We simulate jumps of ocean spreading centers with axial high topography using elastoplastic thin plate flexure models. Processes considered include ridge abandonment, the breaking of a stressed plate on the ridge flank, and renewed spreading at the site of this break. We compare model results to topography at the East Pacific Rise between 15¿25'N and 16¿N, where there is strong evidence of a recent ridge jump. At an apparently abandoned ridge, gravity data do not suggest buoyant support of topography. Model deflections during cooling and melt solidification stages of ridge abandonment are of small vertical amplitude because of plate strengthening, resulting in the preservation of a "frozen" fossil high. The present-day high is bounded by slopes with up to a 40% grade, a scenario very difficult to achieve flexurally given generally accepted constraints on lithospheric strength. We model these slopes by assuming that the height at which magma is accreted increases rapidly after the ridge jumps. This increase is attributed to high overburden pressure on melt that resided in an initially deep magma chamber, followed by a rapid increase in temperature and melt supply to the region shortly after spreading began. The high is widest at the segment center, suggesting that magmatic activity began near the center of the segment, propagated south and then north. The mantle Bouguer anomaly exhibits a "bull's-eye" pattern centered at the widest part of the high, but the depth of the axis is nearly constant along the length of the segment. We reconcile these observations by assigning different cross-axis widths to a low-density zone within the crust.