Although Cr-spinels from the Bushveld Complex of South Africa have been studied for several decades, those from the northern lobe have not received as much attention as Cr-spinels from other lobes of the intrusion. Disseminated Cr-spinels from the Flatreef of the northern lobe have chromium numbers [Cr# = 100 × Cr/(Cr + Al)] that vary widely, from low values of around −32 hosted in altered olivines to high values of ∼55–82 hosted in orthopyroxenes and plagioclase. Few Cr-spinel crystals have almost identical Cr# from rim to core. Almost all Cr-spinels, irrespective of their host silicate types, display slight decreases in Cr# from the core to the rim. These Cr-spinels, however, are characterized by highly variable magnesium numbers [Mg# = 100 × Mg/(Mg 2 + Fe 2+ )], ranging from 37–51 for Cr-spinels in olivines to 1–13 for Cr-spinels in pyroxenes and olivines. The chemistry of the Flatreef Cr-spinels in this study suggests the involvement of crustal contamination and re-equilibration. Plots of oxides such as aluminium, magnesium and titanium are consistent with contamination of the Platreef by floor rocks to the intrusion such as dolomites and argillaceous shales. The low Cr# (<∼20) and enriched Al 2 O 3 (>32 wt%) Cr-spinel chemistry signatures reported in this work appear not to have been previously documented in the Bushveld Complex.

The Rustenburg Layered Suite of the Bushveld Complex, South Africa, contains the world’s largest resource of chrome and platinum group elements (PGEs). Both Cr and PGEs are found in chromitite layers within an approximately 1,000-m-thick section of ultramafic to mafic rocks known as the Critical zone. Neither the process of how the chromitite layers form nor the role that chromite plays in collecting the PGEs is clear. Major and trace element contents of chromite and silicate minerals from each of the 13 chromitite layers, and from chromite in the adjacent peridotites and norites, have been determined. The concentrations of PGEs in both chromite and silicates are less than detection levels (10–20 ppb). Thus, neither are the host of the PGEs in these rocks. The Cr# and Fe# of the chromites from chromitite layers are similar to those found in experiments carried out to model the crystallization of the initial magma (B1) of the Bushveld, with the same decrease in Cr# with increase in Fe#. The fO2 of the experiments Δ 0 FMQ (where FMQ = fayalite-magnetite-quartz buffer) and those of the chromitite chromite calculated from the Fe3+/FeTotal ratios and the V contents of the chromite are similar. Variations in trace element contents of the chromitite chromite can also be modeled using the B1 composition and allowing for ~40% crystal fractionation across the stratigraphy.

Current common usage of chondrite-normalized rare earth element (REE) plots has seen the advent of the disappearance of promethium as a regular phenomenon, in addition to the occasional omission of other analytically compromised lanthanides, and the substitution of Y for similarly sized heavy lanthanides. This communication represents an examination of the underlying principles behind the normalized lanthanide plot, the reasons for increased flexibility in the application of the plot, and recommendations for its use. Specifically, the distinction between the lanthanides and the rare earth elements sensu lato, and the resultant distinction between normalized lanthanides plotted according to their atomic number and REE (or other REE-incorporating) spidergram plots, is highlighted. The distinction is made between the traditional REE plot and what should arguably be described as a modified REE spidergram, which can accommodate chemical interventions. The current actual role of induced Pm as an industrial element is also noted as a justification for the continued recognition of its existence.

