Integral tracer methods consistently imply an annual new primary production of 0.5 ¿ 0.15 mol N m-2 yr-1 for the Sargasso Sea region of the North Atlantic. It has been suggested that as much as half of this may be fueled by the vertical nitrate flux associated with eddy pumping. The key factor in estimates of eddy pumping is the relationship between the time for which water upwelled within eddies remains within the euphotic zone and the rate at which upwelled nutrients are consumed by phytoplankton. The uplift time is strongly influenced by the nature of the eddy, more specifically by its ability to trap waters within it as it propagates. We investigate two scenarios: If eddies propagate as nonlinear features, such that they retain the water within them for their lifetime, only a fraction of eddy events observed at a fixed location actually contribute to the nitrate flux at that position; if eddies propagate as linear features, the efficiency of the pumping process, assumed in current altimetry-based estimates to be 100%, may be very significantly overestimated. Either scenario is shown to result in a major reduction in altimetry-based estimates of the vertical nitrate flux due to eddy pumping. Furthermore, the major contribution of local nitrification to the nitrate recharging of previously uplifted waters, witnessed in our model, raises the possibility that much inferred new production in this area, based on nitrate uptake, is actually regenerated. Our results support the view that mesoscale eddy pumping may not be able to close the Sargasso Sea nitrate budget.