the Tallante Mantle Xenolith Record (Betic Cordillera, SE Spain)

Melt Migration and Intrusion duringExhumation of the Alboran Lithosphere: theTallante Mantle Xenolith Record (BeticCordillera, SE Spain)

E. RAMPONE1*, R. L. M.VISSERS2, M. POGGIO1, M. SCAMBELLURI1

AND A. ZANETTI3

1DIPARTIMENTO PER LO STUDIO DEL TERRITORIO E DELLE SUE RISORSE, UNIVERSITY OF GENOVA, CORSO EUROPA

26, I-16132 GENOVA, ITALY2DEPARTMENT OF EARTH SCIENCES, FACULTY OF GEOSCIENCES, UTRECHT UNIVERSITY, PO BOX 80.021, 3508 TA

UTRECHT, NETHERLANDS3CNR-ISTITUTO DI GEOSCIENZE E GEORISORSE, SEZIONE DI PAVIA, VIA FERRATA 1, I-27100 PAVIA, ITALY

RECEIVED DECEMBER 29, 2008; ACCEPTED AUGUST 12, 2009ADVANCE ACCESS PUBLICATION SEPTEMBER 22, 2009

Microstructural and in situ mineral chemistry studies on mantle

peridotite xenoliths from the Late Neogene alkaline volcanic center

of Cabezo Tallante (SE Spain) reveal an exceptional record of a

multi-stage history of deformation, recrystallization, melt^rock inter-

action and melt intrusion tracking the progressive exhumation of

this lithospheric mantle sector. Xenoliths include porphyroclastic to

equigranular spinel peridotites, impregnated plagioclase peridotites,

and composite xenoliths made up of peridotites intruded first by gab-

bronorite veins and later by amphibole-bearing pyroxenites.The ear-

liest stage involved subsolidus re-equilibration from garnet- to

spinel-facies conditions, represented by rounded opx þ spinel � cpx

clusters indicative of precursor garnet.The spinel-facies equilibration

was followed by development of a porphyroclastic fabric, accentuated

in many xenoliths by spinel trails, in response to shear deformation

that may be related to the early stages of Neogene extension.

Porphyroclastic spinel peridotites subsequently underwent multiple

episodes of reactive porous melt percolation documented by crystalli-

zation of undeformed olivine replacing pyroxene porphyroclasts, and

of undeformed poikilitic orthopyroxene at the expense of both pyrox-

ene porphyroclasts and newly crystallized olivines.The porphyroclas-

tic and melt^rock reaction textures are progressively obliterated

by an equigranular structure developed as the result of static, possi-

bly melt-assisted, annealing recrystallization. Clinopyroxenes in

equigranular peridotites (i.e. the most equilibrated with the percolat-

ing melts) display slight light rare earth element (REE) depletion

and almost flat middle to heavy REE spectra (LaN/YbN¼ 0�37^

0�62; SmN/YbN¼ 0�89^1�23). Computed equilibrium liquids have

an enriched tholeiitic affinity, consistent with the sub-alkaline mag-

matism of the Alboran Domain. Overall, the tectonic and magmatic

stages recorded in spinel peridotites from Tallante are remarkably

consistent with the evolution documented in the Ronda peridotites of

the western Betics. Reactive porous flow and annealing recrystalliza-

tion were followed by an impregnation event, documented by

crystallization of interstitial (plag� opx� ol) aggregates in

porphyroclastic and equigranular xenoliths; this indicates further

exhumation to shallower depths. Diffuse melt percolation was fol-

lowed by intrusion of melts with distinct chemical affinity. The

first event is documented by thin gabbronoritic^noritic veins, show-

ing opx reaction rims against the host peridotite. Comparable gabbro-

norites were previously ascribed to slab-derived melts. The norite

veins are crosscut by centimeter-thick dikelets of amphibole pyroxe-

nite. Geobarometric estimates and the observed crystallization order

(ol^cpxâmph^plag) point to 0�7^0�9 GPa for pyroxenite intru-

sion. Computed melts in equilibrium with clinopyroxene show alka-

line affinity, similar to the host Tallante alkali basalts. Textural

and geochemical features in the xenoliths thus indicate that the

*Correspondingauthor.Telephone: 0039103538315. Fax: 003910352169.E-mail: [email protected]

� The Author 2009. Published by Oxford University Press. Allrights reserved. For Permissions, please e-mail: [email protected]

JOURNALOFPETROLOGY VOLUME 51 NUMBERS 1 & 2 PAGES 295^325 2010 doi:10.1093/petrology/egp061D

ownloaded from

https://academic.oup.com

/petrology/article/51/1-2/295/1457502 by guest on 22 June 2022

progressive uplift of theTallante lithospheric mantle was accompa-

nied by interaction with melts of different sources, reflecting the mag-

matic evolution of the Alboran Domain in response to lithosphere

extension and thinning leading to the formation of the Betic^Rif arc.

