Reduced sapphirine granulite xenoliths from the Lace Kimberlite,South Africa; implications for the deep structure of the Kaapvaal Craton

Sapphirine-bearing granulites, together with sapphirine-free granulites and eclogites, occur as xenoliths in the kimberlite of the Lace diatreme that penetrates the Kaapvaal craton, S. Africa. Absence of (calculated) Fe³⁺ in sapphirine, garnet and sillimanite, together with presence of graphite and sulphides, suggests highly reducing conditions of metamorphism. Chemical considerations and comparisons with experimental investigations suggest metamorphism of a sedimentary (?chlorite-montmorillonite) protolith at 900–1000° C and > 10 Kb; high Cr in the assemblage may point to a basic/ultrabasic precursor. The xenoliths indicate the presence of a very-high-grade granulite terrain, possibly similar to Enderby Land (Antarctica), beneath the Kaapval craton.     

Stable isotope study of the Archaean rocks of the Vredefort impact structure,central Kaapvaal Craton,South Africa

The Vredefort dome in the Kaapvaal Craton was formed as a result of the impact of a large meteorite at 2.02 Ga. The central core of Archaean granitic basement rocks is surrounded by a collar of uplifted and overturned strata of the Witwatersrand Supergroup, exposing a substantial depth section of the Archaean crust. Orthogneisses of the core show little variation in whole-rock δ ¹⁸O value, with the majority being between 8 and 10‰, with a mean of 9.2‰ (n = 35). Quartz and feldspar have per mil differences that are consistent with O-isotope equilibrium at high temperatures, suggesting minimal interaction with fluids during subsequent cooling. These data refute previous suggestions that the Outer Granite Gneiss (OGG) and Inlandsee Leucogranofels (ILG) of the core represent middle and lower crust, respectively. Granulite-facies greenstone remnants from the ILG have δ ¹⁸O values that are on average 1.5‰ higher than the ILG host rocks and are unlikely, therefore, to represent the residuum from the partial melting event that formed the host rock. Witwatersrand Supergroup sedimentary rocks of the collar, which were metamorphosed at greenschist-to amphibolite-facies conditions, generally have lower δ ¹⁸O values than the core rocks with a mean value for metapelites of 7.7‰ (n = 45). Overall, through an ∼20 km thick section of crust, there is a general increase in whole-rock δ ¹⁸O value with increasing depth. This is the reverse of what is normal in the crust, largely because the collar rocks have δ ¹⁸O values that are unusually low in comparison with metamorphosed sedimentary rocks worldwide. The collar rocks have δD values ranging from −35 to −115‰ (average −62‰, n = 29), which are consistent with interaction with water of meteoric origin, having a δD of about −25 to −45‰. We suggest that fluid movement through the collar rocks was enhanced by impact-induced secondary permeability in the dome structure. Electronic supplementary material The online version of this article (doi:) contains supplementary material, which is available to authorized users.     

Eclogite and pyroxenite xenoliths from off-craton kimberlites near the Kaapvaal Craton, South Africa

Mantle xenoliths brought to the surface by kimberlite magmas along the south-western margin of the Kaapvaal craton in South Africa can be subdivided into eclogites sensu stricto, kyanite eclogites and orthopyroxene eclogites, all containing omphacite, and garnet clinopyroxenites and garnet websterites characterised by diopside. Texturally, chemically (major elements) and thermally, we observe an evolution from garnet websterites (T_EG = 742–781 °C) towards garnet clinopyroxenites (T_EG = 715–830 °C) and to eclogites (T_EG = 707–1056 °C, mean value of 913 °C). Pressures calculated for orthopyroxene-bearing samples suggest upper mantle conditions of equilibration (P = 16–33 kb for the garnet websterites, 18 kb for a garnet clinopyroxenite and 23 kb for an opx-bearing eclogite). The overall geochemical similarity between the two groups of xenoliths (omphacite-bearing and diopside-bearing) as well as the similar trace element patterns of clinopyroxenes and garnet suggest a common origin for these rocks. Recently acquired oxygen isotope data on garnet (δ¹⁸O_gnt = 5.25–6.78 ‰ for eclogites, δ¹⁸O_gnt = 5.24–7.03 ‰ for garnet clinopyroxenites) yield values ranging from typical mantle values to other interpreted as resulting from low-temperature alteration or precursors sea-floor basalts and associated rocks. These rocks could then represent former magmatic oceanic rocks that crystallised from a same parental magma as plagioclase free diopside-bearing and plagioclase-bearing crustal rocks. During subduction, these oceanic rock protoliths equilibrated at mantle depth, with the plagioclase-bearing rocks converting to omphacite and garnet-bearing lithologies (eclogites sensu largo), whereas the plagioclase-free diopside-bearing rocks converted to diopside and garnet-bearing lithologies (garnet websterites and garnet clinopyroxenites).     

