A Dissertation submitted to the School of Geosciences, University of the Witwatersrand, Johannesburg, South Africa in fulfilment of the requirements for the degree of Master of Science. May 2015 / The Upper Zone (UZ) of the Bushveld Complex (BC) comprises several magnetitite layers throughout the entire sequence with the most prominent layer, the 2 m thick Main Magnetitite Layer (MML), located towards the base of the sequence. Magnetite mineral separates have been obtained from the UZ with particular focus on the MML in vertical profiles through the MML, Layer 1 and bifurcations of the MML, as well as profiles along the base of the MML and bifurcations. Magnetite mineral separates were also collected from Bierkraal and UCAR mine drill cores. The magnetite mineral separates were analyzed primarily for Cr and V as these two elements have the highest partition coefficients (D>200 and D=20-25 respectively) in magnetite and can be used as magmatic tracers. Electron microprobe data from the Bellevue drill core are also included. The gradational upper contacts of magnetitite layers with overlying anorthosite could be interpreted to suggest that the magnetitite layers accumulated through crystal settling. However, vertical profiles through 1 m of the MML all show an upward exponential decrease in Cr content (12 000-580 ppm) which is inconsistent with crystal settling but better explained by diffusion controlled bottom crystallization. The sharp base of the MML with the underlying anorthosite may suggest that the MML crystallized due to an abrupt event. The MML is not entirely homogeneous as evidenced by lateral heterogeneity along the base of the MML, identified by irregular Cr concentrations along the base of the MML and magnetitite bifurcations. This heterogeneity further supports the contention that the magnetitite layers are a product of diffusion controlled bottom crystallization. Reversals in Cr content, of differing magnitudes, in 3 of 4 vertical profiles above a dome structure interrupting the MML and in 2 of 4 vertical profiles through the MML, are attributed to intermittent convection on various scales bringing primitive undepleted magma into the crystallization zone. The magnitude of the reversals depends on the level to which the convection descends. The feldspar parting, a 10 cm thick horizon with cumulus plagioclase 1 m above the base of the MML, appears at a fairly constant Cr content in magnetite. The lack of a chemical break immediately above the feldspar parting suggests a physical process, such as pressure change, as a mechanism to account for the mineralogical change from the feldspar parting into massive magnetite in the upper portion of the MML. Vanadium, unlike Cr shows no systematic trends. Vanadium content of magnetitite layers is found to be comparable to that of the disseminated magnetite thus ruling out the possibility of a change in fo2 as a mechanism to induce magnetite crystallization. Disseminated magnetite in the UZ is suggested to have re-equilibrated with pyroxene and/or olivine during subsolidus ii cooling resulting in lower MgO contents of the disseminated magnetite compared to that of massive magnetitite layers. Similarities between magnetitite layers in Magnet Heights (eastern lobe); UCAR mine drill core, east of Brits (western lobe); Bierkraal drill core, north of Rustenburg (western lobe) and Bellevue drill core (northern limb) suggest that the different lobes of the BC may be connected.

Thesis (M.Sc.(Geology))--University of Pretoria, 2008. / Abstract in English. Includes bibliographical references (leaves 123-126).

