This study lias initiated with the aim of identifying the existence of any geochemical criteria which may be used to distinguish between the various volcanic formations within the Witwatersrand triad. The Witwatersrand triad comprises three sequences: the Dominion Group at the base, the Witwatersrand Supergroup in the middle, and the Ventersdorp Supergroup at the top. It is underlain by Archaean basement rocks, and covered by rocks of the Transvaal sequence. The Dominion Group consists of the sedimentary Rhenosterspruit quartzite Formation at the base, overlain by a bimodal component of the Syferfontein Porphyry succession of lavas. Basaltic lavas are the major component of the Rhenosterhoek Formation, while the overlying Formation consists primarily of dacitic porphyries. Intercalations of one lava type within the other are common, however, so each formation is not the exclusive domain of only one lava type. The Witwatersrand Supergroup, a predominantly argillaceous and arenaceous sequence, contains two narrow volcanic horizons, one of wbich, the Jeppestown Amygdaloid (now Crown Formation), consisting of tholeiitic andesites, occurs in the study area. The overlying Ventersdorp Supergroup has, at its base, the basaltic Klipriviersberg Group, of which four out of six formations are present in the study area, namely, the Alberton, Orkney, Loraine and Edenville Formations. This group is succeeded unconformably by the PIatberg Group, consisting of the sedimentary Kameel doorns Formation, followed by the (informal) Goedgenoeg, Makwassie Quartz Porphyry and Rietgat Formations. The Goedgenoeg and Rietgat Formations are basaltic, whil e the Mawassie rocks range from basaltic to dacitic, the majority being tholeiitic andesites and andesites . The Pniel sequence at the top of the Ventersdorp Supergroup consists of the sedimentary Bothaville Formation, and the Allarridge Formation, the lavas of which are basaltic with some andesitic tendencies. A well-defined geochemical stratigraphy was found to exist. From the eleven volcanic formations examined, nine distinct geochemical units emerged, as the Loraine and Edenville Formations were found to have the same geochemical characteristics, as did the Goedgenoeg and Rietgat Formations. Despite having undergone law-grade greenschist facies metamorphism, very clear variation patterns with height are displayed by the immobile elements Ti, P, Kb, Zr and Y, and the light rare earth elements La, Ce and Nd. In contrast, much scatter was observed in the variation patterns of Na, K, Mn, Ba and Rb. Three techniques were employed to effect discrimination between formations - orthosonal discrimination, interelement and ratio vs ratio plots, and discriminant analysis. Confidence limits placed on normal probability plots served to isolate outlier samples for further examination by the various discrimination techniques. A successful test of the efficacy of the discrimination techniques was afforded when fourteen samples from an unknown succession were positively identified as representative of the Klipriviersberg Group

Union Section of Rustenburg Patinum Mines is situated in the northwestern part of the Bushveld Complex, some twenty kilometres north of the Pilanesberg Alkaline Complex. The mining lease area covers a roughly triangular segment of Lower, Critical and Main Zone rocks, transgressed to the north and south by magnetite-bearing ferrogabbro of the Upper Zone. The Main Zone at Union Section is the focus of this study. The prime source of sample material for the study is the deep exploration borehole SK2, but additional, supplementary samples were collected on surface and underground, as well as from a second surface exploration borehole, SK4. In line with the recommendations of SACS (1980), the top of the Critical Zone, and therefore the base of the Main Zone, is taken to be the top of the Bastard Cyclic Unit. Sharpe (1985) suggested that the succession from the base of the Main is an isotopically separate entity Zone up to the Pyroxenite Marker from the rest of the Bushveld layered succession. This is not strictly true, as there is evidence that more than one parental magma was involved in the formation of this interval. It is, however, true that there are fundamental differences, particularly in isotopic makeup, between the Main Zone rocks below the pyroxenite Marker and those above (the latter having been assigned by Molyneux (1970) to subzone C of the Main Zone). Kruger et al. (1986, in press) suggested that the Pyroxenite Marker marks the base of the Upper Zone, and this convention is adhered to here. The implication of this is that the rocks which formerly constituted subzone C of the Main Zone are now considered part of the Upper Zone. The Main Zone rocks below the pyroxenite Marker were originally subdivided by Molyneux (1970) into two subzones, A and B. The results of the present study indicate that this subdivision is not justified. Instead, eight units have been distinguished in the Main Zone on geochemical, petrological and mineralogical bases. Each of these units is characterized by a coherent set, or progression, of chemical and petrological characteristics. The specific assignment of genetic connotations to these units