Prior to December 1949, some British rubber industry workers were inadvertently exposed to the human bladder carcinogen beta-naphthylamine, which was present as a contaminant (at 0.25%) in antioxidants used in manufacturing. This study follows a composite cohort of 6450 men employed at a large tyre factory either during the 'at-risk' period or just after it. A group of 2090 at-risk men (employed 1945-1949) and 3038 men, first employed only after January 1950, when the carcinogen had been removed, were followed for their bladder cancer morbidity and mortality experiences. Fifty-eight tumours were registered for those at risk, whereas only 33.9 were expected at national standardized registration rates [SRRN = 171 and 95% confidence interval (CI) = 130-221]. Thirty-nine bladder tumours were reported for the post-1950 intake, whereas 38.3 were expected (SRRN = 102 and 95% CI = 72-139). The use of mortality data did not reveal any underlying hazard because 12 of the 58 at-risk workers with tumours were still alive at the study end date. In only 16 instances was bladder cancer actually certified as the underlying cause of death. Plotting cases by their location of work on a factory plan assisted the interpretation. A statistically significant elevated risk of bladder cancer for the exposed workforce was evident, but this reversed when the carcinogen was removed from processing in October 1949. The use of morbidity (incidence) data in long-term studies of occupational bladder cancer should be the required methodology if the hazard and risk are not to be underestimated.