Evaluation of the potential environmental risk posed by metals depends to a great extent on modeling the fate and mobility of metals with soil-solution partitioning coefficients (Kd). However, the effect of biological cycling on metal partitioning is rarely considered in standard risk assessments. We determined soil-solution partitioning coefficients for 5 metals (Cd, Zn, Pb, Co and Ni) at 46 forested sites that border the Precambrian Shield in central Ontario, where soil pHaq varied from 3.9 to 8.1. Foliage from the dominant tree species and forest floor samples were also collected from each site to compare their metal levels with Kd predictions. Analogous to other studies, log Kd values for all metals were predicted by empirical linear regression with soil pH (r2=0.66-0.72), demonstrating that metal partitioning between soil and soil solution can be reliably predicted for relatively unpolluted forest mineral soils by soil pH. In contrast, whereas the so-called bioavailable water-soluble metal fraction could be predicted from soil pH, metal concentrations in foliage and the forest floor at each site were not consistently related to pH. Risk assessment of metals should take into account the role of biota in metal cycling and partitioning in forests, particularly if metal bio-accumulation and chronic toxicity in the food chain, rather than metal mobility in soils, are of primary concern.