has been deliberately avoided , at least until further studies of the Main Zone prove this to be justified. The demarcation of the eight units is illustrated in the composite diagram (Fig. 34) in the back pocket of this work, and the reasons for the subdivisions are listed in Table 6 (at the end of chapter 7 of this thesis). Until the late 1970's, it was thought that most layered cumulates formed by crystal settling (Wager and Brown, 1968). More recently, there has been a fundamental conceptual change, and many workers now believe that most cumulate rocks formed by in situ crystallization at the floor and walls of the magma chamber (McBirney and Noyes, 1979, Irvine, 1980a; Campbell, 1987). There is, however, some evidence for the physical separation of phases undergoing cotectic crystallization, particularly in the Upper Critical Zone and lower part of the Main Zone (Eales et al., 1986). This process, which has been alluded to in the past by various authors (Ferguson and Botha, 1963; Vermaak, 1976) involves the flotation of early-formed plagioclase crystals due to their positive bouyancy in tholeiitic liquids. The result is an apparent decoupling of the chemistry of pyroxene and plagioclase, as in unit IV of the Main Zone, where plagioclase becomes more anorthitic upwards, whilst pyroxene becomes more iron-rich. There is some substantial evidence, particularly in reversals in the strontium isotope initial ratio and the orthopyroxene Mg/(Mg+Fe) ratio , for multiple intrusion in the Main Zone. Although the largest and most important magma influx in the Main Zone was a high-R₀ aluminous tholeiite, as suggested by Sharpe (1985), the intrusive history of the Main Zone is believed to be far more complex than Sharpe (op. cit.) suggested. Significantly, there is strong evidence for small influxes of Upper Zone-type (Fe-rich tholeiite) magma in the upper reaches of the Main Zone. These are believed to be precursors to the major influx of Upper Zone-type magma at the pyroxenite Marker (Kruger et al, 1986, in press). The fate of intercumulus liquids in cumulate rocks has recently recieved substantial attention (Sparks et al., 1985; Morse, 1986; Barnes, 1986: Campbell, 1987). It is believed that the migration, or at least redistribution, of intercumulus liquids has played a vital role in modifying fractionation trends in the Main Zone. More importantly, the accumulation of late-stage intercumulus liquids is believed to be responsible for the formation of the Fe-rich ultramafic pegmatite bodies that interrupt the layered cumulates in borehole SK2 / Adobe Acrobat 9.53 Paper Capture Plug-in

M.Sc. (Geology) / Nel, B.P. (2013). Petrography and geochemistry of iron formations of the Paleoproterozoic Koegas Subgroup, Transvaal Supergroup, Griqualand West, South Africa. MSc thesis (unpublished), University of Johannesburg, Aucklandpark, pp. 133. The Early Paleoproterozoic Koegas Subgroup comprises a succession of siltstone, mudstone, iron-­‐formation, chert and carbonate rocks that overlies the iron-­‐formations of the Asbestos Hills Subgroup with sharp contact. It is overlain with erosional unconformable contact by glaciogenic diamictites of the Makaganyene Formation. This study focused on the lithostratigraphy, mineralogy and geochemistry of the iron-­‐ formations of the Koegas Subgroup based on fresh diamond drill core samples obtained during the Agouron scientific drilling project in South Africa in 2004. The iron formations the Koegas Subgroup are represented by a few important lithotypes, occurring in distinct sedimentary facies, which formed in unique depositional and diagenetic environments. The iron formations consist essentially of four facies, namely silicate lutite, mixed silicate-­‐siderite lutite, siderite lutite and siderite peloidstone A repetitive sedimentary cycle consisting of fine-­‐grained chemical lithotypes grading upward into reworked chemical lithotypes is evident throughout the Koegas Subgroup iron formations. Silicate lutite formed in deep water settings well below the wave base along a chemocline. Siderite lutite formed in shallower parts of the basin through transformation of primary ferric iron precipitate by iron respiration in presence of organic carbon. Peloidstone formed above normal wave base in shallow water by reworking of earlier siderite lutite deposits. The REE geochemistry provides important clues as to the depositional environment of the iron formation as follows. Depletion in LREE and enrichment in HREE combined with positive Y are typical of ocean water indicate that the iron formations were deposited in a marine environment. Positive Eu anomaly suggest the presence of a hydrothermal component in the ocean water from which the iron formations were deposited. Negative Ce anomalies indicate that somewhere in the marine system Ce3+ was oxidized to Ce4+ oxide, probably in the presence of free oxygen in the ocean water column (Bau and Dulski, 1996). The negative Ce anomalies seen in the Koegas iron formations are the oldest currently known from iron formations. As such the Ce anomalies most probably signify an increase in the oxygenation state of the ocean immediately prior to the rise of atmospheric oxygen as defined by Guo et al. (2009).

