During a 6-month period all lung cancer patients in a university hospital chest clinic were investigated for asbestos exposure by means of personal interview, bronchoalveolar lavage (BAL), roentgenograms, lung function testing, histology and measurement of fibre concentrations in lung tissue samples using scanning electron microscopy (SEM). About one-third of the patients (33%) were classified as having been exposed to asbestos on the basis of the interview. BAL was performed and AB counts were done in 51 patients. Fourteen (27%) BAL specimens had AB counts of 1 or more AB/ml, which is the conventional limit for non-trivial asbestos exposure. For a subgroup of 25 operated lung cancer patients fibre analysis was also available. In six cases (30%) the asbestos-containing samples had asbestos fibre concentrations equal to or more than 1 million fibres/g dry lung. In eight (32%) of the operated lung cancer patients histopathologically confirmed fibrosis was seen; five of these patients were in the two highest exposure classes. Pleural plaques on X-ray films were seen in six (24%) of the operated cases. With each indicator of exposure about 30% of lung cancer patients were found to have been exposed, confirming the "one-third rule"; however, when all the information was collated there were three cases (12%) in which exposure was most obvious according to the different parameters used in this study. In these three cases the cancer could well be attributed to asbestos. Anthophyllite was present in all asbestos-containing samples and anthophyllite was the main fibre type in 61% of these samples.