A cohort of 889 men and 1077 women employed for at least 1 month between 1946 and 1984 by a former Italian leading asbestos (mainly textile) company, characterised by extremely heavy exposures often for short durations, was followed up to 1996, for a total of 53,024 person-years of observation. Employment data were obtained from factory personnel records, while vital status and causes of death were ascertained through municipality registers and local health units. We observed 222 cancer deaths compared with 116.4 expected (standardized mortality ratio, SMR=191). The highest ratios were found for pleural (SMR=4105), peritoneal (SMR=1817) and lung (SMR=282) cancers. We observed direct relationships with duration of employment for lung and peritoneal cancer, and with time since first employment for lung cancer and mesothelioma. Pleural cancer risk was independent from duration (SMR=3428 for employment <1 year, 7659 for 1-4 years, 2979 for 5-9 years and 2130 for > or =10 years). Corresponding SMRs for lung cancer were 139, 251, 233 and 531. Nonsignificantly increased ratios were found for ovarian (SMR=261), laryngeal (SMR=238) and oro-pharyngeal (SMR=226) cancers. This study confirms and further quantifies the central role of latency in pleural mesothelioma and of cumulative exposure in lung cancer.