Mark Little replies

I am grateful for Dr Mossman's comments on my article. As I stated in the penultimate paragraph, sources of uncertainty other than the dosimetry have to be taken into account in deriving risk estimates from the A-bomb data. Constraints of space did not allow me to go into these in detail. I would concur with Dr Mossman that extrapolation of risks to low doses and low dose rates is one of the more substantial of these, others being the extrapolation of risks to the end of life and across populations, for example from a Japanese to a UK population (NCRP Report No. 126, NCRP, Bethesda, 1997).

The various sorts of bias (selection, ascertainment and so on) to which epidemiological studies are prone should also be considered. As discussed in my article, selection bias in A-bomb survivors may be significant, and as with the problems of extrapolation of dose and dose rate, may largely invalidate the A-bomb risk estimates. However, as noted previously, cancer risks derived from A-bomb survivor data are statistically consistent with those observed in groups exposed at moderate-to-low doses and dose rates (M. P. Little et al. Radiat. Res. 151, 218–224; 1999; R. Wakeford et al. Int. J. Radiat. Biol. 79, 293–309; 2003).