Immunophenotyping of different lymphocyte populations was carried out in parallel on 113 consecutively received specimens of human peripheral blood using 2 different data acquisition and analysis systems (EPICS C and 4Cyte-Acmecyte) on the same flow cytometer (EPICS C). The phenotypes analyzed were CD3+, CD4+, CD8+ CD56+ CD16+ CD3-, TCR-yG+CDS-, and TCR-yG+CDg+. Both HIV-and HIV+ specimens were used for this study, including some with CD4 levels as low as 2% of all lymphocytes. Despite differences in gating procedures and shapes of bitmap (rectilinear vs. "amorphous"), the 2 methods agreed to within 2% positive cells in 97% of the cases. Although some statistically significant biases in the methods were observed, these were small and not biologically important. We conclude that both methods of data acquisition and analysis, as employed by experienced operators on the EPICS C flow cytometer, gave essentially equivalent results for lymphocyte subpopulations in peripheral blood preparations.Key terms: Imrnunocytometry, data acquisition, data analysis, flow cytometry, T cell subsets, NK cells, y6 T cells, HIV Flow cytometry has become the accepted method for immunophenotyping clinical specimens. The results obtained using this technology depend critically on both the hardware (e.g., laser, lenses, photomultiplier tubes) and software (e.g., for setting of displays, bitmaps, and gates, and tabulating statistics) used for data acquisition and analysis. In theory, variations in data could arise from either of these independent systerns. However, the lack of instrument bias noted in previous surveys (3, 9) suggests that differences in hardware are not a major source of variation. In contrast, differences in data analysis, particularly the selection of gated populations, can be major contributors to variation in results (4).To evaluate separately the importance of the data acquisition and analysis aspects of the flow cytometer, it is necessary to compare data obtained simultaneously on the same events by two different data acquisition and analysis systems. However, to our knowledge there are no reports in which this has been done. Therefore, the present study was undertaken to determine whether measurements of lymphocyte phenotypes obtained on 2 parallel systems of acquisition hardware and software were equivalent or biased. Specifically, phenotypes of clinical or research importance were examined, and comparisons were made of both overall means and the variability of the measurements obtained with the 2 parallel systems.