We undertook a cohort study of all male pilots employed since January 1, 1950, by CP Air, now Canadian Airlines International. A total of 913 eligible pilots--630 active and 283 no longer employed--contributing 18,060 person-years of observation, were identified through company records. As of October 31, 1988, current status was obtained on 891 (97.6%). Standardized mortality ratio (SMR) and standardized incidence ratio (SIR) were used to compare, respectively, the mortality and cancer incidence of the cohort with that of the British Columbia population. Statistical significance of the SMR and SIR by comparison with the Poisson distribution (p less than 0.05 one-sided) and 90% confidence intervals (CI) were calculated. Excess deaths were observed for aircraft accidents (No. = 23; SMR = 21.29; p less than 0.001; CI 14.60, 30.20), brain cancer (No. = 4; SMR = 4.17, p = 0.017; CI 1.40, 9.50) and rectal cancer (No. = 3; SMR = 4.35; p = 0.033; CI 1.20, 11.20). Excess cancer incidence was noted for non-melanoma skin cancer (No. = 26; SIR = 1.59; p = 0.017; CI 1.10, 2.20), brain cancer (No. = 4; SIR = 3.45; p = 0.030; CI 1.20, 7.90) and Hodgkin's Disease (No. = 3; SIR = 4.54; p = 0.030; CI 1.20, 11.70). These findings, suggesting an excess risk for certain cancers in commercial airline pilots, are based on small numbers and need to be confirmed in larger cohort studies.