M.Sc. (Geology)

The Makganyene Formation forms the base of the Postmasburg Group in the Transvaal Supergroup of the Northern Cape Province. The Makganyene Formation has diamictite as the main rock type, but siltstone, sandstone, shale, and iron-formations are also present. A glacial origin has been proposed in the past due to the presence of dropstones, faceted and striated pebbles. Typically, the Makganyene Formation contains banded iron-formations interbedded with clastic rocks (shale, siltstone, sandstone and diamictites) at the contact with the underlying iron-formations. This transitional zone is generally overlain by massive or layered diamictites which contain poorly sorted clasts (mainly chert) within a shaly matrix. Striated pebbles have been found during field work, and dropstones have been observed in diamictites and banded iron-formations during the study. The top of the Makganyene Formation contains graded cycles interbedded with diamictites and thin layers of andesitic lavas from the Ongeluk Formation. The basal contact of the Makganyene Formation with the underlying Koegas Subgroup was described as unconformable by previous workers. However field work localised in the Rooinekke area shows a broadly conformable and interbedded contact with the underlying Koegas Subgroup. As described above, banded iron-formations are interbedded with the clastic rocks of the Makganyene Formation. Moreover, boreholes from the Sishen area display the same interbedding at the base of the Makganyene Formation. This suggests that no significant time gap is present in the whole succession between the Ghaap and Postmasburg Group. The Transvaal Supergroup in the Northern Cape displays the following succession : carbonates-BIFs-diamictites/ lava-BIFs-carbonates. The Makganyene Formation is thus at the centre of a symmetrical lithologic succession. Bulk rock compositions show that the diamictites have a similar composition to banded iron-formation with regard to their major element contents. Banded iron-formations acted as a source for the diamictites with carbonates and igneous rocks representing minor components. Differences in bulk composition between the Sishen and Matsap areas emphasize that the source of the diamictite was very localised. The Chemical Index of Alteration (CIA) has been calculated, but since the source dominant rock was iron-formation, this index cannot be usefully applied to the diamictites. ACN, A-CN-K, and A-CNK-FM diagrams confer a major importance in sorting processes due to the separation between the fine and coarse diamictites. The interbedded iron-formations display little clastic contamination indicating deposition in clear water conditions. However, dropstones are present in one borehole from the Matsap area, indicating that iron-formation took place under ice cover, or at least under icebergs. Stable isotope studies show that the iron-formations, interbedded towards the base of the Makganyene Formation, have similar values to the iron-formations of the Koegas Subgroup. As a result of the above observations, new correlations are proposed in this study, relating the different Transvaal Supergroup basins located on the Kaapvaal Craton. The Pretoria Group of the Transvaal Basin has no correlative in the Griqualand West Basin, and the Postmasburg Group of the Northern Cape Basin has no lateral equivalent in the Transvaal Basin. These changes have been made to overcome problems present in the current correlations between those two basins. The Makganyene Formation correlates with the Huronian glaciations which occurred between 2.4 and 2.2 Ga ago in North America. Another Precambrian glaciation is the worldwide and well-studied Neoproterozoic glaciation (640 Ma). At each of these glaciations, major banded iron-formation deposition took place with associated deposition of sedimentary manganese in post-glacial positions. The central position of the Makganyene Formation within the Transvaal Supergroup in the Northern Cape emphasizes this glacial climatic dependence of paleoproterozoic banded iron-formation and manganese deposition. However these two Precambrian glaciations are interpreted in paleomagnetic studies as having occurred near to the equator. The controversial theory of the Snowball Earth has been