Case-control studies and a prospective study have suggested a positive relation between serum cholesterol and brain tumors. To examine this association further, mortality from malignant brain tumors among men who participated in the Multiple Risk Factor Intervention Trial (a prospective study, 1973-1986) who indicated they were not black were examined. No relation was seen between age-standardized mortality rates and baseline serum cholesterol. Excluding deaths occurring during the first 5 years or adjusting for median census tract income did not alter this finding. This suggests that no generalizable relation between serum cholesterol and primary malignant brain tumors exists. An environmental factor associated with serum cholesterol in some, but not all populations, may explain the apparently contradictory results.