The survivors of the atomic bombings in Hiroshima and Nagasaki are a general population of all ages and sexes and, because of the wide and well characterised range of doses received, have been used by many scientific committees (International Commission on Radiological Protection (ICRP), United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR), Biological Effects of Ionizing Radiations (BEIR)) as the basis of population cancer risk estimates following radiation exposure. Leukaemia was the first cancer to be associated with atomic bomb radiation exposure, with preliminary indications of an excess among the survivors within the first five years after the bombings. An excess of solid cancers became apparent approximately ten years after radiation exposure. With increasing follow-up, excess risks of most cancer types have been observed, the major exceptions being chronic lymphocytic leukaemia, and pancreatic, prostate and uterine cancer. For most solid cancer sites a linear dose response is observed, although in the latest follow-up of the mortality data there is evidence (p = 0.10) for an upward curvature in the dose response for all solid cancers. The only cancer sites which exhibit (upward) curvature in the dose response are leukaemia, and non-melanoma skin and bone cancer. For leukaemia the dose response is very markedly upward curving, indeed largely describable as a pure quadratic dose response, particularly in the low dose (0-2 Sv) range. Even 55 years after the bombings over 40% of the Life Span Study cohort remain alive, so continued follow-up of this group is vital for completing our understanding of long-term radiation effects in people.In general, the relative risks per unit dose among the Japanese atomic bomb survivors are greater than those among comparable subsets in studies of medically exposed individuals. Cell sterilisation largely accounts for the discrepancy in relative risks between these two populations, although other factors may contribute, such as the generally higher underlying cancer risks in the medical series than in the Japanese atomic bomb survivors. Risks among occupationally exposed groups such as nuclear workforces and underground miners are generally consistent with those observed in the Japanese atomic bomb survivors. In general, consistent patterns of variation of risk with age at exposure are also seen in all studies—risks for all cancer types diminish with increasing age at exposure. There are also excess risks of various types of non-malignant disease in the Japanese atomic bomb survivors, in particular cardiovascular, respiratory and digestive diseases. Indeed, risks are elevated to much the same degree for a number of non-malignant disease endpoints, suggestive of bias. However, in contrast with the cancer data, there is much less consistency in the pattern of risk between the atomic bomb survivors and other exposed groups; for example, radiation-associated respiratory and digestive diseases have not been seen in these other groups. Although cardiovascular risks have been seen elsewhere, particularly in medically exposed groups, in contrast with the cancer data there is much less consistency in risk between studies: risks per unit dose in epidemiological studies vary over at least two orders of magnitude, possibly as a result of confounding factors. In the absence of a convincing mechanistic explanation of epidemiological evidence, at present a cause-and-effect interpretation of the reported statistical associations for cardiovascular disease is unreliable but cannot be excluded. Further epidemiological and biological evidence will allow a firmer conclusion to be drawn. Presented as an invited paper at the 21st LH Gray Conference 'The Radiobiology—Radiation Protection Interface', Royal College of Physicians, Edinburgh 4-6 June 2008.