The growth rate of salt crystals in stone is slower than would be predicted on the basis of growth rates observed in free solution. To account for this observation, it has been proposed that salt crystals have difficulty changing their growth direction at junctions in the pore network. To test this idea, pore channels were prepared in a hexagonal pattern in PDMS and covered with a glass sheet. The pores were filled with sodium sulfate solution and cooled to create supersaturatio, so that crystals would grow in the channels. It was observed that the crystals jumped between junctions but hesitated for long intervals before passing them. This is consistent with the prediction. However, the pressure of crystallization was localized in the most recently invaded pore, and this is believed to be an artifact related to the deformability of the PDMS. To give a true representation of the growth of salt in stone, the experiments must be performed in a rigid host.