Using a modification of the traditional cohort mortality methodology, risk estimates of death due to brain tumors were calculated for production and nonproduction employees of Dow Chemical U.S.A., Texas Division. A total of 25 malignant, benign and unspecified brain tumors were identified using a geographically limited record-linkage process. Expected deaths were extrapolated from 1,666 white males in a 5% sample of the 1940 through 1977 total workforce. The resulting sample-based standardized mortality ratios (which were not evaluated for statistical significance) suggest, at most, only a slight increased risk of mortality from brain tumors for the overall time period. There was, however, a probable elevated risk associated with first employment prior to 1945. To date, diminished risk has been noted for those hired during subsequent years.