Mineralogical and geochemical constraints on environmental impacts from waste rock at Taojiang Mn-ore deposit, central Hunan, China

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

Author's personal copy

Mineralogical and geochemical constraints on the shalloworigin, ancient veining, and multi-stage modification

of the Lherz peridotite

Amy J.V. Riches ⇑, Nick W. Rogers

Department of Earth and Environmental Sciences, The Open University, Milton Keynes MK7 6AA, United Kingdom

Received 29 November 2010; accepted in revised form 25 July 2011; available online 31 July 2011

Abstract

New major- and trace-element data of bulk-rocks and constituent minerals, and whole-rock Re–Os isotopic compositionsof samples from the Lherz Massif, French Pyrenees, reveal complex petrological relationships between the dominant lithol-ogies of lherzolite ± olivine-websterite and harzburgite. The Lherz peridotite body contains elongate, foliation parallel, lith-ological strips of harzburgite, lherzolite, and olivine-websterite cross-cut by later veins of hornblende-bearing pyroxenites.Peridotite lithologies are markedly bimodal, with a clear compositional gap between harzburgites and lherzolites ± olivine-websterite. Bulk-rock and mineral major-element oxide (Mg–Fe–Si–Cr) compositions show that harzburgites are highly-depleted and result from �20-25 wt.% melt extraction at pressures <2 GPa. Incompatible and moderately-compatibletrace-element abundances of hornblendite-free harzburgites are analogous to some mantle-wedge peridotites. In contrast,lherzolites ± olivine-websterite overlap estimates of primitive mantle composition, yet these materials are composite samplesthat represent physical mixtures of residual lherzolites and clinopyroxene dominated cumulates equilibrated with a LREE-enriched tholeiitic melt. Trace-element compositions of harzburgite, and some lherzolite bulk-rocks and pyroxenes have beenmodified by; (1) wide–spread interaction with a low-volume LREE-enriched melt +/� fluid that has disturbed highly-incom-patible elements (e.g., LREEs, Zr) without enrichment of alkali- and Ti-contents; and (2) intrusion of relatively recent, small-volume, hornblendite-forming, basanitic melts linked to modal and cryptic metasomatism resulting in whole-rock and pyrox-ene Ti, Na and MREE enrichment.

Rhenium-Os isotope systematics of Lherz samples are also compositionally bimodal; lherzolites ± olivine-websterite havechondritc to suprachondritic 187Os/188Os and 187Re/188Os values that overlap the range reported for Earth’s primitive uppermantle, whereas harzburgites have sub-chondritic 187Os/188Os and 187Re/188Os values. Various Os-model age calculationsindicate that harzburgites, lherzolites, and olivine-websterites have been isolated from convective homogenisation since theMeso-Proterozoic and this broadly coincides with the time of melt extraction controlled by harzburgite Os-isotope composi-tions. The association between harzburgites resulting from melting in mantle-wedge environments and Os-rich trace-phases(laurite–erlichmanite sulphides and Pt–Os–Ir-alloys) suggests that a significant portion of persistent refractory anomalies inthe present-day convecting mantle of Earth may be linked to ancient large-scale melting events related to wide-spread sub-duction-zone processing.� 2011 Elsevier Ltd. All rights reserved.

1. INTRODUCTION

The origin of orogenic massifs (c.f. Den Tex, 1969; Wyl-lie, 1969), including the Lherz peridotite, has been the sub-ject of much debate and it has been suggested that theselayered assemblages of peridotite and pyroxenite may rep-

0016-7037/$ - see front matter � 2011 Elsevier Ltd. All rights reserved.

doi:10.1016/j.gca.2011.07.036

⇑ Corresponding author. Present address: Department of Earthand Atmospheric Sciences, The University of Alberta, Edmonton,Alberta, Canada T6G 2E3. Tel.: +1 780 492 2676.

E-mail address: [email protected] (A.J.V. Riches).

www.elsevier.com/locate/gca

Available online at www.sciencedirect.com

Geochimica et Cosmochimica Acta 75 (2011) 6160–6182

Author's personal copy

resent; (1) exhumed sub-continental lithospheric mantle(e.g., Menzies and Dupuy, 1991; Reisberg and Lorand,1995; Burnham et al., 1998; Downes, 2001) from themechanical boundary layer (MBL; White, 1988), or (2)inherited heterogeneities from asthenospheric upper-mantle(c.f. Peate et al., 1997; Parkinson and Pearce, 1998; Parkin-son et al., 1998; Brandon et al., 2000; Bizimis et al., 2005;Harvey et al., 2006; Muntener and Manatschal, 2006; Bizi-mis et al., 2007; Liu et al., 2008; Simon et al., 2008; Liuet al., 2009; Warren et al., 2009; Dijkstra et al., 2010; Ishik-awa et al., 2011 for examples of isotopic and mineralogicalheterogeneity present in Earth’s upper-mantle) that mayeven be linked to upwelling diapirs (e.g., Bodinier et al.,1988; Fabries et al., 1991). Several authors have used thechemistry of mantle materials to provide estimates of thecomposition of primitive upper mantle (PUM) and bulk-sil-icate Earth (BSE; e.g., Palme and Nickel, 1985; Zindler andHart, 1986; McDonough and Sun, 1995; Becker et al.,2006), and some authors have integrated information fromultramafic bodies into models of the spatial distribution ofchemical diversity in Earth’s upper mantle (e.g., Allegre andTurcotte, 1986; Morgan and Morgan, 1999).

The Re–Os isotope system is widely considered to recordhigh-temperature processes (e.g., Shirey and Walker, 1998;Carlson, 2005; Pearson et al., 2007; Carlson et al., 2008),and Re–Os isotopic compositions of peridotites have beenused to estimate the timing of major mantle differentiationevents; methods of applying this tool to obtain osmium-model ages (Alumino-chron and Sulphur-chron approaches)were founded on studies of the Lherz peridotite and otherAriege-group massifs (e.g., Reisberg and Lorand, 1995;Burnham et al., 1998). These Os-dating methods have beenapplied to a number of other massif peridotite, ophiolite,abyssal peridotite, and xenolith suites (e.g., Ronda andHoroman massifs, Reisberg et al., 1991; Saal et al., 2001;Southeast Australian Xenoliths, Handler et al., 1997; NorthChina Xenoliths, Gao et al., 2002; Liu et al., 2010, 2011;Gakkel Ridge peridotites, Liu et al., 2008) where mantledepletion ages have tentatively been linked to periods ofsignificant crustal generation and lithosphere stabilisation(c.f. review by Rudnick and Walker, 2009).

As a large, well exposed, relatively fresh expanse of peri-dotite that can be subjected to field investigations on alllength scales up to 1 km2 (e.g., Conquere and Fabries,1984; Bodinier et al., 1988; Fabries et al., 2001; Le Rouxet al., 2007) the Lherz massif has occupied a prominent po-sition in debates concerning mantle evolution. Lherz is oneof 40 ultramafic bodies that crop out in the North PyreneanMetamorphic Zone (NPMZ; Fig. 1), which marks the meet-ing of the Iberian plate to the south and European plate tothe north (e.g., Choukroune, 1992). The peridotite–pyroxe-nite assemblage of Lherz covers a wide-range of major-,minor-, and trace-element abundances (e.g., Bodinieret al., 1988; Burnham et al., 1998; Le Roux et al., 2007),and lithophile-element isotope compositions of this bodyoverlap the spectrum of 147Sm/144Nd, 87Sr/86Sr, and Pb–Pb isotopic values reported in mantle peridotites world-wide (Polve and Allegre, 1980; Downes et al., 1991; Mukasaet al., 1991; Zanetti et al., 1996; Henry et al., 1998; Le Rouxet al., 2009). Within the Lherz massif harzburgites distal

from cross-cutting veins are isotopically enriched with high87Sr/86Sr (0.70475 ± 4), high 87Rb/88Sr (�0.0075), low eNd(�+0.6), and contrast to lherzolites that commonly displaylow 87Sr/86Sr (0.70202 ± 2 to 0.70274 ± 4), low 87Rb/88Sr(0.0011–0.0042), and high eNd (+7.2 to +11.9); these isoto-pic characteristics are not the result of melt extraction pro-cesses alone (Downes et al., 1991; Le Roux et al., 2009).

Two contrasting petrogenetic models may account forthe compositional diversity of the Lherz massif; (1) harz-burgites and lherzolites may result from variable degreesof ancient melt extraction, and isotope systems based onincompatible lithophile-elements may record later eventslinked to thermal perturbation and/or incompatible-ele-ment modification by moderate amounts of exotic meltsand/or fluids (e.g., Burnham et al., 1998; Henry et al.,1998); or (2) Le Roux et al. (2007, 2008, 2009) recentlyproposed that the peridotite–pyroxenite assemblage ofLherz could reflect refertilisation of ancient depleted peri-dotites by substantial volumes of percolating basaltic melt(30–60 wt.% websterite, where these authors view webste-rites as frozen melt-fronts) during the Variscan orogeny.It is important to distinguish between these models as re-cent suggestions of Variscan-age igneous refertilisationmay cast doubt on the significance of osmium model-agedeterminations for the Lherz massif and other global peri-dotite suites. If correct, the model proposed by Le Rouxet al. (2007, 2008, 2009) has important geodynamic impli-cations if mineralogically fertile LREE-depleted lherzolitesare secondary products, and if large-volumes of basalticmelt can ascend through mantle lithosphere by wide-spread percolative flow (defined herein as igneous refertil-isation) that; (1) crystallises substantial amounts of sec-ondary clinopyroxene ± spinel ± sulphide ± amphibole;and (2) is characterised by pronounced fractionation oftrace-elements at a chromatographic melt-front (c.f. Ver-nieres et al., 1997).

In this study we report new field and petrographic char-acteristics, mineral and bulk-rock major-, minor-, andtrace-element abundances, and whole-rock Re–Os concen-trations and isotopic compositions of adjacent harzburgiteand lherzolite bodies (± olivine-websterite) at two sampletraverses within the Lherz massif. We demonstrate thatthe peridotites and (diopside-bearing) olivine-websteritesof the Lherz massif were created in low-pressure environ-ments (spinel–facies) where melt migration was dominatedby channel-flow during the Proterozoic. In addition, weshow that; (1) magnesian harzburgites with low-Ti clinopy-roxenes, analogous to some mantle-wedge peridotites, mayresult from low-pressure (<2 GPa) melting in the presenceof fluid; (2) incompatible-element abundances for this sam-ple suite are decoupled from major-element and Os-isotopesystematics, recording two distinct post-differentiationmetasomatic events linked to; (a) wide–spread interactionwith a LREE-enriched melt +/� fluid; and (b) short-length-scale modal and cryptic modification adjacent tohornblendite-veinlets. These results cast serious doubt onthe igneous refertilisation model of Le Roux et al. (2007,2008, 2009), and indicate that the oldest portions of theLherz massif represent tectonically juxtaposed lithologiesthat formed in a mantle-wedge environment.

Petrogenesis of the Lherz peridotite 6161

Author's personal copy

2. FIELD, PETROGRAPHIC AND MINERALOGICAL

CHARACTERISTICS

2.1. Field constraints

The Lherz peridotite body is a layered ultramafic sequencedominated by foliated peridotite in which the broadly NE-SW striking fabric (Fig. 1b) is defined by elongate silicategrains (olivine + pyroxene). Concordant harzburgites formsubordinate tabular features in direct contact with lherzolitelithologies ± olivine-websterite. Some lherzolites containbands of olivine-websterite (up to 10 cm wide) that parallelthe foliation plane with a broad NE–SW strike. Field rela-tionships indicate that the banded series of harzburgite,lherzolite, and concordant olivine-websterite is old relativeto volumetrically minor anhydrous layered-pyroxenite (±garnet) sequences (up to 4 m thick; Bodinier et al., 1987a)

that cross-cut the peridotite foliation at angles of�20� (Con-quere and Fabries, 1984; Fabries et al., 2001). In addition,Sautter and Fabries (1990) indicated that the peridotiteassemblage, olivine-websterite bands, and layered anhy-drous–pyroxenites are isoclinally folded within Ariege-groupperidotite bodies. Late-stage intrusive features dated at�100 Ma (e.g., Montigny et al., 1986; Henry et al., 1998) in-clude amphibole-bearing pyroxenites and hornblendite-veinsthat cross-cut the dominant foliation within spinel perido-tites at angles of�30� (Conquere and Fabries, 1984; Fabrieset al., 2001). More recent studies by Le Roux et al. (2007,2008, 2009) suggested an alternative view in which the con-tacts between harzburgites and lherzolites are convolutedand the foliation within harzburgites is over-printed by latermineral growth (secondary pyroxene ± spinel ± amphi-bole ± sulphide) linked to the generation of secondary lherz-olites during a pervasive igneous refertilisation process.

Fig. 1. (A) Geological map of the Eastern Pyrenees, with the location of the study area shown in the inset where the grey-shaded areacorresponds to the Alpine mountain chain (after Choukroune, 1992). The Lherz peridotite is associated with metamorphic rocks of the NorthPyrenean Zone (NPZ). NPFT = North Pyrenean Frontal Thrust. (B) A detailed geological map of the Lherz Massif after Fabries et al. (2001)and Le Roux et al. (2007).

6162 A.J.V. Riches, N.W. Rogers / Geochimica et Cosmochimica Acta 75 (2011) 6160–6182

Author's personal copy

In this study, two sample traverses were undertakenacross contacts between harzburgite and lherzolite bodiesthat contain contrasting abundances of pyroxenite. Site Acontains a high pyroxenite abundance evident in the formof numerous olivine-websterite bands (1.5–8 cm wide) with-in the lherzolite body, and this traverse coincides with site 4of Lorand et al. (2010). Site B, which corresponds to tra-verse 2 of Le Roux et al. (2007, 2008, 2009) and site 2 ofLorand et al. (2010), has a low pyroxenite abundance withno clear exposure of olivine-websterite bands (where bandsare continuous features distinct from websteritic pods<30 cm in length) in the lherzolite body over the samplingscale of this work. In contrast to Le Roux et al. (2007,2008, 2009) and Lorand et al. (2010) we observed hornblen-dite–veinlets (<2 cm in width, Fig. A1 of Appendix A) inseveral samples proximal to the harzburgite–lherzolite con-tact at site B. Le Roux et al. (2007) suggested that harzburg-ites record a foliation of N40–60�E that is over-printed by alater and more steeply dipping foliation in lherzolites. Fieldobservations confirm that the foliation of harzburgites andlherzolites proximal to the compositional boundary at siteB is weak, and that the foliation deviates from the N40–60�E orientation that is dominant across the massif. SiteA is structurally distinct from site B (Fig. A2, AppendixA). At site A, the foliation within the harzburgite andbanded-lherzolite body is coincident and conforms to theN40–60�E plane observed at many outcrops across thismassif. Olivine-websterite bands within the lherzolite bodyat site A are generally concordant with the foliation definedby elongate silicates of the adjacent peridotite. In addition,the contact between harzburgite and lherzolite at sites Aand B is sharp (at the 10 cm and thin section scale), andthese observations are consistent with previous field studiesand structural maps of Lherz in which harzburgite andlherzolite lithologies represent elongate lithological strips(e.g., Conquere and Fabries, 1984; Fabries et al, 2001 andreferences therein) juxtaposed during plastic deformation(e.g., Fig. 2 of Sautter and Fabries, 1990).