Stratiform chromitite layers are peculiar to large layered mafic intrusions. The origin of these chromitite layers has been widely debated. Some petrologists suggested that the layers formed as a result of the mixing of two compositionally different magmas whereas others suggest that the chromitite layers formed from changes in pressure. The former hypothesis is widely accepted, and states that chromitite forms when a more evolved magma is injected into the chamber occupied by a more primitive one. To evaluate this hypothesis, a study has been conducted on the silicate textures and major element geochemistry of the silicate-rich layers above and below the MG4 chromitite package in the Critical Zone of the Rustenburg Layered Suite, part of the Bushvel Complex in South Africa.The MG4 chromitite package consists of several chromitite seamsseparated by silicate layers. Orthopyroxene and plagioclase (interstitial plagioclase) are observed in large amounts throughout the silicate layer, with less abundant clinopyroxene and some trace amounts of biotite. Throughout the silicate-rich layers above and below the MG4 chromitite layers (MG4 pyroxenite), the orthopyroxene exhibits no major compositional variation in major elements (Mg#= 1.15-1.25). This is also observed in the clinopyroxene composition throughout the study area. However, plagioclase, which dominates the lower part of the stratigraphy, varies in composition with a decrease in the calcium content (Ca= 0.8-0.5) and a simultaneous increase in the sodium content (Na=0.2-0.5). These similarities between the rocks above and below the MG4 chromitite layers suggest that the chromitite layer originated from a single magma or a mixture of two magmas with similar composition. This model is supported by the observed thin sections where orthopyroxene occurs as euhedral grains throughout the section especially above the 63.13m depth lying above the plagioclase –rich layer. Trace element analysis further suggest that the magma that crystallized the plagioclase-rich lower part mixed with the influx of new magma rich in Mg to crystallize the rocks of the upper sequence dominated by orthopyroxene, clinopyroxene and Na-rich plagioclase. / Dissertation (MSc)--University of Pretoria, 2011. / Geology / unrestricted

M.Sc. (Chemistry) / Please refer to full text to view abstract

The Main Zone is dominated by medium – grained, homogeneous gabbronorite, and anorthosites. The plagioclase compositions of the core is ((Na(0.227 – 0.353), K (0.012 – 0.046), Ca (0.651 – 0.777)) Al (1.630 – 1.752) Si (2.183 – 2.345) O8) and at the rim is ((Na (0.189 – 0.371), K (0.005 – 0.108), Ca (0.651 – 0.777)) Al (1.630 – 1.752) Si (2.183 – 2.345) O8). The composition of orthopyroxene is ((Mg (0.660 – 0.808), Fe (0.206 – 0.309), Ca (0.007 – 0.081)) Si (0.960 – 1.037) O3) and the compositions of clinopyroxene is ((Mg (0.229 – 0.678), Fe (0.092 – 0.427), Ca (0.012 – 0.475) Si (0.776 – 1.012) O3. The Mg# and An# varies with depth, where plagioclase increase in concentration the An# increases and the Mg# decreases. The variations in magma compositions are attributed to interlayering of different lithologies which are the result of fractional crystallisation in the magma chamber. This is supported by linear trends of the major and trace element bivariant plots. The magmatic event forming the Main Zone resulted in lateral expansion of the sheet – like magma chamber. The Main Zone formed through two magmas; first magma forming the lower Main Zone and the second magma, intruded the Main Zone at the level of the Pyroxenite Marker, forming the upper Main Zone. Copyright / Dissertation (MSc)--University of Pretoria, 2012. / Geology / unrestricted

PhD, Faculty of Engineering and the Built Environment, University of the Witwatersrand, 2010 / The aim of this investigation was to provide a proper design procedure for Merensky crush pillars, based primarily on underground measurements. Three sites with a variety of geotechnical conditions were selected. An interaction between the pillars and the rock mass around the stopes was shown by the literature, relevant laboratory tests and numerical modelling. During the investigations, nonlinear rock behaviour was observed at one of the sites. Further studies revealed that nonlinear behaviour also occurred in samples extracted from high stress conditions at the other sites, but the rock mass was not nonlinear at these sites. A methodology for determining stress from strain measured in nonlinear rock was established. The research also established that there is an approximately linear relationship between peak pillar strength and w/h ratio at ratios between about 1.2 and 8. The so called ‘squat’ effect is not observed because pillar failure is not contained within the pillar but extends into the foundations. A linear peak pillar strength formula was established from back analyses of underground pillar failures and was confirmed by numerical modelling. Pillar behaviour was established from underground measurements on one stability pillar and six crush pillars, which included peak and residual strengths. Also, stable and unstable loading conditions were established from an analysis of pillar bursts and the minimum strata stiffness for stable pillar failure was determined. This stiffness is only achieved near the advancing face and pillars that fail in the back areas are likely to burst. For this reason, pillar design needs to include the peak strength as large pillars may be too strong and fail in the back area. The residual strength also needs to be considered as the load-bearing capacity of these pillars needs to satisfy the criterion of 1 MPa across the stope to prevent back-breaks. This translates into a pillar stress of between 8 MPa and 13 MPa if the pillar lines are spaced 30 m apart. The peak and residual requirements have been included in a design chart, and the relationship between w/h ratio and residual strength is provided in a graph for easy design.

