The Lucky Strike segment between 37 and 38N on the Mid-Atlantic Ridge is the focus of the international MoMAR program to monitor seafloor-spreading processes. During the GRAVILUCK cruise in August 2006, hydrographic, velocity and light-scattering data were collected in the rift valley at Lucky Strike in order to investigate the regional dynamics and to provide background information for the monitoring effort. The survey observations reveal a remarkably simple dynamical setting dominated by persistent northward flow transporting ≈0.2 Sv of water along the rift valley. Approximately half of this transport must upwell within a deep basin that occupies the northern half of the segment. In the comparatively shallow segment center, the along-valley transport takes place in two parallel, hydraulically controlled overflows on both sides of an active volcano that rises from the rift-valley floor. Within the better sampled of these overflows instantaneous velocities recorded during the survey were northward more than 95% of the time and occasionally exceeded 20 cm s^-1. Similar to other laterally confined overflows in the deep ocean, the cross-sill density gradients are characterized by a lower layer with streamwise decreasing densities and an upper layer where the densities increase along the path of the flow. This vertical density-gradient dipole is the signature of the buoyancy flux associated with high levels of diapycnal mixing near the sill. Overall, the hydrography and dynamics in the rift valley of the Lucky Strike segment are highly reminiscent of many ridge-flank canyons in the western South Atlantic, where mean along-axial advection of density is balanced by vigorous diapycnal mixing. There is circumstantial evidence from historic hydrographic data suggesting that northward flow below ≈1800m in the rift valley in the MoMAR region is persistent on time scales of years to decades and that it extends more than 200 km to the south. During GRAVILUCK the northward flow at Lucky Strike extended well above 1600m, where two previous one year-long current meters had recorded southward mean flows near the Lucky Strike hydrothermal vent field. While interannual variability can potentially account for this difference, the data also allow for the possibility of a cyclonic re-circulation around an isolated topographic peak east of the vent field, resulting in the southward mean flows observed there. In addition to the light-scattering anomalies associated with plumes rising from the Lucky Strike vent field near the segment center, the GRAVILUCK data also show clear evidence for a separate hydrothermal particle plume emanating from a not-yet-discovered vent field in the southern half of the segment, probably near 2000m.