2.2. Petrographic features

Samples collected during this study are up to20 � 30 � 20 cm in size and are heterogeneous at thehand-specimen scale. All studied samples are coarse-grained peridotites and olivine-websterites (silicates are typ-ically 5–10 mm at the long axis), dominated by olivine,orthopyroxene, clinopyroxene, spinel, minor amphibole,and minor sulphide in an interlocking crystalline matrixwith a porphyroclastic texture (porphyroclasts are up to25 mm at their long axis). Many samples from site A andseveral samples from site B are hybrid formations of peri-dotite + olivine-websterite (bands and pods), and thesecomposite samples are distinct from relatively homoge-neous, mineralogically fertile, peridotites that have beenincorporated into previous studies that addressed the com-position of Earth’s primitive mantle (e.g., McDonough andSun, 1995). Olivine, orthopyroxene, and clinopyroxenecrystals are elongate and show a preferred alignment, defin-ing a clear foliation within the samples. Isolated areas (gen-erally <3 mm wide) of later crystallised

pyroxene ± spinel ± sulphide ± minor amphibole are volu-metrically minor (generally <5 vol.%, particularly in sam-ples distal from cross-cutting hornblendite veinlets).Similar features have been identified in previous petro-graphic and mineralogical studies of the Lherz peridotites(e.g., Lorand, 1989, 1991; Woodland, 1992; Woodlandet al., 1996). Minor trails (<0.5 mm wide) of spinel parallelthe penetrative foliation in some places, but are often in-clined to it, consistent with observations by Woodland(1992), and Woodland et al. (1996). All studied specimenshave experienced relatively low degrees of serpentinisation

Fig. 2. Bulk-rock major-element compositions. Literature data formassif peridotites of the Lherz ultramafic body are from Bodinieret al. (1988) and Burnham et al. (1998). Estimates of primitivemantle compositions shown here were reported by McDonoughand Sun (1995). Hb-hazburgite = hornblendite-bearing harzburg-ite, Hb-Lherzolite = hornblendite-bearing lherzolite, Ol-Webste-rite = olivine-websterite, and C-Lherzolite = composite-lherzolite,the classification scheme is described in the main text.

Petrogenesis of the Lherz peridotite 6163

Author's personal copy

(610 vol.%). Samples are classified on the basis of modalcontent following the IUGS scheme of Streckeisen (1973)and are reported in Appendices A and B.

Harzburgites contain <5 wt.% clinopyroxene, 10–15 wt.% orthopyroxene, up to 75 wt.% olivine, <5 wt.%chromite, minor sulphide, Fe-hydroxide, and calcite. Inaddition, a few samples contain minor amounts of intersti-tial amphibole (generally60.5 mm in maximum dimension).Harzburgites often have a fragmentary/mosaic texture ofheavily-cracked olivine crystals with sub-grain boundariesinclined to the direction of elongation. Minor amounts ofserpentine form along cracks through olivine and at somesub-grain boundaries. Many harzburgites contain a fine-grained groundmass (<0.5 mm) in which heterogeneousshearing is evident, and flow-lines defined by silicate neo-blasts (<0.1 mm, commonly olivine) are present around anumber of silicate porphyroclasts. Silicates (olivine, clino-,and orthopyroxene), particularly olivine, predominantlydisplay undulose extinction and this is evident in porphyro-clastic and matrix grains. In places, orthopyroxene and oliv-ine crystals cross-cut the harzburgite foliation, indicating alater crystallisation phase after the older foliation-formingevent. Harzburgites contain isolated clusters of pyrox-ene + Al-chromite (up to 5 � 5 cm, many of which are sym-plectic), and lesser amounts of Al-chromite forming thintrails (<1 cm wide) generally no more than 2 cm in length.A limited number of harzburgites from site B contain nar-row (0.5–1.5 cm) veinlets of hornblendite (kaersutiticamphibole ± phlogopite ± pyroxene), which are associatedwith modal metasomatism characterised by kaersutite for-mation within the adjacent peridotite matrix (AppendixA). Samples containing visible hornblendite–veinlets areclassified as hornblendite–harzburgites.

Lherzolites contain >5 wt.% clinopyroxene, 10–25 wt.%orthopyroxene, <65 wt.% olivine, minor amounts of Al-chromite (generally <3 wt.%), interstitial amphibole(<0.5 mm), sulphide, calcite, and Fe-hydroxide. Thesecoarse-grained peridotites are often heterogeneous (on 10 scm-scales) with localised areas of low pyroxene abundancecreating portions that are almost harzburgitic (<5 cm diam-eter); the foliation in these domains is parallel to that ofadjacent lherzolitic areas. A limited number of polished sec-tions contain isolated regions (<3 mm wide) where silicatealignment is not well developed, and the dominant foliationis cross-cut in places by later silicate grains that approach anequant character. Lherzolite samples from site A often con-tain bands of olivine-websterite (>60 wt.% clino- and ortho-pyroxene,�3 to 8 wt.% chromite, minor interstitial sulphideand amphibole) that are generally 1.5–4 cm in width, andreach a maximum thickness of 8 cm. These hand-specimenshave a penetrative foliation parallel to the orientation ofolivine-websterite bands, and contain silicates with unduloseextinction. Regions adjacent to olivine-websterite bands arelherzolitic with a gradational edge (<2 cm) displaying amodest increase in clino- and orthopyroxene mode(�5 wt.%), and concomitant decrease in olivine abundanceadjacent to olivine-websterites. Samples containing broadregions (8–20 cm wide) of lherzolite and clearly-definedbands of olivine-websterite are classified as composite-lherz-olites (Appendices A and B).

2.3. Mineralogical characteristics

Major- and minor-element compositions of constituentminerals are reported with trace-element abundances of cli-no- and orthopyroxenes in Appendix B. Porphyroclasticand matrix phases commonly cover a limited range of ma-jor-element contents, and some crystals <0.5 mm in diame-ter display irregular compositional zoning. In general,constituent phases of peridotites and olivine-websteriteswithin the Lherz massif have Mg# (where Mg# = 100Mg/[Mg + Fetotal]) values of the order spinel� olivine <orthopyroxene < clinopyroxene.

2.3.1. Olivine and spinel compositions

Peridotites and olivine-websterites of the Lherz massifcontain forsteritic olivine, with the most magnesian compo-sitions present in harzburgites (Mg# = 91–92, andMg# = 90–91 at sites A and B, respectively), and olivineswith lower Mg# (89–90) present in lherzolites, composite-lherzolites, and olivine-websterites. Harzburgites associatedwith hornblendite-veinlets contain olivine with Mg# values(�91) similar to those of hornblendite-free harzburgites.The Mg# of olivines generally corresponds to the bulk-rockcomposition, and the variation of olivine mode with Mg#in several peridotites of the Lherz massif overlaps the oce-anic trend of Boyd (1989); (Fig. A3 of Appendix A). In de-tail, several harzburgite samples from site A and B containlower modal abundances of olivine and greater concentra-tions of orthopyroxene than predicted by the oceanic trend(Fig. A3, Appendix A), and only one of these samples con-tains visible hornblendite. Nickel contents of olivines of siteA and B generally range from 0.38 to 0.46 wt.% NiO, andextend to lower values (reaching 0.26 wt.%) in composite-lherzolites of site A, with a maximum NiO content(0.54 wt.%) in a lherzolite spatially associated with hornb-lendite at site B (04LH04). This range of Mg# and Ni con-tents is similar to those reported for in olivine-websteritesand peridotites of abyssal peridotites, forearc peridotites,ophiolite sequences, and cratonic xenoliths (e.g., Kelemenet al., 1998; O’Hara and Ishii, 1998; Suhr et al., 2003; Dicket al., 2010; Warren and Shimizu, 2010).

Several types of spinel were identified on the basis oftheir textural locations and major-element compositionsby Woodland et al. (1996), and definitions suggested bythese authors have been adopted and expanded in thiswork. We recognise four spinel-types: (1) P-type spinels(<1 mm) are common, occur interstitially, and are generallyanhedral with dark-brown to mid-brown body colour, (2)S-type spinels have similar crystal form and textural associ-ations to P-type spinels, but are less common and havevarying body colour that is olive-green and red-brown insome cases, (3) rare G-type spinels (<0.5 mm) occur in alimited number of harzburgite samples, have olive-greenbody colour, and often form as isolated crystals at oliv-ine-olivine contacts or adjacent to elongate orthopyroxene,and (4) C-type spinels (0.5–1 mm) occur with clino- andorthopyroxene in symplectic clusters and have irregularlyvarying body colour ranging from olive-green to dark-brown. In general, spinels are of chromite to Al-chromitecomposition and have major-element characteristics similar

6164 A.J.V. Riches, N.W. Rogers / Geochimica et Cosmochimica Acta 75 (2011) 6160–6182

Author's personal copy

to those reported for unmetasomatised abyssal spinel-peri-dotites and orogenic massifs (e.g., Type-1 peridotites ofDick and Bullen, 1984, and spinels of abyssal peridotites re-viewed by Barnes and Roeder, 2001), which overlap therange of spinel compositions suggested for four-phase spi-nel-facies melting residues (olivine + orthopyroxene + spi-nel + clinopyroxene; Dick and Fisher, 1983; Pearce et al.,2000). In detail, spinels of site B harzburgites tend to havelower Cr# (where Cr# = 100Cr/[Cr + Al]) at a given Mg#compared to spinels of site A harzburgites (Fig. A3 ofAppendix A). Spinels of site A harzburgites overlap therange of compositions previously reported in massif perido-tites of the Ariege-group ultramafic bodies. In contrast, spi-nels of site B harzburgites overlap the range of major-element compositions reported for Ariege-group peridotitesthat crop out adjacent to amphibole-pyroxenites (Fig. A3,Appendix A). The olivine and spinel compositions of thestudied samples overlap the range reported for abyssal per-idotites (e.g., Dick and Bullen, 1984), and differ from supra-subduction zone (SSZ) peridotites associated with boninitemagmatism that contain spinels with Cr# >60 (e.g., Parkin-son and Pearce, 1998; Choi et al., 2008).

2.3.2. Pyroxene and amphibole compositions

The studied samples contain diopside and enstatite cover-ing a range of jadeite contents (<1 to 14.3 mol.%). Clinopy-roxene compositions can be divided into three groups onthe basis of their major-element contents. Two clinopyroxenegroups dominate the studied samples and these fall on trend 1of Supplementary Fig. A4 (Appendix A); hornblendite-freeharzburgites contain group-1 clinopyroxenes with the lowestNa2O and highest Al2O3 contents of all studied diopsides,and group-2 pyroxenes hosted by hornblendite-free lherzo-lites and olivine-websterites have moderate Na2O andAl2O3 contents. Lesser amounts of Group-3 clinopyroxenesare present in the studied samples, and these are found inhornblendite-bearing harzburgites and hornblendite-freeharzburgites that crop out proximal to the compositionalboundary at sites A and B; group-3 diopsides have higherNa2O contents at equivalent Al2O3 abundances (trend 2 ofFig. A4, Appendix A) compared to clinopyroxene groups 1and 2. Pyroxenes with relatively high Na2O abundances alsohave high Cr2O3 and lower CaO contents at a given Al2O3.Group-3 clinopyroxenes from each sample site differ whenTi-abundances are compared (Appendix B). Group-3 clino-pyroxenes of site B, often associated with visible hornblen-dite, contain up to 1.2 wt.% TiO2, and are analogous toclinopyroxenes previously reported for Lherz peridotitesthat; (1) crop out adjacent to amphibole-pyroxenites (Wood-land et al., 1996); and (2) have been linked to the refertilisa-tion trend defined by Le Roux et al. (2007). Group-3clinopyroxenes of site A (samples 04LH34A and 04LH35)have much lower Ti-contents (<150 ppm, Appendix B) thanabyssal peridotite clinopyroxenes (Ti >240 ppm; Dick et al.,2010; Warren and Shimizu, 2010) and are analogous to low-Ti clinopyroxenes of subduction-zone peridotites (e.g., Ti<150 ppm; Parkinson et al., 1992; Bizimis et al., 2000).

Harzburgites of the Lherz massif contain minoramounts of interstitial amphibole and previous studies sug-gested that pargasitic compositions are common (e.g.,

Downes et al., 1991; Fabries et al., 1991). Lherzolites, com-posite-lherzolites, and olivine-websterites contain greaterabundances of interstitial amphibole than harzburgites,and distinct compositional groups are present at each sam-ple site. Amphiboles of site A lherzolites ± olivine-webste-rite are generally sodic and of eckermannite composition(Appendix B). In contrast, amphiboles of site B harzburg-ites and lherzolites ± hornblendite are generally calcic withpargasitic to Mg-gedrite compositions. Amphiboles presentin hornblendite–veinlets are kaersutitic, and are opticallyand compositionally analogous to amphiboles reported inprevious studies of hornblendite lithologies within theLherz massif (e.g., McPherson et al., 1996; Fabries et al.,2001; Lorand and Gregoire, 2010).

2.3.3. Geothermometry

Equilibration temperatures calculated from core compo-sitions of pyroxene, olivine, and spinel phases in a numberof textural locations (e.g., porphyroclasts, typical matrix,and pyroxene-spinel symplectites) are reported in AppendixB, and are generally �800–900 �C, �720–900 �C, and�650–900 �C for two-pyroxene, orthopyroxene-spinel, andolivine-spinel thermometry, respectively. This range of equil-ibration temperatures is similar to those reported in earlierstudies of Ca–Al–Fe–Mg thermometry within the Ariege-group peridotites (e.g., Conquere and Fabries, 1984), and re-flects cooling to lithospheric conditions substantially below;(1) the peridotite solidus (c.f. Hirschmann, 2000; Wasylenkiet al., 2003; Herzberg, 2004; for a compilation of anhydrousand H2O under-saturated solidus curves); and (2) the poten-tial temperature thought to be typical of MORB mantle (e.g.,1280 �C; McKenzie and Bickle, 1988). Conquere and Fabries(1984) studied a number of Ariege-group peridotite bodiesthat crop out at different points along the North PyreneanMetamorphic Zone and suggested that two distinct episodesof subsolidus re-equilibration are recorded across this region.The first re-equilibration episode proposed by these authors,R1–D1, has been linked to peridotites with coarse-granulartextures (e.g., Fontete Rouge peridotites), and may reflectcooling to temperatures of 900–1000 �C at pressures of 12–15 kbar. The second re-equilibration event, R2–D2, is associ-ated with porphyroclastic textures (like those reported for theLherz massif), development of a penetrative foliation, forma-tion of cm-thick ultra-mylonite bands in some places (AveLallemant, 1967), and records equilibration extending tolower temperatures and pressures (down to �750 �C and 8–13 kbar; Vetil et al., 1988). The second of these regionallyrecognised mantle re-equilibration events is consistent withthermometric results reported here.

3. WHOLE-ROCK AND PYROXENE CHEMISTRY

3.1. Bulk-rock major-element, Cr, and Ni abundances

The studied sample suite displays highly systematic vari-ations in major-element compositions (Fig. 2 and AppendixA) with pronounced negative correlations between MgO(29.5–46.4 wt.%) and CaO (0.34–7.53 wt.%), Al2O3 (0.47–7.80 wt.%), Na2O (<0.05–0.65 wt.%) and SiO2 (42.9–

Petrogenesis of the Lherz peridotite 6165

Author's personal copy

46.8 wt.%). By contrast, abundances of MgO and FeO arenot well correlated (not shown) in that lherzolites and harz-burgites covering a range of MgO contain up to 8.32 wt.%FeO, while the minimum harzburgite FeO content is 7.67wt.%, and composite-lherzolites and olivine-websterites ex-tend to lower FeO values. (6.33 wt.%). This range of valuescoincides with those reported in previous studies of theLherz massif (e.g., Bodinier et al., 1988; Burnham et al.,1998), and linear correlations between bulk-rock CaO,Al2O3, MgO, and SiO2 contents are a common feature ofperidotite–pyroxenite assemblages in other ultramaficbodies such as Beni-Bousera, Horoman, and Ronda (e.g.,Frey et al., 1985; Bodinier and Godard, 2003). This rangeof compositions overlaps estimated compositions of BSE(e.g., Figs. 2 and 3) and depleted MORB mantle (DMM;Workman and Hart, 2006).

