With the availability of sensitive multielement analytical capability, it is no longer essential to use chemically aggressive soil tests to extract measurable levels of most trace elements. However, the relative abilities of mild and aggressive extractants to assess metal bioavailability in soils have rarely been compared. A greenhouse experiment was carried out to compare a mild soil extractant (hot 0.01MCaCl2) with an aggressive one (Mehlich 3) for predicting accumulation of trace metals by red clover (Trifolium pratenseL.). Clover was grown on large columns of nonacid fine-textured and acid coarse-textured soils that had been amended several years earlier by a heavy application of sewage sludge products, and pH was subsequently adjusted using CaCO3or H2SO4. The soil extractants (CaCl2and Mehlich 3) and clover tissue were analyzed for trace metals (As, Cd, Mo, Cu, Ni, Mn, Pb, and Zn) by axial-view ICP spectrophotometry. Linear regression analyses were performed to relate the concentration of each trace element in the red clover tissue to the concentration extracted from the soil. The results indicate that CaCl2extraction is more reliable than Mehlich 3 extraction when evaluating plant availability of trace elements in soils with a wide range of properties (especially pH). The strongly acidic nature of the Mehlich 3 extractant caused large quantities of metals such as Zn, Cd, Cu, and Ni to be extracted from metal-contaminated soils even when the plant availability of these metals was low because of near-neutral soil pH or high clay and organic matter content. Conversely, in coarse-textured and acid soils containing lower total concentrations of metals, plant-available metals were often relatively high, yet Mehlich 3 frequently extracted smaller quantities of metals from these soils than from the near-neutral soils. We conclude that dilute CaCl2is preferable to Mehlich 3 as a universal soil extractant for estimating short-term trace metal availability to crops.