Abstract Unlike platinum-group elements (PGEs) deposits from the western and eastern lobes of the Bushveld Complex which are generally thought to have been largely formed as a consequence of primary magmatic processes, PGEs from the northern lobe are generally thought to have originated by processes involving contamination of parental magmas by crustal sources. We investigated mineral compositions from the deep Flatreef at the Turfspruit farm where dolomitic calc-silicate rocks form part of the country rocks to determine the role of crustal sources in the formation of PGE mineralization. Orthopyroxene compositions vary widely from En77.9 to En91.7, and compositions of clinopyroxene display wider variations (Ca38.8−67.0Mg24.7−51.6Fe6.4−19.6). Plagioclase similarly varies widely in composition (An1.3−73.7) whereas compositions of olivine (Fo85.9−87.6) are very restricted. Temperatures ranging from ~ 500 to 1300 oC were obtained from two pyroxene geothermometers similarly likely indicating crystallization from contaminated magma and/or hydrothermal alteration. The wide compositional ranges of calcium-bearing phases such as clinopyroxene and feldspars suggest that the Platreef crystallized from parental magmas that were contaminated by dolomitic calc-silicate floor rocks. The very restricted and highly magnesian forsterite contents in olivines are inconsistent with expected primary magmatic compositions. Simple mixing models utilizing existing sulfur isotope data suggest that, in the Turfspruit area, contamination of Bushveld Complex parental magmas with 2–10% contamination by floor rock dolomitic calc-silicates with δ34S values of 32‰ would account for the observed data in the Turfspruit area.

By examining the minor and trace element contents of chromites from three intrusions—the Bushveld Complex (South Africa), the Stillwater Complex (USA), and the Great Dyke (Zimbabwe)—and comparing these chromite compositions with those of magmas from which they could have formed (komatiites and picrites) we conclude that ( i) the variations in Ti, V, Sc, and Ga contents across stratigraphy and across individual layers do not support the model of magma mixing leading to chromite-only crystallization, ( ii) the chromites from the lowest levels of the intrusions could have crystallized from komatiite liquids that were contaminated with continental crust, ( iii) the Great Dyke chromites have the highest Cr# and lowest incompatible element contents and formed from a liquid closest to komatiite, ( iv) all of the chromites, except those of the Dunite Succession of the Great Dyke have equilibrated with a liquid that also had crystallized pyroxene, ( v) the Great Dyke and Stillwater chromites show a narrower range in composition than the Bushveld chromites, and ( vi) Chromites from the western limb of the Bushveld Complex contain much higher V contents than all the other chromites. This requires either, that the oxygen fugacity ( fO2) was lower in the western Bushveld or that the chromites equilibrated with a V-rich magma. We favor a model where chromite and silicate minerals crystallized in cotectic proportions (∼2:98). The chromite, silicates, and transporting liquid are emplaced into the magma chamber. During emplacement the chromite and silicate separated due to viscous particle flow to form a massive chromite layer overlain by silicates.