Submitted in fulfilment of the requirements for a Master of Science degree in Geology, in the Department of Geosciences, University of the Witwatersrand, Johannesburg, South Africa. 2015 / This study undertook to re-evaluate the conventional historic interpretation that accepted the Marginal Zone as representative of the chill phase to the earliest emplacement of Lower Zone magmas. The Clapham Trough preserves a thick sequence of the Marginal Zone rocks, at least 220 m thick. Poor exposures and incomplete stratigraphy of the rock succession that occurs between the floor and the Marginal Zone rocks presented great limitations to earlier studies, and led earlier workers to accepting that the base of the Bushveld Complex is the Marginal Zone norite. This study presents results from the 692 m CH6 drilled core, which intersects the Marginal-Lower Zone boundary in the Clapham Trough. The base of the CH6 drill core consists of melanorite (with less than 40 % cumulus plagioclase), which is conformable with the underlying, thick Basal Ultramafic Sequence (BUS, described in Wilson and Chunnett, 2010; and Wilson, 2012) separating the Marginal Zone rocks with the floor rocks of the Magaliesberg Formation. The amount of cumulus plagioclase in the Marginal Zone increases with increasing stratigraphic height such that the top units of the Marginal Zone are norite-leuconorites (typically 45-65 % cumulus plagioclase), bordering on anorthosite. The progressive changes in the modal variations led to the subdivision of the Marginal Zone norite to a basal Mafic Norite and a xenolith-bearing Shelter Norite. The latter is deemed a correlative of the Xenolithic Norite described at Olifants River Trough. Coupled with the increasing amount of cumulus plagioclase, the An# increases with stratigraphic height. The An# fractionation trend is reversed from that of the co-existing orthopyroxene observed in the same interval (An63-74 vs. En81-70). The reversed An# compositions are an abnormal differentiation trend. The compositional disequilibrium between co-existing orthopyroxene and plagioclase formed from in-situ crystallization with floatation of plagioclase, through convection, separating the cotectic phases. All the data in the Marginal Zone show that these rocks have continuous fractionation trends with no interruptions. The Marginal Zone rocks are cumulus rocks that formed through fractional crystallization in a temporarily closed magma chamber. The present work showed unequivocally that the Marginal Zone is a product of differentiation of earlier emplacement of B1-magma, and cannot be representative of either a chill zone or composite sills. The appropriate (parental) liquid composition of the Marginal Zone formed after 30 % crystallization of the B1-magma. The postulated liquid composition is 6 wt. % MgO and 56.7 wt. % SiO2. The entire Marginal Zone succession would have formed from about 30-54 % crystallization of the B1-magma. The crystallization of the Marginal Zone was ended abruptly by the emplacement of a new batch of B1-magma, which must have mixed with the residual magma that must have ponded atop Marginal Zone cumulates after 54 % crystallization. The mixing of the evolved residual magma and the primitive B1-magma formed the liquid postulated to be parental to the Lower Zone A (10.59 wt. % MgO and 57.10 wt. % SiO2). The Transitional Pyroxenite bears all the evidence of mixing between magmas of contrasting compositions, forming the 10-30 m gradational boundary unit between the Marginal Zone and Lower Zone A (correlative of the Lower Orthopyroxenite Subzone described at Olifants River Trough). The Lower Orthopyroxenite Subzone at the Clapham Trough is almost a mono-mineralic rock succession with generally constant orthopyroxene composition (En87-86), with the exceptions at marker norite horizons (En84-82; An83-81). The constant compositions observed in Lower Zone A are attributed to contemporaneous emplacement of new magma and differentiation, which maintained the composition of the parental magma.

