OBJECTIVE To explore associations between exposure to asbestos cement dust and radiographic findings in lung parenchyma and pleura. METHODS Radiographs from 174 blue collar workers and 29 white collar workers from an asbestos cement plant formed one part of the study. Progression of small opacities was further studied in those 124 blue collar workers, for whom two radiographs taken after the end of employment were available. The median readings from five readers who used the full ILO 1980 classification were used. As exposure indices, time since start of employment, duration of employment, cumulative exposure, and average intensity of asbestos exposure were used. The influence of age and smoking was also considered in multiple logistic regression analyses. RESULTS Small opacities (profusion > or = 1/0) were closely correlated with time related exposure variables, and showed weaker association with intensity based exposure variables. The odds ratio (OR) for small opacities was equal to 2.8 (90% CI 1.2, 6.7) in the > 30 f(fibre)-y/ml group, compared with those in the 0-10 f-y/ml group. Progression of at least two minor ILO categories after the end of employment was seen in 20%. Also, pleural thickening was closely related to time. By contrast, costophrenic angle obliterations were not associated with the time related variables, but closely associated with the intensity of asbestos exposure, and tended to occur during employment. The OR was 4.5 (90% CI 1.3, 15) in the > 2 f/ml group, compared with those in the 0-1 f/ml group. CONCLUSIONS In these workers, exposed mainly to chrysotile but also to small amounts of amphibole, the risk of radiographically visible parenchymal abnormality was substantially increased and strongly dependent on time related exposure variables. Progression was found long after the end of exposure. The findings on costophrenic angle obliterations, supposed to be sequelae of benign pleural effusions, were consistent with an immediate reaction triggered by intense asbestos exposure.