In contrast to the long narrow volcanic chains in the Pacific, Atlantic hotspot tracks, in particular in the South Atlantic, e.g. Tristan-Gough, Discovery, Shona and Bouvet, are irregular and in some cases diffuse and discontinuous. An important question is whether this irregularity results from tectonic dismemberment of the tracks or if it represents differences in the size, structure and strength of the melting anomalies. Here we present new age and geochemical data from volcanic samples from Richardson Seamount, the Agulhas Ridge along the Agulhas-Falkland Fracture Zone (AFFZ) and Meteor Rise. Six samples yielded ages of 83-72 Myr and are 10-30 Myr younger than the underlying seafloor, indicating that they are not on-axis seamounts associated with sea-floor spreading. The incompatible element and Sr-Nd-Pb-Hf isotopic compositions rangefrom compositions similar to those of the Gough domain of the nearby Tristan-Gough hotspot track to compositions similar to samples from the Shona bathymetric and geochemical anomaly along the southern Mid-Atlantic Ridge (49-55°S), indicating the existence of a Shona hotspot as much as 84 Myr ago and its derivation from a similar source region similar to that of the Tristan-Gough hotspot. Similar morphology, ages and geochemistry indicate that Richardson, Meteor and Orcadas guyots originally formed as a single volcano that has been dissected and displaced 3500 km along the AFFZ, providing a dramatic example of how plate tectonics can dismantle and disseminate a hotspot track across an ocean basin.