Background: This retrospective follow-up study evaluated mortality during 1970-1996 among 6,956 employees at a petrochemical research facility in Illinois. Methods: Standardized mortality ratios (SMRs) compared employees' mortality rates with those of the Illinois general population. Poisson regression procedures estimated rate ratios for various subject subgroups compared to other facility employees. Results: Subjects had 267 observed/524 expected deaths (SMR = 51) from all causes combined and a large deficit of deaths from all cancers (76/136, SMR = 56) and from most other major diseases. Other results included fewer than expected brain cancers (1/4.0, SMR = 25) and a slight increase in colorectal cancer (20/14, SMR = 139) that was concentrated in white male scientists employed for one of the three main companies at the facility (SMR = 295, RR = 2.6). Conclusions: The deficit of brain cancer deaths contrasts with an excess incidence seen in a companion study. Subjects' generally favorable mortality experience probably reflects socioeconomic advantages of employees relative to the Illinois general population. © 2001 Wiley-Liss, Inc.