ABSTRACT The wide-reef facies of the Merensky Reef in the eastern part of the western lobe of the Bushveld Complex was sampled in order to better resolve otherwise spatially constrained variation in highly siderophile elements across this geological unit. The platinum group element mineralogy and whole-rock highly siderophile element concentrations were measured across two vertical sections in close proximity. In one section, the Merensky Reef unit was bound by top and bottom platinum group elements-enriched horizons (reefs) with a well-developed pegmatoidal phase in the top third of the intrareef pyroxenite, but with neither a top nor a bottom chromitite present. The other drill core section featured a thin (<1 cm thick) chromitite layer associated with the highest platinum group element concentrations of any rock in this study as the bottom reef, but with a chromitite-absent top reef, and very poor development of the pegmatoid. Primitive mantle-normalized profiles of the main lithological units show relatively flat, primitive mantle-like highly siderophile element abundances (Cr, V, Co, Ni, platinum group elements, Au and Cu) in the Merensky pyroxenite, with modest depletion in Ir-affiliated platinum group elements. The platinum group element-rich top and bottom reefs, and the pegmatoidal upper pyroxenites, display characteristic enrichment in the Pt-affiliated platinum group elements and undepleted Ir-affiliated platinum group elements. The leuconoritic hanging wall and footwall rocks show comparable highly siderophile element profiles, distinguished from one another by relative depletion in the Pt-affiliated platinum group elements of the footwall samples. The vertical variation in highly siderophile element abundances through both sections is characterized by low platinum group element abundances through the lower reef pyroxenite, with platinum group element, Au, and Cu ± Ni concentrations increasing through the upper pegmatoidal pyroxenite, and main enrichment peaks at the top and bottom reefs. Significant localized (centimeter-scale) zones of chalcophile metal depletion are present immediately above the top reef and below the bottom reef. In addition, a wider zone of Pt-affiliated platinum group elements (with Pd more depleted than Pt)-depletion was identified within the pegmatoidal pyroxenite around one meter below the top reef. The platinum group element mineralogy of the bottom reef consists mainly of platinum group element sulfides, with minor arsenides and antimonides. In contrast, the platinum group element mineralogy of the top reef, and the small amount of data from the intrareef pyroxenite, mainly consist of Pt-affiliated platinum group elements-Bi-tellurides. The Pt-sulfides are mainly equant, relatively coarse crystals (many grains between 50 to 100 μm2 area), contrasting with the Pt-affiliated platinum group elements-Sb-As and -Bi-Te minerals that tend be high aspect-ratio grains, occurring in veinlets or as rims on earlier-forming platinum group element phases. These Te-As-Bi-Sb compounds are closely associated with chlorite, actinolite, quartz, and chalcopyrite, consistent with secondary deposition at lower temperatures and association with aqueous fluids. A model is proposed involving the emplacement of the Merensky unit as a magma pulse into at least semi-crystallized host rock, followed by aqueous fluid saturation and local migration, combined with concentration of late magmatic fluids around the top and bottom contacts of the magma pulse. Late remobilization of Pt-affiliated platinum group elements from the zones immediately (centimeter-scale) above the top reef, and from the underlying meter or two of pyroxenite, and from the centimeters underlying the bottom reef, have added additional platinum group elements to the reefs as late platinum group elements-Te-As-Bi-Sb minerals, independent of whether or not chromite is present in the reef initially.

ABSTRACT The Vredefort granophyre dikes have long been recognized as being derived from the now-eroded Vredefort melt sheet. One dike, in particular, the Daskop granophyre dike, is notable for a high abundance of lithic clasts derived from various stratigraphic levels. In this study, we mapped the distribution of the clasts throughout the continuously exposed section of the dike using field mapping and aerial drone photography and attempted to constrain the emplacement mechanisms of the dike. We found that the clasts are not homogeneously spread but instead are distributed between clast-rich zones, which have up to 50% by area clasts, and clast-poor zones, which have 0–10% by area clasts. We examined three models to explain this distribution: gravitational settling of clasts, thermally driven local assimilation of clasts, and mechanical sorting of clasts due to turbulent flow. Of the three models, the gravitational settling cannot be supported based on our field and geophysical data. The assimilation of clasts and turbulent flow of clasts, however, can both potentially result in inhomogeneous clast distribution. Zones of fully assimilated clasts and nonassimilated clasts can occur from spatial temperature differences of 100 °C. Mechanical sorting driven by a turbulent flow can also generate zones of inhomogeneous clast distribution. Both local assimilation and mechanical sorting due to turbulent flow likely contributed to the observed distribution of clasts.

