A Thesis Submitted to the Faculty of Science, University of the Witwatersrand , Johannesburg for the Degree of Doctor of Philosophy / The Phalaborwa Complex, northeastern Transvaal, South Africa, consists of the main body of clinopyroxenites and subordinate phoscorite, carbonatite and syenite which is surrounded by numerous pipe-like bodies of syenitic compositions and rare clinopyroxenites. Clinopyroxenites of the main complex are characterized by cumulus textures formed by separation and accumulation of coprecipitating clinopyroxene, apatite and phlogopite. Potassium feldspar is an intercumulus phase in feldspathic pyroxenite. "Inch-scale" layering of clinopyroxene, apatite and phlogopite formed as in situ cumulus layering near the outer contact of the complex early in the cooling of the magma. Breccias of monomineralic assemblages such as glimmerite and massive pyroxenite reflect breaking up of early formed rocks by magmatic currents. Clinopyroxenes from clinopyroxenites are characterized by Fe/(Fe+Mg) = 0.07-0.29, low T i 02 (0.00-0.25%), A 120 3 (0.00-1.63%), N a 20 (0.00-1.06%), and Cr, and high Wo component and Sr. Micas from pyroxenites have Fe/(Fe+Mg) = 0.12-0.28, low T i 02 (0.17-1.73%) and have reverse and normal pleochroism and increasing A1 with increasing Fe/(Fe+Mg). Mineral compositions among feldspathic, massive and micaceous pyroxenites overlap; no zonation of the complex from outer contact inward is discernible with respect to the Fe and Mg content. However, phlogopites in "inch-scale" layering have low Fe/(Fe+Mg) of 0.12 and have reverse pleochroism due to F e 3+ entry into the Al-deficient tetrahedral site. Micas from phoscorite and carbonatites have reverse pleochroism, Fe/(Fe+Mg) = 0.05-0.58, low T i 02 (0.00-0.84%) and decreasing A1 with increasing Fe/(Fe+Mg). Olivines range from F o79 to F o91 and have very low Ni content (<0.06% N i O ) . Olivines interpreted as xenocrysts have Fo 84 to F o 8 7 . One of the olivine xenocrysts has an NiO content of 0.29%. Minerals from carbonatites have initial 87S r / 86Sr ratios of 0.70393-0.70623 and 0.71022 and minerals from clinopyroxenites have values of 0.71152-0.71242. Smallscale inhomogeneities exist within samples. Postcrystallization processes can account for variations within samples, but cannot account for variations within a rock type or for raising initial 87S r / 86Sr ratios of pyroxene from 0.7039 to 0.7115. Magmas forming pyroxenites and some carbonatites were generated in a high Rb/Sr mantle from isotopically distinct sources. Hence, liquid immiscibility and differentiation are not viable mechanisms for relating these rocks to one another. Mixing of magmas and assimilation of crust may account for isotopic variations within a rock type. U-Pb dating of uranothorianite and baddeleyite from phoscorite and carbonatite yields an age of 2047+11/-8 m.y. Rb-Sr dating of phlogopites gives widely disparate apparent ages. Nine phlogopites yield an isochron of 2012 + 19 m.y. One sample of phlogopite gave ages of 1661-2360 m.y. in nine different determinations. Older micas may be present in the Phalaborwa complex. Based on the two methods, a best age of the Phalaborwa complex is 2030 + 18 m.y. Two of the syenite pipes, Kgopoeloe and Spitskop, show different levels of emplacement. Kgopoeloe is highly brecciated from a fluid derived from the syenite. Spitskop contains minor breccia and represents multiple injection of syenite. At Spitskop, inward crystallization of the second syenite forms a ring syenite and central syenite; both syenites have cumulus enrichment of minerals. Feldspathic pyroxenite of the Guide Copper Mine is cogenetic with the pyroxenites of the main complex and contains clinopyroxenes with oscillatory zoning. Fluctuations in f02 may be attributed to formation of an immiscible sulphide liquid. Multiple intrusion of the Phalaborwa Complex is proposed. Initial injection of potassic, probably ultrabasic, liquid formed the pyroxenites. A second intrusion of low-silica, carbonate-rich magma formed phoscorite and banded carbonatite. A third intrusion of carbonatite liquid formed the transgressive carbonatite. Magmatic, cumulus processes dominated the formation of the main complex. The syenites of Kgopoeloe and Spitskop are not cogenetic with rocks of the main complex. / AC2017

The Kaapvaal Craton was the scene of two major magmatic events around 2.0 Ga, namely the Bushveld Complex and the Phalaborwa Complex. Both complexes have been the subject of numerous palaeomagnetic studies during and prior to the 1980s. Despite these studies, systematic inconstancies for emplacement ages, in particular for the Bushveld Complex, have been found between the palaeomagnetic findings and well constrained ages. The greatest concern with the Bushveld Complex results are the large spread in pole positions previously determined for the different zones. This has been interpreted in prior studies to indicate that the Bushveld Complex was emplaced and cooled below the Curie temperature of magnetite over a time span of 50 my. The results obtained previously for the Phalaborwa Complex appear to be out of position (~16°) with respect to those for the Bushveld Complex. This is of concern because new geochronological data show that the Rustenburg Layered Suite of the Bushveld Complex was emplaced approximately 1 my after the Phalaborwa Complex. These inconsistencies have prompted the current re-investigation of the palaeomagnetic results for both the Bushveld Complex and the Phalaborwa Complex. New palaeomagnetic data collected from all zones of the Rustenburg Layered Suite from the Eastern, Northern and Western Lobes of the Bushveld Complex, yielded palaeomagnetic poles that eliminated the spread in the apparent polar wander path. This observation is in agreement with precise age data, constraining the time period of emplacement of the complex to ~ 6 my. Resulting beddingcorrected high blocking components from all zones produced better groupings, thereby supporting a primary magnetic signature and indicating that the complex was intruded in a near-horizontal position. Dual polarities identified within each zone of the complex and positive reversal tests have identified one of the oldest known reversals of the Earth’s magnetic field. iii Palaeomagnetic data from the Phalaborwa Complex have produced a pole position that is in close proximity to those obtained from the coeval Bushveld Complex. In an attempt to achieve a better understanding of tectonic events occurring in the Kaapvaal Craton a number of dual polarity dykes within the Bushvled and Phalaborwa Complexes were palaeomagnetic analysed. Results revealed that the acquired pole positions are in agreement with ~1.9 Ga dykes, indicating the possibility that the dykes occurring in both complexes are part of the same magmatic event. Palaeopoles generated during this study were used in refining the Kaapvaal Craton apparent polar wander path around 2.0 Ga, and in conjunction with other welldefined 2.0 Ga poles for the Kaapvaal Craton, a robust cratonic pole was produced that was used in Precambrian palaeographic reconstructions with emphasis on the postulated Vaalbara continent and the Columbia supercontinent. Palaeomagnetic reconstruction derived in this study has cast doubt on the existence of the Vaalbara continent at 2.0 Ga. Although, some support is given to the existence of the Columbia supercontinent at the same period.