Chemistry, Morphology and Origin of Magmatic-Reaction Chromite Stringers Associated with Anorthosite in the Upper Critical Zone at Winnaarshoek, Eastern Limb of the Bushveld Complex

The occurrence of numerous chromitite layers within the Rustenburg Layered Suite of the Bushveld Complex, South Africa, has been widely cited in models to explain the origin of the igneous layering. Most hypotheses are based around the principle of episodic replenishment of the magma chamber. Chromitite layers occur in both the Lower Critical Zone (LCZ), which is wholly ultramafic, and the Upper Critical Zone (UCZ), where they are part of repetitive units that include pyroxenite (+/- minor harzburgite), norite and anorthosite. The UCZ also reveals stringers of disseminated chromite, which, despite being only a few millimetres thick, are laterally very persistent. We investigate chromite stringers from the uppermost part of the UCZ at Winnaarshoek in the Eastern Limb of the intrusion, where they are preferentially located on contacts between layers of pyroxenite and anorthosite. Stringers overlain by anorthosite are of particular interest as they are located in the centre of units and are unlikely to have developed from replenishment by basal flows of magma. The uppermost of the two chromite stringers associated with the Merensky Reef is unusual as it is located wholly within a layer of pyroxenite, and does not demarcate a lithological contact. Stringers are categorized as Type III chromite to distinguish them from the thicker layers of chromitite, Type I (in the LCZ) and Type II (in the UCZ). The Cr-spinel in the stringers is characterized by relatively low Cr/Fe ratios and is associated with recrystallized, unusually calcic, plagioclase as the principal silicate phase. Accessory phases include rutile, corundum, zircon and baddeleyite, in addition to base-metal sulphides and PGM. The origin of Type III stringers is ascribed to replenishment by sheets of picritic magma injected, not as basal flows, but as sills into an earlier-formed crystalline substrate dominated by well-defined layers of norite and anorthosite. In the framework of this hypothesis, the units that characterize the UCZ are, therefore, not differentiation cycles: the ultramafic components crystallized from U-type (picritic) magmas and the norite-anorthosite from A-type (tholeiitic) magmas. The two components of the units were emplaced non-sequentially, but the ultramafic rocks (+/- chromitite layers and/or chromite stringers) still occur in stratigraphic sequence relative to each other. Type III stringers developed on either the lower or upper contacts of the picritic magma sheets, dependent on whether they were emplaced above or below a layer of anorthosite. Nucleation of Cr-spinel was triggered by contamination of the picritic magma by partial melting of the anorthosite. Melting of low-temperature oikocrysts of pyroxene and interstitial plagioclase produced a thin boundary layer of melt mush. This boundary layer achieved rapid saturation in Cr-spinel, in part owing to Cr2O3 released from the pyroxene oikocrysts. Heat was insufficient to melt the main framework of plagioclase crystals in the anorthosite or the noritic wall-rocks. The different occurrences of chromite in the Bushveld should not be aggregated into a single overarching hypothesis. Type III stringers are probably an entirely different phenomenon from the thicker layers and provide corroborating evidence of the sill hypothesis.