^a Department of Geological Sciences, University of Cape Town, Rondebosch 7700, South Africa

^b Department of Earth, Atmospheric and Planetary Sciences, Massachusetts Institute of Technology, Cambridge, MA 02139, USA

^c Department of Terrestrial Magnetism, Carnegie Institution of Washington, 5241 Broad Branch Road, N.W., Washington, DC 20015, USA

Xenoliths from both Archean and Proterozoic terranes record conductive limb temperatures characteristic of a “cratonic” geotherm (40 mW m^?2), indicating cooling of Proterozoic mantle following the last major tectonothermal event in the region at 1 Ga and the probability of thick off-craton lithosphere capable of hosting diamond. This inference is supported by U–Pb thermochronology of lower crustal xenoliths Schmitz and Bowring, 2003. Contrib. Mineral. Petrol. 144, 592–618].

The entire region then suffered a protracted regional heating event in the Mesozoic, affecting both mantle and lower crust. In the mantle, the event is recorded at 150 Ma to the southeast of the craton, propagating to the west by 108–74 Ma, the craton interior by 85–90 Ma and the far southwest and northwest by 65–70 Ma. The heating penetrated to shallower levels in the off-craton areas than on the craton, and is more apparent on the southern margin of the craton than in its western interior. The focus and spatial progression mimic inferred patterns of plume activity and supercontinent breakup 30–100 Ma earlier and are probably connected.

Contrasting thermal profiles from Archean and Proterozoic mantle result from penetration to shallower levels of the Proterozoic lithosphere by heat transporting magmas. Extent of penetration is related not to original lithospheric thickness, but to its more fertile character and the presence of structurally weak zones of old tectonism. The present day distribution of surface heat flow in southern Africa is related to this dynamic event and is not a direct reflection of the pre-existing lithospheric architecture.