Considerable differences in the percentage of hypodense eosinophils in the peripheral blood of asthmatics have been reported by different investigators. In these previous studies dextran sedimentation was used for removal of erythrocytes prior to density centrifugation. We hypothesized that the sedimentation procedure might induce the presence of hypodense eosinophils in the peripheral blood of asthmatic patients. In order to test this hypothesis, we compared eosinophil density profiles from peripheral blood of children with asthma and of age-matched healthy controls, using different procedures. In the first method (direct method) blood samples were directly layered on a discontinuous Percoll gradient. Erythrocytes were removed by isotonic lysis. In the second method (dextran sedimentation) erythrocytes were removed by sedimentation with dextran prior to gradient centrifugation. Results of the direct method show no significant difference in percentage of hypodense eosinophils between children with asthma and healthy controls (9.19% and 6.84% respectively). However, after dextran sedimentation, children with asthma had a significantly higher percentage of hypo-dense eosinophils than healthy controls (15.40% and 8.84% respectively; P < 0.05). The percentage of hypodense eosinophils was correlated with the number of eosinophils and with the lung function, measured as the Tiffeneau index (FEV1/VC), in the whole group of subjects when the direct method was used. We conclude that an increased percentage of hypodense eosinophils is not present in the circulation of children with asthma, but can be induced in vitro by dextran sedimentation. Therefore, in vitro generation of hypodense eosinophils in the blood of patients with asthma seems to be related with the primed state of eosinophils.