Partition coefficients are computed for hard spherical solutes in equilibrium with sponge-type matrices containing randomly-placed spherical cavities or pores using grand canonical Monte Carlo simulation. The random pore model is representative of a variety of disordered porous solids including porous glasses and some polymeric materials. The algorithm used brings additional realism to the problem by rigorously distinguishing between accessible and inaccessible pore space. The simulation results display significant concentration effects which are often observed experimentally and are compared to data obtained by Brannon and Anderson (1982) for the partitioning of bovine serum albumin into controlled-pore glass.