Radiologists and radiotherapists were one of the earliest occupational groups to be exposed to ionizing radiation. Their patterns of mortality provide information on the long-term effects of fractionated external radiation exposure. British radiologists who registered with a radiological society between 1897 and 1979 have now been followed-up until 1 January 1997, and the mortality experience examined among those who registered with a society after 1920, when the first radiological protection recommendations were published. The observed number of cancer deaths in those who registered after 1920 was similar to that expected from death rates for all medical practitioners combined (SMR=1.04; 95% CI 0.89-1.21). However, there was evidence of an increasing trend in risk of cancer mortality with time since first registration with a radiological society (p=0.002), such that in those registered for more than 40 years there was a 41% excess risk of cancer mortality (SMR=1.41; 95% CI 1.03-1.90). This is probably a long-term effect of radiation exposure in those who first registered during 1921-1935 and 1936-1954. There was no evidence of an increase in cancer mortality among radiologists who first registered after 1954, in whom radiation exposures are likely to have been lower. Non-cancer causes of death were also examined in more detail than has been reported previously. There was no evidence of an effect of radiation on diseases other than cancer even in the earliest radiologists, despite the fact that doses of the size received by them have been associated with more than a doubling in the death rate among the survivors of the Japanese atomic bombings.