To assess the possible carcinogenicity of phenoxy herbicides and related chlorophenols and dioxins, the International Agency for Research on Cancer is coordinating an international collaborative study of workers exposed to these compounds in their production or use. Four British cohorts of chemical manufacturers which have been recruited to the survey are described. They comprise a total of 2239 men employed during 1963-85. These subjects were traced to 31 December 1987 through the National Health Service Central Register and the National Insurance Index, and their mortality compared with that in the national population. Two deaths were from non-Hodgkin's lymphoma with 0.87 expected. Both deaths occurred more than 10 years after first exposure to phenoxy compounds. One further non-Hodgkin's lymphoma was registered in a living subject with probable exposure to phenoxy compounds. No cases of soft tissue sarcoma or Hodgkin's disease were recorded. A nonsignificant excess of lung cancer (19 deaths observed, 14.2 expected) is probably attributable to chance or a confounding effect of smoking. In one cohort only there was increased mortality from circulatory disease (34 deaths observed, 20.4 expected). A nested case-control study did not point to any occupational cause for this excess, but further evaluation will be needed during continued follow up.