Lung asbestos burden was compared with exposure indices derived from job history interviews in 42 male subjects originating from the Montréal Case-Control Study project, 12 of whom had documented asbestos exposed job histories. Job interview data consisting of a chronological timetable of job histories were translated into detailed exposure indices by an expert group of hygienists and chemists. Total and individual asbestos fibre type concentrations were quantified by transmission electron microscopy with fibre identification by energy dispersive chi ray spectrometry after deparaffinisation of tissue blocks and low temperature plasma ashing. Geometric mean or median asbestos content was higher in subjects with an asbestos exposed job history than those without for retained dose of amosite, total commercial amphiboles, and total asbestos fibre. Except for crocidolite fibre diameter, which was significantly less in the lungs of exposed workers, no consistent differences were found in measurements of fibre dimension for any fibre type. Subgroups of subjects exposed to silica, metals, or smokers and non-smokers without significant occupational exposure showed varying patterns of lung asbestos fibre type deficit compared with the asbestos exposed subgroup. There was an overall trend for higher lung asbestos content proportional to higher exposure indices for asbestos representing concentration, frequency, and reliability. These exposure indices as well as duration of exposure (in years) were independent predictors of total asbestos content in regression analyses when combined in a model with age. Stepwise regression indicated that exposure concentration was the most important variable, explaining 32% of the total variation in total asbestos content. Smoking, whether expressed in ever or never smoked dichotomy or in smoked-years, had no relation to lung asbestos content in this model.