Introduction: Zircon ages are seen as incorruptible and unquestionable recorders of magmatic ages. However, evidence from rocks rendered geological complex by large bolide impacts and related shock metamorphism and crater evolutionary processes, combined with preand post-impact orogenic activities, demonstrate that misleading age-data can be produced, resulting in significantly erroneous geological interpretations of the geological history. In this study, a regionally metamorphosed leucogabbroic sill in volcanosedimentary rocks affected by impact heating by the Palaeoproterozoic Sudbury Igneous Complex (SIC) and localised shearing has produced zircon U-Pb ages both by SHRIMP and TIMS [1] reflecting only impactaged lead loss, leading to the conventional interpretation that the host rock itself represents juvenile, impactinduced magmatism. Furthermore, no evidence of shock deformation had been identified in the previously dated grains. This is believed to be inconsistent with “circumstantial” evidence such as radioisotopic tracer and geochemical evidence, and comparison with other spatially and compositionally apparently equivalent leucogabbro sills regionally [1]. Geological setting: A locally mylonitic constituent of a major regionally-transgressing fault system (the Folsom Lake Fault Zone and South Range Shear Zone) lying within the thermal aureole of the large Palaeoproteroic Sudbury impact structure offers a relatively unique opportunity to examine the effects of superimposed fault-induced deformation, impact-induced shock metamorphism, and melt sheet-induced thermal aureole effects. Recent mapping and petrological and SEM-based study has revealed that the Drury Township leucogabbro hosts mylonitic rocks related to the regional scale South Range Shear Zone in areas where this was not previously recorded as such. Furthermore, the mylonite has been locally remelted synto post-deformationally, characterised by centimetreto decimetre-scale melt veinlets (Fig. 1). Within or proximal to these veinlets, populations of very small diameter zircon grains have been identified which display a variety of textures, in contrast to those previously extracted for geochronology, including clear evidence of shock metamorphism and subsequent new growth of zoned igneous zircon, which has been subsequently cataclastically deformed. The association of thorite with the zircon is also suggestive of thermal breakdown of existing zircon. These textures clearly require the existence of magmatic zircon prior to fault deformation and high grade thermal metamorphism, precluding the possibility that the intrusion is impactrelated or -induced in origin. This also has implications for the prospective ability of large terrestrial impact craters to induce or access mantle partial melting.

SUMMARYThe Bushveld Complex has continued to serve as the basis for study into the fundamental nature of petrological processes for layered intrusion formation and for oxide and sulphide hosted Platinum Group Element (PGE)–Cu–Ni ore deposits. These studies have included discoveries in terms of the physical extent of Bushveld magmatism, both laterally and internally. Lateral variations in the mafic to ultramafic Rustenburg Layered Suite of the Northern Lobe of the complex have also revealed petrologically distinctive Upper Critical Zone equivalent rocks (the so-called Flatreef) with enhanced contamination and mineralization traits that reflect a transition between Eastern and Western Lobe equivalent stratigraphy and Platreef-style complexity. Traditional magma mixing models have been re-examined in light of radiogenic isotopic evidence for crustal involvement early in the chromite precipitation or formation process, combined with evidence for associated heterogeneous fluid contents, cryptic layering profiles, and textural evidence. A wide variety of alternative ore-genesis models have been proposed as a consequence. The fundamental mechanics of magma chamber processes and the existence of the magma chamber as an entity have been called into question through various lines of evidence which have promoted the concept of progressive emplacement of the complex as a stack of not-necessarily-quite-sequentially intruded sills (with or without significant quantities of transported phenocrysts), emplaced into variably crystallized and compacted crystal-liquid mush mixtures, modified by compaction-driven late magmatic fluid (silicate and aqueous) activity. Alternatively, petrological and geochemical observations have been used to discount these interpretations in favour of more conventional cooling and gravity-driven accumulation of silicate and ore minerals in a large, liquid-dominated system.RÉSUMÉLe complexe de Bushveld a demeuré à la base d’études sur la nature fondamentale des processus pétrologiques de formation d’intrusions litées et des gîtes des éléments du groupe platine (ÉGP)-Cu-Ni hébergés dans les oxydes et les sulfures. Ces études ont comporté des découvertes sur l’étendue physique, à la fois latérale et interne, du magmatisme de Bushveld. Les variations latérales de la suite stratifiée et mafique à ultramafique Rustenburg du lobe nord du complexe ont également révélé des roches équivalentes pétrologiquement distinctes de la zone critique supérieure (le communément désigné Flatreef) avec des traits de contamination et de minéralisation accrus qui reflètent une transition entre la stratigraphie équivalente des lobes est et ouest et la complexité de type Platreef. Les modèles traditionnels de mélanges magmatiques ont été réexaminés à la lumière de preuves isotopiques radiogéniques indiquant une implication de la croûte au début du processus de précipitation ou de formation de la chromite, combinées à des preuves de contenu fluide hétérogène associé, de profils de litage cryptique et de preuves texturales. Ainsi, une grande variété de modèles alternatifs de genèse de minerai a été proposée. La mécanique fondamentale des processus de la chambre magmatique et l'existence de la chambre magmatique en tant qu'entité ont été remises en question au moyen de divers éléments de preuve qui ont mis en avant le concept de mise en place progressive du complexe sous forme d'un empilement non-nécessairement séquentiel de sills injectés (avec ou sans quantités significatives de phénocristaux transportés) mis en place dans des mélanges de bouillie cristaux/liquide à cristallisation et compaction variable, modifiés par une activité tardive de fluide magmatique (silicaté et aqueux) induite par la compaction. Alternativement, des observations pétrologiques et géochimiques ont été utilisées pour écarter ces interprétations en faveur d'un processus plus conventionnel de refroidissement et d’accumulation de minérais silicatés et minéralisés induite par la gravité dans un vaste système à dominance liquide.