Single zircon ages for felsic to intermediate rocks from the Pietersburg and Giyani greenstone belts and bordering granitoid orthogneisses, northern Kaapvaal Craton, South Africa

Previous models for the temporal evolution of greenstone belts and surrounding granitoid gneisses in the northern Kaapvaal Craton can be revised on the basis of new single zircon ages, obtained by conventional U---Pb dating and Pb---Pb evaporation. In the Pietersburg greenstone belt, zircons from a metaquartz porphyry of the Ysterberg Formation yielded an age of 2949.7±0.2 Ma, while a granite intruding the greenstones, and deformed together with them, has an age of 2853 + 19/−18 Ma. These data show felsic volcanism in this belt to have been coeval with felsic volcanism in the Murchison belt farther east, and the date of 2853 Ma provides an older age limit for deformation in the region. In contrast, a meta-andesite of the Giyani greenstone belt has a zircon age of 3203.3±0.2 Ma, while a younger and cross-cutting feldspar porphyry has an emplacement age of 2874.1±0.2 Ma. The meta-andesite is intercalated with various mafic and ultramafic rocks and, therefore, the age of 3.2 Ga appears plausible for the bulk of the Giyani greenstones.Granitoid gneisses surrounding the Pietersburg and Giyani belts vary in composition from tonalite to granite and texturally from well-layered to homogeneous but strongly foliated. These rocks yielded zircon ages between 2811 and 3283 Ma. The pre-3.2 Ga gneisses are polydeformed and may have constituted a basement to the Giyani greenstone sequence, while the younger gneisses are intrusive into the older gneiss assemblage and/or into the greenstones. The Giyani and Pietersburg belts probably define two separate crustal entities that were originally close together but were later displaced by strike-slip movement.     

Comments on the Eu and Th geochemistry of the archaean continental crust of the Kaapvaal Craton,South Africa

Chemical data from sedimentary and igneous rocks of the Kaapvaal Craton have been used to trace the occurrence of highly fractionated K₂O-rich granites with pronounced negative Eu-anomalies back to the early Archaean. From the sediments of the Fig Tree Group (about 3.4 Ga-old) and granite pebbles in the overlying Moodies Group it was estimated that the igneous and metamorphic part of the continental crust at that time already contained up to about 35% of such granites. Their formation requires the presence of a thick continental crust to allow intracrustal partial melting with plagioclase as a residual phase, and the intrusion of the resulting granites at high levels within the crust. Save for some small remnants these Eu-depleted Archaean granites have since been removed by erosion.Shales of the 3.0 Ga-old Pongola Supergroup reflect only small source areas and contain considerable amounts of material derived from residual soils. The Eu-anomalies of the Pongola shales suggest up to 30% Eu-depleted granites in the source area. The shales of the Witwatersrand Supergroup (about 2.6 Ga-old) — with Eu-anomalies as low as 0.45 — reflect a source area already having typical post-Archaean crustal compositions.The Th-concentrations in South African shales support this conclusion and typical post-Archaean concentrations of above 10 ppm are found already in the early Archaean Sheba shales. Uranium and thorium mineralization has also occurred on the Kaapvaal Craton since the early Archaean.The South African late Proterozoic and Phanerozoic sedimentary record of Eu-anomalies and Th-concentrations confirms the trend observed on other cratonic areas and supports the concept of worldwide uniformity in crustal composition since the end of the Archaean. On the Kaapvaal Craton crustal growth and consolidation began 500–600 Ma earlier than in other parts of the world. The Kaapvaal Craton thus appears to represent a ‘cold spot’ on the face of the Earth.     

