The success of risk assessment of metal contaminated soils depends on how precisely one can predict the bio-availability of metals in soil and transfer to the human food chain. In the present investigation, we tested several formulations of the 'free-ion activity model (FIAM)' to predict uptake of Cd, Zn and Cu by perpetual spinach (Beta vulgaris, Cicla) grown on a range of soils amended with sewage sludge. The model was parameterised using data measured on samples of pore water extracted by centrifugation and with porous Rhizon samplers installed within the rhizosphere of the growing plants. Free ion activities (M 2+ ) were estimated following speciation of solution data using version 6 of the 'Windermere Humic Aqueous Model (WHAM-VI). For all three metals, the best formulation of the FIAM appeared to require only one hypothetical root sorption site without competition from protons. Values of (M 2+ ) could also be predicted satisfactorily from a pH-dependent Freundlich relation. Thus, from a combined FIAM-Freundlich relation and population dietary information, it was possible to estimate risk (hazard quotients) to consumers from very simple soil measurements: extractable metal content (0.05 M EDTA (Zn and Cu) or 1 M CaCl 2 (Cd)), soil humus content and pH. The role of increased soil organic matter content and soil pH, in reducing risk to consumers, is illustrated for Cd in a hypothetical soil at the current UK statutory Cd limit for sludge application to agricultural land.