The offset dykes of the Sudbury Igneous Complex comprise two distinct main magmatic facies, a high-temperature inclusion-free quartz diorite (QD), and a subsequently intruded lower temperature, mineralized, and inclusion-rich quartz diorite (MIQD). The MIQD facies was emplaced after QD dykes had solidified. Key controlling factors of the two injection phases were (1) the development of a coherent roof, which confined the melt sheet; and (2) the periodic increase of melt and fluid pressure within the melt sheet. For the injection of QD melt, the melt pressure exceeded the normal stress acting on fracture surfaces. For the later refracturing of QD dykes and the injection of MIQD melt, the melt pressure increased further, exceeding the tensile strength of, and the normal stress acting on, QD dykes. We associate the melt pressure increase required for both injection episodes with degassing and devolatilization of cooling melt close to the roof. Within the hydraulically connected melt column, the related pressure increase was transmitted to the base of the melt sheet where QD and MIQD melt was extracted into dykes. Residual core to rim thermal gradients in the QD dykes produced tensile strength gradients, accounting for the typically central location of MIQD dykes within QD dykes.

The Uitkomst intrusion is a tubular mafic-ultramafic layered body that hosts one of South Africa's largest Ni-Cu-Cr-PGE deposits, Nkomati. The sulphide ore occurs in the form of massive lenses in the immediate quartzitic footwall and as disseminations within peridotite. The chromite ore forms an up to ∼10-m-thick layer in the lower portion of the intrusion. Uitkomst has generally been interpreted as a magma conduit, possibly related to the Bushveld event. Here, we present a new high-precision U-Pb zircon date of 2057.64 ± 0.69 Ma that overlaps with the age of the Merensky Reef of the Bushveld Complex and thus demonstrates a coeval relationship between the intrusions. Based on incompatible trace elements as well as O- and Nd isotope data (εNd −4.5 to −6.2), we show that the Uitkomst parent magmas were contaminated with up to 20% Archean upper crust prior to emplacement, and with up to 15% dolomitic country rock during emplacement. Ore formation at Nkomati was critically aided by substantial devolatisation and removal of dolomitic floor rocks leading to hydrodynamic concentration of sulphide and chromite during slumping of crystal mushes into the trough-like centre of the subsiding intrusion and its footwall.

