A nested case-control study using conditional logistic regression was conducted to evaluate the exposure-response relationship between external ionizing radiation exposure and leukemia mortality among civilian workers at the Portsmouth Naval Shipyard (PNS), Kittery, Maine. The PNS civilian workers received occupational radiation exposure while performing construction, overhaul, repair and refueling activities on nuclear-powered submarines. The study age-matched 115 leukemia deaths with 460 controls selected from a cohort of 37,853 civilian workers employed at PNS between 1952 and 1992. In addition to radiation doses received in the workplace, a secondary analysis incorporating doses from work-related medical X rays and other occupational radiation exposures was conducted. A significant positive association was found between leukemia mortality and external radiation exposure, adjusting for gender, radiation worker status, and solvent exposure duration (OR = 1.08 at 10 mSv of exposure; 95% CI = 1.01, 1.16). Solvent exposure (including benzene and carbon tetrachloride) was also significantly associated with leukemia mortality adjusting for radiation dose, radiation worker status, and gender. Incorporating doses from work-related medical X rays did not change the estimated leukemia risk per unit of dose.