A petroleum-contaminated soil remediation study was conducted in a greenhouse. The system consisted of 36 pots, 12 were vegetated with squash, 12 were vegetated with fescue grass and the last 12 units served as unvegetated controls. For each group, three treatments were applied, 1) the addition of single dose of nutrients, 2) the addition of double dose of nutrients and 3) the addition of double dose of nutrients and acclimated bacteria to the irrigation water. The two plants were selected to represent extremely different species in terms of transpiration potential and root density in order to better understand the mechanisms involved in phytoremediation. Clay sandy soil (3: 1, by weight) was spiked with Fuel oil No.2 and allowed to weather for 1 week before it was placed in the pots. Under all study treatments, units vegetated with fescue grass had significantly less TPH concentration than the unvegetated controls after 10 weeks. Units vegetated with squash had significantly less TPH concentration than the un vegetated controls after 10 weeks only under treatment 3. Squash significantly accumulated TPH in the shoot under all treatments while grass shoot accumulated TPH significantly only under treatment 1. The mechanisms most important in phytoremediation seemed to include plant uptake of TPH, desorption and enhanced bioavailability by transpiration-induced water movement in the rhizosphere and root stimulation of microbial degradation.