The Olifantshoek Group in southern Africa contains Paleoproterozoic red beds that are exceptionally well preserved, lying unconformably atop a regionally extensive lateritic paleoweathering profile. We studied the basal unit of this succession, known as the Gamagara or Mapedi Formation, and the lateritized substrate (so-called Drakenstein or Wolhaarkop paleosol) on which it developed. Two ancient magnetic components are observed. One (INT), usually with a distributed unblocking spectrum between 300¿ and 600¿C but occasionally persisting to >675¿C, is directed shallowly southward or northward. A mesoscale fold test at South Sishen Mine indicates that this component was acquired during deformation; similarity of the direction to previous results suggests that it was acquired at ~1240 Ma, during early Namaqua orogenesis. Combining our INT results with existing data from the Namaqua eastern zone (NEZ), we calculate the NEZ pole at (44.9¿N, 021.5¿E, K = 23.2, A95 = 12.8¿, Q = 5). The most stable component from our data set (HIG), always persisting as a nonzero endpoint to demagnetization at >665¿--680¿C, is observed in 32 samples from South Sishen and Beeshoek Mines. Directed moderately east-downward (Sishen) or west-upward (Beeshoek), this component predates the mesoscale fold at Sishen. More importantly, a conglomerate test at Beeshoek indicates that HIG was acquired prior to Paleoproterozoic deposition of the Gamagara/Mapedi Formation. The concordance between directions from the paleoweathering zone and immediately overlying red beds indicates that HIG is a primary magnetic remanence for the basal Gamagara/Mapedi (BGM) Formation. The dual-polarity BGM paleomagnetic pole (02.2¿N, 081.9¿E, dp = 7.2¿, dm = 11.5¿, Q = 6) lies neatly between previous Kaapvaal poles with ages of ~2220 (Ongeluk lavas) and 2060 Ma (Bushveld complex). Our data thus support recent correlations of the Gamagara/Mapedi Formation with pre-Bushveld sediments of the Pretoria Group. A pre-Bushveld age for BGM is also consistent with its substantial distance from a new, albeit less reliable, paleomagnetic pole from the ~1930-Ma Hartley lavas, higher within the Olifantshoek succession (12.5¿N, 332.8¿E, K = 18.6, A95 = 16.0¿, Q = 3). Our conglomerate test at Beeshoek confirms previous allegations that the intense hematitization observed in the Drakenstein-Wolhaarkop paleosol occurred during Paleoproterozoic weathering under a highly oxygenated atmosphere.