We compared the cancer incidence of male United States Air Force (USAF) aircrew (342 cancers, 532,980.97 man-years) with non-flying Air Force officers (827 cancers, 1,084,370.08 man-years) between 1975-89. METHODS Incident cancer cases for both aviators and non-flying officers were obtained from USAF hospitalization records. Age-adjusted standardized incidence ratios (SIR's) were calculated for aircrew using data from the National Cancer Institute's Surveillance Epidemiology and End Results (SEER) program. Aviator age-adjusted cancer rate ratios were also obtained using non-flying officers as an internal comparison group. RESULTS We observed statistically significant excesses of aircrew cancers for all sites, testis, and urinary bladder. All other aviator cancer classifications were not significantly different from the comparison cohort; most notably, cancers of the colon and rectum, skin (both malignant melanoma and non-epithelial), brain and nervous system, Hodgkin's Disease and leukemias. CONCLUSIONS Previous studies of commercial pilots that demonstrated excesses of these cancers may have been biased by the use of external comparison groups. We used an internal comparison population to reduce selection bias, information bias and confounding. From these data we detected notable excess aircrew cancer risk for cancers of the testis, urinary bladder, and all sites combined.