We made 29 in situ doorstopper strain relaxation measurements distributed among eight sites spaced on a 35-km transect running from the foothills of the San Gabriel Mountains, across the San Andreas fault, into the western Mojave desert southeast of Palmdale, California. This was a pilot study to test if such measurements can detect the regional stress field and any modification of it in a tectonically complex area. Strain was measured by overcoring strain gauge rosettes which had been bonded to the flattened bottom of a shallow borehole. Measurements of stress were repeatable at each site. We found NNE trending maximum compressive stress(&sgr;l) at our sites farthest from the fault. This orientation is parallel to the &sgr;l inferred from fault plane solutions of major southern California earthquakes and compares favorably with deep hydrofracture stress measurements made near our sites. Nearer the San Andreas fault, the orientation of &sgr;l was approximately east-west north of the fault, and northwest-southeast to north-south south of the fault. The repeatability of measurements at a site and the favorable comparison of our measurements with Tullis' <1977> near-surface stress measurements suggest that we reliably determined the stress field orientation present at a site. It was not possible, using the available data, to distinguish the stress caused by residual or topographic effects from tectonically applied stress or to account for modification of the stress field by decoupling across fractures.