Cryopreserved cells obtained by leucopheresis of normal donors were found to be useful as quality control reagents in long-term studies of human lymphocytes. Enough cells were obtained from one donor on a single day to test aliquots in parallel with patient samples for periods of up to 10 months. The results of E-rosette assays of two such cryopreserved panels, expressed as the mean percentage of lymphoid cells±s.d., were 63·7±8·4 (panel 1) and 52·3±7·8 (panel 2). The results of EAC-rosette assays on these cells were 6·2±1·7 and 11·3±6·0. These data were comparable to the values of 69·3±2·3 for E-rosettes and 6·5±4·9 for EAC-rosettes obtained using fresh cells from normal individuals. The long-term variation in rosette assays was also comparable to that in replicate samples counted on a single day. Detailed analysis of 155 serial tests showed that much of the variation was inherent in the microscope counting procedures and sampling errors. The cryopreserved cells were also used in lymphocyte stimulation assays with the mitogens concanavalin A, phytohaemagglutinin P and pokeweed. In mitogen assays, the principal sources of variation were found to be initial cell viability and incubation conditions. The major limitation of cryopreserved cells was that they did not control for the variation encountered in the gradient procedures used to separate lymphocytes from blood. Nevertheless, cryopreserved cells have been shown to be valuable for defining the limits of reproducibility in these assays and their consistent use has increased our confidence in the interpretation of day to day results on patient samples.