Abstract
It is important for both radiation protection and scientific reasons to understand the age–time patterns of radiation cancer risk. This is surprisingly difficult even for acute exposures and much more so for prolonged exposures. I shall provide current information on this for solid cancers among atomic-bomb survivors, pointing out some of the difficulties in description and interpretation. I shall then take up some stochastic considerations regarding accumulation of mutations, which may help in dealing with these difficulties. These considerations are highly idealised, and their consequences should mainly be used only for guidance rather than as a primary basis for descriptive analyses. They are particularly suitable for this because they provide insights fairly independent of parameter values in the stochastic models involved.