The separation of South America from Africa during the Cretaceous is poorly understood due to the long period of stable polarity of the geomagnetic field, the Cretaceous Normal Superchron (CNS, lasted between ∼121 and 83.6 Myr ago). We present a new identification of magnetic anomalies located within the southern South Atlantic magnetic quiet zones that have arisen due to past variations in the strength of the dipolar geomagnetic field. Using these anomalies, together with fracture zone locations, we calculate the first set of magnetic anomalies-based finite rotation parameters for South America and Africa during that period. The kinematic solutions are generally consistent with fracture zone traces and magnetic anomalies outside the area used to construct them. The rotations indicate that seafloor spreading rates increased steadily throughout most of the Cretaceous and decreased sharply at around 80 Myr ago. A change in plate motion took place in the middle of the superchron, roughly 100 Myr ago, around the time of the final breakup (i.e., separation of continental–oceanic boundary in the Equatorial Atlantic). Prominent misfit between the calculated synthetic flowlines (older than Anomaly Q1) and the fracture zones straddling the African Plate in the central South Atlantic could only be explained by a combination of seafloor asymmetry and internal dextral motion (<100 km) within South America, west of the Rio Grande fracture zone. This process has lasted until ∼92 Myr ago after which both Africa and South America (south of the equator) behaved rigidly. The clearing of the continental–oceanic boundaries within the Equatorial Atlantic Gateway was probably completed by ∼95 Myr ago. The clearing was followed by a progressive widening and deepening of the passageway, leading to the emergence of north–south flow of intermediate and deep-water which might have triggered the global cooling of bottom water and the end for the Cretaceous greenhouse period.