M.Sc.(Geology) / The Mesoarchean Witwatersrand Supergroup is a remarkably well preserved siliciclastic dominated cratonic platform succession located on the Kaapvaal Craton in South Africa. The vast gold resources which have been mined since 1886 make it relevant for study. The study aimed to identify significant provenance shifts throughout the depositional life of the basin which should be reflected in the in heavy mineral populations and the geochemical composition of the siliciclastic rocks. The study identified major changes in the source rock compositions through the basin lifespan and inferred major tectonic events during the life of the basin. It was found that the mechanical effects of sorting in different depositional environments tended to obscure provenance shifts, but with careful evaluation of the various factors in play significant provenance shifts could be identified. It was found that these provenance shifts corresponded closely with major unconformity sequence boundaries identified by Beukes (1995). These major provenance shifts are a record of a major tectonic event during the development of the basin. The Hospital Subgroup records a passive trailing margin, fed by a combination of felsic and ultra-mafic source rocks. Within the Hospital Hill Subgroup, there is a trend of increasing ultramafic components in the source area with increasing stratigraphic height. This trend is believed to reflect progressive unroofing of tonalite and greenstone belt complexes over the life of the Hospital Hill Subgroup. At the base of the Promise Formation a basin wide unconformity is present, which marks a shift from mature shallow marine and outer shelf sediments of the Hospital Hill Subgroup to immature fluvial quartzites for the Government and Jeppestown Subgroups (Beukes, 1995). In addition to the major change in depofacies that was recognised by Beukes (1995), this study found evidence for a shift in provenance to generally more fractionated source rocks, that were heterogeneous, but well mixed. The presence of lithoclasts indicates a possible metamorphic component was also present in the source area. This is consistent with a source area containing granitoid batholiths, and granite plutonism which is associated with early subduction tectonics and volcanic arc formation during the deposition of the Government and Jeppestown Subgroups (Wronkiewicz and Condie, 1987 and Poujol, et al., 2003, Kositcin and Krapez, 2004). Another important basin wide unconformity is present at the base of the Johannesburg Subgroup, and marks another major provenance change. These rocks are chemically more mature than the Government and Jeppestown Subgroups and represent a shift to an immature fluvial depositional setting related to basin closure (Beukes, 1995). A shift to moderate Th:Sc and La:Sc suggests a less fractionated mix of source rocks. The disappearance of the lithoclasts indicates that the metamorphic source rocks no longer supplied material to the basin. A small increase in the chromite to zircon ratio also suggests that some unfractionated source rocks were present. The narrow range in Th:Sc, La:Sc, Nb:Y ratios suggests that a homogeneous source area is present, but this is contradicted by the highly variable zircon ages measured by Kositcin and Krapez (2004), so the narrow spread might indicate that the rocks are very well mixed. Zircon populations measured by Kositcin and Krapez (2004) suggest that source terrain of the Johannesburg Subgroup probably consisted of a mixture of the granitoid batholiths from which the Government and Jeppestown Subgroups are a derived as well as some intermediate igneous material with ages of 3000-2870 ma. This would reflect incorporation of syntectonic granitoid plutons into the source areas, Kositcin and Krapez, (2004). The Turffontein Subgroup rocks are very coarse and chemically mature, but they display poor to moderate sorting and rounding. The rocks were deposited in a fluvial environment but marine quartzites are not uncommon. It is believed that these rocks were transported in a high energy environment, but the duration of transportation was short. This allows for effective winnowing but insufficient time for physically mature rocks with well-rounded grains to develop, explaining the mature chemical composition but immature physical composition. The source rocks of the Turffontein Subgroup were probably the same as the Johannesburg Subgroup with the higher energy mode of transportation responsible for the observed increase in Zr:Ti ratio. It would also explain the scarcity of feldspars and chlorite in the Turffontein Subgroup. Th:Sc and Nb:Y ratios suggest highly fractionated source rocks, but care must be taken because the mature nature and coarse grainsize of these rocks make trace element analyses unreliable. The zircon population indicates the presence of 3090-3060ma (Kositcin and Krapez, 2004) granite batholiths, as well as 3000-2870 Ma (Kositcin and Krapez, 2004) syntectonic granite plutons, as well as ancient granitoid gneiss (Kositcin and Krapez, 2004) in the source area. This study has provided new support for a foreland basin origin of the Witwatersrand Supergroup, proposed by Beukes (1995), Beukes and Nelson (1995) and Nhleko (2003), resulting from orogenic collision of the Witwatersrand and Kimberley blocks along the western margin of the Witwatersrand block. The Amalia, Kraaipan and Madibe greenstone belts and Colesberg Magnetic Anomaly are probably the only remaining remnants of this orogeny today.

This dissertation is based largely on data that was accumulated during the execution of an exploration program by Iscor Ltd in the Northern Transvaal. The program included geological mapping, geophysical surveys and drilling, on Precambrian iron-formations in the Central Zone of the Limpopo Belt. The structure, stratigraphy, metamorphism, and economic importance of the magnetite quartzites and associated lithologies of the Moonlight prospect are discussed. The lithologies underlying the Moonlight prospect area consist of various pink- and grey-banded gneisses and pink granulite, together with a variety of metasedimentary supracrustal rock-types and concordant serpentinite bodies. The gneissic rock-types consist of chlorite-quartz-feldspar gneiss, chlorite-quartz-feldspar augen gneiss, hornblende-quartz-feldspar gneiss, biotite-quartz-feldspar gneiss, felsic and mafic granulite, and foliated amphibolite. The metasedimentary lithologies are represented by calc-silicates and marble, white quartz-feldspar granulite, magnetite quartzite, metaquartzite and garnet-bearing granulite and gneiss (metapelites). The concordant ultramafic bodies consist of serpentinite with lesser amphibolite, dunite, and chromitite. Intrusive pegmatites and diabase dykes are also present in the prospect area. Metamorphism reached granulite-facies, and more than one retrqgrade metamorphic event is recognized . Amphibolite-facies assemblages are present, but it is uncertain whether they represent another retrograde event . Polyphase deformation has produced intense and complex folding , resulting in irregular magnetite quartzite orebodies. The high metamorphic grades have resulted in medium- grained recrystallization of the magnetite-quartzites with a loss of prominent banding often associated with these rock-types . The magnetite quartzite occurs as three seperate but related ore zones, consisting of one or more ore-bands seperated by other lithologies. All three zones form poor outcrops and suboutcrops in a generally flat lying and sand covered area. · Although representing a low-grade iron ore (32% total Fe), the magnetite quartzite deposits at Moonlight are regarded as potentially viable due to the large opencast tonnages available at low stripping ratios, and the relatively cheap and easy beneficiation process needed to produce a magnetite concentrate with 69-70% total Fe.