The Bushveld Complex (Fig. 5.4) is not only the largest and most extensively studied of the known layered intrusions, it is also the richest in terms of mineral deposits. In addition to deposits of chromite, platinum-group minerals, Ni-Cu sulfide, and titaniferous and vanadiferous magnetite associated with the mafic and ultramafic layered rocks, the granitic rocks of the complex are associated with endogenetic and exogenetic deposits of tin and fluorite. Asbestos is mined from serpentinized parts of ultramafic layers in the Complex and exploitation of andalusite in hornfels adjacent to the complex accounts for nearly one-half of the global production of aluminosilicates. The emplacement of the complex may also be responsible for the sparse lead-zinc-fluorite and gold mineralization in the surrounding rocks of the Transvaal Supergroup. The vast amount of published literature on Bushveld includes several review articles (e.g., Willemse 1969a, von Gruenewaldt 1979, von Gruenewaldt et al. 1985, Hatton & von Gruenewaldt 1987, Cawthom & Lee 1998), which discuss the complexity of its geology and the diversity of ideas regarding its evolution.
The Bushveld Complex, with an exposed area of 65,000 km2 in the Kaapvaal craton of southern Africa (see Fig. 2.10), was emplaced about 2100 m.y. ago (Hamilton 1977). The country rocks belong to the early Proterozoic Transvaal Supergroup, an 11 km-thick succession of sediments with four intercalated volcanic horizons. Ages of the Transvaal Supergroup rocks are uncertain, but the basal rocks may be as old as 2.5 Ga (Walraven et al. 1990). The Complex includes three major rock groups: the andesitic to rhyolitic Rooiberg Felsite and Dullstroom Volcanics; the ultamafic to mafic Rustenburg Layered Suite; and the Bushveld Granites. The most precise and reliable ages on these rock suites cluster round 2.06 Ga (Walraven et al. 1990). It has been suggested (e.g., Hamilton 1970, Rhodes 1975) that the emplacement of the Complex was the result of a major meteorite impact, which also gave rise to the Vredefort Dome to the south of the Complex, but shock metamorphic features, such as shatter cones, present around the Vredefort Dome have not been found in the rocks associated with the Bushveld Complex. The Complex is located at the intersection of several structural trends, suggesting that its emplacement was largely controlled by the structural framework of the Kaapvaal craton.
The mafic-ultramafic rocks of the Complex, the Rustenburg Layered Suite, occur in three major compartments (Fig. 5.4): (a) the western Bushveld, including the' Nietverdiend area in the far west; (b) the eastern Bushveld, which contains the best exposure of the Layered Suite; and (c) the Potgietersrus lobe or northern Bushveld. The succession of rock types within the three compartments is broadly similar, indicating that the bulk composition of magmas and the conditions of fractional crystallization were comparable. It appears that the magmas for the three compartments were derived from a centrally located master magma chamber and emplaced simultaneously through a system of feeders (von Gruenewaldt et al. 1985).
The Rustenburg Layered Suite, which ranges in thickness from about 7.5 km in the western compartment to about 9 km in the eastern compartment, has been subdivided into five informal zones (Fig. 5.5): Marginal (Basal) Zone, Lower Zone, Critical Zone, Main Zone, and Upper Zone. The Lower Zone is composed of alternating layers of bronzitite and harzburgite, with a few layers of dunite. The Critical Zone consists of two subzones: a lower pyroxenitic subzone, composed essentially of pyroxenites; and an upper anorthositic subzone dominated by norites and anorthosites, marking the incoming of cumulus plagioclase. Norites, anorthosites, and gabbronorites constitute the main rock types of the Main Zone. The Upper Zone is distinctly Fe-rich and contains at least 30 vanadiferous magnetite seams interlayered with ferrogabbros and ferrodiorites. Variation in the Mg-number [Mg/(Mg + Fe) atomic ratio] of bronzite and olivine indicates an overall upward Fe-enrichment in the Complex, consistent with fractional crystallization, although in detail there are numerous reversals and large intervals over which no systematic compositional variation can be discerned (Naldrett et al. 1987). The different zones and subzones vary greatly in thickness throughout the Complex, and are absent in some areas. The Marginal Zone marks the contact between the cumulate sequence of the Complex (the four zones mentioned above) and the floor rocks (Transvaal Supergroup). In eastern Bushveld, the Marginal Zone rocks show a broad correspondence with the overlying cumulate rocks: a predominantly pyroxenitic group borders the Lower Zone and the lower part of the Critical Zone; a predominantly gabbroic group borders the feldspathic upper portion of the Complex. Many samples of marginal rocks represent quenched liquid compositions and such liquids may have been parental to certain parts of the layered cumulate pile (Harmer & Sharpe 1985). The recent field excursion guidebook compiled by Cawthom and Lee (1998) includes an excellent summary of the petrologic and isotopic characteristics of the Layered Suite.
