A greenhouse experiment was performed to evaluate effectiveness of goosegrass (Eleusine indica) in phytoremediation of soil contaminated with 8,247 mg kg −1 of total petroleum hydrocarbons (TPH). We determined seed germination toxicity, soil microbial viable counts, catalase activity, dehydrogenase activity (DHA), and the concentrations of TPH and 14 polycyclic aromatic hydrocarbons (PAHs) in soil and plant tissue. After 5 months, the initial level of contamination was reduced by 47% in planted soil, whereas it was only reduced by 11% in nonplanted soil. Bacterial numbers were 72 times greater in the rhizosphere treated soil than in the unvegetated treatment at the end of the study. There was no correlation between microbial counts or DHA and catalase activity, and the correlation between microbial counts and DHA was weak. Significant chemical reduction of H 2 O 2 , caused by the soil fabric, was observed in the determination of catalase activity. In case of vegetated treatment, 32% of PAHs was removed, but only 5% of PAHs was dissipated in the unvegetated pots. Gas chromatography/mass spectrometry analysis of plant tissue indicated that a low amount of PAHs (25.50 mg kg −1 dry biomass) was detected in goosegrass roots growing in the contaminated soil, and no uptake into the shoots was occurring.