In many epidemiological and medical studies, a number of cancer mortalities in categorical classification may be considered as having Poisson distribution with person-years at risk depending upon time. The cancer mortalities have been evaluated by additive or multiplicative models with regard to background and excess risks based on several covariances such as sex, age at the time of bombings, time at exposure, or ionizing radiation, cigarette smoking habits, duration of smoking habits, etc. An interest herein is to examine an additive, synergistic, or antagonistic relationship between radiation exposures and cigarette smoking habits for cancer mortalities. The results revealed a highly significant antagonistic in uence for cancer mortalities from all non-hematologic findings, lung and respiratory system with negative interaction between radiation exposures and cigarette smoking amounts.