Data from regional and mine-scale mapping, alteration zonation, and ore mineral paragenesis studies are given for eight mafic-ultramafic, shear zone-hosted, gold-bearing quartz deposits within the 3.1- to 3.5-Ga Barberton greenstone belt, South Africa. Fluid properties and light stable isotope geochemistry of the ore-bearing fluid are also discussed. The studied deposits occur mostly in an arc near the Kaap Valley tonalite-greenstone belt contact, over a distance of approximately 60 km. Each deposit is structurally controlled, typified by host D3 (deformation) shear zones that are either located within, or in close proximity to, major D2 and/or D3 tectonic breaks, inferred to be thrust faults. The deposits have a distinctly similar structural style and alteration zonation with a noticeable correlation between alteration type and proximity to the mineralized D3 shear(s). From the least to the most intensely deformed rocks, the following alteration assemblages are recorded: talc-carbonate, quartz-carbonate, fuchsite-quartz-carbonate +/- sulfides, and sericite-quartz-sulfides +/- carbonate +/- fuchsite. Gold appears late in the overall alteration sequence, although it is relatively early within the sulfide paragenetic sequence. Gold has been introduced syndeformationally, during a period of intense shear development. U-Pb analyses of zircon and rutile show that gold mineralization is bracketed between 3126 +/- 21 and 3084 +/- 18 Ma, i.e., approximately 100 to 140 m.y. after the intrusion of the nearby Kaap Valley tonalite (3227 +/- 1 Ma). The Kaap Valley tonalite may thus have acted as a relatively impermeable barrier to the hydrothermal fluid. The gold-bearing fluid(s) also postdate the main thrusting event in the north-central part of the greenstone belt by 100 to 140 m.y. but utilized the D2 (3229-3227 Ma) and later D3 (< 3164 Ma) thrusts as zones of weakness for fluid flow. Fluid inclusions and light stable isotopes show evidence for fluid homogeneity over a minimum scale of approximately 60 X 10 km and for phase separation on a local (mine) scale. Evidence for phase separation is noted by spatially related, primary, H2O-CO2 inclusions having variable H2O/CO2 ratios (type I), coexisting with monophase, primary, CO2-rich (type III) inclusions. Wt percent NaCl equiv values for the dominant type I inclusions are typically in the 5 to 6 percent range for all the mines studied. Variable CO2 densities are noted which are also ascribed to phase separation. Fluid inclusion homogenization temperatures for type I inclusions are 230-degrees to 310-degrees-C, with 290-degrees to 310-degrees-C considered representative for the main-stage fluid. The confining pressure for type I inclusions is approximately 890 bars which can be converted to a minimum depth of formation of approximately 3,400 m. Fluid inclusion volatiles, from heated crushing-gas chromatographic analysis, show dominant H2O (approximately 90 mole %) and CO2 (approximately 10 mole %) with minor CH4 (approximately 0.06 mole %) and N2 (approximately 0.04 mole %) and traces of COS, C2H6, C3H8, and other hydrocarbons, A combined H2O-CO2-NaCl "average" fluid for the gold deposits studied is approximately 88.5 mole percent H2O, approximately 9.8 mole percent CO2, and approximately 1.7 mole percent NaCl equiv. Average results of light stable isotope studies on alteration minerals associated with the quartz-carbonate +/- gold veins (including adjacent wall rock) are as follows: delta-C-13(carbonate) = approximately 4.5 to -2.0 per mil, xBAR = approximately -2.5 per mil; delta-O-18(carbonate) = approximately 11 to 13 per mil; delta-O-18(quartz) = approximately 12 to 13 per mil; delta-D(fluid inclusions) = approximately -20 to approximately -60 per mil, xBAR = approximately -40 per mil; and delta-S-34(sulfides) = 1 to per mil, xBAR = approximately 2.5 per mil. The delta-(OH2O)-O-18 values, calculated at 300-degrees-C for quartz and carbonate, range from 4.7 to 5.8 per mil. The fluid chemistry results presented here are very similar to those given for Archean Au quartz vein deposits in both Canada and Western Australia, suggestive of Archean gold genetic processes which were recurrent in both time and space were replicated within relatively narrow geochemical limits. It is suggested that the phase separation of an H2O-CO2-NaCl parental fluid was the principal factor in producing the Barberton gold deposits; this fluid probably originated from a fluid source(s) external to the greenstone belt, with shear zones focusing the fluid mainly near the Kaap Valley tonalite-greenstone belt contact. Various potential fluid sources are assessed in the light of the data base presented here. The possibility exists that the onset of mineralization in the Barberton greenstone belt occurred synchronously with the early development of the Witwatersrand basin (3074 +/- 6 Ma).
