Discrepancies have been reported in differential cell counts according to the diverse processing methods used in bronchoalveolar lavage (BAL) fluid management. The differences have proved to be mainly the result of selective lymphocyte loss, while the exact mechanisms of the phenomenon remain controversial. Observing a similar variation in differentials from differently stained identical smears, we quantified the cell loss due to staining procedures from 45 consecutive satisfactory BAL procedures. To do this, we compared relative lymphocyte recovery on neat pooled lavage in a hemocytometer with that from smears and cytopreps fixed and stained in different ways. We found (1) A significant lymphocyte loss (p < 0.05) whatever the staining method. (2) Different methods of fixation and staining lead to considerable variation in differentials from slides otherwise identically managed. The loss is higher during air-drying fixation followed by staining with an aqueous medium such as Diff-Quik than on spray-fixed slides stained in an alcohol medium such as Papanicolaou stain. (3) The effect of lymphocyte loss on differentials is more important when the initial lymphocytosis is less than 35%, and decreases to nonsignificance when it exceeds 70%. The role of cytocentrifugation or other manipulations in cell loss probably has been overestimated because unknown effects of staining methods were also attributed to these manipulations. We suggest that lymphocyte loss could arise from poor adherence on slides, which is exacerbated during aqueous staining if no artifice (e.g., spray fixation), is used to hold them. Thus, the definition of the long-awaited standard procedure for an accurate differential count of BAL fluid must take into account fixation and staining methods.