A model of crystallization in pore networks must account for the fact that the stone expands as the crystals grow, indicating that pressure is exerted on the pore walls. If growth were under diffusion control, the crystal/liquid interface would be at the solubility limit, so no pressure could be tolerated. Therefore, the growth must be controlled by interface attachment kinetics. Comparison of the growth rates of sodium sulfate in free solution and in pore networks suggests that the crystals have difficulty branching through intersecting pores. A model is presented that allows for these factors and which can provide a good fit to experimental data, but the assumptions remain to be supported by direct observation of growth in pore networks.