proposed which proposes that the Earth was entirely frozen from pole to pole. Results from field work, sedimentology, petrography and geochemistry were integrated in a proposed depositional model of the Makganyene Formation occurring at the symmetrical centre of the lithologic succession of the Transvaal Supergroup. At the beginning of the Makganyene glaciation, a regression occurred and glacial advance took place. The diamictites are mostly interpreted as being deposited from wet-based glaciers, probably tidewater glaciers, where significant slumping and debris flows occurred. Any transgression would cause a glacial retreat by rapid calving, re-establishing the chemical sedimentation of banded iron-formations. These sea-level variations are responsible for the interbedding of these different types of rocks (clastic and chemical). The end of the Makganyene glacial event is characterised by subaerial eruptions of andesitic lava of the Ongeluk Formation bringing ashes into the basin. Banded iron-formation and associated manganese accumulations are climate-dependant. Glacial events are responsible for the build up of metallic ions such as iron and manganese in solution in deep waters. A warmer climate would induce a transgression and precipitation of these metallic ions when Eh conditions are favourable. In the Transvaal Supergroup, the climatic variations from warm to cold, and cold to warm are expressed by the lithologic succession. The warm climates are represented by carbonates. Cold climates are represented by banded iron-formations and the peak in cold climate represented by the diamictites of the Makganyene Formation. These changes in climate are gradual, which contradict the dramatic Snowball Earth event: a rapid spread of glaciated areas over low-latitudes freezing the Earth from pole-to-pole. Therefore, to explain low-latitude glaciations at sea-level, a high obliquity of the ecliptic is most likely to have occurred. This high obliquity of the ecliptic was acquired at 4.5 Ga when a giant impactor collided into the Earth to form the Moon. Above the critical value of 54° of the obliquity of the ecliptic, normal climatic zonation reverts, and glaciations will take place preferentially at low-latitudes only when favourable conditions are gathered (relative position ofthe continents and PC02 in the atmosphere).

Sixty one basalt samples from the Drakensberg Subgroup in the Barkly East area were analysed for major elements and fourteen trace elements viz. Sr, Rb, Zr, Y, Nb, Zn, Cu, Ni, Co, V, Cr, La, Ce and Nd . These data confirm the tholeiitic nature of the Drakensberg Subgroup. Geochemical differences in the major element compositions are small within the different units of the Drakensberg Subgroup. Small differences are also evident between these units. The trace element variation between the units and within the Lesotho Formation are more pronounced than the variation of the major elements. The Lesotho Formation samples show a minor vertical increase in value for the top half of the Naudes Nek section in Fe, Ti, P, Zr, Nb, Y, La, Ce and Nd . A decrease in value from the same position in Mg, Ni and Co was observed. These patterns of variation are interpreted as representing low-pressure fractionation of plagioclase, olivine and clinopyroxene. Differences between the units of the Drakensberg Subgroup are examined by using absolute trace element contents and interelement ratios. Ratios of incompatible elements differ for the different units leading to the suggestion that a heterogeneous mantle is the most likely explanation for these differences. A unique unit of flows designated the Omega Formation is examined. The basal massive unit displays an unusual pattern of variation with height which suggests an origin through two different differentiation mechanisms. The data from this thesis are compared with those of Cox and Hornung (1966) on the central Lesotho basalts, Bristow (1976) on the southern Lebombo basalts and Robey (1976) on the Karroo dolerites of the Eastern Cape. The geochemical comparison between the lavas of the Lesotho Formation in the Barkly East area, central Lesotho and the Karroo dolerites show no major differences. However the southern Lebombo basalts show an enrichment in Sr and depletion in Cr.

