Continuous extraction of squalene from shark liver oil using supercritical carbon dioxide was carried out in both laboratory and pilot scale plant. The shark liver oil contained around 50% by weight squalene, which was recovered as the main extract stream. The other major components in the oil were triglycerides, which were recovered as raffinate, and pristane, which was recovered as a second extract stream. Separation performance was determined as a function of temperature; pressure; oil to carbon dioxide flow rate ratio, packed height and type of packing; and reflux ratio. The pressure, temperature, and feed oil concentration of squalene determined the maximum loading of oil in carbon dioxide. The oil to carbon dioxide ratio determined the squalene concentration in both the product stream and raffinate stream. The ratio of oil flow rate to the flow rate of squalene required to just saturate carbon dioxide was found to be a useful correlating parameter for the oil loadings and product compositions. Of the three packings investigated, wire wool gave the best separation efficiency and Raschig rings the worst efficiency. Mass transfer correlations from the literature were used to estimate the number of transfer units NTU from experimental data and literature correlations. NTU's from the experimental data were comparable to predictions at a pilot scale but were underpredicted at the laboratory scale. The use of reflux at the pilot scale enabled the concentration of squalene in the product stream to be increased from 92% by mass to a maximum of 99% by mass at fractionation conditions of 250 bar and 333 K.