Thirty-five necropsied cases of diffuse malignant mesothelioma of pleura and six cases of mesothelioma of peritoneum are reported. Occupational histories from 40 cases showed that 32 cases (80%) were definitely exposed to asbestos; six cases (15%) were probably exposed and two (5%) had no known exposure. The mean latent interval was 34 years. Occupational histories of 56 control patients matched in pairs for age and sex with 28 of the mesothelioma cases revealed definite exposure to asbestos in 12 controls (21%), probable exposure in 11 (20%), and no evidence of exposure in 33 (59%). The difference in incidence of exposure between mesotheliomas and controls is statistically highly significant. Six-micron histological sections of lung, available in 39 mesothelioma cases, contained coated asbestos fibres in 36 cases (92%). Sixteen cases of mesothelioma showing fibres were free of asbestosis; asbestosis was slight in 14 cases, moderately severe in five cases, and severe in one case. The incidence of asbestos bodies in sections from mesothelioma cases was significantly higher than in a previously reported series of routine necropsies examined by lung smear. Counts of coated and uncoated asbestos fibres were performed on samples of lung tissue from 33 mesothelioma cases. No fibres were identified in one case and only occasional coated fibres in a second case. The remainder gave counts ranging from 154 thousand to 684 million fibres per gram dry weight. Uncoated fibres were invariably present and usually comprised 50 to 80% of the count. No relationship was found between total fibre count and the latent period of the mesothelioma. Fibre counts were also performed on lung samples from 18 smear-positive and 30 smear-negative routine necropsies. Cases with numerous coated fibres in the smear gave total fibre counts similar to those of mesothelioma cases without asbestosis. Routine cases with only one fibre in the smear usually yielded only occasional fibres on analysis. Coated and uncoated fibres were found in 13 out of 30 smear-negative cases, indicating a true incidence of 56% of exposure to asbestos in the whole necropsy series. When compared with the mesothelioma series the difference in incidence of fibres remains statistically significant. The data suggest that, on Tyneside, exposure to asbestos sufficient to cause an appreciable risk of developing mesothelioma has usually occurred through occupational exposure rather than by general environmental pollution by asbestos.