The ±260Ma Panzhihua mafic layered intrusion is one of a suite of intrusions related to the Emeishan Large Igneous Province (ELIP), SW China. The Panzhihua intrusion hosts a large (±60m thick) FeTi oxide ore body at the base of the intrusion. This study provides new constraints on the genesis of FeTi oxide ore layers at the Panzhihua intrusion using: geochemistry, petrography, and modeling of parent magma crystallisation with variation in fO2 and H2O content.Whole-rock major element geochemical trends are controlled by the modal abundance of FeTi oxides (Ti-magnetite and ilmenite). The lower ±270m of the intrusion is clearly dominated by a Ti-magnetite accumulation trend whereas above this level the geochemical variation is consistent with a Ti-magnetite and ilmenite accumulation trend. This suggests that the lower ±270m of the intrusion crystallised at higher fO2 conditions relative to that above ±270m. Detailed petrographic analysis of exsolution microtextures of the Ti-magnetite from within ore layers and gabbroic host rocks also indicates a relative increase in fO2 within the ore rocks. Modeling of Panzhihua parent magma shows that Ti-magnetite crystallises late for dry (<0.5wt.%) starting compositions and early for wet (>1.5wt.% H2O) compositions. This corresponds to a distinct variation in plagioclase crystallisation temperature, which decreases with increasing H2O content of the parent magma. The initial plagioclase composition varies from An55 for a dry magma up to An73 for a magma containing 3wt.% H2O. The average plagioclase composition for the Panzhihua intrusion is An58, indicating an initial magma with low H2O content. Textures within the FeTi oxide ore rocks clearly indicate that Ti-magnetite crystallised after the silicate phases (plag.+cpx.) and disequilibrium textures indicate that the FeTi oxide ores were not in equilibrium with the enclosed silicates.We present a model for FeTi oxide ore formation in an open system by multiple replenishments of magma with variable H2O contents, Ti-magnetite crystal load and volume. Intruding H2O- and crystal-rich magmas effectively thermo-chemically erode previously formed gabbroic cumulates forming the footwall and incorporates previously crystallised silicate grains. These grains are consumed by the magma due to the high H2O content and higher temperature resulting in the association of consumed silicate primocrysts enclosed in FeTi oxide ore layers.

Magmatic oxide deposits in the w260 Ma Emeishan Large Igneous Province (ELIP), SW China and northern Vietnam, are important sources of Fe, Ti and V. Some giant magmatic Fe-Ti-V oxide deposits, such as the Panzhihua, Hongge, and Baima deposits, are well described in the literature and are hosted in layered mafic-ultramafic intrusions in the Panxi region, the central ELIP. The same type of ELIPrelated deposits also occur far to the south and include the Anyi deposit, about 130 km south of Panzhihua, and the Mianhuadi deposit in the Red River fault zone. The Anyi deposit is relatively small but is similarly hosted in a layered mafic intrusion. The Mianhuadi deposit has a zircon U-Pb age of w260 Ma and is thus contemporaneous with the ELIP. This deposit was variably metamorphosed during the Indosinian orogeny and Red River faulting. Compositionally, magnetite of the Mianhuadi deposit contains smaller amounts of Ti and V than that of the other deposits, possibly attributable to the later metamorphism. The distribution of the oxide ore deposits is not related to the domal structure of the ELIP. One major feature of all the oxide deposits in the ELIP is the spatial association of oxide-bearing gabbroic intrusions, syenitic plutons and high-Ti flood basalts. Thus, we propose that magmas from a mantle plume were emplaced into a shallow magma chamber where they were evolved into a field of liquid immiscibility to form two silicate liquids, one with an extremely Fe-Tirich gabbroic composition and the other syenitic. An immiscible Fe-Ti-(P) oxide melt may then separate from the mafic magmas to form oxide deposits. The parental magmas from which these deposits formed were likely Fe-Ti-rich picritic in composition and were derived from enriched asthenospheric mantle at a greater depth than the magmas that produced sulfide-bearing intrusions