The Critical Zone is by far the most important interval for chromite, sulfide, and platinum-group-element (PGE) concentrations in the Complex. Except in the Potgietersrus area (northern Bushveld), the chromitite layers are restricted to the Critical Zone. Correlation between the western and eastern Bushveld has established the presence of three groups of chromitite layers of varying thickness and composition (Table 5.3): the Lower Group (LG), the Middle Group (MG), and the Upper Group (UG). The Lower Group comprises a maximum of seven chromitite layers (LG-1 to LG-7) that occur essentially within the lower pyroxenite subzone of the Critical Zone. The Middle Group chromitite layers (MG-1 to MG-4) straddle the boundary between the lower pyroxenite and the upper norite-anorthosite subzones of the Critical Zone. The Upper Group chromitite layers (UG-1 to UG-3) occur near the top part of the Critical Zone. The Merensky Reef could be denoted UG-4 as its sulfides typically accompany two very thin (1-5 cm) chromite-rich layers, which bound a roughly l-m*thick pegmatitic feldspathic pyroxenite and/or harzburgite unit (see Fig. 7.2). The most important chromitite layer for the purpose of mining is the LG-6 layer of the Lower Group. It has a strike length of about 70 km in the western Bushveld (where it is also called the Main Seam) and about 90 km in the eastern Bushveld (where it is often referred to as the Steelport Seam), a thickness of 3 cm to 2.5 m, and an estimated reserve of 752 million tonnes to a depth of 300 m (Buchanan 1979). The other chromitite layers are commonly 5-30 cm thick, and LG-1, 2, 3, 4 and MG-1 have been mined locally. All the chromitite layers have anomalous PGE values and UG-2 contains sulfides and ore-grade concentrations of PGE (Lee & Tredoux 1986). The Critical Zone hosts the PGE-bearing Merensky Reef, which also contains significant Cu-Ni sulfides, and platiniferous pipelike bodies of dunite (see Ch. 7).
The Lower Zone in eastern and western Bushveld is characterized by the absence of chromitite layers and low chromium content of the rocks. In contrast, the Lower Zone in the Potgietersrus area contains several well-developed chromitite layers associated with very magnesian (olivine-rich) rocks. Two of these layers, locally known as the lower and upper layers, are being mined at the Grasvally Chrome Mine. The thickness of the layers is less than 50 cm, but the Cr203 contents (54-55%) and Cr:Fe ratios (2.5- 3) are much higher than any of the chromitite layers in western or eastern Bushveld and
TABLE 5.3. Thickness (in meters) and composition (in wt %) of chromitite layers in the Critical Zone of eastern and western Bushveld Complex
Number of layers
4 (UG-1 to 3a)
4 (MG-1 to 4)
7 (LG-1 to 7)
0.35 - 2.23
8 - 14
1.1 - 1.4
Source of data: compilation by Hatton & Von Gruenewaldt (1987).
resemble more closely the Great Dyke chromitite (Hatton & von Gruenewaldt 1987).
