This paper reports the mortality experience froin 1948 to 1989 of 2,504 maintenance employees who had a minimum of one year of employment in jobs with potential exposure to asbestos at a Texas refinery and petrochemical plant. For the purposes of this study, "potential exposure" is equated with those jobs or crafts having the greatest direct potential proximity to, or which worked directly with, asbestos-containing materials, especially asbestos-containing thermal insulation. Appro.xiniate1.y one-half of the study population had I0 years or longer potential exposure, and 80% had their first potential exposure before 1970. The total population exhibited significantly lower mortality for all causes, the standardized mortality ratio (SMR = 77); and for all cancer (SMR = SS), as compared to residents in the surrounding communities. Statistically significant deficits in mortality were also observed in a number of noncancerous diseases such as heart disease (SMR = 78; 95% CI = 69-88), nonmalignant respiratory disease (SMR = 70; 95% CI = S0-95), and cirrhosis of the liver (SMR = 44; 95% CI = 22-79). Mortality among employees who had 20 vears or longer since their first potential exposure was also examined: the pattern of mortality was similar to that exhibited by the total cohort, with a slight increase in the SMR for most of the causes. The only statistically significant excess of mortality ,found was a fourfold increase in mesothelioma (5 observed and 1.2 expected deaths); the SMR wav 428 (95% CI = 139-996) for the total coliort and was 469 (95% CI = 152-1093) for those who had 20 years or more since first potential exposure. hi contrast to asbestos industry worker studies, mortality for lung cancer was substantitilly lower than the general population (SMR = 81; 9.5% CI = 63-103). The observed number of deaths for cancer of the larynx was virtually the same as expected ( 3 observed vs. 2.8 expected). This study also showed decreased mortality for cancers of gastrointestinal organs such as the esophagus (SMR = 78), stomach (SMR = 63), large intestine (SMR = 91). rectum (SMR = 55), or pancreas (SMR = 90)--cancers that have been reported to be elevated in studies of various industry workers directly exposed to asbestos. 0 I996 Wilt>l-LlSb, IIIC.