Abstract The Witwatersrand goldfields in South Africa have been, and promise to remain, the world’s largest goldproducing province, at leat for the medium-term future. The majority of gold occurs, together with pyrite, uraninite, and local bitumen, on degradational surfaces of fluvial conglomerates that were laid down in the 2.90 to 2.84 Ga Central Rand basin. Integration of recent data on the tectonic evolution of the Kaapvaal craton with data on sediment provenance and paleocurrent directions indicate a strong tectonic control on sediment supply from vanished sources, whose equivalents are now exposed mainly in the Amalia-Kraaipan and Murchison granitoid-greenstone belts to the west and north, respectively. Sedimentation took place first in a foreland setting relative to collision between the Witwatersrand and Kimberley crustal blocks, with a subsequent change to a retroarc position in consequence of oceanic-basin closure to the north of the craton. Most of the gold is located within postdepositional microfractures and hydrothermal phases. Microtextural, mineralogical, geo-chemical, and isotopic data indicate that this hydrothermal gold was derived from the local mobilization of detrital particles, supporting a modified paleoplacer model for the genesis of these deposits. The distribution of heavy minerals with which gold is associated, the correlations between elements of detrital minerals, and the isotopic characteristics of uraninite, pyrite, and gold point to erosion of both granitic and mafic and/or ultramafic rocks in highly variable proportions, with the latter having contributed most of the detrital gold. The global and temporal distribution of orogenic gold deposits and Witwatersrand-type paleo-placer deposits suggests that extraction of gold from the mantle into the continental crust has decreased considerably since Mesoarchean times, probably because of decreased mantle heat-flux and partial melting rates. The apparent uniqueness of the Witwatersrand, in terms of total gold content, can be explained by a combination of intense reworking of sediments derived from a rapidly exhumed Paleo- to Mesoarchean greenstone-dominated hinterland by active braided streams in the absence of vegetation and intense chemical weathering under an acidic, reactive Archean atmosphere. Preservation resulted in the first instance from basin capping by a thick pile of extensive 2.71 Ga flood basalts. Discovery of Witwatersrand-style gold deposits in ter-ranes younger than Paleoproterozoic is highly unlikely. Only older foreland and/or retroarc basin fills on stable cratons with a significant greenstone component are considered good targets for future exploration. The expected gold content in Eoproterozoic source rocks is at least an order of magnitude less than in their Archean counterparts.

Abstract Archean mesothermal gold mineralization is commonly located in brittle-ductile structures active during the late deformation events of greenstone belt evolution, and in several cases, reactivation of earlier, mechanically weak structures was a controlling factor. Intensely mineralized greenstone belts in both Western Australia and southern Africa are characterized by an early compressional to oblique-compressional event. Within these belts, zones of low-strain greenstones are bounded by narrow high-strain zones of up to hundreds of kilometers in strike length, which are in turn linked to the brittle-ductile structures hosting gold mineralization. This structural pattern ensured that there was strongly focused fluid flux and probably accounts for the generally high productivity of Archean greenstone belts. In Western Australia, the most mineralized greenstone belt (the Norseman-Wiluna belt in the Yilgarn block) is one in which deformation followed shortly after volcanism and sedimentation and probably resulted from the closure of an extensional basin. Other Yilgarn greenstone belts do, however, host large gold deposits, probably in structures synchronous with and related to those in the Norseman-Wiluna belt. In southern Africa, heterogeneously deformed greenstone belts with high-strain zones of reverse fault-thrust regimes related to externally imposed tectonics are highly prospective. Other important sites are reverse shear zones close to or at granitoid dome margins, which may be due to jostling of these rigid granitoids and differential movement of greenstones during externally imposed deformation.

Abstract Carbon and oxygen isotope studies of carbonate minerals from the two carbonate alteration styles that predate regional metamorphism and gold mineralization in the Norseman-Wiluna greenstone belt of Western Australia have significance for the sources of carbon and oxygen in ore fluids that produce Archean mesothermal gold deposits. The mantle-derived carbon reservoir in regional carbonation zones is the most likely source for auriferous ore fluids rather than the seawater-derived carbon reservoir in altered basalts. The wide range of δ 13 C values for gold-related carbonates reflects dissolution of different carbonate species and/or a variable input of organic and seawater-derived carbon. The oxygen isotope compositions of auriferous ore fluids in the Norseman-Wiluna belt (δ 18 O = 4-9‰) indicate that these fluids (1) could have been in equilibrium with sea floor-altered basalts at 500° to 600°C, (2) derived from regional carbonation zones or interacted with them, or (3) derived from felsic porphyries or granitoids only if modified by subsequent fluid-rock interaction or changing P-T-X conditions. There is a positive correlation between δ 13 C and δ 18 O for carbonates from gold-related alteration (both district and mine scale), with data for the two largest Archean mesothermal gold deposits in the world (Kalgoorlie, Western Australia; and Timmins, Ontario) plotting at the most positive end of this trend. Correlated δ 13 C and δ 18 O variations in gold-related carbonates on a district scale probably relate to source-conduit heterogeneities and host-rock compositional controls, whereas on a mine scale a model involving finite reservoir effects for carbon, temperature-controlled oxygen isotope fractionation, and possibly phase separation can account for the observed trends.