The aim of this paper, based on a cross-sectional study of 129 patients with nonallergic chronic nasal symptoms and 40 healthy controls, was to examine the leucocyte differential count in nasal secretions as a diagnostic test. Nasal secretions were collected using preweighed suction glass canulas under controlled conditions (-100Pa, 30 sec). Leucocyte and differential counts were performed using a Thoma hemocytometer and on cytospin slides after May-Grünwald-Giemsa staining. The percentage of eosinophils (Eo) was significantly higher in patients (mean +/- SEM: 15.1 +/- 2.3%) than in controls (5 +/- 2.6%) (p < 0.04). Comparison of the frequency distribution of the percentage of Eo in patients and controls clearly showed a subgroup of patients presenting with nasal secretion hypereosinophilia, and allowed us to set the positivity criterion at Eo = 20%. Diurnal variations in Eo count in 11 controls and 8 patients confirmed the value of the cutoff point. In 28 patients with nasal polyposis who underwent surgery, a correlation was found between secretion and tissue eosinophelia (r = 0.58, p = 0.001). Patients with nasal secretion hypereosinophilia had no more leucocytes in their secretions than healthy controls, the increase in eosinophils being balanced by a decrease in neutrophils. In patients without hypereosinophilia, the number of leucocytes per milligram of secretion was four times higher (8672 +/- 2521) than in the controls (2020 +/- 823) (p = 0.06) (cut-off point = 2500 leu/mg). These data show that the nasal cytogram can be modified either in qualitative or quantitative way, probably depending on the underlying inflammatory process.