A dissertation submitted to the Faculty of Science, University of the Witwatersrand, in fulfilment of the requirements for the degree of Master of Science. Johannesburg, 2015 / The Bushveld Complex (BC) is the largest known layered intrusion. This suite of rock crop out in northern South Africa to form the Western, Eastern and Northern Limbs. Most research carried out focuses on the mineralized horizons in the Rustenburg Layered Suite (RLS) of the BC. This study presents a large database of wireline geophysical logs across a substantive part of the stratigraphy of the RLS. These consist of density and magnetic susceptibility datasets sampled at 1 cm. The major lithologies of the RLS intersected in the boreholes presented are gabbro, gabbronorite, norite and anorthosite whose density histograms reveal that they are predominantly normally distributed, with density averages of 2.86-2.91 g/cm3. The lithologies consist of mainly two minerals, pyroxene and plagioclase. In general, the average density increases with an increase in pyroxene. The distribution of the magnetic susceptibility for these lithologies has a large variation from SI to 13.2 SI, which is typical of layered intrusions. Susceptibility distributions are also multi-modal, asymmetric and not normally distributed, which makes the average magnetic susceptibilities less representative of the lithologies. Cross-correlation plots between density and magnetic susceptibility for several boreholes show that the above-mentioned lithologies form clusters (circular to elliptical), which typically overlap. This has been further investigated using k-means classification, to automatically detect these clusters in the cross-correlation plots and to compare these with those created by lithologies. The comparison shows some degree of correlation, implying that physical properties can be used to identify lithologies. This is particularly true for the Eastern Limb. However the classification has not been effective in all of the boreholes and often becomes complicated and an inaccurate representation of lithology log. This occurs in boreholes in which there is an overlap in the physical properties of the abovementioned lithologies. Analysis on the density and magnetic susceptibility data has also been carried out using wavelet analysis at individual locations across the BC. This has revealed multi-scale cyclicity in all of the boreholes studied, which is attributed to subtle layering created by variations in modal proportions between plagioclase and pyroxene. In addition to this, since layering is generally ubiquitous across layered intrusions, this cyclicity can be assumed to be present across the entire BC. This technique may become increasingly important should the cyclicity in physical property data correlate with reversals in fractionation trends since this may suggest zones of magma addition, whose thickness or III volumes can be quantified using wavelet analysis. This could be an important contribution since the current perspective on magma addition in the RLS is that four major additions have formed this 8 km thick suite of rocks, as opposed to smaller periodic influxes of magma. Wavelet-based semblance analysis has been used to compare the wavelengths at which the cyclicity occurs across boreholes. A comparison of wavelengths of this cyclicity shows that boreholes in the northern Western Limb show positive correlation in the density data at wavelengths >160 m and 20-60 m, while those further south show correlations at wavelengths of 120-200 m and 60-80 m. Boreholes of the Eastern Limb show positive correlation in the density and magnetic susceptibility data at wavelengths of 10-20 m, 20-30 m and 5m. These positive correlations across boreholes in density and magnetic susceptibility respectively, may imply that cyclicity may be produced by a chamber-wide process for several kilometres of the BC.

The main magnetite seam of the Upper Zone of the Rustenburg Layered Suite (SACS, 1980) on the Bushveld Complex is known to host the world‘s largest vanadium bearing titaniferous iron ores. The vanadiferous titanomagnetites, contain vanadium in sufficient concentrations (1.2 - 2.2 per cent V₂O₅) to be considered as resources and vanadium has been mined historically by a number of companies among them Anglo-American, Highveld Steel and Vanadium and VanMag Resources as well as currently by Evraz Highveld Steel and Vanadium Limited of South Africa. The titanomagnetites contain iron ore in the form of magnetite and titanium with concentrations averaging 50-75 per cent FeO and 12-21 per cent TiO₂. The titaniferous iron ores have been historically dismissed as a source of iron and titanium, due to the known difficulties of using iron ore with high titania content in blast furnaces. The economic potential for the extractability of the titaniferous magnetites lies in the capacity of the ores to be separated into iron rich and titanium rich concentrates usually through, crushing, grinding and magnetic separation. The separatability of iron oxides and titanium oxides, is dependent on the nature in which the titanium oxide occurs, with granular ilmenite being the most favourable since it can be separated from magnetite via magnetic separation. Titanium that occurs as finely exsolved lamellae or as iron-titanium oxides with low titania content such as ulvospinel render the potential recoverability of titanium poor. The Upper Zone vanadiferous titanomagnetites contain titanium in various forms varying from discrete granular ilmenite to finely exsolved lamellae as well as occurring as part of the minerals ulvospinel (Fe₂TiO₄) and titanomagnetite (a solid solution series between ulvospinel and magnetite) . Discrete ilmenite constitutes between 3-5 per cent by volume of the massive titanomagnetite ores, and between 5-10 per cent by volume of the magnetite-plagioclase cumulates with more than 50 per cent opaque oxide minerals. The purpose of this research was to investigate the mineralogical setting and distribution of the iron and titanium oxides within the magnetitite layers from top to bottom as well as spatially along a strike length of 2 000m to determine the potential for the titanium to be extracted from the titanomagnetite ores. The titanomagnetites of the Upper Zone of the Bushveld Complex with particular reference to the Northern Limb where this research was conducted contains titanium oxides as discrete ilmenite grains but in low concentrations whose potential for separate economic extraction will be challenging. The highest concentration of titanium in the magnetite ores is not contained in the granular ilmenite, but rather in ulvospinel and titanomagnetite as illustrated by the marked higher concentration of TiO₂ in the massive ores which contain less granular ilmenite in comparison to the disseminated ores which contain 3 to 8 percentage points higher granular ilmenite than the massive ores. On the scale of the main magnetite seam, the TiO₂ content increases with increasing stratigraphic height from being completely absent in the footwall anorthosite. The V₂2O₅ content also increases with stratigraphic height except for in one of the 3 boreholes where it drops with increasing height. The decrease or increase patterns are repeated in every seam. The titanomagnetites of the main magnetite seam display a variety of textures from coarse granular magnetite and ilmenite, to trellis ilmenite lamellae, intergranular ilmenite and magnesian spinels and fine exsolution lamellae of ulvospinel and ferro-magnesian spinels parallel to the magnetite cleavage. The bottom contact of the main magnetite seam is very sharp and there is no titanium or vanadium in the footwall barely 10cm below the contact. Chromium is present in the bottom of the 4 layers that constitute the main magnetite seam and it upwards decreases rapidly. In boreholes P21 and P55, there are slight reversals in the TiO₂ and V₂O₅ content towards the top of the magnetite seams.

