In this experiment, three grasses, bermuda grass (Cynodon dactylon), bent grass (Agrostis palustris Huds.), lawn grass (Zoysia japonica), and a shallow-rooted legume, white clover (Trifolium repens L) were planted into uncontaminated soil and dibenzofuran (DBF)-contaminated soil. The germination rates of all plants were investigated using contaminated soils to evaluate their sensitivities to DBF. During 2 months of growth, the root biomass and heterotrophic microbial numbers were measured in order to evaluate the potential of remediation. Furthermore, the number of DBF-degrading bacteria was counted to evaluate plants that enhance the microbial DBF degradation potential in contaminated soil. The DBF-removal performance of four plant species was also compared. Regardless of the contamination of DBF, white clover had not only the highest root biomass, but also the highest DBF-degrading bacterial number!; compared to those of the other three grasses. Moreover, white clover-planted contaminated soil exhibited the highest rate of DBF removal among all tested plants. These results suggest that microbial populations capable of degrading DBF were selectively increased by the addition of DBF in the rhizosphere, and also indicate that the presence of plants significantly enhances the reduction of DBF in soils. Based upon these results, white clover was selected for the further investigation of the phytoremediation of dioxin-contaminated soil. (C) 2009 Elsevier B.V. All rights reserved.