The uncoated and coated fibre load in bronchoalveolar lavage (BAL) fluid was assessed using light microscopy, scanning electron microscopy, and x ray microanalysis in 15 subjects with previous, unprotected exposure to asbestos, including three with clinical and radiological evidence of asbestosis, and in 13 urban dwelling control subjects with no known occupational exposure to asbestos. The mean ferruginous body count per ml BAL fluid in asbestos exposed subjects as determined by light microscopy was 52 (range 0-333). No ferruginous bodies were detected in control subjects. The mean fibre count per ml BAL fluid in asbestos exposed subjects as determined by electron microscopy was 793 (133-3700), significantly greater than 239 (44-544) in controls (p less than 0.05). Electron microscopic counts correlated with duration of previous exposure to asbestos (r = 0.47, p less than 0.05) and with percentage neutrophil counts (r = 0.53, p less than 0.025). There was no relation between electron microscopic fibre counts and light microscopic ferruginous body counts. In 11 asbestos exposed cases x ray microanalysis confirmed the presence of asbestos and in six the asbestos fibre type was clearly identified. Of five subjects showing no asbestos bodies by light microscopy, all showed fibres by electron microscopy, and in three cases the presence of asbestos was confirmed by microanalysis. Among control subjects, fibres were either large organic fibres or smaller particles which microanalysis showed were not asbestos. In only one control case were a few fibres identified which were confirmed as asbestos fibres on microanalysis. Electron microscopic examination of BAL fluid may confirm past exposure to asbestos and probably gives a crude quantitative estimate of asbestos load.