A case control study of primary brain tumor cases among employees of a petrochemical production facility (SIC-2911) was conducted. Seventeen gliomas were identified from the records of employees who had died from brain tumors between 1950 and 1977. For each case, a pool of comparisons was selected and matched by race, sex, age, and employment. Employee residence location was also included as a factor. For data collection, facility personnel completed coding sheets containing demographic data, factory department codes, job codes, dates of employment, and vital status. Chemical exposures were determined from records showing production dates and departmental locations, and facility engineers characterized feed stocks. Tissue specimens of deceased workers, where available, were examined to confirm tumor type. Highest glioma risk was associated with first employment in the 1940s and 1950s and with residence in a community that had a chemical dump listed by the Environmental Protection Agency as a high priority site for emergency cleanup. The greatest apparent factory risks are associated with exposure to carbon-dioxide (124389), diethyl-sulfate (64675), diethylene-glycol (111466), ethanol (64175), ethylene (74851), isopropanol (67630), methane (74828), tetraethylene-glycol (112607), and vinyl-acetate (108054). The authors conclude that, in the absence of supporting data, such as in the case of carbon-dioxide, some of the compounds can be eliminated from consideration as potential carcinogens. Residential associations appear to be stronger than factory associations. Further studies in this area are recommended, with attention to local use of pesticides and other possible factors.