Bronchoalveolar lavage has been widely used to sample the lower respiratory tract. Most of the material recovered with this technique represents alveolar contents. A number of modifications have been suggested in order to obtain samples relatively enriched for bronchial material. In order to be able to use a standard technique for bronchoalveolar lavage to sample both airways and "routine" alveolar material, a simple modification of the technique as described by Reynolds and Newball was used: five sequential 20-ml aliquots were infused into the lower respiratory tract, and each aliquot was immediately aspirated. The return from the first aliquot was processed separately from the return from the subsequent four aliquots. These last four aliquots were pooled. Analysis of the first aliquot revealed it to be enriched for ciliated epithelial cells when compared with the subsequent aliquots. There were also differences in inflammatory cell composition with the bronchial sample containing relatively more neutrophils and relatively less lymphocytes. Aspiration during transoral bronchoscopy was documented by quantifying salivary amylase in the bronchial and alveolar lavage fluids. It was estimated, however, that the aspiration was not of quantitative significance in the vast majority of subjects studied. Finally, with the technique of fractional processing of bronchoalveolar lavage samples, it was possible to compare the protein concentrations in bronchial and alveolar lavages. Most prominent among the differences was a marked relative enrichment in the bronchial samples for immunoglobulin A. The technique of fractional processing of bronchoalveolar lavage samples provides a simple means to obtain samples enriched for bronchial and alveolar components. This should facilitate analysis of lower respiratory tract specimens in airway disease.