Thesis (MSc)--Stellenbosch University, 2012. / ENGLISH ABSTRACT: Pt mineralization within the Bushveld Complex is strikingly focused on the chromitite reefs, despite these horizons being associated with low volumes of base metal sulphide relative to Pt grade. Partitioning of Pt (Dsil/sulp) from silicate magma into immiscible sulphide liquid appears unable to explain Pt concentrations in chromitite horizons, due to the mismatch that exists between very large R factor required and the relevant silicate rock volume. Consequently, in this experimental study we attempt to gain better insight into possible Pt grade enhancement processes that may occur with the Bushveld Complex (BC) sulphide magma. We investigate the wetting properties of sulphide melt relevant to chromite and silicate minerals, as this is a key parameter controlling sulphide liquid percolation through the cumulate pile. Additionally, we have investigated how fractionation of the sulphide liquid from mono-sulphide-solid-solution (Mss) crystals formed within the overlying melanorite might affect sulphide composition and Pt grades within the evolved sulphide melt. Two sets of experiments were conducted: Firstly, at 1 atm to investigate the phase relations between 900OC and 1150OC, within Pt-bearing sulphide magma relevant to the BC; Secondly, at 4 kbar, between 900OC to 1050OC, which investigated the downwards percolation of sulphide magma through several layers of silicate (melanorite) and chromitite. In addition, 1atm experiments were conducted within a chromite dominated chromite-sulphide mixture to test if interaction with chromite affects the sulphide system by ether adding or removing Fe2+. Primary observations are as follows: We found sulphide liquid to be extremely mobile, the median dihedral angles between sulphide melt and the minerals of chromitite and silicate layers are 11O and 33O respectively. This is far below the percolation threshold of 60O for natural geological systems. In silicate layers sulphide liquid forms vertical melt networks promoting percolation. In contrast, the extremely effective wetting of sulphide liquid in chromitites restricts sulphide percolation. Inter-granular capillary forces increase melt retention, thus chromitites serve as a reservoir for sulphide melt. Sulphide liquid preferentially leaches Fe2+ from chromite, increasing the Fe concentration of the sulphide liquid. The reacted chromite rims are enriched in spinel end-member. This addition of Fe2+ to the sulphide magma prompts crystallization Fe-rich Mss, decreasing the S-content of sulphide melt. This lowers Pt solubility and leads to the formation of Pt alloys within the chromitite layer. Eventually, Cu-rich sulphide melt escapes through the bottom of the chromitite layer. These observations appear directly applicable to the mineralized chromitite reefs of the Bushveld complex. We propose that sulphide magma, potentially injected from the mantle with new silicate magma injections, percolated through the silicate cumulate overlying the chromitite and crystallized a significant volume of Fe-Mss. Chromitite layers functioned as traps for percolating, evolved, Cu-, Ni- and Pt-rich sulphide liquids. This is supported by the common phenomenon that chromitites contain higher percentages of Ni, Cu and Pt relative to hanging wall silicate layers. When in contact with chromite, sulphide melt is forced to crystallize Mss as it leaches Fe2+ from the chromite, thereby further lowering the S-content of the melt. This results in precipitation, as Pt alloys, of a large proportion of the Pt dissolved in the sulphide melt. In combination, these processes explain why chromitite reefs in the Bushveld Complex have Pt/S ratios are up to an order of magnitude higher that adjacent melanorite layers. / AFRIKAANSE OPSOMMING: