Processes in the rhizosphere of metal hyperaccumulator species are largely unknown. We investigated root-induced changes of Ni biogeochemistry in the rhizosphere of Thlaspi goesingense Ha´la´csy in a rhizobox experiment and in related soil chemical and Ni uptake studies. In the rhizobox, a root monolayer was separated from rhizosphere soil by a nylon membrane. Rhizosphere soil was then sliced into 0.5 mm layers and analyzed for changes in soluble (water-extractable, Ni S ) and labile (1 M NH 4 NO 3 -extractable, Ni L ) Ni pools. Ni L in the rhizosphere was depleted due to excessive uptake in T. goesingense. Ni S in the rhizosphere increased in contrast to expectations based on the experimental Ni desorption isotherm. Mathematical simulations following the Tinker-Nye-Barber approach overestimated the depletion of the Ni L and predicted a decrease of Ni S in the rhizosphere. In a hydroponic experiment, we demonstrated that T. goesingense takes up Ni 2+ but excludes metal-organic complexes. The model output was then improved in later versions considering this finding. A sensitivity analysis identified I max and K m , derived from the Michaelis-Menten uptake kinetics experiment to be the most sensitive of the model parameters. The model was also sensitive to the accuracy of the estimate of the initial Ni concentration (C Si ) in soil solution. The formation of Ni-DOM complexes in solution could not explain the poor fit as in contrast to previous field experiments, the correlation between soluble Ni and dissolved organic carbon (DOC) was weak. Ion competition of Ni with Ca and Mg could be ruled out as explanation of enhanced Ni solubility in the rhizosphere as the molar ratio of Ni/(Ca + Mg) in solution was not affected. However, a decreased Vanselov coefficient Kv near the root plane indicated (an apparent) lower selectivity of the exchange complex for Ni, possibly due to adsorption of oxalate exuded by T. goesingense roots or associated rhizosphere microbes. This conclusion is supported by field data, showing enhanced oxalate concentrations in the rhizosphere of T. goesingense on the same experimental soil. The implications for phytoextraction and bio-available contaminant stripping (BCS) as well as for future modeling and experimental work are discussed. * FAX No: +43-1-47654-3130.

SummarySorption curves are commonly used to predict the partitioning of ions between the soil solid and solution in ion transport and rhizosphere models. We review methods for measuring metal desorption for such purposes, and we illustrate the problems involved by means of measurements of Ni desorption from a contaminated soil as an example. We find that measurements of metal concentration in shaken soil suspensions at a range of solution:soil ratios give similar desorption curves to the more tedious method of repeated extraction at a constant, smaller and therefore more realistic solution:soil ratio. We also find that desorption is sensitive to the nature and concentration of the background electrolyte used, and to interactions between these and additions of organic chelating agents of the sort excreted by roots and rhizosphere microbes.