Short-term bioassay techniques that measure root development of seedlings in soil can be used to evaluate the applicability of chemical methods to predict the Al toxicity of a soil. Wheat (Triticum aestivumcv. Hart) seedlings were grown for 3 d in 20 acidic soil horizons, and root length was related to amounts and forms of Al in soil solution and Al extracted by 0.01MCaCl2, 1MKCl, 0.5MCuCl2, and 0.33MLaCl3. With the exception of Al extracted by 0.01MCaCl2, soil extraction methods were not effective in predicting root growth. Root growth was not closely related to total soil solution Al or soil solution Al reacting with 8-hydroxyquinoline, pyrocatechol violet, or aluminon. These colorimetric reagents reacted with nontoxic forms of Al, including organically bound Al, and overestimated the toxic component. Activities of Al species in soil solution were estimated from soil solution composition using the GEOCHEM program. The calculated Al3+activity ({Al3+}) ranged from 0 to 119 μMand taken alone was not a good predictor of root growth. It is likely that calculated {Al3+} overestimated actual {Al3+} due to the difficulty of measuring all Al-complexing ligands in soil solution. Root growth in these soils was best described by a multiple regression equation containing soil solution pH, Ca, and Al. The results of this study illustrate the limitations associated with the current methods used to estimate toxic Al levels. New methods are needed to directly measure the toxic Al component or components in soil solution. Until better chemical or instrumental techniques are developed, the short-term bioassay technique should serve as a useful means of assessing the potential Al toxicity of a soil.