In more detail, the studied sample suite is broadly bimo-dal with harzburgite MgO abundances generally between45 and 47 wt.%, whereas lherzolites, composite-lherzolitesand olivine-websterites contain <40 wt.% MgO. Theseharzburgites have low Al2O3 (61 wt.%), CaO (61 wt.%),and high Ni contents (2142–2379 ppm) and Mg# (91.7–92.1), consistent with P70 wt.% olivine and <5 wt.% clino-pyroxene resulting from �20-25 wt.% melt removal(Appendix A). In contrast, lherzolites ± olivine-websteritebands have higher Al2O3 (1.5–5.8 wt.%), CaO (2.0–5.6 wt.%), and lower MgO (34–43 wt.%), Ni (1542–1918 ppm) contents, and Mg# (90.3 to 91.6) correspondingto higher pyroxene abundances. Olivine-websterite bandshave the highest Al2O3 (6.6–7.8 wt.%), CaO (6.9–7.5 wt.%), lower MgO (31–29 wt.%), Mg# (90.2–91.5) with-in the range of enclosing lherzolites, and the lowest Niabundances (1412–1381 ppm) consistent with P60 wt.%pyroxene and 630 wt.% olivine. The majority of studiedlherzolites ± olivine-websterite bands are composite sam-ples with Al2O3 values in excess of primitive mantle esti-

mates, and these specimens are not simple residues ofmelt extraction.

Despite the strong systematic co-variations in abun-dances of many major-element oxides (e.g., Fig. 2a), thereare significant differences in the major-element systematicsof the two sample sites that are exemplified by plots ofMg/Si against Fe/Si, and Cr/Al against distance from thecompositional boundary (Figs. 2c and 3). Harzburgiteand lherzolite samples generally have Cr/Al values that de-fine distinct groups related to each lithology. Harzburgites

Fig. 3. Bulk-rock Fe/Si and Mg/Si values. Magnesium-Fe-Sisystematics of the studied samples are compared to the compositionof residues resulting from isobaric anhydrous melting experiments(a) performed with fertile peridotite starting materials over apressure range of 1–5 GPa (Herzberg (2004) after Walter (1998)).Binary mixing analogues of igneous refertilisation (b) were derivedfrom the constant addition of experimental melts (analogous toprimary mantle malts) reported by Hirose and Kushiro (1993) andWalter (1998) to harzburgite 04LH33 of this study. Representativemineral compositions (c) were derived from those reported inAppendix B, and error bars represent 2rstdev calculated directlyfrom the population analysed (in each case n� 30). The range ofobserved compositions is also compared to; (1) estimates ofprimitive mantle composition, J79 = Jagoutz et al. (1979),G79 = Green et al. (1979), HZ86 = Hart and Zindler (1986),R91 = Ringwood (1991), MS95 = McDonough and Sun (1995),A95 = Allegre et al. (1995); and (2) a representative CI-chondritevalue, TM85 = Taylor and McLennan (1985). The thick grey arrowdelineates the linear-array of Mg/Si–Fe/Si compositions reportedfor lherzolite ± Ol-websterite of site A, and this array of data tendstoward the low Mg/Si–Fe/Si values of Ol-websterites. Hb-hazburg-ite = hornblendite-bearing harzburgite, Hb-Lherzolite = hornblen-dite-bearing lherzolite, Ol-Websterite = olivine-websterite, and C-Lherzolite = composite-lherzolite, refer to the main text for details.

"

6166 A.J.V. Riches, N.W. Rogers / Geochimica et Cosmochimica Acta 75 (2011) 6160–6182

Author's personal copy

commonly have high Cr/Al values that range from 0.50 to0.85, whereas lherzolites, composite lherzolites, and olivine-websterites generally have Cr/Al values <0.20. A number ofsamples at site B have Cr/Al values that are intermediate(0.28–0.45) between harzburgite and lherzolite end-mem-bers; these generally crop out close (<1.5 m) to the compo-sitional boundary between adjacent lherzolite andharzburgite bodies at site B. Samples with intermediateCr/Al contain visible, or crop out in close proximity tomm- to cm-scale hornblendite veins. Samples associatedwith hornblendite-veinlets at site B have Mg#’s that areintermediate between; (1) distal harzburgite and lherzolitesamples of site B; and (2) all samples of site A.

On the basis of Mg–Fe–Si systematics peridotites of siteA are separated into two distinct compositional groups cor-responding to harzburgite and lherzolite bodies (Fig. 3).Harzburgites define a tightly-clustered group with a narrowrange of Mg/Si (1.298–1.326) and Fe/Si values (0.286–0.298). Lherzolites and olivine-websterites of site A definea striking linear trend with Mg/Si varying from 0.809 to1.099 as Fe/Si increases from 0.224 to 0.309. This trendcoincides with the estimated composition of BSE given byMcDonough and Sun (1995), and the DMM-value ofWorkman and Hart (2006). In contrast, peridotites of siteB do not show such a clear division between lherzolitesand harzburgites and cover a narrower range of Fe/Si val-ues. Harzburgites define a clustered group with Fe/Si values

between 0.298 and 0.314 and Mg/Si values of 1.341–1.389.Site B lherzolites cover a range of Fe/Si and Mg/Si values(0.273–0.314 and 1.031–1.271, respectively) defining anelongate field sub-parallel to the linear correlation reportedfor the websterite-banded lherzolite body of site A (Fig. 3).Taken together, the harzburgite samples of site A and site Bdescribe an elongate field that is sub-parallel to the lherzo-lite field of site B.

3.2. Bulk-rock and pyroxene trace-element characteristics

Lherz peridotites (± olivine-websterite) and their con-stituent pyroxenes cover a range of trace-element abun-dances (Figs. 4 and 5, Appendices A and B), and therare-earth-element (REE) profiles of bulk-rocks are re-flected by the REE-patterns of constituent clinopyroxenes.In general, lherzolites and olivine-websterites are light-rare-earth-element (LREE) depleted and have higher abun-dances of heavy- and middle-rare-earth-elements (HREEand MREE) than harzburgite samples. Europium anoma-lies ([Eu/Eu*]N, where [Eu*]N = [(Gd + Sm)/2]N, and N de-notes normalisation to CI-chondrite) of the studied samplesare 0.44–0.74 (Appendix B), and constitutent pyroxenesgenerally have (Eu/Eu*)N of 0.24–0.32; these relativelysmall negative values differ from positive Eu-anomalies(>1) reported for bulk-rocks and clinopyroxenes of selectedophiolite peridotites from the Eastern Central Alps (e.g.,

Fig. 4. Rare-earth-element (REE) abundances (with an interpolated space on the x-axis for Pm) of bulk-rocks (BR) and constituentclinopyroxenes (CPX) normalized to the CI-chondrite estimate of McDonough and Sun (1995). Hb-hazburgite = hornblendite-bearingharzburgite, Hb-Lherzolite = hornblendite-bearing lherzolite, Ol-Websterite = olivine-websterite, and C-Lherzolite = composite-lherzolite,the main text contains details of lithological classifications.

Petrogenesis of the Lherz peridotite 6167

Author's personal copy

Totalp and Platta ultramafics; Muntener et al., 2010).Hornblendite-free harzburgites of site A are characterisedby U-shaped REE-patterns, and samples with the greatestdegree of LREE-enrichment also have elevated chondrite-normalised U and Th relative to Ba and Nb. The trace-ele-ment systematics of peridotites at site B, differ from those ofsite A, and show a wide-range of REE-patterns, includingconvex-upward and sinusoidal shapes (e.g., Fig. 4); the lat-ter are generally found in hornblendite-bearing rocks orspatially associated samples, whereas U-shaped REE-pat-terns are present in samples that lack hornblendite and havelow bulk Ti-contents (<30 ppm). In general, the studiedsamples lack the selective enrichment of large-ion-litho-phile-elements (LILE; Cs, Rb, Ba, and Sr) relative to

high-field-strength-elements (HFSE; Th, U, Nb, Ta, Zr,and Hf) reported for SSZ-peridotites of the Izu–Bonin–Mariana arc sequence (Fig. 5), but hornblendite-free harz-burgites have chondrite-normalised trace-element patternssimilar to harzburgite xenoliths recovered from the Avachavolcano, Kamchatka arc (Ionov, 2010).

The range of trace-element abundances in the studiedbulk-rocks and clinopyroxenes is similar to the composi-tional range reported in previous investigations of theLherz massif (e.g., Bodinier et al., 1988, 1990; McPhersonet al., 1996; Burnham et al., 1998; Bodinier et al., 2004;Le Roux et al., 2007, 2009). No extremely depleted clinopy-roxenes, analogous to that reported for harzburgite 06LI15(REE 6CI-chondrite abundances; Le Roux et al., 2009) are

Fig. 5. Spider-diagram of bulk-rock trace-element abundances normalized to the CI-chondrite value of McDonough and Sun (1995).Proximal harzburgites of site A crop out within 25 cm of the contact between harzburgite and lherzolite bodies, whereas distal samples cropout >1 m from this compositional boundary. For comparison, the range of bulk-rock trace-element compositions reported for Torishimaforearc samples and peridotites of the Conical seamount, Izu–Bonin–Mariana arc (Parkinson and Pearce, 1998), are shown by the grey field.SSZ = supra-subduction zone. Hb-hazburgite = hornblendite-bearing harzburgite, Hb-Lherzolite = hornblendite-bearing lherzolite, Ol-Websterite = olivine-websterite, and C-Lherzolite = composite-lherzolite, details of the classification scheme applied here are given in the text.

6168 A.J.V. Riches, N.W. Rogers / Geochimica et Cosmochimica Acta 75 (2011) 6160–6182

Author's personal copy

present in our sample set. Bulk-rock HREE contents andratios determined in this study overlap the range previouslyreported for Lherz peridotites and for peridotite and pyrox-enite assemblages of other Ariege-group massifs, Ronda,Horoman, and the Beni-Bousera ultramafic complexes(Fig. 6a).

Trace-element partition coefficients between co-existingclino- and orthopyroxene (KdCPX/OPX) phases are broadlyanalogous to those reported in previous studies of mantleperidotites (e.g., Stosch, 1982; Bodinier et al., 1988; Egginset al., 1998; Witt-Eickschen and O’Neill, 2005; Lee et al.,

2007; Harvey et al., 2010; Appendix B). Absolute valuesof REE KdCPX/OPX are relatively high in the studied samplesuite that has equilibration temperatures of �800 �C or less.The elevated REE KdCPX/OPX values of Lherz samples withlower equilibration temperatures than reported for severalxenolith suites (refer to Appendix B) probably reflect thetemperature dependence of subsolidus REE redistributionthat is thought to be linked to Ca-exchange between clino-and orthopyroxene during cooling (e.g., Witt-Eickschenand O’Neill, 2005 and references therein), and this is thesubject of ongoing study.

Fig. 6. Moderately-incompatible (Dy, Yb, V), incompatible (Ti), and highly incompatible (Zr) trace-element compositions of bulk-rocks(Bulk; (a) and (b) and constituent clinopyroxenes (CPX; (c) and (d) Bulk-rock models of spinel- (Sp) and garnet- (Gt) facies fractional meltingassume a starting composition equivalent to that of primitive mantle (PUM; McDonough and Sun, 1995) and incorporate melting modes andpartition coefficients reported by Johnson et al. (1990) that are relevant for anhydrous conditions. These melting curves reflect dynamicmelting in which a critical porosity retains 1% of the melt in all residues (after McKenzie, 1985). The melting modes of Kinzler (1997) are usedin our calculations. The blue curve in (a) describes compositions expected for melting residues produced under hydrous conditions in themantle-wedge where HREEs are expected to be more incompatible (Gaetani et al., 2003; McDade et al., 2003). Vanadium is multi-valent onEarth, occurring as V3+, V4+, and V5+; models of bulk-rock abundances resulting from variable degrees of fractional melting are redoxsensitive and are shown over a range of oxidation conditions (FMQ � 1 to FMQ + 1; after Parkinson and Pearce, 1998). Data fields delineatethe range of compositions reported for spinel-bearing peridotites devoid of garnet ± plagioclase (Bodinier et al., 1987b; Bodinier et al., 1988;Fabries et al., 1989; Bodinier et al. 1990; McPherson et al., 1996; Van der Wal and Bodinier, 1996; Burnham et al., 1998; Fabries et al., 1998;Parkinson and Pearce, 1998; Garrido et al., 2000; Lenoir et al., 2000, 2001; Downes, 2001; Beccaluva et al., 2004; Bianchini et al., 2007; Ionov,2010), and clinopyroxene compositions determined by in-situ methods (Parkinson et al., 1992; Bizimis et al., 2000; Johnson et al., 1990;Johnson and Dick, 1992; Warren and Shimizu, 2010). SSZ = supra-subduction zone, MOR = mid-ocean ridge. Hb-hazburgite = hornblen-dite-bearing harzburgite, Hb-Lherzolite = hornblendite-bearing lherzolite, Ol-Websterite = olivine-websterite, and C-Lherzolite = composite-lherzolite, see the main text for details. FMM = fertile MORB mantle (Pearce and Parkinson, 1993), DMM = depleted MORB mantle(Workman and Hart, 2006).

Petrogenesis of the Lherz peridotite 6169

Author's personal copy

3.3. Bulk-rock S, Cu, Re and Os concentrations and Re–Os

isotope compositions

Bulk-rock S, Cu, Re, and Os abundances and Re–Osisotopic compositions generally fall into distinct composi-tional groups with no clear gradation across the composi-tional boundary between harzburgites and lherzolites atsites A and B (Fig. 7). Harzburgites have low bulk S andCu contents (35–52 ppm and 1.12–2.90 ppm, respectively),whereas lherzolites and composite-lherzolites have high Sand Cu contents (163–415 ppm and 16.2–42.1 ppm, respec-tively). Olivine-websterites measured in this study have Scontents (334 ppm; 04LH38) within the range observed inspatially associated lherzolites, but higher Cu concentra-tions (77.6 ppm; 04LH14A) than adjacent peridotites. Sul-phur and Re abundances define broad negativecorrelations with increasing bulk-MgO content (Fig. A8of Appendix A), and this range of values coincides withthose previously reported for the Lherz peridotite body(Reisberg and Lorand, 1995; Burnham et al., 1998), passingclose to estimates of primitive mantle composition (Appen-dices A and B). A limited number of samples (04LH07A,04LH37B) have relatively high S contents at a given MgOconcentration, and these rocks also have elevated Re con-tents. The olivine-websterite analysed in this work(04LH38) has low S and Re abundances when comparedto the dominant negative correlation between these ele-ments and bulk-MgO (Fig. A8, Appendix A). Harzburgitesgenerally have high osmium contents (4012–5312 ppt) withfew Lherz harzburgites containing <4100 ppt (an exceptionincludes 04LH34A of site A with 3017 ppt Os). Lherzolitesand composite-lherzolites have Os concentrations that gen-erally range from 3624 to 4112 ppt, and these lherzolitesrarely have Os abundances in excess of 4100 ppt. Lherzolite04LH07A, associated with hornblendite-veinlets and char-acterised by relatively high Re and S abundances, has thehighest Os content (4538 ppt) of the studied sample suite.Olivine-websterite 04LH38 has the lowest osmium concen-tration (2830 ppt) of all studied samples, and this value issignificantly greater than Os concentrations generally re-ported for basaltic materials (e.g., 1–50 ppt range ofMORB; Shirey and Walker, 1998). Measured Os abun-dances in this sample suite coincide with the range of valuesreported in other investigations of Lherz peridotites, andthese data do not define a strong positive correlation withbulk-MgO (Appendices A and B).

