The opening of the South Atlantic is a classical example for a plume related continental breakup. Flood basalts are present on both conjugate margins as well as aseismic ridges connecting them with the current plume location at Tristan da Cunha. To determine the effect of the proposed plume head on the continental crust, we acquired wide-angle seismic data at the junction of the Walvis Ridge with the African continent and modelled the P-wave velocity structure in a forward approach. The profile extends 430 km along the ridge and continues onshore to a length of 720 km. Crustal velocities beneath the Walvis Ridge vary between 5.5 km/s and 7.0 km/s, a typical range for oceanic crust. The crustal thickness of 22 km, however, is approximately three times larger than of normal oceanic crust. The continent-ocean transition is characterized by 30 km thick crust with strong lateral velocity variations in the upper crust and a high-velocity lower crust (HVLC), where velocities reach up to 7.5 km/s. The HVLC is 100 to 130 km wider at the Walvis Ridge than it is farther south, and impinges onto the continental crust of the Kaoko fold belt. Such high seismic velocities indicate Mg-rich igneous material intruded into the continental crust during the initial rifting stage. However, the remaining continental crust seems unaffected by intrusions and the root of the 40 km-thick crust of the Kaoko belt is not thermally abraded. We conclude that the plume head did not modify the continental crust on a large scale, but caused rather local effects. Thus, it seems unlikely that a plume drove or initiated the breakup process. We further propose that the plume already existed underneath the continent prior to the breakup, and ponded melt erupted at emerging rift structures providing the magma for continental flood basalts.