Folds commonly form along transform fault systems. In the San Andreas strike-slip fault, two mechanisms of folding have been described that are regarded as incompatible: (1) wrench tectonics, or distributed shear, and (2) strain partitioning, shortening being perpendicular to the fault. The compatibility and relative importance of these two mechanisms are not well understood, in part because folds are the result of lengthy deformation. This paper synthesizes paleomagnetic data, sedimentary thickness maps, fault patterns, and seismic reflection profiling data along the San Andreas transform fault system that indicate a 20°–30° clockwise (map view) reorientation of Pliocene en echelon folds to their present position, subparallel to the San Andreas fault. The folds continue to develop today as a result of shortening perpendicular to the fault. Folds apparently rotated with continuing strain in the wrench zone and with increasing shortening across the shear zone. Rotation likely occurred when changing Pacific plate motions increased the component of convergence across the fault system. This field example documents deformation by these two different mechanisms in the same fold belt, which has broad implications for interpreting other fold belts formed in similar settings. The geologic record suggests that the opposing theories of wrench deformation and strain-partitioned shortening perpendicular to the fault are both important and compatible processes that act at different times and places.