Emplacement,crystallization and alteration of spinifex-textured komatiitic basalt flows in the Archaean Nondweni greenstone belt,southern Kaapvaal Craton,South Africa

A well exposed succession of spinifex-textured komatiite flows is reported from the Archaean Nondweni greenstone belt located near the southern margin of the Kaapvaal Craton. The flows are relatively thin (1–5 m) compared to similar occurrences in other greenstone belts. They are characterised by well developed cone structures of highly elongate amphibole crystals (after augite) which fan downwards from the tops of the flows. Extreme development of coned spinifex has not been reported from other greenstone belts and points to specific thermal conditions prevailing in the Nondweni environment. The zones of bladed spinifex are contained between layers of random spinifex and overlie a lower cumulus layer originally of augite, orthopyroxene and minor olivine. The observed major and trace element distributions through a 1.7 m thick spinifex-textured flow are consistent with a model involving concentration of phenocryst phases resulting in significant fractionation upwards in the flow. Approximately 40% of the spinifex-textured phenocrysts grew in situ after the lithological units were established. Collapse and displacement of the coned crystal networks, originally attached to the top of the flow, are shown to have influenced the distribution of liquid within the flow and accentuated the fractionation. Associated with the spinifex-textured units are massive aphyric and brecciated flows which show distinct chemical cycles through the succession. The brecciated zones have compositions with <18% MgO and are characterised by ovoid bodies that are not pillows and may represent magmatic reworking and movement of a partly congealed flow. Post-solidus alteration is considered to have caused early hydration of the original mineralogy and also introduced SiO₂ and Na₂O into the upper part of the flow by way of microfractures. The observed alteration is different to that of Mid-Ocean Ridge basalts, and a subaerial/shallow water environment is suggested.     

Sedimentology of the Kuruman and Griquatown Iron-formations,Transvaal Supergroup,Griqualand West,South Africa

Iron-formation was deposited as a distal facies of ferruginous carbonate turbidites in an open shelf environment in front of a shallow-water carbonate platform at the time of deposition of the Campbellrand carbonate sequence. A subsequent transgression resulted in the deposition of open shelf iron-formation on top of the Campbellrand carbonate platform. Progradational sedimentation coupled with shoaling followed, and an iron-formation sequence represented by the Kuruman and Griquatown Iron-formations was deposited. This sequence consists from the base upwards of stacked open-shelf cycles of altered volcanic ash stilpnomelane lutite beds and autochthonous banded ferhythmite units; toe-of-slope greenalite—siderite rhythmites; platform slope greenalite—siderite rhythmites with grainflow bands; platform edge sideritic orthochemical and allochemical iron-formations; epeiric sea orthochemical and allochemical sideritic, hematitic and greenalitic iron-formations; supratidal disclutites and lacustrine banded greenalite lutite. Landwards, the lacustrine felutites were followed by deltaic chloritic claystone, siltstone and quartz wacke of the Koegas Subgroup.Autochthonous ferhythmites of the Kuruman Iron-formation reach a maximum development in a basin near Prieska, whereas orthochemical and allochemical units are more abundant on the Kaapvaal craton. Iron mineral and chert microbanding in the ferhythmites is attributed to seasonal changes in Eh and pH in the depository and may be related to biological activity. Chert mesobanding in the iron-formations is essentially of an early diagenetic origin.     

Garnet Lherzolites from the Kaapvaal Craton (South Africa): Trace Element Evidence for a Metasomatic History

Kimberlites from the Kaapvaal craton have sampled numerous mantlegarnet lherzolites in addition to garnet harzburgites. Traceelement characteristics of constituent clinopyroxenes allowtwo groups of garnet lherzolites to be distinguished. Traceelement compositions of all clinopyroxenes are characterizedby enrichment in light rare earth elements (LREE) and largeion lithophile elements and by a relative depletion in Ti, Nb,Ta, and to a lesser extent Zr and Hf. However, the LREE enrichmentand the depletion in Nb and Zr (Hf) are less in the Type 1 clinopyroxenesthan in the Type 2 clinopyroxenes. Our study suggests that thetwo melts responsible for the metasomatic imprints observedin the two garnet lherzolite groups are highly alkaline maficsilicate melts. Type 1 clinopyroxenes that have trace elementsimilarities to those of PIC (Phlogopite–Ilmenite–Clinopyroxene)rocks appear to have crystallized from, or been completely equilibratedwith, the same melt related to Group I kimberlite magma. TheType 2 clinopyroxenes have trace element similarities to thoseof MARID (Mica– Amphibole–Rutile–Ilmenite–Diopside)rocks and are therefore probably linked to melt related to GroupII kimberlite magma. KEY WORDS: garnet lherzolites; Kaapvaal craton; mantle xenoliths; mantle metasomatism; trace elements     