The Grahamstown clay deposits extend in a broad belt from 26°23 to 26°50 East longitude and from 33°15 to 33°22 South latitude along two distinct geomorphological features, the Grahamstown Peneplane (650 m) and the Coastal Plain (520m). The clay deposits traverse four different lithologies including the Bokkeveld Shale, Witteberg Shale, Dwyka Tillite and Ecca Shale. The two planes invariably have a covering of silcrete which is also present over most of the clay deposits except where erosion has taken place. X-ray fluorescence analysis shows that chemically there is a fairly wide variation between and witnin the deposits. The greatest variation is in the Si0₂/Al₂0₃ ratio which appears to be controlled by the parent lithology and to some extent by the amount of leaching. K²0 shows an increase in concentration with depth and therefore indicates the limits of hydrolysis and leaching and of the clay. X-ray diffraction study shows the Peneplane and Coastal Plain deposits to be quite distinct. The Peneplane deposits consist of kaolinite, illite and quartz and the Coastal Plain deposits of kaolinite, illite, quartz and pyrophyllite. The presence of pyrophyllite is not fully understood as there is no indication of major faulting, metamorphism or pyrophyllite in the parent rock. The pyrophyllite most probably represents a transformation product of kaolinite. The kaolinite from the various deposits shows a considerable variation in crystallinity in both the X-ray diffraction traces and electron photomicrographs. The most poorly crystalline kaolinites are from the Coastal Plain deposits and the difference in crystallinity is most probably due to differences in the degree of hydrolysis and the parent rock material in the case of the tillite. Genetically all of the deposits are residual types generated by hydrolysis and subsequent leaching of micas and feldspars. The principal elements leached are silicon, iron and potassium. The hydrolysis and leaching took place over a long period of time in the flat lying areas of the Peneplane and Coastal Plain. The deposits are exploited economically and the clay is used principally in the tile, pottery and whiteware industries with some usage in the paper, refractory and brickmaking industries. The price commanded by raw kaolin is not very high and as a result the clay industry in Grahamstown is not as viable economically as it could be.

One of the most remarkable features of the layered sequence of the Bushveld Complex is its lateral consistency in lithology. This work has established a geochemical and lithological correlation along 170 km of strike of the interval between the UG2 chromitite and the Merensky Reef within the Upper Critical zone of the western limb of the Bushveld Complex. The correlation is based on geochemical investigations of 10 borehole intersections and lithological comparisons of more than 20 borehole intersections around the western lobe of the complex. The basic data presented include 123 whole-rock analyses for major and 12 trace elements, 97 analyses for ' 12 trace elements, and ca. 5500 microprobe analyses of all major phases. Patterns of cryptic variation are established. Some layers (the UG2 chromitite and pyroxenite) show considerable consistency with regard to geochemistry and lithology. Others can be traced along most of the investigated strike length, such as the Lone Chrome Seam, the Footwall Marker anorthosite and the immediate anorthosite footwall to the Merensky Unit. Most of the distinguishable members within the study section, however, show great variation along strike (i.e., the Lower and Upper Pseudoreef Markers, the central noritic sequence in the southern arm of the western limb and parts of the immediate Merensky Reef footwall succession). Several models have been evaluated to interpret the geochemical and lithological data. The author comes to the conclusion that the degree of lithological consistency depends on the variability of magmatic parameters within different parts of the chamber. The most important of these parameters are: (i) the size of fresh primitive influxes and consequently the heat flux, (ii) the composition of the residual liquid, and (iii) the frequency of the influxes. Fresh influxes of more or less similar composition thus spread out along the floor if the residual liquid was less dense than the fresh primitive liquid, but intruded the chamber as a plume where plagioclase had crystallized for some time and the residual liquid had become relatively dense. The size of the influx may be regarded as a measure of the amount of heat flux from the feeder into the chamber. A large influx created uniform physicochemical conditions in the chamber whereas a smaller influx created a strong lateral gradient of physicochemical parameters in the chamber, with subsequent differences in viscosity, density, convection currents, yield strength and thus different mixing behaviour of different liquids. Furthermore, a persistent heat flux from the feeder may have delayed crystallization of successive phases in those parts of the chamber proximal to the feeder . Therefore, new influxes would have been deposited on a footwall of varying thickness and lithology in response to different degrees of crystallization and accumulation along strike. The development of a normal cyclic unit (chromititeharzburgite-pyroxenite-norite (+anorthosite?)) may thus have been interrupted at various stages in different parts of the chamber. The ability to correlate anorthosites over great strike distances implies that their formation did not follow entirely random processes but was dependent on specific magmatic conditions which prevailed over laterally extensive portions of the chamber at certain stages during the evolution of the crystallizing liquid.

