The quantitative assessment of airway inflammation in patients with apparently similar clinical severity of asthma has yielded variable results. The aim of this study was to assess the variability of inflammatory cell counts and airway structure in large and small airways from subjects with and without asthma, and to calculate the number of cases required to detect significant differences between disease groups. Three serial transverse sections, 20 microns apart, were examined from cartilaginous and membranous bronchioles. Airway dimensions measured were the areas and perimeters defined by the luminal surface of the airway epithelium, basement membrane, outer border of airway smooth muscle, and surrounding border of the airway adventitia. Airway wall components measured were the areas of airway smooth muscle, mucous glands and cartilage. The numbers (per mm of basement membrane) of lymphocytes, eosinophils, neutrophils and plasma cells were counted using haematoxylin and eosin staining and mast cells were counted in sections stained with Toluidine blue. Measurements were repeated by two or three observers. The coefficient of variation (CV) was used to assess intraairway, intraobserver and interobserver variability. For airway dimensions the mean (+/- SD) CV within airways was 3.8% +/- 2.1%, within observers was 3.7% +/- 1.8% and between observers, 3.4% +/- 1.8%. For airway wall components, the CV within airways was 10.7% +/- 4.6%, and between observers was 8.8% +/- 2.2%. For cell counts, the CV within airways was 20.8% +/- 12.5%, within observers was 8.3% +/- 3.7% and between observers, 9.6% +/- 4.6%. There were no significant differences in variability between cases with and without asthma. Intra-airway variability was higher (p < 0.05) in membranous airways (5.6% +/- 1.7%) than in large cartilaginous airways (2.3% +/- 1.1%). We estimate that between 5 and 15 cases in each group would be required to detect differences of 50-100% between groups, depending on the parameter being compared. These findings suggest that airway structure and inflammatory cell numbers are uniform throughout the bronchial tree in normal and asthmatic cases, and that small samples of large or small airways are likely to be representative when comparing cases.