A stressed fluid-filled porous system was modeled by hollow cylinders of St. Peter sandstone subjected to various combinations of pore and confining pressure at 270 degrees to 280 degrees C for up to four weeks. Large reductions in porosity, up to more than 50 percent, were produced purely by pressure solution without grain crushing. Most of the porosity reduction occurred early in the experiments and in samples with the finer of two grain sizes. Experiments with the same pore pressure, but different confining pressures, and experiments with the same effective stress, but different stress magnitudes showed that a simple effective stress law does not hold for pressure solution, and that the amount of porosity reduction depends on pore fluid pressure. However, nonhydrostatic stress appears to be necessary for rapid porosity reduction because experiments with hydrostatic pressure produced very little change in porosity. Also, experiments with the same confining pressure but different pore pressures showed that the amount of porosity loss is dependent on both pore pressure and effective stress. Pore pressure appears to place an upper limit on the rate of porosity reduction, while nonhydrostatic stress appears to be necessary for rapid porosity reduction. A dry control experiment showed that fluid must be present for porosity reduction at the temperatures and pressures in our study.The porosities of many of the samples in this study were determined both gravimetrically and by point counts on cathodoluminescent micrographs. Cathodoluminescence is useful in studying pressure solution because the intergranular relationships and pore spaces are very distinct. However, in examining natural samples caution is required when relying solely on the luminescence to determine pressure solution, because luminescent characteristics change with time.