Boreholes NG1 and NG2 were drilled on the farm Nooitgedacht 406 KQ to intersect the lower Critical and lower Zones of the western Bushveld Complex. The aim of this study is to describe the textural features and chemical characteristics of the olivine-bearing rocks in the intersections, as determined by petrographic studies, XRF analysis and microprobe analysis. The olivine-bearing rocks are dunites, harzburgites and olivine pyroxenites. They comprise olivine and orthopyroxene, with minor chromite, clinopyroxene and plagioclase, and their textures vary between adcumulate, mesocumulate and poikilitic. The sequence intersected can be broadly correlated with that in the eastern Bushveld Complex. Of the whole-rock inter-element ratios, the MMF (MgO)/[MgO+FeO])ratio is the clearest indicator of cyclicity. The olivine-rich rocks are more primitive than the associated rocks, and seem to become more primitive with height in most intervals. The plagioclase in the olivine-bearing rocks is unusually sodic in corrposition, having a maximum Na₂0 content of 8.12%. A comparison of olivine and plagioclase compositions with those in other intrusions has revealed that the only other major intrusion with sodic plagioclase is the Kiglapait intrusion of Canada. In the Kiglapait intrusion the sodic plagioclase occurs in conjunction with fayalitic olivine as opposed to the forsteritic variety of this study. Chemical variations in the rocks sampled indicate that periodic replenishment of the magma from which the rocks crystallised must have occurred. In some of the olivine-bearing intervals where little fractionation is evident, replenishment seems to have been continuous. In other intervals fractionation appears to have continued uninterrupted for significant periods, prior to rejuvenation by fresh influxes of magma.

The New Amalfi Sheet is a highly differentiated tholeiitic intrusion which is situated between the towns of Matatiele and Swartberg in East Griqualand. It lies within the Central area of the Karoo Igneous Province. Rock types range from dolerites at the base and top through to a highly differentiated granophyre which is found as a 'sandwich horizon' within the top half of the sheet. The most evolved granophyre represents 15.86% of the initial liquid, which was found to be very similar in composition to the average Lesotho-type magma of Marsh and Eales (1984). The paragenetic sequence was found to be chromite → olivine → plagioclase → pigeonite and augite. Cumulus magnetite and ilmenite enter the paragenetic sequence together with immiscible sulphide droplets after 35% crystallization. In the late stages of crystallization, augite changes composition towards ferrohedenbergite. The reappearance of iron-rich olivine coincides with the disappearance of pigeonite and apatite appears as a cumulus phase for the first time after 70% crystallization. Granophyric intergrowth, which contains coarse perthitic Kfeldspar, becomes the most abundant modal entity within the most evolved granophyres. Differentiation was dominated by fractionation of plagioclase and pyroxene, with subordinate olivine and opaque-oxide fractionation. A minor amount of assimilation of country rock occurred within the topmost granophyres. The intrusion has been dated, using the Rb-Sr isochron method, at 178.37 ± 5.52 Ma. Extensive subsolidus deuteric alteration has resulted in the formation of a complete series of hydrothermally altered clinopyroxenes which are enriched in CaO but depleted in TiO₂ compared to the unaltered magmatic clinopyroxenes. It has also resulted in the formation of abundant vermiform ilmenite in the most evolved rocks, recognized by the fact that this phase is enriched in MnO compared to magmatic ilmenites. The very iron-rich orthopyroxene, ferrohypersthene, was found to have crystallized, during cooling of the sheet from the intercumulus liquid. Olivine in the dolerite re-equilibrated with the intercumulus liquid, becoming more iron-rich in composition.