The continental divide in the Appalachian Mountains is widely thought to have initiated east of its current location and subsequently migrated westward, elongating Atlantic drainages at the expense of the headwaters of Mississippi drainages. Several lines of evidence such as barbed tributaries suggest this process continues today; however, the common notion that the process is driven by steeper slopes and a shorter distance to base level along the Atlantic flank is not supported by the nearly equal declivities of headwater streams on both sides of the divide. U-Th/He and AFT thermochronology reveal a denudation pattern consistent with a recent shift in the drainage divide. In the central Appalachians, U-Th/He and AFT closure ages are virtually identical for the Ridge and Valley and suggest very modest amounts of unroofing in the past ~ 200 Ma. In contrast, the U-Th/He ages are systematically younger than AFT ages to the east in the Piedmont and Blue Ridge suggesting more and continuous unroofing through Alleghenian orogenesis and Mesozoic rifting. The problem with these data, particularly in Pennsylvania, lies in the fact that the largest amount of unroofing coincides with the lowest contemporary erosion rates and lowest standing topography, whereas the least amount of unroofing is associated with the highest contemporary erosion rates and the most rugged and highest standing topography. This apparent paradox may be explained by a drainage divide that forms on the rift flank in the Mesozoic and remains pinned until the late Cenozoic when it jumps westward to its current position. The jump leaves a dissected and rugged topography in its wake, but has not yet generated enough unroofing to lock-in Cenozoic U-Th/He ages. The rapid, westward shift in the divide may actually be rolling southward along orogen strike as suggested by: (1) the location of the highest rates of Appalachian stream incision, (2) considerable separation of the crest of the range and the crest of the long-wavelength topography (defined as the topography filtered at the flexural wavelength), and (3) a southwest sweep of Miocene to Quaternary depocenters on the Atlantic Coastal Plain. These results suggest that even in "dead" orogens like the Appalachians, the continental divide is localized geodynamically and the integration of drainage transverse to the orogen may be strongly delayed.