Harzburgites generally have low 187Re/188Os(0.023 ± 0.002 to 0.066 ± 0.007) and sub-chondritic187Os/188Os values that are within uncertainty of one an-other (�0.117; Fig. 7 and Appendix B), and higher187Os/188Os values than reported for 3 of the 4 low-S harz-burgites studied by Luguet et al. (2007). Few studied harz-burgites (04LH34A of site A is an exception) haveconcomitant elevations of 187Re/188Os and 187Os/188Osabove values of 0.1 and 0.120, respectively. Lherzolitesand composite-lherzolites have 187Re/188Os compositionsranging from 0.32 ± 0.003 to 0.55 ± 0.006, and chondriticto supra-chondritic 187Os/188Os values (0.1246–0.1324).Sample 04LH07A, associated with hornblendite-veinlets,

Fig. 7. Bulk-rock Re–Os isotope compositions, bulk- Al2O3

abundances, and distance from compositional boundaries betweenharzburgite and lherzolite bodies at sites A and B. Correlations arecalculated using the isoplot 3.00 tool (Ludwig, 2001) and allliterature data is combined with that reported here duringregression calculations that assume a model-1 fit (where assigneduncertainties are 2r internal precision values). Model-ages calcu-lated via the ‘aluminochron’ method compare an initial 187Os/188Osvalue (taken from the Os-isotope value of the linear regression atan Al2O3 content of 0.7 wt.%) to a mantle evolution curve thatassumes a 187Os/188Os Solar System initial of 0.09524 ± 0.00011(IIIA irons, 95% confidence interval; Smoliar et al., 1996), a Re-decay constant of 1.666 � 10�11 (Smoliar et al., 1996) and apresent-day chondritic value of 0.1274 ± 0.0034 (Walker et al.,2002). PME = Proterozoic melt extraction, AME = Achaean meltextraction, after Becker et al. (2006). Data reported for Lherzsamples of previous studies were taken from by Reisberg andLorand (1995), Burnham et al. (1998), Becker et al. (2006), andLuguet et al. (2007). Primitive mantle estimates are taken fromMcDonough and Sun (1995) and Meisel et al. (2001). Ol-Webste-rite = olivine-websterite, and C-Lherzolite = composite-lherzolite,refer to the main text for classification details.

6170 A.J.V. Riches, N.W. Rogers / Geochimica et Cosmochimica Acta 75 (2011) 6160–6182

Author's personal copy

has unusual Re–Os isotope compositions compared toother lherzolites analysed in this work; the bulk 187Re/188Osvalue of 0.36 ± 0.004 falls within the range of other lherzo-lite samples, but 04LH07A has the lowest 187Os/188Os com-position (0.11258 ± 0.00010) of all studied specimens. Thehighest 187Re/188Os (0.73 ± 0.007) and 187Os/188Os(0.1387) isotope compositions yet reported for the Lherzmassif (Fig. 7) were determined for olivine-websterite04LH38.

Regional studies of Ariege-group peridotites identifiedbroad linear correlations between 187Os/188Os and187Re/188Os, and relatively strong positive correlations be-tween 187Os/188Os, S and Al2O3 contents (Reisberg andLorand, 1995; Burnham et al., 1998) that overlap the rangeof Os–Al values reported here. New results for the Lherzmassif alone show that bulk-rock 187Os/188Os values ofthe studied sample suite do not correlate linearly with187Re/188Os, S, Al2O3, MgO, Yb (e.g., Fig. 7 and AppendixB). Harzburgites generally cluster at low 187Os/188Os,187Re/188Os, S, Al2O3, MgO, and Yb values, whereas lherz-olites ± olivine-websterite form a cluster at higher S, Al2O3,MgO, and Yb contents with 187Os/188Os values overlapping

the range of Os-isotope compositions reported for recentlyexhumed abyssal peridotites that sample present-day con-vecting upper mantle (e.g., Fig. 9b), and similar observa-tions have been described for Middle-Atlas peridotitexenoliths (Wittig et al., 2010).

4. PETROGENETIC MODELS

The Lherz body lacks dunite and contains relativelysmall volumetric proportions of harzburgite, thus it differsfrom harzburgite dominated massifs associated with ophio-lite sequences that generally contain cross-cutting dunitechannels, which have been linked to melt generation at highmelt-flux rates (e.g., Canyon Mountain, Oregan; Josephine,Oregan, Semail, Oman; Troodos, Cyprus, Red Mountain,New Zealand; Dick and Sinton, 1979; Boudier and Nicolas,1985; Kelemen et al., 1995). The entire range of bulk-rockand spinel major-element compositions reported for Lherzdiffer from SSZ-peridotites that may have equilibrated withboninitic magmas (e.g., spinel Cr# >60 reported for perido-tites of the Izu–Bonin–Mariana Arc, Parkinson and Pearce,1998) and their experimental equivalents (e.g., Gaetani and

Fig. 8. Models of bulk-rock compositions resulting from binary mixing between a basaltic melt akin to MORB picrite and harzburgites04LH33 are constructed at selected points in time (t), and the proportion of melt (wt.%) added is marked by vertical dashed lines. Modelssimulate incremental addition MORB-like melt (16 wt.% Al2O3, 10 ppt Os, 2000 ppt Re, and a present-day 187Os/188Os value of 0.134, afterShirey and Walker, 1998 and BVSP, 1981). Data retrieved from previous studies of Lherz peridotites were taken from Reisberg and Lorand(1995), Burnham et al. (1998), Becker et al. (2006), and Luguet et al. (2007). Primitive mantle estimates are taken from McDonough and Sun(1995) and Meisel et al. (2001). Ol-Websterite = olivine-websterite, and C-Lherzolite = composite-lherzolite, classification details are given inthe main text.

Petrogenesis of the Lherz peridotite 6171

Author's personal copy

Grove, 1998; Parman and Grove, 2004; Grove et al., 2006).However, highly-depleted (whole-rock Al2O3 <0.85 wt.%)hornblendite-free harzburgites of site A, with bulk-rocktrace-element contents analogous to mantle-wedge perido-tites (e.g., Avacha peridotite xenoliths; Ionov, 2010), con-tain low-Ti clinopyroxenes that are similar to thosereported for subduction-zone peridotites of the Hellenicophiolite and Izu–Bonin–Mariana arc peridotite suites(Fig. 6c and d). Many of the studied peridotites from theLherz body have bulk-rock and mineral major-, minor-,and trace-element compositions that are transitional be-tween abyssal and SSZ-peridotites (e.g., Fig. 6; AppendicesA and B), and for these reasons the following discussionexamines whether the studied range of peridotite ± oliv-ine-websterite compositions can be accounted for by; (1)percolative refertilisation of refractory peridotites withbasaltic melt; (2) variable degrees of melt depletion fol-lowed or accompanied by physical mixing of peridotiteand pyroxenite during mechanical mingling (± short-lengthscale melt-interaction at pyroxenite margins); and (3) short-length scale metasomatism unrelated to postulated Ca, Al,Ti, Fe enrichment envisaged in models of igneous-refertili-sation. We also evaluate the tectonic setting in whichmelting and olivine-websterite formation may have takenplace.

4.1. Assessment of percolative igneous refertilisation

Models of igneous refertilisation constructed by two-component mixing between harzburgite 04LH33 and exper-

imental melts (analogous to primary melts that could beenvisaged in a zone-refining scenario) produce a broad fieldof products overlapping the range of lherzolite Mg–Fe–Sicompositions at site B, and the majority of compositionsobserved in the banded-lherzolite body at site A (Fig. 3b).However, the orientation of the elongate fields of harzburg-ite at site A and B, and the field of site B lherzolite Mg–Fe–Si compositions cross-cut the trajectory of these igneousrefertilisation models. In this context, the range of meltcompositions required to satisfy the observed bulk-rockcompositions does not follow a single refertilisation trajec-tory and requires interaction with substantial melt volumes(P5-10 wt.% of the product) ranging from basanite to tho-leiite. The requirement of significant melt volumes of dis-tinct major-element compositions coexisting in closelyspaced samples (10 s cm scales) during a percolative igneousrefertilisation process is difficult to envisage as a physicallyrealistic aspect of melt infiltration. Hand-specimen, miner-alogical, and bulk compositional data show that site Alherzolites represent physical mixtures of lherzolite + oliv-ine-websterite, and field observations indicate that this pro-cess is structurally old and took place under conditionswhere plastic-deformation operated. Site B lherzolites areheterogeneous and may also contain a websteritic compo-nent, but the interpretation of this group of samples is com-plicated by modal-metsomatism that formed secondarypyroxene ± spinel ± amphibole ± sulphide accompaniedby cryptic-metasomatism that elevated LREE and MREEabundances of bulk-rocks and clinopyroxenes during theformation of late-stage cross-cutting hornblendite-veinlets.

Fig. 9. Whole-rock Al2O3 contents and 187Os/188Os compositions (a) and a histogram of Os-isotope compositions (b). These diagramsincorporate data for two European continental xenolith suites (Massif Central, Harvey et al., 2010; and Spitsbergen, Svalbard Archipelago,Choi et al., 2010), peridotite xenoliths of the Middle-Atlas, Morocco (Wittig et al., 2010), and four oceanic peridotite suites; Izu–Bonin–Mariana SSZ-peridotites (Parkinson et al., 1998); Mid-Atlantic Ridge abyssal periodtites, ODP Site 920 (Brandon et al., 2000) and ODP Hole1274A (Harvey et al., 2006); and abyssal peridotites of Gakkel Ridge (Liu et al., 2008). The primitive mantle estimate corresponds to thatreported by Meisel et al. (2001). Literature data are sourced from Reisberg and Lorand (1995), Burnham et al. (1998), Becker et al. (2006), andLuguet et al. (2007). SSZ = supra-subduction zone, MOR = mid-ocean ridge. Ol-Websterite = olivine-websterite, and C-Lherzolite = com-posite-lherzolite, refer to the main text for details.

6172 A.J.V. Riches, N.W. Rogers / Geochimica et Cosmochimica Acta 75 (2011) 6160–6182

Author's personal copy

4.2. Spinel-facies harzburgite formation

Bulk-rock MgO, Mg#, Al2O3, and HREE abundancesof the studied peridotites correlate with one another andtrace the extent of melt extraction in relatively homoge-neous samples unaffected by recent interaction with basan-itic magmas associated with hornblendite formation. Forthese reasons moderately-compatible and compatible-ele-ment compositions of harzburgites reflect the conditionsof melt generation. The highest degrees of melting are re-corded by harzburgites (up to �25 wt.% melt loss; Appen-dix B), which have Mg–Fe–Si abundances and relativelyhigh olivine/orthopyroxene values (olivine/orthopyrox-ene = 2.08–5.76) that cannot be accounted for by anhy-drous melting alone (Fig. 3a and Fig. A3 of Appendix A).Low-Al harzburgites of site A contain clinopyroxenes withlow-alkali contents; the low-Ti abundances of these clino-pyroxenes differ from residual clinopyroxenes of abyssalperidotites (Fig. 6c and d), but are broadly analogous toclinopyroxenes of SSZ-peridotites. However, clinopyroxeneand bulk-rock trace-element data for these rocks lack pro-nounced enrichment of Cs, Rb, Ba, and Pb, suggesting thatthese materials did not form in an environment that experi-enced high-rates of fluid flux from subducted lithosphere(i.e., not within �100 km of a slab-wedge interface; Pearceand Parkinson, 1993; though this parameter is dependenton nature of subducting materials, the angle of dip of thesubducting slab, and the maturity of the subduction-zonesystem). The lack of correlation between harzburgiteincompatible-element abundances and orthopyroxene/oliv-ine values (not shown) suggests that it is unlikely that orth-opyroxene is of secondary origin, and significant degrees ofinteraction with Si-bearing melts/fluids are doubtful. Min-eralogical affinities between peridotite bands of the Lherzmassif, abyssal (Fig. A3, Appendix A) and SSZ-peridotites(e.g., Fig. 6), which are thought to be derived from meltingover an interval that does not exceed 2 GPa (c.f. Kinzler,1997; Walter, 2003 and references therein), lead us to exam-ine trace-element constraints on the depth at which meltingtook place to produce the peridotites of the Lherz massif.

The studied sample set displays a wide-range of trace-element characteristics, which indicates that Lherz perido-tites have a complex metasomatic history (appraised in Sec-tion 4.3). We focus on constraining the genesis of site Aharzburgites as the origin of site B harzburgites is obscuredby Al, Fe, Ti, Na, LREE, and MREE enrichment resultingfrom interaction with hornblendite forming magmas. Mod-erately-incompatible element contents of peridotites may besignificantly less susceptible to post-melting modificationthan incompatible elements, and for this reason HREEsare used to constrain melting style. Previous authors haveused bulk-rock Tm–Yb compositions to suggest that theLherz peridotites result from variable degrees of garnet-fa-cies melting (i.e., in excess of 2 GPa; Bodinier et al., 1988;Burnham et al., 1998). New trace-element data for theLherz peridotites, combined with that of previous studies,show that harzburgites generally have a positive slope be-tween HREEs, and bulk-rock Dy–Yb and Lu–Hf contents(Fig. 6 and Fig. A7 of Appendix A) define arrays analogousto residue compositions expected for spinel-facies melting.

Hornblendite-free harzburgites and constituent clinopyrox-enes generally have positive MREE–HREE slopes, and thiscontrasts to negative MREE–HREE slopes of clinopyrox-enes of abyssal peridotites exhumed at slow-spreadingridges where the earliest stages of melting take place inthe presence of garnet followed by significant degrees offractional melting in spinel-facies mantle (e.g., Johnsonet al., 1990; Johnson and Dick, 1992; Hellebrand et al.,2002, 2005). For these reasons Lherz harzburgites may haveexperienced melting under slightly shallower conditions (orover a narrower melting interval) than many of the abyssalperidotites studied to date.

Models were constructed for fractional melting of a prim-itive-mantle (McDonough and Sun, 1995) source, and theprecise composition of predicted residues is sensitive to theselected partition coefficients, melting modes, and startingcomposition. The range of Dy–Yb compositions in Lherzperidotites defines a fractionation curve sub-parallel to melt-ing curves calculated for residue compositions expected forfractional melting of fertile peridotite (analogous to primitivemantle (McDonough and Sun, 1995), fertile-MORB mantle(FMM; Pearce and Parkinson, 1993), or DMM (Workmanand Hart, 2006)) in the spinel-facies (Fig. 6a). Compared toresidue compositions expected for spinel-facies melting un-der anhydrous conditions, Lherz harzburgites have consis-tently lower Yb-abundances over a range of Dy/Yb values.Experimental studies have suggested that HREEs may be lesscompatible during fluid-present melting (e.g., Gaetani et al.,2003; McDade et al., 2003, and for this reason the relativelylow HREE-contents of Lherz harzburgites indicate that theyare residual after fluid-present melting (Fig. 6a). Abundancesof V–Yb–Sc in these harzburgites are broadly consistent withmodels of melt depletion under oxidation conditions of FMQto FMQ+1 (Fig. 6b and Fig. A7c of Appendix A), but thehighest V-abundances determined for site A harzburgites,while overlapping the range reported for subduction-zoneperidotites of Avacha and Izu–Bonin–Mariana, correspondto samples displaying the greatest degree of LREE-enrich-ment that may be linked to secondary processes. Earlier stud-ies of oxidation conditions, which used olivine-orthopyroxene-spinel major-element equilibria (Woodland,1992; Woodland et al., 1996, 2006), suggested that spinel-per-idotites of the Lherz massif have oxidation states rangingfrom QFM � 1.5 to QFM + 1; which are transitional be-tween fo2 values reported for abyssal peridotites and conti-nental xenolith suites. Differences between fo2 conditionsindicated by V-abundances determined in this study com-pared to previous studies of olivine–orthopyroxene–spinelequilibria (e.g., Woodland, 1992; Woodland et al., 1996,2006) may reflect partial or complete resetting of oxidationrecords after melt extraction, and resolving the cause of thesedifferences is beyond the scope of this work. Importantly, re-sults of modelling the behaviour of moderately compatibletrace-element compositions (Dy, Yb, Lu, Hf, V, and Sc) dur-ing melting are consistent with bulk-rock major-element andmineralogical characteristics that suggest harzburgites resultfrom�20–25 wt.% melt removal (Appendix B) in the absenceof residual garnet (Fig. 6a and b, Fig. A7b and c of AppendixA), and a starting composition broadly analogous to FMMand/or DMM is in agreement with previous results for melt-

Petrogenesis of the Lherz peridotite 6173

Author's personal copy

ing in mantle-wedge environments, particularly those locatedbeneath relatively thin lithosphere such as that expected forback-arc basin settings (e.g., Pearce and Parkinson, 1993;Woodhead et al., 1993; Peate et al., 1997; Ewart et al.,1998; Parkinson et al., 1998; Langmuir et al., 2006).

