To test the hypothesis that a parent's job exposure to radiation affects his or her child's risk of cancer, the authors compared this exposure during the year before the child's birth for parents of children with and without cancer. Parents of children with cancer were no more likely to have worked in occupations, industries, or combined occupations and industries with potential ionizing radiation exposure. Bone cancer and Wilms' tumor occurred more frequently among children of fathers in all industries with moderate potential ionizing radiation exposure. Children with cancer more often had fathers who were aircraft mechanics (odds ratio (OR) = infinity, one-sided 95% lower limit = 1.5; P = 0.04). Although four of these six were military aircraft mechanics, only children whose fathers had military jobs with potential ionizing radiation exposure had an increased cancer risk (OR = 2.73; P = 0.01). Four cancer types occurred more often among children of fathers in specific radiation-related occupations: rhabdomyosarcoma among children whose fathers were petroleum industry foremen; retinoblastoma among children whose fathers were radio and television repairmen; central nervous system cancers and other lymphatic cancers among children of Air Force fathers. Because numbers of case fathers are small and confidence limits are broad, the associations identified by this study need to be confirmed in other studies. Better identification and gradation of occupational exposure to radiation would increase the sensitivity to detect associations.