The crustal architecture of the Kaapvaal crustal block South Africa,between 3.5 and 2.0 Ga

Following terrane amalgamation of early oceanic lithosphere, the southern and central parts of the Kaapvaal Craton were a coherent unit by 3.1 Ga. Juxta-position of the northern and western granitoid-greenstone terranes including the Murchison Island Arc was the result of terrane accretion that started at 3.1 Ga. The culmination of these events was the collision of the Kaapvaal Craton, the pre-cratonic Zimbabwe block and the Central Zone to generate the Limpopo granulite gneiss terrane. Coeval with these orogenic events the central Kaapvaal Craton underwent extension to accommodate the development of the Dominion, Witwatersrand/Pongola and Ventersdorp basins. The craton scale Thabazimbi-Murchison Lineament development during the 3.1 Ga accretion event and continued to influence the tectonic evolution of the Kaapvaal block throughout the period under review as indicated by the syn-sedimentary tectonics of the > 2.64 Ga Wolkberg Group, overlying Black Reef Formation and the Transvaal Sequence. The Transvaal and Griqualand West basins developed in the Late Archaean (> 2.55 Ga) with basin dynamics influenced by far field stresses related to the Limpopo Orogeny. During this period the Thabazimbi-Murchison Lineament lay close to the northern margin of the depository. Reactivation of the Lineament between 2.4 and 2.2 Ga resulted in inversion of the Transvaal Basin and formation of the northward verging Mhlapitsi fold and thrust belt. The half-graben setting envisaged for the deposition of the Pretoria Group was influenced by the Thabazimbi-Murchison Lineament as was the emplacement and subsequent deformation of the Bushveld Complex.     

The timing and isotopic character of regional hydrothermal alteration and associated epigenetic mineralization in the western sector of the Kaapvaal Craton (South Africa)

The Kaapvaal Craton of South Africa comprises an Archaean core of ≈3.5 Ga lithospheric and crustal rocks surrounded by younger accreted terrains of ≈3.0–2.7 and ≈2.1–1.9 Ga. The craton is covered by relatively undeformed 3.0–2.4 Ga supracrustal rocks, which show the effects of thermal and hydrothermal interaction. Part of this activity is manifested by a large number of epigenetic Pb–Zn (±Ag, Au, Cu, F) deposits in the cover rocks of the Kaapvaal Craton. These include small volcanic and breccia hosted deposits in mafic and felsic volcanic rocks of the 2.7 Ga Ventersdorp Supergroup and the Mississippi Valley-type (MVT) deposits in the carbonates of the Transvaal Supergroup.MVT mineralization at the Pering (and other Zn–Pb deposits) is hosted in fracture-generated N–S breccia bodies in the Paleoproterozoic carbonate succession of the western Kaapvaal Craton. The fluids carrying the metals were focused in vertical bodies within the fracture zones (FZ), the metals and the sulphur being carried together and precipitated in organic-rich sectors of the basin. Two small Pb–Zn deposits within mafic rocks of the Ventersdorp Supergroup, stratigraphically below the basin-hosted MVTs on the southwestern part of the Kaapvaal Craton have secondary chlorite which is extremely Rb-rich, associated with the mineralization. This chlorite and the associated altered basaltic host rocks give a Rb–Sr date of ≈1.98 Ga, and the associated galena Pb isotope data plot on the same array as those of other Pb–Zn deposits, the radiogenic intercept giving a date of ≈2.0 Ga. We interpret these data to indicate a craton-wide epigenetic fluid-infiltration event, which exploited the Maquassie Quartz Porphyry (MQP) as the aquifer and metal source.Sr isotopic results for the ore-zone gangue minerals show highly radiogenic ⁸⁷Sr/⁸⁶Sr ratios (>0.710) which support earlier models that the origin of radiogenic Sr isotopic composition in the calcite cements is the felsic tuffs (MQP) of the Ventersdorp Supergroup occurring at deeper levels within the basin. Relationships between δ¹⁸O and δ¹³C performed on carbonate cements within the aquifers are complex: the range in δ¹³C for some of the cements represents a mixture from two sources and with a progression from heavy carbon in the host to somewhat lighter carbon in the cements. Similarly, the lighter δ¹⁸O values have a narrow range indicative of rapid exchanges between hydrous fluid and rock.     

