Geodetic shear strains are determined for a triangulation network in the southwest corner of the South Island of New Zealand, which extends from Haast Pass to the west coast across south Westland, a distance of about 50 km. The Alpine fault crosses the network. In order to examine the variation of maximum (engineering) shear as a function of distance across the plate boundary zone, I have reduced the observations into a series of small networks which span at least half of the actively deforming part of the Pacific-Australian plate boundary. The azimuth of the principal axis of relative compression has a relatively constant value of 105¿ across the plate boundary, which is consistent with strike-slip movement on this part of the Alpine fault. The Maximum (engineering) shear strains have a relatively constant value of about 0.5 μrad/yr over most of the study area; area however, two localized high strain anomalies, both with peak strains of 1.2 to 1.4 μrad/yr, were observed. One of these appears to be restricted to a localized zone, a few kilometers in extent adjacent to the Alpine fault, while the second exists 30 km to the east of the Alpine fault near Haast Pass. The spatial extent of the second anomaly cannot be determined. I have modeled the anomaly using a dislocation model for a dipping strike-slip fault. I can produce a reasonable fit between this simple model and the observed stains if I assume that the brittle-ductile transition is shallow (about 4 km deep) and that the slip rate on Alpine fault is 20 mm/yr. I cannot match the observed strains if the slip rate equals the 40 mm/yr relative Pacific-Australian plate velocity; however, suggesting that only about half the relative plate motion is accomodated on the Alpine fault. This model does not match strains of 0.5 μrad/yr which persist 10 km west of the Alpine fault or the high strain zone observed near Haast Pass which suggests that a significant amount of the extra deformation is accomodated near the Alpine fault. ¿American Geophysical Union 1991