Neutrophils are considered fragile cells, easily damaged by improper handling. Assays of neutrophil function frequently require preliminary isolation procedures that may be potentially harmful. The authors did an extensive investigation of the effects of currently used isolation procedures on a broad spectrum of functional assays that included: 1) phagocytosis; 2) bacterial killing by a culture technique and tritiated thymidine labeling; 3) candidacidal activity by methylene blue dye exclusion and differential Giemsa staining; 4) quantitative unstimulated and histochemical stimulated nitroblue tetrazolium dye reduction; 5) chemiluminescence; 6) random locomotion; and 7) chemotaxis by the agarose method. In addition, cells were examined morphologically by electron microscopy, and the granule density was quantitated by morphometric analysis. Isolation techniques included erythrocyte sedimentation, hypotonic lysis, gradient density separation using Ficoll-Hypaque, and counterflow centrifugal elutriation. The purest neutrophil suspensions were obtained by density gradient separation and counterflow centrifugal elutriation with mean neutrophil percentages of greater than or equal to 94%. Regardless of the isolation procedure, neutrophils were similar in all groups (P greater than .05) in the following studies: phagocytosis, nitroblue tetrazolium dye reduction, bacterial killing, candidal killing, inhibition of candidal germ tube formation, and cytoplasmic granulation. Differences were noted in assays of neutrophil migration and chemiluminescence. Neutrophil suspensions isolated by counterflow centrifugal elutriation and Ficoll-Hypaque had the highest scores for random migration and chemotaxis. These differences can be related to the purity of the neutrophil suspension rather than the harmful effects of the isolation procedures. Erythrocyte contamination affected both the slope and the time to peak response in the chemiluminescence assay. Exposure of neutrophils to cold temperatures (0-4 degrees C) for 1.5 hours impaired both random locomotion and chemotaxis. Current recommendations for the storage, transport, and preparation of leukocytes for neutrophil function studies need to be reassessed.