The economic mineral potential of the mid-Proterozoic Waterberg Group, northwestern Kaapvaal craton, South Africa

The 1900–1700 Ma Waterberg Group in the main Waterberg fault-bounded basin consists of dominantly coarse siliciclastic red beds with minor volcanic rocks. The sedimentary rocks were deposited mainly by alluvial fans, fluvial braidplains and transgressive shallow marine environments, with lesser lacustrine and aeolian settings. Uplifted, largely granitic source areas were located along the Thabazimbi-Murchison lineament (TML) fault system in the south, and along the Palala shear zone in the northeast. Palaeoplacer titanomagnetite-ilmenite-zircon heavy mineral deposits, best developed in the Cleremont Formation in the centre of the basin, reflect initial fluvial reworking and subsequent littoral marine concentration. Coarse alluvial cassiterite placer deposits are found in the Gatkop area in the southwest of the basin, and appear to have been derived from stanniferous Bushveld Complex lithologies south of the TML. Hydrothermal zinc and U-Cu mineralisation in the Alma lithologies in the same area appears to be related to the TML fault system. Small manganese deposits and anomalous tungsten values occur in the south of the basin, where they are again closely spatially associated with the TML. Copper-barium mineralisation is found associated with dolerite dykes, and in stratigraphically controlled, inferred syngenetic settings. The most interesting of these apparently syngenetic occurrences is found within green coloured reduced mudrocks and inferred volcanic rocks, at an unconformity developed within the overall red bed sequence of the Waterberg Group, adjacent to the TML in the southwest of the basin. The most important potential mineralisation in the main Waterberg basin thus encompasses shoreline placer Ti and the possibility of substantial sediment-hosted copper deposits. Received: 31 May 1996 / Accepted: 17 February 1997     

The copper-bearing basic rocks of Namaqualand,South Africa

Copper mineralisation is associated with rocks of basic/intermediate composition which occur as a widespread swarm of relatively small irregularly-shaped bodies within the gneissic terrain of the Okiep District. These basic rocks comprise the so-called Koperberg noritoid suite of Namaqualand and constitute one of the three major copper producers in South Africa. While a metasomatic origin was briefly favoured for the noritoids, most researchers have emphasized the cross-cutting nature of the noritoids relative to their host rocks and prefer an origin involving magmatic injection. High noritoid ⁸⁷Sr/⁸⁶Sr initial ratio values indicated to some workers that mobilisation of suitable crustal materials was more attractive than intrusion of mantle-derived melts to account for the origin of the noritoid suite. The present study has shown that the structure, texture and mineralogy of the noritoids do not lend themselves to interpretation of the noritoids as pristine intrusive rocks, and indicate that the noritoids have undergone high-grade metamorphism after emplacement. The geochemistry of the noritoid suite reveals a pronounced alkaline affinity and mineralogical control of the chemistry characterises the rocks as cumulates. The dimensions of the noritoids appear unfavourable for development of cumulate rocks and it is considered possible that the present configuration of the noritoids results from dismemberment of larger intrusive bodies as a result of shear deformation during high-grade metamorphism. It has not been possible to identify a magma parent to the noritoid suite, but we believe the parent to have been an alkaline mantle-derived magma which suffered contamination by granitic anatects in the lower crust. Recognition of the metamorphic history of the Namaqualand noritoids renders a comparison between them and the Caraiba noritoids of interest.     

Kimberlites of the Man craton, West Africa

The Man craton in West Africa is an Archaean craton formerly joined to the Guyana craton (South America) that was rifted apart in the Mesozoic. Kimberlites of the Man craton include three Jurassic-aged clusters in Guinea, two Jurassic-aged clusters in Sierra Leone, and in Liberia two clusters of unknown age and one Neoproterozoic cluster recently dated at 800 Ma.

All of the kimberlites irrespective of age occur as small pipes and prolific dykes. Some of the Banankoro cluster pipes in Guinea, the Koidu pipes in Sierra Leone and small pipes in the Weasua cluster in Liberia contain hypabyssal-facies kimberlite and remnants of the so-called transitional-facies and diatreme-facies kimberlite. Most of the Man craton kimberlites are mineralogically classified as phlogopite kimberlites, although potassium contents are relatively low. They are chemically similar to mica-poor Group 1A Southern African examples.

