI gradient yielded the lowest CD5+, CD4+ and CD25+ cell percentages. On the other hand AC lysis produced higher percentages of T cells and lower percentages of B cells than the other methods studied. The percentages obtained after BDL or CQP methods for T lymphocyte subsets and B cells were found to be reproducible. The commercial methods (BDL and CQP) are faster but rather expensive, whereas AC lysis and FI gradient separations are cheap and particularly useful when there is a requirement to culture the cells.