The findings of studies investigating whether exposures to ionizing radiation before birth, either pre- or post-conception, increase the risk of childhood cancer have provoked much scientific controversy. An epidemiological association between the abdominal exposure of pregnant women to diagnostic X-rays and childhood cancer was first reported in the 1950s, while an association between the recorded dose of radiation received occupationally by fathers before the conception of their offspring and childhood leukemia was reported only recently in 1990. The scientific interpretation of these particular statistical associations is by no means straightforward, but the latest analyses of intrauterine irradiation and childhood cancer indicate that a causal inference is likely. Scientific committees have adopted risk coefficients for the intrauterine exposure of somatic tissues, which for childhood leukemia are comparable to those accepted for exposure in infancy, although questions remain about the level of risk of childhood solid tumors imparted by exposure to radiation in utero and shortly after birth. In contrast, the association between paternal preconceptional radiation dose and childhood leukemia has not been confirmed by studies using objectively determined doses. The original association has been found to be restricted to children born in one village, it does not extend to cancers other than leukemia, and it is markedly inconsistent with the established body of knowledge on radiation-induced hereditary disease. A causal interpretation of this association has effectively been abandoned by scientific authorities.