Pt mineralisasie in die Bosveld Kompleks is kenmerkend gefokus op die chromatiet riwwe, alhoewel die riwwe geassosieer is met lae volumes basismetaal sulfiedes relatief tot Pt graad. Verdeling van Pt (Dsil/sulp) vanaf silikaat magma in onmengbare sulfiedvloeistof is klaarblyklik onvoldoende om Pt konsentrasies in chromatiet lae te verduidelik, a.g.v. die wanverhouding wat bestaan tussen ‘n baie groot R-faktor wat benodig word en die relatiewe silikaat rots volumes. Gevolglik, in die eksperimentele studie probeer ons beter insig kry oor moontlike Pt graad verhogingsprosesse wat plaasvind in die BK sulfied magma. Ons ondersoek die benattingseienskappe van sulfied vloeistof relevant tot chromiet- en silikaat minerale, omdat dit die sleutel maatstaf is vir die beheer van sulfied vloeistof deursypeling deur die kumulaat opeenhoping. Addisioneel het ons ook ondersoek hoe die fraksionering van sulfied vloeistof vanaf MSS kristalle, gevorm binne die hangende melanoriet muur, moontlik die sulfied samestelling en Pt graad binne ontwikkelde sulfied smelt kan beïnvloed. Twee stelle van eksperimente is gedoen: Eerstens, by 1 atm om ondersoek in te stel oor fase verwantskappe tussen 900OC en 1150OC, binne ‘n Pt-verrykte sulfied magma samestelling relevant tot die BK; Tweedens, by 4 kbar, tussen 900OC tot 1050OC, wat die afwaartse deursypeling van sulfied magma deur veelvuldige lae van silikaat minerale en chromatiet. Addisionele 1 atm eksperimente is gedoen binne ‘n chromiet gedomineerde chromiet-sulfied mengsel, om te toets of interaksie met chromiet die sulfied sisteem affekteer deur Fe2+ te verwyder of by te dra. Primêre observasies is soos volg: Ons het bevind sulfiedsmelt is uiters mobiel, die mediaan dihedrale hoek tussen sulfiedsmelt en minerale van chromiet en silikaat lae is 11O en 33O onderskydelik. Dit is ver onder die deursypelings drumpel van 60O vir natuurlike geologiese stelsels. In silikaatlae vorm die sulfiedsmelt vertikale netwerke wat deursypeling bevorder. Inteendeel, uiters effektiewe benatting van sulfiedsmelt binne chromatiete vertraag sulfied deusypeling. Tussen kristal kapilêre kragte verhoog smelt retensie, dus dien chromatiete as ‘n opgaarmedium vir sulfiedsmelt. S oorversadigte sulfied vloeistof loogsif Fe2+ vanuit chromiet en veroorsaak ‘n verhoging in Fe-konsentraie. Die gereageerde chromiet buiterante is daarvolgens verryk in Cr-spinêl eind-ledemaat. Die addisionele byvoeging van Fe2+ aan sulfied magma veroorsaak die kristalisasie van Fe-ryke Mss en verlaag dus die S-konsentrasie van die sulfied smelt. Dit verlaag Pt oplosbaarheid en lei tot die formasie van Py allooie binne-in chromatiete. Ten einde, ontsnap Cu-ryke sulfied smelt deur die onderkant van die chromatiet lae. Die observasies is direk van toepassing op die gemineraliseerde chromatiet riwwe van die Bosveld Kompleks. Ons stel voor dat sulfied magma, potensiaal ingespuit vanuit die mantel saam nuwe inspuitings van silikaat magma, deur die hangende silikaat kumulaat bo chromatiet lae deurgesypel het en ‘n betekenisvolle volume Fe-Mss gekristalliseer het. Chromatiet lae het gefunksioneer as lokvalle vir afwaartsbewegende, ontwikkelde, Cu-, Ni-, en Pt-ryke sulfied vloeistowwe. Dit word ondersteun deur die algemene verskynsel dat chromatiete hoër persentasies van Ni, Cu en Pt relatief teenoor die hangende muur silikaat lae het. Wanneer sulfied smelt in kontak is met chromiet, word dit geforseer om Mss te kristalliseer soos Fe2+ geloogsif word, waarvolgens die smelt se S konsentrasie verder verlaag word. Dit veroorsaak die presipitasie, as Pt allooie, van groot proporsies opgeloste Pt vanuit sulfied smelt. Deur die prosesse te kombineer, kan dit moontlik verduidelik word hoekom chromatiet riwwe in die Bosveld Kompleks Pt/S verhoudings veel hoër is as aanrakende melanoriet lae.