Degradation of naphtalene-benzene-UL-14C (Sigma, 31.3 mCi/mmol) to 14CO2 was studied in two oxisol samples in Caratinga, MG State, Brazil, from monitored areas which have been contaminated with oil spills. One site is located just below a heavily contaminated soil, and another one was collected above the contaminated site, with less direct influence of runoff. An apparatus was developed to analyse the CO2 evolved from degradation of naphathalene-14C and consisted of 12 flasks containing the soil samples connected to a CO2–free air source. Air from the flasks containing the treatments was bubbled through a CO2 trapping solution of NaOH 0.5 M. Samples of 15g of soil from the two sites, taken from 0-10cm and 10-20cm depths, and 4 ml of sterile distilled water, were added to the flasks, using 3 replicates of each treatment. Naphtalene-benzene-UL-14C was added to each flask at a radioactivity concentration of 1.32x10-2 ìCi.g-1of soil). The measurement of radioactivity in the CO2 trap solution was carried out after 1, 8, 16 and 23 days of incubation at room temperature, using a Liquid Scintillation Counter (Beckman Instruments). There were no significant differences between the two soil sites, but there were differences between soil depths. The 0-10cm soil depth showed higher 14CO2 evolution than the 10-20cm soil depth in both sites. After 8 days of incubation approximatly 25% of the naphtalene-benzene-UL-14C added was evolved as 14CO2 from surface soil, and after 16 days of incubation no more evolution of 14CO2 was observed. These results showed that a minor part of naphtalene-benzene-UL-14C is quickly degraded to CO2 in these oxisols, and a greater amount of naphthalene either remains in the soil, adsorbed by soil particles or metabolized by the soil microbial population, or is leached away downwards.