It has been argued that the evolution of the Layered Suite, including the generation of cyclic units, involved the mixing of derivatives of two parental magma types in varying proportions at different levels (Cawthom et al. 1981, Irvine & Sharpe 1982,1986, Todd et al. 1982, Irvine et al. 1983, Harmer & Sharpe 1985, Cawthorn & McCarthy 1985, von Gruenewaldt et al. 1985). The two-magma concept is dictated by the observation that the Layered Suite contains substantial units of both ultramafic and anorthositic one-mineral cumulates. There is simply no straightforward way by fractional crystallization alone that a single magma crystallizing mafic minerals can switch to crystallizing only plagioclase. Moreover, Sharpe (1981) has distinguished two suites of syn-Bushveld sills that appear to have formed from different magmas: an earlier suite of orthopyroxene-rich sills, representing offshoots of magma that was introduced during the formation of bronzite-rich cumulates; and a later suite of gabbroic sills (with practically no orthopyroxene) associated with the formation of anorthositic and gabbroic cumulates in the Layered Suite. The two-magma concept has also been applied to the Stillwater Complex, although based on a different line of argument (Todd et al. 1982, Irvine et al. 1983). These two types of magmas have been termed ‘If (Ultramafic) and ‘A’ (Anorthositic) by Todd et al. (1982) to emphasize their olivine ± pyroxene-rich and plagioclase-rich early differentiates, respectively. The magma mixing hypothesis is consistent with variations of 87Sr/86Sr ratio in the Bushveld Complex cumulate zones, but there is some controversy as to whether the ‘A’ magma was a distinct magma (e.g., Sharpe 1985) or a residual magma that evolved in the magma chamber (e.g., Eales et al., 1990).
Primitive compositions of these magmas, designated as U0 and A0, are not known precisely, but have been estimated from the syn-Bushveld sills representing variously fractionated derivatives (designated as t/„ U2, ... and A,, A2, ...). The estimated Bushveld A0 melt is a silica-undersaturated, high-alumina basalt, not unlike common mid-oceanic ridge tholeiites but somewhat richer in normative plagioclase; the estimated U0 parent has no volcanic equivalent, but the chilled (/, derivatives are compositionally similar to olivine boninites (Irvine & Sharpe 1982, Irvine et al. 1983, Sharpe & Hulbert 1985). Judged from the products, the U liquids were rich in Si02 (52-56%), MgO (12-16%), compatible elements (e.g., Cr 800-2,000 ppm), and incompatible elements (e.g., Rb 20-50 ppm, Zr 150-400 ppm), and had an initial 87Sr/86Sr ratio of 0.703-0.705. In contrast, the A liquids had lower concentrations of Si02 (48-50%) and MgO (8-10%), lower concentrations of compatible and incompatible elements, and a higher initial 87Sr/86Sr ratio of 0.707-0.708 (von Gruenewald et al. 1985). As will be discussed later, the mixing of U and A magmas, as they underwent crystallization differentiation, is the most widely accepted explanation for the origin of chromitite layers in layered intrusions.
Melting experiments (Sharpe & Irvine 1983) on Bushveld chilled margin compositions (t/, and A,) have shown that under conditions of fractional crystallization, the crystallization order for the t/, melt (olivine + minor chromite => orthopyroxene => plagioclase) is appropriate to yield dunitic, harzburgitic, and noritic cumulates of the Bushveld Lower and Critical zones and that for the A, melt (plagioclase + chromite => olivine => Ca-rich pyroxene => Ca-poor pyroxene) should produce cumulate units of anorthosite, troctolite, olivine gabbro, and two-pyroxene gabbro, of which the first and the last are well represented in the upper (anorthositic) subzone of the Critical Zone and in the Main Zone. It follows that anorthosite layers in the Layered Suite record major influxes of A magma and chromitite layers, if they are formed by the mixing of A and U magmas, should be associated with anorthosites. This happens to be the case generally in the Bushveld Complex.
The chronology of magma addition during the evolution of the Layered Suite is yet to be worked out. In general terms, the Lower Zone rocks containing abundant ultramafic cumulate layers appear to have crystallized from melts in which the U component was dominant. The upward increase in the abundance of plagioclase, in the initial 87Sr/86Sr ratio, and in the thickness of chromitite layers indicate an increasing influx of A liquids during the lower part of the Critical Zone. Crystallization of the Upper Critical Zone was marked by frequent reversals of mineral compositions and initial 87Sr/86Sr ratio, representing multiple inputs of A magmas. The marked increase in the initial 87Sr/86Sr ratio at the Merensky Reef level at the top of the Critical Zone (as well as the Cu-Ni and PGE mineraliztion associated with the Merensky Reef) is believed to be related to a major influx of A liquids (Kruger & Marsh 1985). The Main Zone and Upper Zone crystallization was dominantly from A liquids, and at least the uppermost 500 m of the Layered Suite has trace element geochemistry consistent with its derivation from a single homogeneous magma (Cawthom & McCarthy 1985). The final major A magma input probably occurred at the level of the “pyroxenite marker” of the Main Zone (Cawthorn et al. 1981, von Gruenewaldt et al. 1985, Kruger 1994).