A laboratory incubation study was conducted on a petroleum-contaminated soil from Macquarie Island in sub-Antarctic Australia to develop a target O 2 level for bioventing. The soil was amended with NH 4 NO 3 (175mg Nkg −1 soil) and 14 C-hexadecane (250mgkg −1 soil) and placed in sealed respirometry chambers. The headspaces in the chambers were adjusted to 0, 1, 2.6, 5.2, 10.5, and 20.9% O 2 . Each chamber was connected to an NaOH CO 2 trap and to an O 2 feed line (except the 0% O 2 chambers were connected to an N 2 feed line). Chambers were supplied with O 2 in response to pressure drop resulting from CO 2 trapping. Soils were incubated at 6°C for 12weeks. O 2 consumption and petroleum degradation were maximized in chambers with 10.4% O 2 . There was a slight decline in both O 2 consumption and petroleum degradation at 20.9% O 2 . As O 2 concentrations declined below 10.4% O 2 both O 2 consumption and petroleum degradation declined markedly. 14 C collected in the CO 2 traps did not follow this pattern, but was greater in the 1% O 2 chambers than in 2.6 or 5.2% O 2 chambers. Nitrogen remaining at the conclusion of the study indicated that nitrate was completely consumed in the 0, 1, 2.6, and 5.2% O 2 chambers. nC 17 :pristane and nC 18 :phytane ratios in the soil at the conclusion of the incubation were significantly lower in the 10.4% O 2 chambers than in those with 20.9% O 2 , and more petroleum hydrocarbons were consumed in the 10.4% chambers. Preferential degradation of pristane and phytane in the presence of limited O 2 may be the result of denitrification, evidenced by lower residual nitrate levels in the 10.4 than the 20.9% O 2 environment. Ten percent O 2 is suggested as a target for O 2 enhanced bioremediation.