Data on the gastrointestinal absorption of 12 elements have been reviewed. In each case, absorption is expressed as the fraction of the ingested element absorbed to blood, referred to as the f1 value, applying to intakes of unspecified chemical form by average population groups. The level of confidence in individual absorption values has been estimated in terms of lower and upper bounds, A and B, such that there is judged to be roughly a 90% probability that the true central value is no less than A and no greater than B. Ranges are proposed for intakes by adults, 10-year-old children and 3-month-old infants. Uncertainty in f1 values (B/A) ranged from 10% to factors of 100-400. The lowest uncertainties were for the well absorbed elements, H, I and Cs, for which there are good data, and the greatest uncertainties were for less well absorbed elements for which few data are available, particularly Zr and Sb. Ranges were generally wider for children and infants than for adults because of the need to allow for the likelihood of increased absorption with only limited data in support of the proposed values. The largest ranges were for 3-month-old infants, reflective lack of knowledge on the time-course and magnitude of possible increased absorption in the first few months of life. For each age group, ICRP values of absorption tend towards the upper bound of the ranges, indicating a degree of conservatism in th calculation of ingestion dose coefficients. Examination of the effect of the proposed confidence intervals for f1 values on uncertainties in dose coefficients for ingested radionuclides showed that there was no direct relationship. For some radionuclides, uncertainties in effective dose were small despite large uncertainties in f1 values while for others the uncertainties in effective doses approached the corresponding values for uncertainty in f1 values. These differences reflect the relative contributions to effective dose from cumulative activity in the contents of the alimentary tract, which in many cases is insensitive to uncertainties in f1, and cumulative activity of the absorbed radionuclide in systemic tissues, which is proportional to f1. In general, uncertainties in effective close for children and infants exceeded those in adults as a result of greater uncertainties in f1 values for the younger age groups. However, this effect was reduced in some cases by shorter retention times of absorbed nuclides in body tissues and organs.