Abstract : Understanding compaction processes in sediments or rocks is important for instance for the characterisation of compaction in sedimentary basins or for sealing of active fault. The aims of the present study are firstly to separate and quantify the relative role of mechanical and chemical compaction in carbonate sediments. Secondly to better understand chemical compaction processes acting on sediments. The potential for porosity loss by mechanical compaction of platforms carbonate strata was investigated by carrying out K0 triaxial tests. Eleven samples cemented with low--Mg calcite and five dolomitized samples from the Marion plateau, offshore northeast Australia (ODP (ocean drilling program) Leg194) were uniaxially compacted at effective stresses up to 70 MPa. Early cementation of bioclastic carbonate samples created a stable cemented framework with a high degree of over--consolidation and low compressibility. Water saturation of the samples produces weakening of the mechanical strength and greater scatter in the correlation of P--wave velocity versus porosity. Most of the tested samples were already so strongly cemented at 30--400 meters that further porosity loss during burial up to 4--5 km depth must occur mainly by chemical rather than mechanical processes. To study chemical processes two other types of experiments were carried out. Pressure solution is the main chemical compaction mechanism affecting sediments during burial, therefore the rate of calcite deformation by pressure solution creep at a single contact was studied. The results enable the identification of the relative importance of pressure solution driven by normal load, and free surface dissolution driven by strain energy. Two different processes occur during pressure solution of calcite crystals at the grain scale. In one case, diffusion of the dissolved solid takes place in the pore fluid present along a rough interface between calcite and the indenter. In the second case, diffusion occurs through cracks that propagate from the contact toward the less stressed part of the crystal. Strain rates are higher for experiments in which crack propagation occurred. Overall it seems strain rates are not really stress dependent but rather dependent on whether crack propagation occurs or not. Eventually, both mechanical and chemical compaction processes were studied on aggregates of calcite and bioclastic carbonate sands. Experimental compaction showed that compaction of carbonates sands should be modelled as a function of both mechanical and chemical compaction also at relatively shallow depth and low temperature. In all cases, the nature of the fluid, the initial grain packing, and the grain size represent important control parameters of the final strain and the strain rates at a given stress. Samples saturated with non--reactive fluids, e.g. air or decane, show less strain than samples saturated with reactive fluids at the same effective stress since the compaction was only mechanical. During the loading phase, chemical compaction occurs by pressure solution creep which is enhanced by the presence of cracks at the grain--to--grain contacts. This is also supported by the identification of compaction related microstructures in thin--sections. During creep tests, the samples compressibility is controlled by, in order of importance, grain size, stress, and water saturation. Low ultrasonic velocities are especially observed in samples saturated with reactive fluids. Dissolution and transport affecting the grain--to--grain contacts geometry and crack propagation are likely to be the reason for such velocity alteration. In conclusion, porosity loss in carbonate sediments is mostly due to chemical compaction and very little to mechanical compaction. Chemical compaction processes are pressure solution and pressure solution enhanced by subcritical crack growth. The predominance of one or the other mechanism is to be related to the fluid in presence and to the nature of the grains.