4.3. Modification of Lherz peridotites by melt- and fluid-

metasomatism

Incompatible- and highly-incompatible element contents(including LREE, Zr, and Sr) of bulk-rocks and pyroxenescannot be accounted for by models of melt extraction, andmay have been modified after an earlier melting event. Pre-vious authors have suggested that the range of trace-ele-ment abundances reported for Lherz peridotites may beexplained by hydrous, carbonatite, and silicate-melt meta-somatism that may reflect two distinct processes; (1) mul-ti-stage metasomatic interaction with several generationsof compositionally distinct melts/fluids (e.g., Woodlandet al., 1996; Burnham et al., 1998; Fabries et al., 1998,2001); or (2) pervasive melt-infiltration accompanied by achromatographic melt-front (e.g., Bodinier et al., 1988;Bodinier et al., 1990; Le Roux et al., 2007, 2008, 2009).New trace-element data for harzburgite–lherzolite outcropswith contrasting pyroxenite and structural characteristicsare consistent with the occurrence of two distinct metaso-matic events. The first metasomatic event is characterisedby cryptic metasomatism of harzburgites at site A and B;these samples display LREE-enrichment relative to MREEand HREE, have relatively high chondrite-normalised Uand Th contents compared to Ba and Nb (pronounced insite A harzburgites). These samples lack pronouncedenrichment of Cs, Ba, and Pb, that is generally associatedwith SSZ-melts that impart a distinctive trace-element sig-nature linked to fluid release close to the slab-wedge inter-face (e.g., Pearce and Parkinson, 1993; Parkinson andPearce, 1998). The lack of Ti-enrichment (Figs. 5 and 6cand d), absence of apatite, and the lack of a linear correla-tion between clinopyroxene abundances, orthopyroxene/olivine values, and LREE-abundances and LREE/HREEvalues (of bulk-rocks and constituent pyroxenes; not shownrefer to Appendix B), combined with the range of bulk-rockand pyroxene Zr abundances (Appendix B) suggest thatLREE-enrichment in hornblendite-free harzburgites is theresult of post-melting interaction with; (1) a low-volumeSi-poor fluid phase, (2) a low-volume volatile-bearing melt,or (3) a combination of these (c.f., Rudnick et al., 1993;Schiano and Clocchiatti, 1994; Schiano et al., 1995; Xuet al., 2003; Bouvier et al., 2010a,b for relevant compari-sons). No fluid inclusion data are currently available forthe Lherz massif, but at least two generations of CO2–H2Obearing fluids thought to have equilibrated at temperaturesof �950 �C and pressures of 6–7 kbar and 8–9 kbar, respec-tively, have been reported in this region for amphibole-bear-ing peridotites of the Cassou massif (Fabries et al., 1989;Bilal, 1978), and fluids of broadly similar compositionsmay be responsible for a significant portion of LREE, andSr, enrichment in hornblendite-free harzburgites.

The second metasomatic event identified in this studypostdates earlier LREE-enrichment linked to Si-poor flu-

ids/melts, and involves modal (amphibole ± phlogo-pite ± sulphide ± secondary pyroxene) and crypticmetasomatism (MREE-enrichment, intermediate Cr/Alvalues, and relatively high bulk-rock and pyroxene Ti andNa contents in the absence of secondary minerals) linkedto the formation of hornblendite-veinlets at site B. Thesemineralogical and chemical enrichment characteristics areconsistent with results reported in earlier studies of melt–rock interaction adjacent to amphibole-bearing veins (e.g.,Bodinier et al., 1987b; McPherson et al., 1996; Woodlandet al., 1996; Bodinier et al., 2004). Outcrop relationshipssuggest that hornblendite-vein formation is a relatively re-cent event, and previous studies indicated that hornblen-dites crystallised at �100 Ma (e.g., Henry et al., 1998) atrelatively low-pressures (e.g., <1.3 GPa; Fabries et al.,2001; Lorand and Gregoire, 2010). Conquere and Fabries(1984) indicated that hornblendites intrude along composi-tional boundaries between harzburgite and lherzolite, andour own observations suggest that this may be the case atsite B. In addition, new spatial information related toLREE-enrichment at site A suggests that the lithologicalboundary between adjacent harzburgite and lherzolitebodies represents a conduit that facilitated post-meltingfluid ingress, and both site A and B demonstrate that pre-existing structural features can focus fluid and/or melt flowin mantle materials.

4.4. Petrogenesis of Ol-websterite bands and associated

Lherzolites

The majority of the studied lherzolites are compositesamples with distinct textures at the two study sites. Sam-ples of the banded-lherzolite body of site A contain web-sterites that are generally thick (up to 8 cm wide),foliation parallel, and laterally extensive (10 s of metres).Websterite bands of site A plot on an extension ofbulk-rock major-element covariation trends defined byLherz peridotites (e.g., Fig. 2a). Major-element composi-tions (e.g., Mg#) of site A olivine-websterites do not so-lely reflect equilibrium and fractional crystallisation of abasaltic melt (c.f. crystallisation experiments of Villigeret al., 2004, and references therein for a discussion ofexperimental constraints on basalt crystallisation pro-cesses), and differ from crystallisation products derivedfrom H2O-undersaturated melts (up to 5 wt.% H2O) at1.2 GPa (e.g., Muntener et al., 2001). Although modalabundances and mineralogical compositions of olivine-websterites cannot be accounted for by products of crys-tallisation experiments performed to date, publishedexperimental work has generally used starting materialswith Mg# 670 (c.f. Muntener et al., 2001; Villigeret al., 2004, 2007 and references therein). Major-elementcompositions of olivine-websterites tend toward the com-position of clinopyroxene (e.g., Fig. 3c), do not lay on atie-line between clino- and orthopyroxene, and while theyprobably represent cumulate bands in the broadest sense(c.f. Dantes et al., 2007 and a review by Downes, 2007),it is unclear if pyroxene-segregation during channelledmagma flow (e.g., Irving, 1980) significantly influencedtheir compositions.

6174 A.J.V. Riches, N.W. Rogers / Geochimica et Cosmochimica Acta 75 (2011) 6160–6182

Author's personal copy

Equilibrium melts calculated from trace-element com-positions of clinopyroxenes indicate that lherzolites andolivine-websterites are in equilibrium with a LREE-en-riched melt (Appendix B). The absolute concentrationsof REE in the calculated melt are dependent upon thechoice of partition coefficients (and the conditions forwhich they are relevant), but all calculated melts are con-sistently LREE-enriched. This melt composition is distinctfrom that of bulk-rock olivine-websterites, and REE con-tents of equilibrium melts are broadly analogous to E-MORB (Sun and McDonough, 1989), and continentaltholeiites parental to Triassic Pyrenean dolerites (e.g., Ali-bert, 1985) when anhydrous partition coefficients are ap-plied. Downes (2001) used lithophile-element isotopecompositions to show that no direct link between thelherzolites of Lherz and Pyrenean dolerites exists (e.g.,Downes, 2001).

LREE-enriched basalts and parental melts have beenidentified in back-arc basin environments (e.g., Pearcyet al., 1990; Stern et al., 1990), and for this reason REEcompositions do not provide unique information to con-strain the environment in which lherzolites and olivine-web-sterites formed. The presence of amphibole in the studiedspecimens, and major and trace-element evidence of frac-tionation controlled by olivine + clinopyroxene in the ab-sence of garnet (e.g., Figs. 3c and 6a) may suggest thatpartition coefficients determined for mantle-wedge environ-ments are more appropriate. The application of partitioncoefficients reported for mantle-wedge conditions (e.g.,Pearce and Parkinson, 1993; Gaetani et al., 2003; McDadeet al., 2003) indicates that melts in equilibrium with lherzo-lite and olivine-websterite pyroxenes have Cr and Y abun-dances that overlap the compositional range reported forisland-arc tholeiites (IAT), back-arc basin basalts (BABB),and MORB produced by melting of a relatively fertile man-tle source (c.f. Fig. 8 of Pearce et al., 1984, Appendix B).Titanium–vanadium ratios of these equilibrium melts are<20, and this is consistent with Ti/V values expected forback-arc basin settings (c.f. Shervais, 1982 for a discussionof Ti/V variation in basalts of distinct provenance). Nio-bium contents calculated for these equilibrium melts (gener-ally 1–4 ppm) are broadly analogous to those of N-MORB,and are lower than Nb concentrations typical of EMORBand OIB (2.33 ppm, 8.30 ppm and 48 ppm, respectively;Sun and McDonough, 1989). The similarity of equilibriumliquid compositions among olivine-websterites and adjacentlherzolites suggests that these rocks are genetically related.The simplest explanation of the current major- andtrace-element data set in closely-spaced lherzolites and oliv-ine-websterites is that olivine-websterites representpyroxene-rich cumulates of a tholeiitic LREE-enriched meltthat crystallised in melt-flow channels created during mod-erate degrees of melting (615 wt.% melt removal; Bodinieret al., 1988; Burnham et al., 1998), and this differentiationevent may also have formed residual lherzolites. In thissense site A banded-lherzolites may be broadly analogousto thin-layer peridotites and pyroxenites of the Horomanmassif where pyroxenite forming melts are not thought tocause significant refertilisation (i.e., Al, Ca, Ti addition)of adjacent peridotites (Malaviarachchi et al., 2010).

4.5. Re–Os: timing of lithosphere stabilisation or inherited

upper-mantle heterogeneity?

Samples of the Lherz massif are characterised by Re andOs abundance variations with bulk-MgO content that indi-cate Re and Os behave as moderately–incompatible andcompatible elements, respectively, and this observation isbroadly consistent with previous studies of Re–Os behav-iour in basaltic silicate melt systems (e.g., Pearson et al.,2004). Rhenium-Os isotope compositions of Lherz perido-tites ± olivine-websterite do not define a statistically mean-ingful isochron (e.g. Fig. 7a and b; large MSWD� 1), nordo they coincide with mixing curves that approximate igne-ous refertilisation (constant addition of basaltic melt torefractory peridotite) recently, at 100–500 Ma (correspond-ing to the time of hornblendite formation, and including theperiod of late-Variscan thermal events), and at 0.5 Ga inter-vals between 1 Ga and 3.5 Ga (Fig. 8). The large differencebetween the relatively low Os contents of basalts (typically10–500 ppt) compared to the high Os abundances in mantlematerials (generally on the order of 3000–5000 ppt), wherethe majority of the osmium is hosted by Os-rich sulp-hides ± Os–Ir-alloys associated with olivine (e.g., Luguetet al., 2007), mean that 187Os/188Os compositions of perido-tites will not be significantly disturbed by small to modestdegrees of melt–rock interaction, particularly when interac-tion has taken place within the last 1 Ga (Fig. 8). For thesereasons, osmium isotope compositions of individual bulk-rock peridotites, while strictly reflecting the sum of a mixedphase population (e.g., Lorand, 1991; Burton et al., 1999;Alard et al., 2000; Harvey et al., 2006, 2010, 2011; van Ack-en et al., 2008, 2010; Lorand et al., 2010; Lorand and Alard,2011), probably record the approximate time of melt-deple-tion as volumetrically minor Os-rich trace-phases may ac-count for >90% of the bulk-rock Os content (c.f. Harveyet al., 2010 for a study of the effects of variable degrees ofmelt/fluid metasomatism on Os-isotope compositions).

The lack of an isochronous bulk-rock Re–Os isotoperelationship reflects the fact that the Lherz body comprisesintercalated units of distinct lithologies where harzburgitesmay not be strictly coeval with lherzolite ± olivine-webste-rite, and these materials may have evolved with variable ini-tial 187Os/188Os compositions from �1.6 Ga onwards.Additionally, a comparison of Re–Os isotope data to a1.6 Ga reference line (Fig. 7) indicates that some sampleshave higher Al2O3 and Re abundances at a given187Os/188Os composition, and this probably reflects recentRe and Al disturbance that may be linked to hornblen-dite-interaction that has not significantly disturbed187Os/188Os values. The 187Os/188Os isotope compositionsof harzburgites generally yield TRD ages (representing aminimum differentiation age; Shirey and Walker, 1998) of�1.4 to 1.5 Ga (Fig. 9b). The period of isolation requiredto generate the observed 187Os/188Os values in composite-lherzolites and olivine-websterites with suprachondriticRe/Os values is in broad agreement with TRD ages of harz-burgites, assuming that these materials have experienced asingle-stage evolution with respect to Os.

Osmium isotope compositions do not offer highly-pre-cise geochronological information (c.f. review by Rudnick

Petrogenesis of the Lherz peridotite 6175

Author's personal copy

and Walker, 2009), but do provide broad constraints on thetiming of ancient differentiation. Meso-Proterozoic Os-agesdetermined here are; (1) similar to model ages derived fromthe Os–Al correlations generated from our study focusedon constraining the petrogenesis of the oldest portions ofthe Lherz massif (foliation parallel lithological strips); (2)within error of Os-model ages reported for peridotites thatinclude ultramafic samples from an ESE–WNW traverse of>300 km of the Pyrenean chain (Reisberg and Lorand,1995; Burnham et al., 1998); and (3) overlap Nd-model agesreported for the oldest portions of western European crustwhere structurally divided I- and S-type granites have beenlinked to ancient subduction systems (e.g., �1.7 to 1.4 Ga;Liew and Hofmann, 1988). Crustal materials of central andwestern Europe for which Liew and Hofmann (1988) re-ported Meso-Proterozoic Nd-model ages do not have a di-rect genetic link to ultramafic bodies of the Pyrenees, butthese authors suggested that crustal materials of the Mal-donubian zone (including central and southern France)may be linked to a Proterozoic continental mass to thesouth, and this crustal mass could potentially be relatedto the genesis of Ariege-group massifs (c.f. Burnhamet al., 1998). Further studies are required to constrain thecomposition, origin, and evolution of the oldest crustalmaterials within the Pyrenean chain, and this information,combined with further knowledge of peridotite emplace-ment mechanisms, may then be used to provide a more rig-orous assessment of crust–mantle relationships in thisregion.

Osmium-Al values reported for the Lherz peridotitebody overlap and extend to higher and lower 187Os/188Osvalues than those reported for other Ariege-group ultra-mafic massifs that define a broad positive correlation withbulk-rock Al2O3 (Fig. 9a), and such trends, when broadlysupported by correlations between 187Os/188Os composi-tions and Re/Os abundance and isotopic ratios (Appendi-ces A and B), are generally considered to reflect long-termisolation from the homogenising environment of the con-vecting upper mantle (e.g., Reisberg and Lorand, 1995;Burnham et al., 1998). However, the recent discovery of abroad Os–Al correlation in modern convecting upper man-tle (Fig. 9a, spinel-facies peridotites exhumed at the ultra-slow-spreading Gakkel Ridge (1.4–0.7 cm/year; Michaelet al., 2003) that yields an ancient model-age (�2 Ga; Liuet al., 2008) increases uncertainty about the geologicalmeaning of Os-ages determined for ultramafic materialsof Proterozoic and Phanerozoic terranes. The range of187Os/188Os isotope compositions reported for samplesfrom the Lherz massif overlaps the spread of abyssal peri-dotite Os-isotope compositions, but Os–Al correlationsand 187Os/188Os frequency distributions of Lherz samplesdiffer from abyssal peridotites in detail (Fig. 9b). Mid-Atlantic Ridge (Brandon et al., 2000; Harvey et al., 2006)and Izu–Bonin–Mariana peridotites (Parkinson et al.,1998) form discrete data clusters covering a range of187Os/188Os compositions with limited variation in bulk-rock Al2O3 abundances, and this is distinct from the distri-bution of Os-isotope and Al2O3 compositions for samplesof the Lherz massif. When compared to all available187Os/188Os data for abyssal peridotites (Fig. 9b) a greater

proportion of 187Os/188Os compositions reported for theLherz massif are subchondritic, clustering at TRD ages of�2 to 1.5 Ga. The broad correlation between 187Os/188Osisotope compositions and bulk-rock Al2O3 abundances thatencompasses Lherz and many other Pyrenean ultramaficmassifs (sampling an area >300 km in length) exhumed ina mountain belt with a significant collisional history (c.f.McCann, 2008a,b; Garcia-Sansegundo et al., 2011), leadus toward a preferred model in which the oldest portionsof the Lherz massif were created during Meso-Proterozoicmelting linked to the formation of overlying crust. Newmineralogical and trace-element data for Lherz perido-tites ± olivine-websterite indicate that the oldest portionsof this massif represent materials from mantle-wedge envi-ronments, and fluid-present spinel-facies melting may havetaken place in a back-arc basin setting. For these reasonssubduction zone processes may have been active at theboundary between the Iberian and European plates duringthe Meso-Proterozoic, and this may be a regionally impor-tant aspect of crustal growth.

