Abstract New seismic and well data in the deep-water basins of Campos, Santos, South Kwanza and Benguela, supported by plate reconstructions, help answer fundamental questions on the rifting history of the central South Atlantic, specifically on the amount and effect of fault-related deformation, and on when and where sea-floor spreading started. The Paraná mantle plume played a fundamental role – dynamically raising the plate, prolonging continental rifting by heat-softening the crust and, after break-up, delaying the onset of marine conditions. Previous discrepancies in extension and subsidence have been solved, and the location and age of the continent–ocean boundary can now be determined. Rifting involved approximately 450 km of homogeneous pure shear, equivalent to a β factor (lithosphere stretching factor) of 4.5. Break-up occurred at 123 Ma (Barremian–Aptian boundary), 7–8 Ma later than the southern South Atlantic but 6 Ma before widespread salt deposition. The mid-Atlantic ridge was initially subaerial, marked by a volcanic high. Sea-floor spreading was at a rate of 24 mm year −1 , similar to syn-rift deformation prior to break-up. Transcontinental strike-slip shear zones are not evident but a major NW–SE lithospheric lineament associated with a failed triple junction arm had a major influence on the magmatic history, both prior to and after break-up. Supplementary material: A4 versions of the seismic sections shown in Figures 6 & 7 are available at http://www.geolsoc.org.uk/SUP18563 .