In using geodetic networks to detect crustal motion there is always a question about what to use as a reference. If the stations in a subset of the network can be shown to remain fixed relative to one another, then this subset can be used as a reference for determining the motion of the remaining stations in the network. We derive such a solution for the displacement of a geodetic network, referred to as a stationary solution. The advantage of this technique over other solutions to the same problem is that it neither depends on a single station nor makes assumptions about the movement of all the stations of the network. Application of the technique to a geodetic network near Hollister, California, indicates that stations between the faults are stationary, although for the block between the Calaveras and San Andreas faults the stations are only marginally stationary relative to one another. The technique is then extended to give the total block motion between the faults.