The investigation of hydrocarbon fractionation between oil and gas phases is of interest for several purposes in reservoir exploitation. In reservoir geochemistry, the evolution of light hydrocarbon fractions of oils may explain some migration phenomena. In gas injection projects, the preferred dissolution of some components in gas may alter the composition as well as the properties of the oil. Underground gas storage in depleted oil reservoirs may also be concerned by these problems. Results of several IFP studies are described here to illustrate and to quantify the phenomenon. Two of them, using real reservoir fluids, concern reservoir geochemistry, while the third, which is a swelling test, aimed to study gas injection, investigated a synthetic reservoir fluid with hydrocarbon components up to C-30. Two pieces of equipment were used: a sapphire cell with a maximum pressure rating of 400 bar and a high pressure apparatus called Hercule with a maximum pressure of 1500 bar. For each fluid, the saturation pressure was measured. For various pressure levels below saturation, the coexisting liquid and gas phases were sampled at constant pressure, and subsequently analyzed by gas chromatography. In the gas injection study, sampling was repeated with different quantities of injection gas. Compared to a n-paraffin with the same number of carbon atoms, aromatic hydrocarbons appear to stay preferentially in the liquid phase, as do cycloalkanes to a lesser extent. The gaseous phase is slightly enriched in isoalkanes. These fractionation effects are less pronounced near the critical region. These phenomena have been modeled with a cubic equation of state combined with a group contribution mixing rule.