Quantitative analysis of surface molecule expression on viable alveolar macrophages (AM) by use of flow cytometry is hampered by non-specific antibody binding to various AM FcIgG receptors as well as extensive and heterogeneous autofluorescence of this cell type. The following approaches were undertaken to circumvent these obstacles. FcIgG receptors were blocked by excess human immunoglobulin. The use of a long wave-emitting dye (phycoerythrin/cyanine-5 tandem conjugate) permitted avoidance of the peak (green) AM autofluorescence range. Moreover, a cell-by-cell compensation for the remaining red autofluorescence background was employed. This was based on two facts: (i) strict correlation between green (F488/530) and red autofluorescence (F488/660) for all AM populations investigated; and (ii) neglectable overlap of the antibody-associated red fluorescence into the 530 nm autofluorescence detection wavelength. A fraction of the green autofluorescence (F488/530; channel 1) was then subtracted from the red fluorescence (F488/660; channel 2) on a cell-by-cell basis using standard two colour fluorescence compensation circuits. The validity of this FACS technique was confirmed by comparison with immunocytochemical staining and a reverse rosetting method. On AM lavaged from carcinoma-bearing but otherwise disease-free human lungs, the pattern of surface antigen expression was assessed with a panel of monoclonal antibodies. When applying to complex mixtures of bronchoalveolar lavage cells, the autofluorescence was employed to separate AM from granulocytes and lymphocytes. In conclusion, the presently described FACS technique allows quantitative immunostaining of surface molecules on AM, even when present in low copy numbers on highly autofluorescent cells originating from smokers.