A cadmium/zinc hyperaccumulator extracted metals from four contaminated soils over three years in a glasshouse experiment. Changes in plant metal uptake and soil total (aqua regia-extractable) and available metals were investigated. Plant Cd concentrations in a high-Cd acid soil and plant Zn concentrations in two acid soils decreased during repeated phytoextraction and were predicted by soil available metal concentrations. However, on repeated phytoextraction, plant Cd concentrations remained constant in lightly Cd-polluted acid soils, as did plant Cd and Zn in alkaline soils, although soil available metal concentrations decreased markedly. After phytoextraction acid soils showed much higher total metal removal efficiencies, indicating possible suitability of phytoextraction for acid soils. However, DGT-testing, which takes soil metal re-supply into consideration, showed substantial removal of available metal and distinct decreases in metal supply capacity in alkaline soils after phytoextraction, suggesting that a strategy based on lowering the bioavailable contaminant might be feasible.