Soil samples were contaminated with several solutes, two means of contamination being evaluated to obtain homogeneous samples, in order to simulate real samples. The soils were allowed to age and periodically sampled. Letting the samples dry in a hood for less than 1 day resulted in a dramatic decrease in the recoveries, owing to evaporation of the spiking solvent. The extraction became generally more difficult as the ageing time increased, but the nature of both the solute and the soil had a strong influence on the results. Hence, a phosphonate and a phosphate were always quantitatively extracted, whereas aromatics were strongly retained in a very organic soil. Increasing the temperature, at constant pressure, greatly enhanced the extraction of all the investigated solutes. Similar results were obtained by adding methanol (10%, v/v) to the supercritical carbon dioxide; however, in that event, the dynamic time must be chosen with care, otherwise part of the extracted solutes is eluted from the trap by the polar modifier.