The general acceptance of plate tectonics as a process in which whole continents travel large distances across the globe over geological time appears to have given licence to ideas that small islands and submarine plateaus in the oceans may move large distances - and even rotate - independently. If it is not erupted on pre-existing oceanic crust, the only way such an edifice can be left isolated in oceanic crust is if it travels first attached to one contintental plate, then rafts off that plate and joins another. The mechanism to achieve this is called a ridge jump.

Classic examples of fragments isolated by ridge jumps are (a) Madagascar (first follows India then joins the Africa plate) and (b) Sri Lanka (first follows Antarctica, then joins the India plate). A ridge jump often occurs when an active mid-ocean ridge becomes distant from the plume head and its role is taken over by a new rift closer to the action. This is part of the process that keeps active mid-ocean ridges close to the constellation of plume heads (illustrated in Animation J).

Our model presents the tectonic history of the features off the coast of SE Africa that includes the Mozambique Ridge as such a piece of crust that was first attached to Antarctica, growing in area under widespread magmatism, then being progressively left behind on the Africa plate. Interaction with other small fragments in its vicinity produces a story that is more complex than that of, say, Sri Lanka.

Small fragments are prone to raft off the margins of large plates when in the vicinity of plume heads. In the case of the Bouvet plume, we will show Animation N that the small continental fragments involved never succeed in crossing the plume head but defect from one plate margin to another once in its vicinity. Similarly, the early mid-ocean ridge that separated the Malvinas plateau from the Maurice Ewing bank at about 165 Ma (way ahead of the earliest rifting in the southernmost South Atlantic ocean) lost momentum 130-110 Ma and was replaced by a new active ridge (initially two ridges) off the KwaZulu-Natal coast of South Africa, directly above the plume head. The more-westerly of the two new ridges was to prove a long-lived part of the ridge system in the South Atlantic ocean, south of the Falklands-Agulhas fracture zone and north of the Bouvet triple junction.

The Bouvet triple junction is central to the story of Gondwana disruption and dispersion but Gondwana did not split cleanly in its vicinity. So we have to work out the movements of several small fragments whose allegiance was often first to one plate and then another. Ridge jumps are often progressive, so there were multiple triple junctions to consider at times before the relatively simple three-plate solution (Africa-Antarctica-South America) was established by about 113 Ma, at the start of the Albian (final stage of the Early Cretaceous).

Finding a defensible geometrical model for so many fragments demands careful interpretation of the ocean-floor topopgraphy and the marine magnetic anomalies, where available. Until recently, the belief that the Mozambique Plains and the Agulhas plateau were of continental - as opposed to magmatic - origin was a major obstacle to the logic suggested by simple geometry. Matching ocean fracture zones in reconstruction and interpreting ridge jumps in a geometrically consistent manner has been a challenge for many years. I now feel confident to offer a consistent solution at the heart of the wider story of Gondwana dispersal.

The animation at the top of this page illustrates the solution graphically. The animation may be followed in an MP4 file with stop-start facility here. Note the two implicit timescales, namely the stages of the Lower Cretaceous are coloured appropriately in the ocean crust growing within the Africa-Antarctica Corridor (AAC) and that the westward progress of the Malvinas Plateau along the Agulhas fault zone turns out to be very regular, starting at about 165 Ma (at least 30 My before the onset of ocean growth in the southernmost South Atlantic ocean) and reaching full pace very early in the Cretaceous. This conforms with evidence of marine sedimentation progressing westwards in the coastal and offshore Outeniqua basins of South Africa. The model also explains rifting (Xai-Xai) below the plains of Southern Mozambique dated from late Jurassic to Barremian (Macgregor & Reeves, 2025).

Note the way Antarctica to rotates clockwise, away from India, after about 142 Ma, as it skirts around the Bouvet plume head. Small circles drawn at 5 degree intervals around the Africa-Antarctica interval poles highlight this and show a clear relationship to fracture zones as far afield as Australia. The Maurice Ewing Bank remains (mostly) compressed against southernmost Africa, while all the short and short-lived mid-ocean ridges in the vicinity of the plume head show ocean growth until they become extinct.

This animation highlights the main transform faults in the system. Note particularly the alignment of the Lebombo FZ (highlighted in green) with the new mid-ocean ridge initiated in the Weddell Sea at about 142 Ma and with the Xai-Xai grabens active in southern Mozambique at this time.

The detailed marine magnetic anomaly observations of Mueller and Jokat (2018) and Konig and Jokat (2010) in conjugate parts of the AAC enable the pre-M0 movements of Aantarctica against Africa to be closely defined with help from fracture-zone matching. Matching fracture zones alone and assuming a steady pace (58 km/My) through the Cretaceous Quiet Zone (121.4 to 83.64 Ma) works well within the broader context of Gondwana disruption and we define a period of 15-20 My, centred around 125 Ma, in which Antarctica swung steadily clockwise by about 20 degrees as Madagascar and India came to rest, east of Africa, and South America started to leave from its west.

We follow a notional fragment of continental crust having its origin in the Limpopo delta which we call Limpopia. In the interval 140-120 Ma approximately, the active transform and ridges west and north of Limpopia die out as a new join between the mid-ocean ridge in the AAC and that in the actively spreading Weddell Sea is forged along an 'easier' path south and east of Limpopia. The westernmost transform of the AAC becomes a leaky transform allowing the emplacement of the middle section of the Mozambique, probably co-eval with the Astrid Ridge off Antarctica.

The initial (142 Ma) ridge between Limpopia and Antarctica jumped several times, getting closer to Limpopia each time, until a symmetrically-growing ocean panel was establishde immediately south of Limpopia at about 122 Ma. Note the transfer of two small continental fragments from one plate to anoether: Limpopia passes from Antarctica (with extensive submarine magmatic crust - the Explora Wedge - created off Dronning Maud Land) to Africa (to become the Mozambique Ridge) and the Maurice Ewing Bank (MEB) passes from Africa to become the extremity of South America, consolidated in its present configuration by about 110 Ma.

After about 92 Ma (and the creation of the magmatic Agulhas Bank) the three major continental plates (Africa, Antarctica and South America) were intact and a relatively simple rift-rift-rift triple junction has existed between them to the present day. A longer history of this triple junction is shown in Animation O. The steady mid-ocean spreading maintained in the Africa-Antarctica corridor (AAC) is demonstrated in Animation K.

It is important to understand how the main transform-offset linking the AAC to the Weddell Sea (highlighted in green) migrated from north and west of Limpopia (i.e. close to the Lebombo fault zone) in the Jurassic to south and east of it in the interval 135-125 Ma (Hauterivian-Barremian). We see Limpopia coming to rest against Africa and the morphing of the westerly transform into the active Africa-South America ridge from this time. More rapid separation of Limpopia from Antarctica occurred in the same interval; the model predicts 150 km of separation between Limpopia and Antarctica already by 130 Ma.

An earlier model was presented at the European Geophysical Union in Vienna on 27 April 2023 - download the abstract -. A second presentation was given at the GESGB/GSH 'Africa' meeting in London on September 20 News item and a webinar was given for the Yorkshire Geological Society on 2023 November 1. Further small improvements to the smooth operation and internal consistency of all elements of the model, including the behaviour of all the small mid-ocean ridges, have been made in the first half of 2026.

A more detailed animation of the area around southern Mozambique is shown in Animation L.