As part of the "IARC International Register of Persons Exposed to Phenoxy Herbicides and Contaminants," a cohort of workers who manufacture and prepare chlorophenoxy herbicides was recruited in The Netherlands. The cohort comprised 2,310 workers from two plants, operated by different companies, who were followed during the periods 1955-1985 and 1965-1986, respectively. In 1963, there had been an industrial accident in one factory with concomitant release of dioxin into the environment. Loss to follow-up was 3%. Mortality data on 963 exposed and 1,111 nonexposed men were evaluated by external and internal comparison. Compared with national rates, total mortality (94 deaths, standardized mortality ratio [SMR] = 101; 95% confidence interval [CI], 82-124) and cancer mortality (31 deaths, SMR = 107; 95% CI, 73-152) for exposed workers were not significantly increased. A statistically insignificant increase was observed for non-Hodgkin's lymphoma (2 deaths, SMR = 299; 95% CI, 36-1,078). No cases of soft-tissue sarcoma were encountered. There was no increase in either total mortality (25 deaths, SMR = 111; 95% CI, 72-163) or cancer mortality (10 deaths, SMR = 137; 95% CI, 66-252) among the 139 workers probably exposed to dioxins during the 2,4,5-trichlorophenol production accident or the subsequent clean-up operations. Compared with nonexposed workers, exposed workers did not exhibit a higher total mortality (rate ratio [RR] = 1.28; 95% CI, 0.89-1.82). Mortality due to all cancers (RR = 1.7; 95% CI, 0.9-3.4) and respiratory cancer (RR = 1.7; 95% CI, 0.5-6.3) was insignificantly elevated. These findings suggest that the increases in cancer mortality among workers exposed to phenoxy herbicides and chlorophenols may be attributable to chance. Lack of power prevented evaluation with respect to specific cancers.