OBJECTIVE To assess and quantify the occurrence of pleural malignant mesotheliomas in people who neither experienced occupational exposure to asbestos nor were married to (or known to live with) workers exposed to asbestos in the workplace. The study was conducted in the area of the local health authority of Casale Monferrato, in north western Italy, where a large factory that produced asbestos cement was active up to 1985. No other major activities related to asbestos have ever been present in the area. METHODS A retrospective survey covering the period 1980 to 1991 identified 126 incident pleural malignant mesotheliomas histologically diagnosed among residents in the local health authority (population at the 1981 census 98,000). Submission of 83 of 95 cases diagnosed during 1980-9 for revision by a panel of five expert pathologists led to the exclusion of 21. The 31 cases diagnosed in 1990-1 were not submitted for revision. For 64 of the 105 retained cases, information derived from different sources (rosters of the employees in the asbestos cement factory dated back to 1907, list of their spouses, clinical records) did not suggest occupational or paraoccupational exposure to asbestos. RESULTS Incidence excludes cases for which there was some suggestion of occupational or paraoccupational exposure to asbestos. Incidence of histologically confirmed malignant mesothelioma among residents in the local health authority (annual x 100,000; age adjusted) was 4.2 in men and 2.3 in women (based on 26 and 18 cases respectively). In both sexes, rates in 1985-9 were higher than in the previous quinquennium. Corresponding estimates for 1990-1 (based on unrevised diagnoses) suggest similar rates in men and women. CONCLUSIONS Rate ratios which are four to six times those measured by conventional Italian cancer registries can hardly be totally explained by bias produced by lack of recognition of occupational or paraoccupational exposure. The problem of proving this type of negative data is common to other circumstances of alleged cancer clusters of environmental (non occupational) origin.