The Jurassic kimberlites are considered to represent one province of kimberlites that track from older bodies in Guinea (Droujba 153 Ma) to progressively younger kimberlites in Sierra Leone (Koidu, 146 Ma and Tongo, 140 Ma). The scarcity of diatreme-facies kimberlites relative to hypabyssal-facies kimberlites and the presence of the so-called transitional-facies indicate that the pipes have been eroded down to the interface between the root and diatreme zones. From this observation, it is concluded that extensive erosion (1–2 km) has occurred since the Jurassic. In addition to erosion, the presence of abundant early crystallizing phlogopite is considered to have had an effect on the relatively small sizes of the Man craton kimberlites.     

Paleoproterozoic Kimberlites of the Lake Kimozero Area,Karelian Craton: Ore Mineralization in Kimberlites and Fault Zones

Syngenetic and epigenetic ore mineralization was studied in Paleoproterozoic metakimberlites in the area of Kimozero Lake. In the Kimozero structure, redeposited ore mineralization is constrained to fracture and shear zones and consists of Fe-vaesite, Fe–Co-polydymite, millerite, Ni-pyrrhotite, pentlandite, chalcopyrite, Zn-bearing copper, galena, and Ni-pyrite. The composition of this mineralization is analogous to that of syngenetic mineralization in pyroclastic and coherent kimberlite, and its likely source was the kimberlite itself.     

The Timing of the Palaeoproterozoic Great Oxidation Event using Dykes,Sills and Bolcanics of the Ongeluk Large Igneous Province,Kaapvaal Craton

正The Neoarchaean to Palaeoproterozoic Transvaal Supergroup of the Kaapvaal Craton,southern Africa,is one of the best-preserved and most complete stratigraphic records across a critical in juncture in the Earth’s history.     

Geochemistry of intermediate to siliceous volcanic rocks of the Rooiberg Group, Bushveld Magmatic Province, South Africa

The volcanic Rooiberg Group represents the earliest phase of Bushveld-related magmatism and comprises, in some areas, the floor and roof rocks of the mafic-ultramafic intrusive units of the Bushveld Complex. The lower to middle Dullstroom Formation is composed of two interbedded series of low Ti and high Ti volcanic strata, which are predominantly basaltic andesites. Volcanic units above these strata range from andesites to dacites in the upper Dullstroom Formation and to predominantly rhyolites in the overlying Damwal and Kwaggasnek Formations. Compositional data suggest that these intermediate to siliceous volcanic rocks are petrogenetically related to the low Ti volcanic suite and suggest that the low Ti magmas resided in a shallow magma chamber where they experienced fractional crystallization and assimilation of crustal material. In contrast, the high Ti volcanic suite is petrogenetically unrelated. These data confirm previous suggestions that Bushveld-related magmas experienced significant amounts of assimilation of continental crust.     

Identification and Genesis of the Mixed-Layer Lizardite–Saponite Phase in a Kimberlite Pipe,South Africa

In autolithic breccia of a kimberlite pipe in South Africa, the ordered mixed-layer lizardite–saponite phase was first identified in rocks containing a mechanical mixture of lizardite and saponite or only saponite in association with lizardite. The mixed-layer phase consists of 52% lizardite and 48% saponite layers. The trioctahedral type of this phase indicates its formation in a closed system and reducing environment, whereas the high content of expanding layers suggests low PT parameters.     

A Chemical and Multi-Isotope Study of the Western Cape Olivine Melilitite Province, South Africa: Implications for the Sources of Kimberlites and the Origin of the HIMU Signature in Africa

We present major and trace element and Sr–Nd–Pb–Hf–Osisotopic data for the 76–58 Ma Western Cape melilititeprovince, an age-progressive magmatic lineation in which primitiveolivine melilitite intrusives and alkali basalt lavas have beenemplaced on the southwestern margin of South Africa. The magmasrange from alkali basalts with strong HIMU isotopic and traceelement affinities on the continental shelf to melilitites withkimberlite-like incompatible element compositions and EM 1 isotopicaffinities on thick Proterozoic lithosphere (i.e. ⁸⁷Sr/⁸⁶Sr_i= 0·7029–0·7043,