Kinematic data from two segments of the Coast Range fault are used to test the hypothesis that the Franciscan subduction complex was exhumed by extensional faulting. The data reveal a consistent geometry for the principal directions of brittle strain. The maximum extension direction has an east-west orientation and lies subperpendicular to the present attitude of the Coast Range fault, which dips steeply to the east. The maximum shortening direction is subvertical and lies at a low angle to the present down-dip direction of the fault. Given that the Coast Range fault probably formed with a gentle dip, our results indicate subhorizontal crustal shortening in a northeast direction, which is at variance with interpretations that invoke regional-scale extension between the Franciscan subduction complex and the overlying Coast Range ophiolite and Great Valley Group. We favor a model involving out-of-sequence thrust faults. Such a model is consistent with fault-parallel contraction and the pronounced discontinuity in metamorphic grade across the Coast Range fault, as well as substantial thinning of the Coast Range ophiolite. This interpretation implies that erosion, not extensional faulting, was the major process that exhumed the Franciscan subduction wedge during the Late Cretaceous.