The Limpopo Mobile Belt is a zone of predominantly high-grade gneisses located between the Rhodesian and Kaapvaal cratons in southern Africa. The last major period of metamorphism and deformation in the Belt occurred ∼2600 m.y. ago, but numerous Rb-Sr mineral ages indicate that final cooling occurred at 2000 to 1950 m.y. B.P. Paleomagnetic investigations have been carried out on samples from 78 sites in the metamorphic rocks and 6 sites in post-tectonic dikes that intruded at 1876 (±68) m.y. B.P. Detailed demagnetization of the metamorphic sites revealed three stable magnetizations. Component “A” (D = 352°, I = 60°, α95 = 7°; corresponding pole 26°N, 22°E) is thought to have been acquired during cooling of the Belt at ∼ 1980 m.y. B.P., and, together with a result previously published, suggests diachronous cooling across the Belt with a polarity change occurring during the cooling interval. Component “B” (D = 41°, I = −7°, α95 = 16°; corresponding pole 46°N, 99°E) is probably a later chemical remanent magnetization (CRM), but its exact age is unknown. Component “C” (D = 357°, I = −66°, α95 = 10°) is very close to the Earth's present field direction and is probably of recent origin. The post-tectonic dikes yielded a stable magnetization: D = 318°, I = 73°, α95 = 6° and corresponding pole 3°N, 9°E and is thought to have been acquired at the time of intrusion at ∼1876 m.y. B.P. The new data assist in refining the apparent polar wander (a.p.w.) path for the Kalahari Shield for the interval circa 2050 to circa 1800 m.y. ago. The data also suggests that over this time interval (a) it is difficult to draw firm conclusions about the relative positions of the various African shields and the South American Guayana Shield; and (b) the paleomagnetic data argue against any proposed supercontinent reconstruction of Africa and North America, and that at least the Kalahari Shield was in strong relative motion with respect to North America.