An important outcome of this study is new composi-tional information that suggests a link between high degreesof melting in shallow, subduction-influenced environmentsthat may be responsible for the formation Pt–Ir–Os alloysidentified in harzburgites produced by 20–25 wt.% melt re-moval (e.g., those studied by Luguet et al., 2007). Pt–Ir–Osalloys ± laurite-erlichmanite sulphides may form as thepoint of S-exhaustion is approached during mantle melting(c.f. Luguet et al., 2007). Complete removal of S duringlow-pressure, subduction influenced melt generation is con-sistent with experimental evidence that indicates such envi-ronments probably produce primary melts with high Scontents at sulphur saturation (>1300 ppm) capable of rel-atively high-degrees of S-removal from the residue whencompared to melting at pressures >2 GPa (e.g., Mavroge-nes and O’Neill, 1999; Holzheid and Grove, 2002; Jugo,2009). Isolated minerals with high Os contents created dur-ing melting to the point of S-exhaustion are essentially de-void of Re, have relatively high-melting temperatures(>1300 �C; e.g., Andrews and Brenan, 2002), and may pre-serve records of ancient mantle differentiation in materialsthat have a multi-stage history of magmatism and solid-state mixing (Luguet et al., 2007). Thus, the geochemicalproperties of Os mean that bulk-rock, Pt–Ir–Os alloy, andprimary laurite-erlichmanite Os-isotope compositions pro-vide a powerful record of ancient, large-volume differentia-tion events, and statistical assessment of such information(e.g., Pearson et al., 2007), in the context of well-definedpetrological constraints, is of great importance for deter-mining; (1) the timing and mechanisms of local, regional,and global crust–mantle evolution; and (2) the proportionof refractory material present in Earth’s heterogeneousupper mantle.

5. SUMMARY AND IMPLICATIONS

New compositional data for closely spaced samplesrecovered from two traverses of adjacent harzburgite andlherzolite bodies ± olivine-websterite show that percolativeigneous refertilisation during the Variscan is not a viable

6176 A.J.V. Riches, N.W. Rogers / Geochimica et Cosmochimica Acta 75 (2011) 6160–6182

Author's personal copy

process to explain the range of major- and trace-elementabundances, and Re–Os isotopic compositions reportedfor the Lherz massif. Field, textural, and compositionaldata indicate that the Lherz massif is composed of elongate,foliation parallel, lithological strips that were juxtaposedduring plastic deformation. Major-element oxide,bulk-rock and clinopyroxene trace-element compositionsshow that highly-depleted harzburgites result from�20 – 25 wt.% melt extraction, are analogous to some man-tle-wedge peridotites (e.g., Avacha xenoliths), and were cre-ated at pressures <2 GPa. In contrast, the studiedlherzolites ± olivine-websterite are heterogeneous and rep-resent physical mixtures of residual materials and clinopy-roxene-dominated cumulates equilibrated with a LREE-enriched tholeiitic melt that formed in the absence ofresidual garnet. Two metasomatic events have modifiedperidotite compositions; the first involves wide-spread per-colation of a Si-poor melt/fluid; and the second is evident atsite B where relatively recent, small-volume melt interactionis linked with the intrusion of hornblendite-veinlets at thecompositional boundary between adjacent harzburgiteand lherzolite bodies. In addition, new mineral major-ele-ment data indicate that the intercalated peridotite + pyrox-enite assemblages of the Lherz ultramafic body equilibratedat lithospheric conditions (temperatures of �800–900 �C orless), and previous studies of lithophile-element isotopecompositions suggested that cooling and recrystallisationoccurred during the Cretaceous (Henry et al., 1998 and ref-erences therein).

Rhenium-Os systematics in this suite of samples showthat Os behaves compatibly and 187Os/188Os compositionshave not been significantly disturbed by small-volume meltinteraction. Osmium-isotope compositions, combined withliterature data for other Ariege-group ultramfic bodies, de-fine a broad positive correlation with bulk-rock Al2O3

abundances indicating that harzburgites, lherzolites, andolivine-websterites have been isolated from convectivehomogenisation since the Meso-Proterozoic and thisbroadly coincides with the time at which melting createdthe residual materials. The association between harzburg-ites resulting from spinel-facies melting in mantle-wedgeenvironments and residual, Os-rich, laurite-erlichmanitesulphides and Pt–Os–Ir-alloys suggests that a substantialproportion of persistent refractory anomalies in the pres-ent-day convecting mantle of Earth may be linked to an-cient large-scale melting events that may in turn berelated to subduction processing.

ACKNOWLEDGMENTS

Dr. B.L.A. Charlier is thanked for his guidance during the doc-toral study of A.J.V. Riches. In addition, we are grateful to Dr. A.Gannoun for support during Re–Os isotope analyses, Dr. S. Ham-mond for guidance during measurement of trace-element concen-trations, Dr. John Watson for his assistance during XRFanalyses, and we are indebted to Michelle Higgins and Kay Greenfor the preparation of many fine polished sections. Dr. H. Downesand Dr. I.J. Parkinson are thanked for their encouragement and re-marks on a previous version of this work, and Dr. M. Bizimis, twoanonymous reviewers, and the Associate Editor, Dr. S. Huang, arethanked for their astute comments on an earlier draft of this man-

uscript. This work was supported by a NERC studentship (GrantNumber NER/S/A/2004/13014) to A.J.V. Riches.

APPENDIX A. SUPPLEMENTARY DATA

Supplementary data associated with this article can befound, in the online version, at doi:10.1016/j.gca.2011.07.036.

REFERENCES

Alard O., Griffin W. L., Lorand J. P., Jackson S. E. and O’Reilly S.Y. (2000) Non-chondritic distribution of highly siderophileelements in mantle sulphides. Nature 407, 891–894.

Alibert C. (1985) A Sr–Nd isotope and REE study of late-Triassicdolerites from the Pyrenees (France) and the Messejana Dyke(Spain and Portugal). Earth Planet. Sci. Lett. 73(1), 81–90.

Allegre C. J. and Turcotte D. L. (1986) Implications of a two-component marble-cake mantle. Nature 323, 123–127.

Allegre C. J., Poirier J.-P., Humler E. and Hofmann A. W. (1995)The chemical composition of the Earth. Earth Planet. Sci. Lett.

134(3–4), 515–526.

Andrews D. R. and Brenan J. M. (2002) The solubility ofruthenium in sulfide liquid: implications for platinum groupmineral stability and sulfide melt-silicate melt partitioning.Chem. Geol. 192, 163–181.

Ave Lallemant G. (1967) Structural and petrographic analysis ofan “Alpine-type” peridotite: the lherzolite of the FrenchPyrenees. Leidse. Geol. Mededel. 42, 1–57.

Barnes S. J. and Roeder P. L. (2001) The range of spinelcompositions in terrestrial mafic and ultramafic rocks. J. Petrol.

42(12), 2279–2302.

Beccaluva L., Bianchini G., Bonadiman C., Siena F. and VaccaroC. (2004) Coexisting anorogenic and subduction-related meta-somatism in mantle xenoliths from the Betic Cordillera (south-ern Spain). Lithos 75, 67–87.

Becker H., Horan M. F., Walker R. J., Gao S., Lorand J.-P. andRudnick R. L. (2006) Highly siderophile element compositionof the Earth’s primitive upper mantle: constraints from newdata on peridotite massifs and xenoliths. Geochim. Cosmochim.

Acta 70, 4528–4550.

Bianchini G., Beccaluva L., Bonadiman C., Nowell G., Pearson G.,Siena F. and Wilson M. (2007) Evidence of diverse depletionand metasomatic events in harzburgite–lherzolite mantle xeno-liths from the Iberian plate (Olot, NE Spain): implications forlithosphere accretionary processes. Lithos 94, 25–45.

Bilal A. (1978) Les inclusions fluides des roches de la croute

continentale et du manteau superieur. These Doctorat d’Etat,Paris, p311.

Bizimis M., Salters V. J. M. and Bonatti E. (2000) Trace and REEcontent of clinopyroxenes from supra–subduction zone perido-tites. Implications for melting and enrichment processes inisland arcs. Chem. Geol. 165, 67–85.

Bizimis M., Sen G., Salters V. J. M. and Keshav S. (2005) Hf–Nd–Sr isotope systematics of garnet pyroxenites from Salt LakeCrater, Oahu, Hawaii: evidence for a depleted component inHawaiian volcanism. Geochim. Cosmochim. Acta 69, 2629–

2646.

Bizimis M., Griselin M., Lassiter J. C., Salters V. J. M. and Sen G.(2007) Ancient recycled mantle lithosphere in the Hawaiianplume: Osmium–Hafnium isotopic evidence from peridotitemantle xenoliths. Earth Planet. Sci. Lett. 257, 259–273.

Bodinier J. -L. and Godard M. (2003) Orogenic, Ophiolitic, andAbyssal Peridotites. In Treatise on Geochemistry 2: The Mantle

Petrogenesis of the Lherz peridotite 6177

Author's personal copy

and Core (eds. H. D. Holland and K. K. Turekian). Elsevier,Pergamon, pp. 103-170.

Bodinier J.-L., Guirand M., Fabries J., Dostal J. and Dupuy C.(1987a) Petrogenesis of layered pyroxenites from the Lherz,Freychinede and Prades ultramafic bodies (Ariege, FrenchPyrenees). Geochim. Cosmochim. Acta 51(2), 279–290.

Bodinier J.-L., Fabries J., Lorand J.-P., Dostal J. and Dupuy C.(1987b) Geochemistry of amphibole pyroxenite veins from theLherz and Freychinede ultramafic bodies (Ariege, FrenchPyrenees). Bull. Mineral. 110(4), 345–358.

Bodinier J.-L., Dupuy C. and Dostal J. (1988) Geochemistry andpetrogenesis of Eastern Pyrenean peridotites. Geochim. Cosmo-

chim. Acta 52(1–2), 2893–2907.

Bodinier J.-L., Vasseur G., Verneires J., Dupuy C. and Fabries J.(1990) Mechanisms of mantle metasomatism: geochemicalevidence from the Lherz orogenic peridotite. J. Petrol. 31(3),

597–628.

Bodinier J.-L., Menzies M. A., Shimizu N., Frey F. A. andMcPherson E. (2004) Silicate, hydrous and carbonate metaso-matism at lherz, France: contemporaneous derivatives ofsilicate melt–harzburgite reaction. J. Petrol. 45(2), 299–320.

Boudier and Nicolas (1985) Harzburgite and lherzolite subtypes inophiolitic and oceanic environments. Earth Planet. Sci. Lett.

76(1–2), 84–92.

Bouvier A. S., Deloule E. and Metrich N. (2010a) Fluid inputs tomagma sources of St. Vincent and Grenada (Lesser Antilles):new insights from trace-elements in olivine-hosted melts inclu-sions. J. Petrol. 51(8), 1597–1615.

Bouvier A. S., Metrich N. and Deboule E. (2010b) Light-elements,volatiles, and stable-isotopes in basaltic melt inclusions fromGrenada, Lesser Antilles: inferences for magma genesis. Geo-

chem. Geophys. Geosyst. 11, Q09004. doi:10.1029/

2010GC003051.

Boyd F. R. (1989) Compositional distinction between oceanic andcratonic lithosphere. Earth Planet. Sci. Lett. 96(1–2), 15–26.

Brandon A. D., Snow J. E., Walker R. J., Morgan J. W. and MockT. D. (2000) 190Pt–186Os and 187Re–187Os systematics of abyssalperidotites. Earth Planet. Sci. Lett. 177(3–4), 319–335.

Burnham O. M., Rogers N. W., Pearson D. G., van Calsteren P.W. and Hawkesworth C. J. (1998) The petrogenesis of theeastern pyrenean peridotites: An integrated study of their wholerock geochemistry and Re–Os isotope composition. Geochim.

Cosmochim. Acta 62(13), 2293–2310.

Burton K. W., Schiano P., Birck J.-L. and Allegre C. J. (1999)Osmium isotope disequilibrium between mantle minerals in aspinel-lherzolite. Earth Planet. Sci. Lett. 172(3–4), 311–322.

BVSP (Basaltic Volcanism Study Project) (1981) Basaltic Volca-

nism on the Terrestrial Planets. New York, Pergamon.Carlson R. W. (2005) Application of the Pt–Re–Os isotopic

systems to mantle geochemistry and gechronology. Lithos 82(3–

4), 249–272.

Carlson R. W., Shirey S. B. and Schonbachler M. (2008)Applications of PGE radioisotope systems in geo- and cosmo-chemistry. Elements 4(4), 239–245.

Choi S. H., Shervais J. W. and Mukasa S. B. (2008) Supra-subduction and abyssal mantle peridotites of the Coast Rangeophiolite, California. Contrib. Mineral. Petrol. 156, 551–576.

Choi S. H., Suzuki K., Mukasa S. B., Lee J.-I. and Jung H. (2010)Lu–Hf and Re–Os systematics of peridotite xenoliths fromSpitsbergen, western Svalbard: implications for mantle-crustcoupling. Earth Planet. Sci. Lett. 297, 121–132.

Choukroune P. (1992) Tectonic evolution of the pyrenees. Ann.

Rev. Earth Sci. 20, 143–158.

Conquere F. and Fabries J. (1984) Chemical disequilibrium and itsthermal significance in spinel-peridotites from the Lherz andFreychinede ultramafic bodies. In Kimberlites II: The Mantle

and Crust–Mantle Relationships (ed. J. Kornprobst). FrenchPyrenees, Ariege.

Dantes C., Ceuleneer G., Gregoire M., Python M., Freydier R.,Warren J. and Dick H. J. B. (2007) Pyroxenites from theSouthwest Indian Ridge, 9-16 �E, Cumulates from incremen-tal melt fractions produced at the top of a cold meltingregime. J. Petrol. 48(4), 647–660.

Den Tex E. (1969) Origin of ultramafic rocks, their tectonic settingand history: a contribution to the discussion of the paper“origin of ultramafic and ultrabasic rocks” by P.J. Wyllie.Tectonophysics 7(5–6), 457–488.

Dick H. J. B. and Sinton J. M. (1979) Compositional layering inalpine peridotites: evidence for pressure solution creep in themantle. J. Geol. 87(4), 403–416.

Dick H. J. B. and Fisher R. L. (1983). Mineralogic studies of theresidues of mantle melting: abyssal and alpine-type peridotites.In Kimberlites II: The mantle and crust-mantle relationships (ed.J. Kornprobst).

Dick H. J. B. and Bullen T. (1984) Chromian spinel as a petrogeneticindicator in abyssal and alpine-type peridotites and spatiallyassociated lava. Contrib. Mineral. Petrol. 86(1), 54–76.

Dick H. J. B., Lissenberg C. J. and Warren J. M. (2010) Mantlemelting, melt transport, and delivery beneath a slow-spreadingridge: The Paleo-MAR from 23 150N to 23 450N. J. Petrol.

51(1–2), 425–467.

