Results from a wide-angle seismic and gravity study between the Rockall Bank and the Iceland Basin in the North Atlantic are presented. Crustal and sedimentary structures are resolved in the Hatton Basin and across the Hatton continental margin (HCM) east of magnetic anomaly 24. The structure of the oceanic crust west of the anomaly is also determined. Gravity data support the seismic model in areas of good seismic coverage and are used to control the model where the wide-angle seismic data are poor. A two-layer sedimentary sequence is present both in the Hatton Basin and across the continental margin. The lower layer, with P wave velocity of about 4 km/s, is interpreted as pre-Eocene synrift sediments and is up to 3.5 km thick. A younger and thinner (1--2.5 km) postrift sequence, with a velocity of about 2 km/s, defines a strong velocity contrast, which suggests an erosional unconformity surface. The sedimentary structure is distinctly different from that in the Rockall Trough, where a third intermediate layer (Vp≈3 km/s) occurs. The three-layer crust, characterized by two intracrustal reflections (PiP1 and PiP2) varies from 30 km thick under the Rockall Bank to about 15 km below the Hatton Basin, where it is stretched by a factor of 2 relative to onshore Ireland. The crust is thinnest below the Hatton Bank, where the presence of a single intracrustal reflection indicates that the lower crustal layer thins to below the seismic resolution limit. Below the HCM a region of thick lower crust with anomalously high velocity (Vp≈7.2 km/s) is resolved by the seismic and gravity data. It is connected (west of anomaly 24) to a region of oceanic crust, which is thicker than in the Iceland Basin. These relationships between the thick lower crust below the HCM and the oceanic crust in the Iceland Basin are interpreted as evidence for magmatic underplating, consistent with previous models for the HCM. The inferred unconformity surface between the synrift and postrift layers may be due to regional uplift driven by upwelling of hot asthenosphere before anomaly 24 (early Eocene) time. ¿ 1998 American Geophysical Union