BACKGROUND Cancer incidence and mortality were evaluated among 4581 aniline dye production workers in Moscow. METHODS A historical cohort was assembled and followed-up from 1 January 1975 to 31 December 1989. Moscow district oncologic dispensary registries furnished case ascertainment and employer records provided job exposure data. Expected cancers and deaths were calculated based on gender-, age-, and calendar time-specific incidence and mortality rates for the Moscow general population applied to the cohort's person-years of follow-up. Disease-specific standardized mortality and incidence values were derived from ratios of observed to expected cancers. RESULTS Men experienced elevated total cancer mortality (standardized mortality ratio [SMR] = 125; 95% CI: 110-142) and urinary bladder cancer mortality (SMR = 279; 95% CI: 192-391), and increased all malignancy (standardized incidence ratio [SIR] = 142; 95% CI: 125-160), oesophageal (SIR = 203; 95% CI: 108-347), respiratory tract (SIR = 154; 95% CI: 120-194) and bladder (SIR = 394; 95% CI: 268-559) cancer incidence. Women had elevated oesophageal (SMR = 313; 95% CI: 124-664) and bladder (SMR = 311; 95% CI: 149-571) cancer mortality and elevated all malignancy (SIR = 124; 95% CI: 106-144), oesophageal (SIR = 348; 95% CI: 140-719), and bladder (SIR = 861; 95% CI: 458-8002) cancer incidence. Bladder cancer rate increased with employment duration and younger age first hired. Rate estimates were highest among beta-naphthylamine exposed workers but was also increased among workers with other chemical exposures. A cancer prevention and control effort that limited benzidine exposure to < or = 3 years was apparently unsuccessful as indicated by a significant excess of bladder cancer (SIR = 1773; 95% CI: 356-5180) among these workers. CONCLUSIONS Relative rates of oesophageal, lung, and stomach cancer were also elevated among all workers, but did not increase with total years worked, age first hired, or year first hired, suggesting a non-occupational aetiology.