Dijkstra A. H., Sergeev D. S., Spandler C., Pettke T., Meisel T. andCawood P. A. (2010) Highly refractory peridotites on Mac-quarie Island and the case for anciently depleted domains in theEarth’s mantle. J. Petrol. 51(1–2), 469–493.

Downes H. (2001) Formation and modification of the shallow sub-continental lithospheric mantle: a review of geochemicalevidence from ultramafic xenolith suites and tectonicallyemplaced ultramafic massifs of western and central Europe. J.

Petrol. 42(1), 233–250.

Downes H. (2007) Origin and significance of spinel and garnetpyroxenites in the shallow lithospheric mantle: Ultramaficmassifs in orogenic belts in Western Europe and NW Africa.Lithos 99, 1–24.

Downes H., Bodinier J. -L., Thirlwall M. F., Lorand J. -P. andFabries J. (1991) REE and Sr–Nd isotopic geochemistry ofeastern pyrenean peridotite massifs: sub-continental lithospher-ic mantle modified by continental magmatism. J. Petrol. 97–115(Special Lherzolites Issue).

Eggins S. M., Rudnick R. L. and McDonough W. F. (1998) Thecomposition of peridotites and their minerals: a laser ablationICP–MS study. Earth Planet. Sci. Lett. 154, 53–71.

Ewart A., Collerson K. D., Regelous M., Wendt J. I. and Niu Y.(1998) Geochemical evolution within the Tonga–Kermadec–Lau arc-back-arc systems: the role of varying mantle wedgecomposition in space and time. J. Petrol. 39(3), 331–368.

Fabries J., Bodinier J. L., Dupuy C., Lorand J.-P. and BenkerrouC. (1989) Evidence for modal metasomatism in the orogenicspinel lherzolite body from Cassou (Northeastern Pyrenees,France). J. Petrol. 30(1), 199–228.

Fabries J., Lorand J.-P., Bodinier J.-L. and Dupuy C. (1991)Evolution of the upper mantle beneath the Pyrenees: evidencefrom orogenic spinel lherzolite massifs. J. Petrol. 2, 55–76

(Special Orogenic Lherzolites Issue).

Fabries J., Lorand J.-P. and Bodinier J.-L. (1998) Petrogeneticevolution of orogenic lherzolite massifs in the central andwestern Pyrenees. Tectonophys 292, 145–167.

Fabries J., Lorand J.-P. and Guiraud M. (2001) Petrogenesis of theamphibole-rich veins from Lherz orogenic lherzolite massif(Eastern Pyrenees, France): a case study for the origin oforthopyroxene-bearing pyroxenites in the lithopsheric mantle.Contrib. Mineral. Petrol. 140, 383–403.

6178 A.J.V. Riches, N.W. Rogers / Geochimica et Cosmochimica Acta 75 (2011) 6160–6182

Author's personal copy

Frey F. A., Suen C. J. and Stockman H. W. (1985) The Ronda hightemperature peridotite: geochemistry and petrogenesis. Geo-

chim. Cosmochim. Acta 49, 2469–2491.

Gaetani G. A. and Grove T. L. (1998) The influence of water onmelting of mantle peridotite. Contrib. Mineral. Petrol. 131(4),

323–346.

Gaetani G. A., Kent A. J. R., Grove T. L., Hutcheon I. D. andStolper E. M. (2003) Mineral/melt partitioning of traceelements during hydrous periodtite partial melting. Contrib.

Mineral. Petrol. 145, 391–405.

Gao S., Rudnick R. L., Carlson R. W., McDonough W. F. and LiuY.-S. (2002) Re–Os evidence for replacement of ancient mantlelithosphere beneath the North China craton. Earth Planet. Sci.

Lett. 198(3–4), 307–322.

Garcia-Sansegundo J., Poblet J., Alonso J. L. and Clariana P.(2011) Hinterland-foreland zonation of the Variscan orogen inthe Central Pyrenees: comparison with the northern part of theIberian Variscan Massif. Geol. Soc. Lond. Spec. Publ. 349, 169–

184.

Garrido C. J., Bodinier J.-L. and Alard O. (2000) Incompatibletrace element partitioning and residence in anhydrous spinelperiodtites and websterites from the Ronda orogenic peridotite.Earth Planet. Sci. Lett. 181(3), 341–358.

Green D. H., Hibberson W. O. and Jaques A. L. (1979)Petrogenesis of mid-ocean ridge basalts. In The Earth: Its

Origin, Structure and Evolution (ed. M. W. McElhinny). pp.265–299.

Grove T. L., Chatterjee N., Parman S. W. and Medard E. (2006)The influence of H2O on mantle wedge melting. Earth Planet.

Sci. Lett. 249, 74–89.

Handler M. R., Bennett V. C. and Esat T. M. (1997) Thepersistence of off-cratonic lithospheric mantle: Os isotopicsystematics of variably metasomatised southeast Australianxenoliths. Earth Planet. Sci. Lett. 151(1–2), 61–75.

Hart S. R. and Zindler A. (1986) In search of a bulk Earthcomposition. Chem. Geol. 57, 247–267.

Harvey J., Gannoun A., Burton K. W., Rogers N. W., Alard O.and Parkinson I. J. (2006) Ancient melt extraction from theoceanic upper mantle revealed by Re–Os isotopes in abyssalperidotites from the Mid-Atlantic ridge. Earth Planet. Sci. Lett.

244, 606–621.

Harvey J., Gannoun A., Burton K. W., Schiano P., Rogers N. W.and Alard O. (2010) Unravelling the effects of melt depletionand secondary infiltration on mantle Re–Os isotopes beneaththe French Massif Central. Geochim. Cosmochim. Acta 74(1),

293–320.

Harvey J., Dale C. W., Gannoun A. and Burton K. W. (2011)Osmium mass balance in peridotite and the effects of mantle-derived sulfides on basalt petrogenesis. Geochim. Cosmochim.

Acta 74(19), 5443–5462.

Hellebrand E., Snow J. E., Hoppe P. and Hofmann A. W. (2002)Garnet-field melting and late-stage refertilization in ‘residual’abyssal perdiotites from the Central Indian Ridge. J. Petrol. 43,

2305–2338.

Hellebrand E., Snow J. E., Mostefaoui S. and Hoppe P. (2005)Trace element distribution between orthopyroxene and clino-pyroxene in peridotites from the Gakkel Ridge: a SIMS andNanoSIMS study. Contrib. Mineral. Petrol. 150, 486–504.

Henry P., Azambre B., Montigny R., Rossy M. and Stevenson R.K. (1998) Late mantle evolution of the Pyrenean sub-continen-tal lithospheric mantle in light of new 40Ar–39Ar and Sm–Ndages on pyroxenites and peridotites (Pyrenees, France). Tec-

tonophysics 296(1–2), 103–123.

Herzberg C. (2004) Geodynamic information in peridotite petrol-ogy. J. Petrol. 45(12), 2507–2530.

Hirose K. and Kushiro I. (1993) Partial melting of dry peridotitesat high pressures, Determination of compositions of meltssegregated from peridotite using aggregates of diamond. Earth

Planet Sci. Lett. 114(4), 477–489.

Hirschmann M. M. (2000). Mantle solidus: experimental con-straints and the effects of peridotite composition. Geochem.

Geophys. Geosyst. 1(10), 1042, doi: 10.1029/2000GC000070.Holzheid A. and Grove T. L. (2002) Sulfur saturation limits in

silicate melts and their imlications for core formation scenariosfor terrestrial planets. Am. Min. 87(2–3), 227–237.

Ionov D. A. (2010) Petrology of mantle wedge lithosphere: newdata on supra-subduction zone peridotite xenoliths from theAndesitic Avacha Volcano, Kamchatka. J. Petrol. 51(1–2),

327–361.

Irving A. J. (1980) Petrology and geochemistry of compositeultramafic xenoliths in alkalic basalts and implications formagmatic processes within the mantle. Am. J. Sci. 280(A), 289–

426.

Ishikawa A., Pearson D. G. and Dale C. W. (2011) Ancient Osisotope signatures from the Ontong Java Plateau lithosphere:tracing lithospheric accretion history. Earth Planet. Sci. Lett.

301(1–2), 159–170.

Jagoutz E., Palme H., Baddenhausen H., Blum K., Cendales M.and Dreibus G. (1979) The abundances of major, minor andtrace elements in the Earth’s mantle as derived from primitiveultramafic nodules. 10th Lunar Planet Sci. Conf. 2, 2031–2050.

Johnson K. T. M. and Dick H. J. B. (1992) Open system meltingand temporal and spatial variation of peridotite and basalt atthe Atlantis II fracture zone. J. Geophys. Res. 97(B6), 9219–

9241.

Johnson K. T. M., Dick H. J. B. and Shimizu N. (1990) Melting inthe oceanic upper mantle: an ion microprobe study of diopsidesin Abyssal peridotites. J. Geophys. Res. 95(B3), 2661–2678.

Jugo P. J. (2009) Sulfur content at sulfide saturation in oxidizedmagmas. Geology 37(5), 415–418.

Kelemen P. B., Shimizu N. and Salters V. J. M. (1995) Extractionof mid-ocean-ridge basalt from the upwelling mantle by focusedflow of melt in dunite channels. Nature 375, 747–753.

Kelemen P. B., Hart S. R. and Bernstein S. (1998) Silica enrichmentin the continental upper mantle via melt/rock reaction. Earth

Planet. Sci Lett. 164, 387–406.

Kinzler R. J. (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition: application to mid-ocean ridge basalt petrogenesis. J. Geophys. Res. 102(B1), 853–

874.

Langmuir C., Bezos A., Escrig S. and Parman S. W. (2006)Chemical systematics and hydrous melting of the mantle inback-arc basins. Geophys. Mono. 166, 87–146.

Lenoir X., Garrido C. J., Bodinier J.-L. and Dautria J.-M. (2000)Contrasting lithospheric mantle domains beneath the MassifCentral (France) revealed by geochemistry of peridotite xeno-liths. Earth Planet. Sci. Lett. 181(3), 359–375.

Lenoir X., Garrido C. J., Bodinier J.-L., Dautria J.-M. andFernando G. (2001) The recrystallization front of the Rondaperidotite: evidenc for melting and thermal erosion of subcon-tinental lithospheric mantle beneath the alboran basin. J.

Petrol. 42(1), 141–158.

Le Roux V., Bodinier J.-L., Tommasi A., Alard O., Dautria J.-M.,Vauchez A. and Riches A. J. V. (2007) The lherz spinellherzolite: refertilised rather than pristine mantle. Earth Planet.

Sci. Lett. 259(3–4), 599–612.

Le Roux V., Tommasi A. and Vauchez A. (2008) Feedbackbetween melt percolation and deformation in an exhumedlithosphere–asthenosphere boundary. Earth Planet. Sci. Lett.

274(3–4), 401–413.

Petrogenesis of the Lherz peridotite 6179

Author's personal copy

Le Roux V., Bodinier J.-L., Alard O., O’Reilly S. Y. and Griffin W.L. (2009) Isotopic decoupling during porous melt flow: a case-study in the Lherz peridotite. Earth Planet. Sci. Lett. 279(1–2),

76–85.

Lee C.-T., Harbert A. and Leeman W. P. (2007) Extension oflattice strain theory to mineral/mineral rare-earth elementpartitioning: an approach for assessing disequilibrium anddeveloping internally consistent partition coefficients betweenolivine, orthopyroxene, clinopyroxene and basaltic melt. Geo-

chim. Cosmochim. Acta 71(12), 481–496.

Liew T. C. and Hofmann A. W. (1988) Precambrian crustalcomponents, plutonic associations, plate environment of theHercynian fold belt of central Europe: indications from a Ndand Sr isotopic study. Contrib. Mineral. Petrol. 98(2), 129–138.

Liu C.-Z., Snow J. E., Hellebrand E., Brugmann G., von der HandtA., Buchl A. and Hofmann A. W. (2008) Ancient, highlyheterogeneous mantle beneath Gakkel ridge, Arctic Ocean.Nature 452, 311–316.

Liu C.-Z., Snow J. E., Brugmann G., Hellebrand E. and HofmannA. W. (2009) Non-chondritic HSE budget in Earth’s uppermantle evidenced by abyssal peridotites from Gakkel ridge(Arctic Ocean). Earth Planet. Sci. Lett. 283(1–4), 122–132.

Liu J., Rudnick R. L., Walker R. J., Gao S., Wu F. and Piccoli P.M. (2010) Processes controlling highly siderophile elementfractionations in xenolithic peridotites and their influence onosmium isotopes. Earth Planet. Sci. Lett. 297(1–2), 287–297.

Liu J., Rudnick R. L., Walker R. J., Gao S., Wu F.-Y., Piccoli P.M., Yuan H., Xu W.-L. and Xu Y.-G. (2011) Mappinglithospheric boundaries using Os isotopes of mantle xenoliths:an example from the North China Craton. Geochim. Cosmo-

chim. Acta 75(13), 3881–3902.

Lorand J.-P. (1989) Abundance and distribution of Cu–Fe–Nisulfides, sulfur, copper and platinum-group elements in oro-genic-type spinel lherzolite massifs of Ariege (northeasternPyrenees, France). Earth Planet. Sci. Lett. 93(1), 50–64.

Lorand J.-P. (1991) Sulphide petrology and sulphur geochemistryof orogenic lherzolites: a comparative study of the Pyreneanbodies (France) and the Lanzo Massif (Italy). J. Petrol. 2, 1111–

1140 (Special Orogenic Lherzolites Issue).

Lorand J.-P. and Gregoire M. (2010) Petrogenesis of Fe–Ti oxidesin amphibole-rich veins from the Lherz orogenic peridotite(Northeastern Pyrenees, France). Contrib. Mineral. Petrol.

160(1), 99–113.

Lorand J.-P. and Alard O. (2011) Pyrite tracks assimilation of crustalsulfur in Pyrenean peridotites. Mineral. Petrol. 101(1), 115–128.

Lorand J.-P., Alard O. and Luguet A. (2010) Platinum-groupelement micronuggets and refertilization process in Lherzorogenic peridotite (northeast Pyrenees, France). Earth Planet.

Sci. Lett. 289(1–2), 298–310.

Ludwig K. R. (2001) Isoplot 3.0-A geochronological Toolkit for

Microsoft Excel: Special Publication No. 4. Berkeley Geochro-nological Center, California, 71.

Luguet A., Shirey S. B., Lorand J.-P., Horan M. F. and Carlson R.W. (2007) Residual platinum-group minerals from highlydepleted harzburgites of the Lherz massif (France) and theirrole in HSE fractionation of the mantle. Geochim. Cosmochim.

Acta 71(12), 3082–3097.

Malaviarachchi S. P. K., Makishimia A. and Nakamura E. (2010)Melt-peridotite reactions and fluid metasomatism in the uppermantle, revealed from geochemistry of peridotite and gabbrofrom the horoman peridotite massif, Japan. J. Petrol. 51(7),

1417–1445.

Mavrogenes J. A. and O’Neill H. St. C. (1999) The relative effectsof pressur, temperature, and oxygen fugacity on the solubilityof sulfide in mafic magmas. Geochim. Cosmochim. Acta 63(7–8),

1173–1180.

McCann, T. (ed.) (2008a) The Geology of Central Europe, Vol. 1,

Precambrian and Palaeozoic. Geological Society of London.McCann, T. (ed.) (2008b) The Geology of Central Europe, Vol. 2,

Mesozoic and Cenozoic. Geological Society of London.McDade P., Blundy J. D. and Wood B. J. (2003) Trace element

partitioning between mantle wedge peridotite and hydrousMgO-rich melt. American Min. 88(11-12), 1825–1831.

McDonough W. F. and Sun S.-s. (1995) The composition of theEarth. Chem. Geol. 120, 223–253.

McKenzie D. (1985) The extraction of magma from the crust andmantle. Earth Planet. Sci. Lett. 74(1), 81–91.