The Panzhihua intrusion in southwest China is part of the Emeishan Large Igneous Province and host of a large Fe-Ti-V ore deposit. During emplacement of the main intrusion, multiple generations of mafic dykes invaded carbonate wall rocks, producing a large contact aureole. We measured the oxygen-isotope composition of the intrusions, their constituent minerals, and samples of the country rock. Magnetite and plagioclase from Panzhihua intrusion have δ18O values that are consistent with magmatic equilibrium, and formed from magmas with δ18O values that were 1–2‰ higher than expected in a mantle-derived magma. The unmetamorphosed country rock has high δ18O values, ranging from 13.2‰ (sandstone) to 24.6–28.6‰ (dolomite). The skarns and marbles from the aureole have lower δ18O and δ13C values than their protolith suggesting interaction with fluids that were in exchange equilibrium with the adjacent mafic magmas and especially the numerous mafic dykes that intruded the aureole. This would explain the alteration of δ18O of the dykes which have significantly higher values than expected for a mantle-derived magma. Depending on the exact δ18O values assumed for the magma and contaminant, the amount of assimilation required to produce the elevated δ18O value of the Panzhihua intrusion was between 8 and 13.7 wt.%, assuming simple mixing. The exact mechanism of contamination is unclear but may involve a combination of assimilation of bulk country rock, mixing with a melt of the country rock and exchange with CO2-rich fluid derived from decarbonation of the marls and dolomites. These mechanisms, particularly the latter, were probably involved in the formation of the Fe-Ti-V ores.

Magmatic oxide deposits in the ∼260 Ma Emeishan Large Igneous Province (ELIP), SW China and northern Vietnam, are important sources of Fe, Ti and V. Some giant magmatic Fe-Ti-V oxide deposits, such as the Panzhihua, Hongge, and Baima deposits, are well described in the literature and are hosted in layered mafic-ultramafic intrusions in the Panxi region, the central ELIP. The same type of ELIP-related deposits also occur far to the south and include the Anyi deposit, about 130 km south of Panzhihua, and the Mianhuadi deposit in the Red River fault zone. The Anyi deposit is relatively small but is similarly hosted in a layered mafic intrusion. The Mianhuadi deposit has a zircon U-Pb age of ∼260 Ma and is thus contemporaneous with the ELIP. This deposit was variably metamorphosed during the Indosinian orogeny and Red River faulting. Compositionally, magnetite of the Mianhuadi deposit contains smaller amounts of Ti and V than that of the other deposits, possibly attributable to the later metamorphism. The distribution of the oxide ore deposits is not related to the domal structure of the ELIP. One major feature of all the oxide deposits in the ELIP is the spatial association of oxide-bearing gabbroic intrusions, syenitic plutons and high-Ti flood basalts. Thus, we propose that magmas from a mantle plume were emplaced into a shallow magma chamber where they were evolved into a field of liquid immiscibility to form two silicate liquids, one with an extremely Fe-Ti-rich gabbroic composition and the other syenitic. An immiscible Fe-Ti-(P) oxide melt may then separate from the mafic magmas to form oxide deposits. The parental magmas from which these deposits formed were likely Fe-Ti-rich picritic in composition and were derived from enriched asthenospheric mantle at a greater depth than the magmas that produced sulfide-bearing intrusions of the ELIP.

