Linkage with records of the Connecticut Tumor Registry was used to determine cancer incidence in a cohort of workers (n = 984) at a benzidine manufacturing facility. Compared to the findings for the Connecticut population, there was a statistically significant excess of bladder tumor among male cohort members [standardized incidence ratio (SIR) = 343; 95% confidence limits (CL) = 148, 676; n = 830], which was confined to those with the highest estimated level of benzidine exposure (SIR = 1,303; CL = 479, 2,839; n = 105). No significantly elevated risks were found for cancers at other anatomic sites in men or at any anatomic sites in women; nor was there any pattern of increasing risk with increasing benzidine exposure for sites other than bladder. In addition, the elevated bladder cancer risk was greater for men first employed during the earliest years of the plant, namely, 1945-49 (SIR = 976; CL = 262, 2,498) as compared to those first employed in 1950-54 (SIR = 213; CL = 3, 1,184) after equalization of duration of follow-up. These results suggest that the major preventive measures instituted around 1950 may have reduced bladder cancer risk in this plant.