Pressure solution experiments were carried out, using a quartz knife-edge 0.26 mm wide on halite single crystals in halite saturated solutions, to observe the detailed development of pressure solution contacts and the rates of pressure solution. A rate of about 3 μm/day was observed for initial knife-edge stresses ranging from 4.5 to 15 MPa. Close examination of the contact leads to the conclusion that the mechanism of pressure solution is a combination of plastic deformation at the contact and free surface pressure dissolution near its periphery. Free surface pressure dissolution increases the contact stress to about 18 MPa, high enough to cause plastic deformation, by changing the area of contact. This mechanism differs from a water film diffusion mechanism, previously suggested by many authors, but is similar in some ways to the undercutting hypothesis of Bathurst (1958). We infer a steady state plastic deformation instead of catastrophic grain crushing at the contact. Free surface dissolution plus the plastic deformation mechanism may be primarily responsible for pressure solution in relatively porous rocks.