The Panzhihua intrusion (260 Ma) is part of the Emeishan Large Igneous Province (ELIP), SW China. Layered intrusions related to the ELIP have recently received much attention due to their economic potential for Ni-Cu-PGE and Fe-Ti oxide ore deposits. ELIP layered intrusions vary in composition from ultramafic to mafic, which are interpreted to have crystallised from two distinct magma series through plume-related partial melting of the mantle: (1) high-Ti basalt and (2) low-Ti basalt.We present here new trace element, PGE, and Sr-Nd isotope data for the Panzhihua intrusion in order to constrain the evolution of the parent magma from source to current chamber. REE profiles are generally characterised by flat LREE profiles with low (La/Sm)(N) < 2 and steeper HREE with (Gd/Yb)(N) - general < 2.5. Primitive mantle-normalised trace element profiles are characterised by negative Th-U, La-Ce and Zr-Hf anomalies. The Panzhihua intrusion is PGE-depleted relative to related volcanic rocks, with total PGE concentrations generally << 5 ppb. Cu/Pd (24,000-250,000) and Pd/Pt (0.3-8.3). (Sr-87/Sr-86) i and epsilon(Ndi) value ranges are 0.7043-0.7054 and 0.61-3.71 respectively with distinct enriched values for marginal zone rocks. The Panzhihua magma is likely related to a picritic parent magma. Sr-Nd isotope data indicate a common source for the gabbroic and Fe-Ti oxide ore rocks with only minor or localised crustal contamination. The highly PGE-depleted and elevated Cu/Pd nature of the Panzhihua rocks suggests early segregation of a sulphide liquid. Positive correlation of Cr and Ni with MgO suggest fractionation of olivine and chromite. Enriched Sr-Nd isotopic ratios indicate a proportion of assimilation of crustal material at depth, which would aid S-saturation. S-saturation was likely caused by a combination of fractional crystallisation and assimilation of crustal material. Modelled fractionation of picritic parent magma, using Pt, Pd and Cr indicate that approximately 25% fractional crystallisation of 96% olivine, 4% chromite and 0.15% sulphide is required to produce the PGE and Cr depletion observed.The Panzhihua intrusion represents a feeder/plumbing system related to ELIP volcanism. Picritic magmas fractionate at depth forming the parental basaltic magmas of the Panzhihua intrusion. Furthermore, the main Fe-Ti oxide ore layer, within the section of the intrusion, is not in situ and interpreted to be emplaced along the footwall contact as a result of gravitational instability and slumping of crystal-rich slurries. (C) 2013 Elsevier Ltd. All rights reserved.

Recent work on the Panzhihua intrusion has produced two separate models for the crystallisation of the intrusion: (1) low-Ti, high CaO and low H2O (0.5 wt.%) parent magma (equivalent to Emeishan low-Ti basalt) at FMQ; and (2) high-Ti, low CaO and higher H2O (>1.5 wt.%) parent magma (equivalent to Emeishan high-Ti basalt) at FMQ + 1.5. Modelling of these parent magma compositions produces significantly different results.

Progressive leaching of plagioclase for Sr isotopes and microdrilling for Sr and Pb isotopes from grains of plagioclase and orthopyroxene from the Critical Zone and the Lower Zone indicates that these minerals are not in isotopic equilibrium. Leaching suggests Critical Zone plagioclase either lost Rb or had a more radiogenic Sri rim relative to the core, whereas plagioclase from an Upper Zone sample is isotopically homogeneous for Sri. Microdrilling analyses of plagioclase from the Lower and Critical Zones consistently have a higher initial 87Sr/86Sr (Sri) and a less radiogenic modeled 238U/204Pb composition (μ2) than coexisting orthopyroxene. The range of calculated Sri for plagioclase and orthopyroxene is 0.70506–0.70662(34) and 0.70290–0.70654(36), respectively. The average difference in Sri between mineral pairs was 0.00095. The range of calculated μ2 for plagioclase and orthopyroxene is 9.42–10.30 (average 9.7) and 9.83–15.75 (average 10.1), respectively. The range of measured 208Pb/206Pb for plagioclase and orthopyroxene is 34.757–36.439(33) and 36.669–41.845(85), respectively. One orthopyroxenite without evidence for more than one population of crystal size distribution, nonetheless had Sri = 0.70654 (36) with calculated μ2 of 10.32 for larger grains as compared with Sri = 0.70290 (32) and calculated μ2 of 9.97 for smaller grain-size fractions. Isotopic results from this study demonstrate that whole-rock isotopic data may not provide the appropriate level of detail necessary to address some processes in the Bushveld Complex. However, systematic changes have the potential to elucidate the timing of contamination with regard to other processes (crystal aging, compaction-driven recrystallization, and mineral exsolution) occurring within a slowly cooled crystal–liquid–vapor mush system.