McKenzie D. and Bickle M. J. (1988) The volume and compositionof melt generated by extension of the lithosphere. J. Petrol.

29(3), 625–679.

McPherson E., Thirlwall M. F., Parkinson I. J., Menzies M. A.,Bodinier J.-L., Woodland A. and Bussod G. (1996) Geochem-istry of metasomatism adjacent to amphibole-bearing veins inthe Lherz peridotite massif. Chem. Geol. 134, 135–157.

Meisel T., Walker R. J., Irving A. J. and Lorand J.-P. (2001)Osmium isotopic compositions of mantle xenoliths: a globalperspective. Geochim. Cosmochim. Acta 65(8), 1311–1323.

Menzies M. A. and Dupuy C. (1991) Orogenic massifs: protolith,process and provenance. J. Petrol. 1–16 (Special LherzolitesIssue).

Michael P. J., Langmuir C. H., Dick H. J. B., Snow J. E., GoldsteinS. L., Graham D. W., Lehnert K., Kurras G., Jokat W., MuheR. and Edmonds H. N. (2003) Magmatic and amagmaticseafloor generation at the ultraslow-spreading Gakkel ridge,Arctic Ocean. Nature 423, 956–961.

Montigny R., Azambre B., Rossy M. and Thuizat R. (1986) K–Arstudy of Cretaceus magmatism and metamorphism in thePyrenees: age and length of rotation of the Iberian Peninsula.Tectonophysics 129, 257–273.

Morgan J. P. and Morgan J. W. (1999) Two-stage melting and thegeochemical evolution of the mantle: a recipe for mantle plum-pudding. Earth Planet. Sci. Lett. 170(3), 215–239.

Mukasa S. B., Shervias J. W., Wilshire H. G. and Nielson J. E.(1991) Intrinsic Nd, Pb, and Sr isotopic heterogeneities exhib-ited by the Lherz alpine peridotite massif, French Pyrenees. J.

Petrol. 117–134 (Special Orogenic Lherzolites Issue).Muntener O. and Manatschal G. (2006) High degrees of melt

extraction recorded by spinel harzburgite of the Newfoundlandmargin: the role of inheritance and consequences for theevolution of the southern North Atlantic. Earth Planet. Sci.

Lett. 252, 437–452.

Muntener O., Kelemen P. B. and Grove T. L. (2001) The role ofH2O during crystallisation of primitive arc magmas underuppermost mantle conditions and genesis of igneous pyroxe-nites: an experimental study. Contrib. Mineral. Petrol. 141, 643–

658.

Muntener O., Manatschai G., Desmurs L. and Pettke T. (2010)Plagioclase peridotites in Ocean-continent transitions: refertil-ized mantle domains generated by melt stagnation in theshallow mantle lithosphere. J. Petrol. 51(1–2), 255–294.

O’Hara Y. and Ishii R. (1998) Peridotites from the southernMariana forearc: heterogeneous fluid supply in mantle wedge.Island Arc 7(3), 541–558.

Palme H. and Nickel K. G. (1985) Ca/Al ratio and composition ofthe Earth’s upper mantle. Geochim. Cosmochim. Acta 49(10),

2123–2132.

Parkinson I. J. and Pearce J. A. (1998) Peridotites from the Izu–Bonin–Mariana Forearc (ODP Leg 125): evidence for mantlemelting and melt–mantle interaction in a supra-subductionzone setting. J. Petrol. 39(9), 1577–1618.

Parkinson I. J., Pearce J. A., Thirlwall M. F., Johnson K. T. M.and Ingram G. (1992) Trace element geochemistry of perido-

6180 A.J.V. Riches, N.W. Rogers / Geochimica et Cosmochimica Acta 75 (2011) 6160–6182

Author's personal copy

tites from the Izu–Bonin–Mariana Forearc, Leg 125. Proc.

Ocean Drilling Prog. Sci. Results 125, 487–506.

Parkinson I. J., Hawkesworth C. J. and Cohen A. S. (1998) Ancientmantle in a modern arc: osmium isotopes in Izu–Bonin–Mariana forearc peridotites. Science 281(5385), 2011–2013.

Parman S. W. and Grove T. L. (2004). Harzburgite melting withand without H2O: experimental data and predictive modelling.J. Geophys. Res. 109, BO2201, doi: 10.1029/2003JB002566.

Pearce J. A. and Parkinson I. J. (1993) Trace element models formantle melting: application to volcanic arc petrogenesis. Geol.

Soc. Lond. Spec. Publ. 76, 373–403.

Pearce J. A., Lippard S. J. and Roberts S. (1984) Characteristicsand tectonic significance of supra-subduction zone ophiolites.Geol. Soc. Lond. Spec. Publ. 16, 77–94.

Pearce J. A., Barket P. F., Edwards S. J., Parkinson I. J. and LeatP. T. (2000) Geochemical and tectonic significance of perido-tites from the South Sandwich arc-basin, South Atlantic.Contrib. Mineral. Petrol. 139(1), 36–53.

Pearcy L. G., DeBari S. and Sleep N. H. (1990) Mass balancecalculations for two sections of island arc crust and implicationsfor the formation of continents. Earth Planet. Sci. Lett. 96, 427–

442.

Pearson D. G., Irvine G. J., Ionov D. A., Boyd F. R. and DreibusG. E. (2004) Re–Os isotope systematics and platinum groupelement fractionation during mantle melt extraction: a study ofmassif and xenolith peridotite suites. Chem. Geol. 208, 29–59.

Pearson D. G., Parman S. W. and Nowell G. M. (2007) A linkbetween large mantle melting events and continent growth seenin osmium isotopes. Nature 449, 202–205.

Peate D. W., Pearce J. A., Hawkesworth C. J., Colley H., EdwardsC. M. H. and Hirose K. (1997) Geochemical variations inVanuatu arc lavas: the role of subducted material and a variablemantle wedge composition. J. Petrol. 38(10), 1331–1358.

Polve M. and Allegre C. J. (1980) Orogenic lherzolite complexesstudied by 87Rb–87Sr: a clue to understand the mantle convec-tion processes? Earth Planet. Sci. Lett. 51(1), 71–93.

Reisberg L. and Lorand J.-P. (1995) Longevity of sub-continentalmantle lithosphere from osmium isotope systematics in oro-genic peridotite massifs. Nature 376, 159–162.

Reisberg L., Allegre C. J. and Luck J.-M. (1991) The Re–Ossystematics of the Ronda ultramafic complex of southern Spain.Earth Planet. Sci. Lett. 105(1–3), 196–213.

Ringwood A. E. (1991) Phase transformations and their bearing onthe constitution and dynamics of the mantle. Geochim. Cosmo-

chim. Acta 55(8), 2083–2110.

Rudnick R. L., McDonough W. F. and Chappell B. W. (1993)Carbonatite metasomatism of the Tanzanian mantle: Petro-graphic and geochemical characteristics. Earth Planet. Sci Lett.

111(4), 463–475.

Rudnick R. L. and Walker R. J. (2009) Interpreting ages from Re–Os isotopes in peridotites. Lithos 112(2), 1083–1095.

Saal A. E., Takazawa E., Frey F. A., Shimizu N. and Hart S. R.(2001) Re–Os isotopes in the horoman peridotite: evidence forrefertilization? J. Petrol. 42(1), 25–37.

Sautter V. and Fabries J. (1990) Cooling kinetics of garnetwebsterites from the Freychinede orogenic lherzolite massif,French Pyrenees. Contrib. Mineral. Petrol. 105(5), 533–549.

Schiano P. and Clocchiatti R. (1994) Worldwide occurrence ofsilicate-rich melts in sub-continental and sub-oceanic mantleminerals. Nature 368, 621–624.

Schiano P., Clocchiatti R., Shimizu N., Maury R. C., Jochum K. P.and Hofmann A. W. (1995) Hydrous, silica-rich melts in thesub-arc mantle and their relationship with erupted arc lavas.Nature 377, 595–600.

Shervais J. W. (1982) Ti-V plots and the petrogenesis of modernand ophiolitic lavas. Earth Planet. Sci. Lett. 59, 101–118.

Shirey S. B. and Walker R. J. (1998) The Re–Os isotope system incosmochemistry and high-temperature geochemistry. Ann. Rev.

Earth Planet. Sci. 26, 423–500.

Simon N. S. C., Neumann E.-R., Bonadiman C., Coltorti M.,Delpech G., Gregoire M. and Widom E. (2008) Ultra-refrac-tory domains in the oceanic mantle lithosphere sampled asmantle Xenoliths at Ocean Islands. J. Petrol. 49(6), 1223–1251.

Smoliar M. I., Walker R. J. and Morgan J. W. (1996) Re–Os agesof group IIA, IIIA, IVA, and IVB iron meteorites. Science 271,

1099–1102.

Stern R. J., Lin P.-N., Morris J., Jackson M. C., Fryer P., BloomerS. H. and Ito E. (1990) Enriched back-arc basin basalts fromnorthern Mariana Trough: implications for magmatic evolutionof back-arc basins. Earth Planet. Sci. Lett. 100, 210–225.

Streckeisen A. (1973) Plutonic rocks: classification and nomencla-ture recommended by the IUS subcommission on the system-atics of igneous rocks. Geotimes 8(10), 26–30.

Stosch H.-G. (1982) Rare earth element partitioning betweenminerals from anhydrous spinel peridotite xenoliths. Geochim.

Cosmochim. Acta 46, 793–811.

Suhr G., Hellebrand E., Snow J. E., Seck H. A. and Hofmann A.W. (2003). Significance of large, refractory dunite bodies in theupper mantle of the Bay of Islands Ophiolite. Geochem.

Geophys. Geosyst. 4, 2001GC000277.Sun S. -s. and McDonough W. F. (1989) Chemical and isotopic

systematics of oceanic basalts: implications for mantle compo-sition and processes. In: Magmatism in the Ocean Basins (eds.A. D. Saunders and M. J. Norry). Geol. Soc. Lond. Spec. Publ.42, 313–345.

Taylor S. R. and McLennan S. M. (1985) The Continental Crust:

Its Composition and Evolution. Blackwell Sci. Publ., Oxford,330.

van Acken D., Becker H. and Walker R. J. (2008) Refertilization ofJurassic oceanic periodties from the tethys ocean–implicationsfor the Re–Os systematics of the upper mantle. Earth Planet.

Sci. Lett. 268(1–2), 171–181.

van Acken D., Becker H., Walker R. J., McDonough W. F.,Wombacher F., Ash R. D. and Piccoli P. M. (2010) Formationof pyroxenite layers in the Totalp ultramafic massif (Swiss Alps)– insights from highly siderophile elements and Os isotopes.Geochim. Cosmochim. Acta 74(2), 661–683.

Van der Wal D. and Bodinier J.-L. (1996) Origin of therecrystallisation front in the Ronda peridotite by km-scalepervasive porous melt flow. Contrib. Mineral. Petrol. 122, 387–

405.

Vernieres J., Godard M. and Bodinier J. –L. (1997) A plate modelfor the simulation of trace-element fractionation during partialmelting and magma transport in the Earth’s upper mantle. J.

Geophys. Res. 102(B11), 24771–24784.

Vetil J. Y., Lorand J. P. and Fabries J. (1988) Conditions de miseen place des filons de pyroxenites a amphibole du massifultramafique de Lherz (Ariege, France). C. R. Acad. Sci. Paris

Ser. II 307, 587–598.

Villiger S., Ulmer P., Mutener O. and Thompson A. B. (2004) Theliquid line of descent of anhydrous, mantle-derived, tholeiiticliquids by fractional and equilibrium crystallization – anexperimental study at 1.0 GPa. J. Petrol. 45(12), 2369–2388.

Villiger S., Ulmer P. and Muntener O. (2007) Equilibrium andfractional crystallisation experiments at 0.7 GPa; the effect ofpressure on phase relations and liquid compositions of tholeiiticmagmas. J. Petrol. 48(1), 159–184.

Walker R. J., Horan M. F., Morgan J. W., Becker H., GrossmanJ. N. and Rubin A. E. (2002) Comparative 187Re–187Ossystematics of chondrites: implcations regarding early solarsystem processes. Geochim. Cosmochim. Acta 66(23), 4187–

4201.

Petrogenesis of the Lherz peridotite 6181

Author's personal copy

Walter M. J. (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere. J. Petrol. 39, 29–60.

Walter M. J. (2003) Melt extraction and compositional variabilityin mantle lithosphere. In The Mantle and Core, ed. R. W.

Carlson, Vol. 2 Treatise on Geochemistry (eds. H. D. Hollandand K. K. Turekian), pp. 363–394.

Warren J. M. and Shimizu N. (2010) Cryptic variations in abyssalperidotite compositions: evidence for shallow-level melt infil-tration in the oceanic lithosphere. J. Petrol. 51(1–2), 395–423.

Warren J. M., Shimizu N., Sakaguchi C., Dick H. J. B. andNakamura E. (2009) An assessment of upper mantle heteroge-neity based on abyssal peridotite isotopic compositions. J.

Geophys. Res. 114 (B12203), doi:10.1029/2008JB006186.

Wasylenki L. E., Baker M. B., Kent A. J. R. and Stolper E. M.(2003) Near-solidus melting of shallow upper mantle: partialmelting experiments on depleted peridotite. J. Petrol. 44(7),

1163–1191.

White R. S. (1988) The Earth’s crust and lithosphere. J. Petrol.

Special Lithosphere Issue, 1–10.Witt-Eickschen G. and O’Neill H. St. C. (2005) The effect of

temperature on the equilibrium distribution of trace elementsbetween clinopyroxene, orthopyroxene, olivine and spinel inupper mantle peridotite. Chem. Geol. 221(1–2), 65–101.

Wittig N., Pearson D. G., Baker J. A., Duggen S. and Hoernle K.(2010) A major element, PGE and Re–Os isotope study ofMiddle Atlas (Morocco) peridotite xenoliths: evidence forcoupled introduction of metasomatic sulphides and clinopy-roxene. Lithos 115, 15–26.

Woodhead J., Eggins S. and Gamble J. (1993) High field strengthand transition element systematics in island arc and back-arc

basin basalts: evidence for multi-phase melt extraction and adepleted mantle wedge. Earth Planet. Sci. Lett. 114, 491–504.

Woodland A. B. (1992) Oxygen thermobarometry of orogeniclherzolite massifs. J. Petrol. 33(1), 203–230.

Woodland A. B., Kornprobst J., McPherson E., Bodinier J.-L. andMenzies M. A. (1996) Metasomatic interactions in the litho-spheric mantle: petrological evidence from the Lherz massif,French Pyrenees. Chem. Geol. 134(1–3), 83–112.

Woodland A. B., Kornprobst J. and Tabit A. (2006) Ferric iron inorogenic lherzolite massifs and controls of oxygen fugacity inthe upper mantle. Lithos 89, 222–241.

Workman R. K. and Hart S. R. (2006) Major and trace elementcompositions of the depleted MORB mantle (DMM). Earth

Planet. Sci. Lett. 231, 53–72.

Wyllie P. J. (1969) The origin of ultramafic and ultrabasic rocks.Tectonophysics 7(5–6), 437–455.

Xu Y.-G., Menzies M. A., Thirlwall M. F., Huang X.-L, LiuY. and Chen X.-M. (2003) Reactive harzburgites fromHuinan, Ne China: Products of the lithosphere-astheno-sphere interaction during lithospheric thinning? Geochim.

Cosmochim. Acta. 67(3), 487–505.

Zanetti A., Vannucci R., Bottazzi P., Oberti R. and Ottolini L.(1996) Infiltration metasomatism at Lherz as monitored bysystematic ion-microprobe investigations close to a hornblen-dite vein. Chem. Geol. 134(1–3), 113–133.

Zindler A. and Hart S. (1986) Chemical geodynamics. Ann. Rev.

Earth Planet. Sci. 14, 493–571.

Associate editor: Shichun Huang

6182 A.J.V. Riches, N.W. Rogers / Geochimica et Cosmochimica Acta 75 (2011) 6160–6182