Post-1906 geodetic surveys in the region of the great San Francisco earthquake have provided some details on the pattern and mechanism of the strain accumulation now occurring between 37¿ and 38¿ N on the San Andreas fault system. Tensor shear strain rates have averaged 0.3¿10-6 yr-1 during 1907--1963 in a ~70-km-wide region around San Francisco Bay. Postearthquake effects on the San Andreas fault persisting for ~30 years following 1906 can account for most of the early observations. However, since ~1940 and perhaps earlier, faults to the east of the San Andreas have played an important role in the strain accumulation. Since about 1950, shear straining in the San Francisco Bay area has within the data uncertainties been uniform across an 80-km-wide region immediately to the east of the San Andreas. The maximum shear strain direction is approximately parallel to the Calaveras fault and is distinctly different from both the strike of the San Andreas fault and the local direction of relative plate motion. At Point Reyes, ~70 km northwest of San Franciso, shear strains are concentrated near the San Andreas and Rogers Creek faults and are significantly smaller than strains observed further southeast, averaging only 0.17¿10-6 yr-1 (1938--1961) across a 70-km-long triangulation arc extending northeast from Point Reyes. Taken together, all of these observations indicate the importance of fault slip at depth in the strain accumulation process but also suggest that the San Andreas plate boundary may become progressively more locked northwest of San Francisco.