The mortality pattern of taxi drivers in Rome as possibly exposed mainly to gasoline engine exhausts was evaluated by means of a historical cohort study. A total of 2,311 male subjects registered as taxi drivers between 1950 and 1975 was followed from 1965 through 1988. The overall mortality was lower than expected on the basis of regional (Latium) reference rates (692 deaths, standardized mortality ratio [SMR] = 0.89, 95% confidence interval [CI] 0.82-0.96), whereas the number of recorded deaths for malignant neoplasms was about the expected (205 deaths, SMR = 0.99, 95% CI 0.86-1.13). Mortality from circulatory and respiratory diseases was lower than expected. Diabetes was significantly increased (42 deaths, SMR = 1.73, 95% CI = 1.25-2.34). An increased SMR appeared for respiratory cancer (SMR = 1.23, 95% CI = 0.98-1.50), mainly due to lung cancer (observed [O] = 76, SMR = 1.23, 95% CI = 0.97-1.54); two pleural cancers were also recorded. The excess of lung cancer deaths was present only among those enrolled in the most recent period (1965-1975) (45 deaths, SMR = 1.40, 95% CI = 1.02-1.87), especially among those of younger age (< 65 years) (SMR = 1.86); there was no relation between lung cancer mortality and latency since first enrollment in the cooperatives or duration of membership. There are difficulties in interpreting the excess of lung cancer on the basis of occupational exposures; however, the increased risk observed among workers employed in more recent calendar periods may be due to heavier exposure in the last decades; further follow-up of the cohort may elucidate whether there is an increasing lung cancer risk among taxi drivers.