During a 1984 Sea Beam and Magnetics survey in the North Pacific we collected data in the area north of the Mendocino fracture zone where a change in the direction of seafloor spreading occurred between magnetic anomalies 24 and 21 and also conducted detailed surveys of the bends in the Surveyor and Mendocino fracture zones associated with this change in plate motion. We have analyzed these data combined with all available magnetics and bathymetric data from previous cruises to determine whether the ridge reorientation and reorganization occurred by way of rift propagation or rift rotation through differential asymmetric spreading. The reorganization began just after anomaly 24 to the north of the Surveyor fracture zone but did not begin between the Mendocino and Surveyor fracture zones until the time of anomaly 22, a 2.5-m.y. delay produced by the old Mendocino transform trend constraining the Farallon plate lithosphere between the Mendocino and Surveyor transforms to move in the old spreading direction until the fracturing through of a new Mendocino transform fault in the new orientation. Most of the reorientation of spreading rift segments occurs within 2--3 m.y.; the entire reorganization requires about 7 m.y. We find direct evidence for seven propagation episodes accounting for the primary reorientation of 75% of the length of the ridge system, and nowhere is the data more consistent with rift rotation than with rift propagation as the principal means of reorientation. We conclude that the primary mechanism by which the reorganization occurred was rift propagation. Assuming that rift propagation can account for all of the ridge reorganization, we present forward models of magnetic anomaly profiles which succeed in fitting the data well using 11 episodes of both northward and southward propagation. ¿ American Geophysical Union 1988