Mercury concentrations were quantified in Halimione portulacoides (roots, stems and leaves) as well as in sediments from eight Portuguese estuarine systems, covering seventeen salt marshes with distinct degrees of mercury contamination. The concentration of mercury in the sediments ranged from 0.03 to 17.0 microg g(-1). The results show that the accumulation of mercury differed according to the organ of the plant examined and the concentration of mercury in the sediments. Higher mercury concentrations were found in the roots (up to 12.9 microg g(-1)) followed by the leaves (up to 0.12 microg g(-1)), while the stems had the lowest concentrations (up to 0.056 microg g(-1)). A linear model explained the relation between the concentrations of mercury in the different plant organs: roots and stems (R(adj)(2)=0.75), stems and leaves (R(adj)(2)=0.85) and roots and leaves (R(adj)(2)=0.78). However, the results show that the variation of mercury concentration in the roots versus mercury concentration in the sediments was best fitted by a sigmoidal model (R(adj)(2)=0.89). Mercury accumulation in the roots can be described in three steps: at a low range of mercury concentrations in the sediments (from 0.03 up to 2 microg g(-1)), the accumulation of mercury in roots is also low reaching a maximum concentration of 1.3 microg g(-1); the highest rates of mercury accumulation in the roots occur in a second step, until the concentrations of mercury in the sediments reach approximately 4.5 microg g(-1); after reaching this maximum value, the rate of mercury accumulation in the roots slows down leading to a plateau in the concentration of mercury in the roots of about 9.4 microg g(-1), which corresponds to a mercury concentration in the sediments of about 11 microg g(-1). A linear model explained also the accumulation of mercury in leaves versus the mercury concentration in the sediments (R(adj)(2)=0.88). Differences in responses of roots and leaves are explained by the dynamics of the plant organs: old roots are mineralised in situ close to new roots, while leaves are renewed. Previous studies have already shown that H. portulacoides is a bioindicator for mercury and the results from this work sustain that H. portulacoides may also be used as a biomonitor for mercury contamination in salt marshes. Nevertheless, caution should be taken in the application of the models, concerning the life cycle of the species and the spatial variability of the systems.