A mini-round-robin study of a supercritical fluid extraction (SFE) method for polynuclear aromatic hydrocarbons (PAHS) from soil samples was conducted. Three laboratories participated in the study, and each laboratory extracted three real-world samples in triplicate. The cryogenically milled samples were extracted at 350 atm (1 atm = 101 325 Pa) and 90°C for 20 min in the dynamic mode using supercritical carbon dioxide at a flow-rate of 1 to 1.5 ml/min and the extracted material was collected in 10 ml dichloromethane, which was then subjected to silica chromatography. The SFE method accuracy (percent recovery) was determined relative to the sonication extraction since the true levels of PAHs in these samples are not known. The PAHs were recovered quantitatively (recovery ⪢ 80%) by SFE when present at concentrations of 1 mg/kg or higher. The interlaboratory method precisions (overall R.S.D.s) appear to be concentration-dependent; at concentrations above 1 mg/kg, they were 27% or lower; at concentrations below 1 mg/kg, they ranged from 19 to 80%. From these results, we concluded that the method appears quite rugged, and the interlaboratory data compare well with other SFE interlaboratory studies.