KEY WORDS: mantle xenoliths; Alboran Domain; lithosphere

extension; reactive porous flow; melt impregnation

I NTRODUCTIONSince the latest Oligocene the western Mediterraneanregion has experienced a complex geodynamic evolution,involving lithosphere extension and development of theAlboran Sea whilst the bounding African and Eurasianplates formed an essentially convergent setting.Extensional thinning of the Alboran lithosphere was mostprobably accompanied by the ascent of hot asthenosphereand consequent lithosphereâsthenosphere interaction.This tectonic context invites geological and geochemicalstudies of the pertinent upper mantle rocks, aiming to elu-cidate the details of these processes and their bearing onthe geodynamics of the region.The Betic Cordillera of southern Spain, together with

the Rif, Tell and Kabylies chains of Morocco, Algeria andTunisia, form part of a tight arc-shaped mountain beltmaking up the westernmost part of the Alpine orogenicsystem (Fig. 1). The external parts of the belt represent theSouth Iberian and North African passive marginsequences, which were strongly deformed during Neogenefolding and thrusting (Garc|¤a-Herna¤ ndez et al., 1980;Banks & Warburton, 1991). The inner parts of the BeticCordillera and Rif Mountains are made up of anallochthonous pile of intensely deformed, mostly metamor-phic rocks; they are considered as the relics of an earlyAlpine collisional system formed during Late Mesozoic toTertiary convergence between Africa and Iberia, subse-quently exhumed and strongly dismembered duringNeogene late-orogenic extension (Platt & Vissers, 1989;Garc|¤a-Duen‹ as et al., 1992; Lonergan & White, 1997;Comas et al., 1999; Jolivet & Faccenna, 2000). These rocksof the Betic and Rif internal zones, intensely affected byNeogene extension, are often referred to as the AlboranDomain (Garc|¤a-Duen‹ as et al., 1992). The present-day crus-tal thickness in the internal parts of this domain variesfrom 20^25 km to less than 10 km, with minimum thick-nesses towards the east at the transition from the AlboranBasin to the South Balearic Basin (Torne et al., 2000). Inresponse to Neogene lithospheric extension, the AlboranSea region has been affected by widespread magmaticactivity involving eruption of tholeiitic, calc-alkaline andshoshonitic magmas, followed by Late Neogene alkalinebasalts. The region thus records the post-collisional transi-tion from subduction- to intraplate-type magmatism

(Turner et al., 1999; Coulon et al., 2002; Duggen et al.,2003, 2004, 2005, 2008).Two major groups of models have been proposed to

explain the complex geodynamic and magmatic evolutionof the western Mediterranean region: (1) non-subductionmodels, involving delamination or convective removal ofgravitationally unstable, thickened subcontinental litho-sphere beneath the Alboran Domain (Platt & Vissers,1989; Docherty & Banda, 1995; Platt et al., 1996; Comaset al., 1999; Turner et al., 1999; Doblas et al., 2007); (2) sub-duction models, involving subduction of Tethyan oceaniclithosphere associated with slab roll-back and steepeningand/or detachment of the subducted slab (Royden, 1993;Lonergan & White, 1997; Zeck et al., 1998; Hoernle et al.,1999; Wortel & Spakman, 2000; Coulon et al., 2002;Gutscher et al., 2002; Duggen et al., 2003, 2004, 2005, 2008;Gill et al., 2004). This latter group of models is supportedby tomographic studies (Wortel & Spakman, 2000;Gutscher et al., 2002; Spakman & Wortel, 2004) providingevidence of an east-dipping slab of cold oceanic lithospheredescending from the Atlantic Domain beneath theAlboran Sea. According to Gutscher et al. (2002) andSpakman & Wortel (2004), Paleogene subduction of theLigurian ocean in the present-day western part of theMediterranean region occurred in a northwesterly direc-tion underneath the Balearic Islands, Sardinia andCorsica. During the latest Oligocene to early Miocene,this subduction system became divided into two distinctsegments after collision between Iberia and North Africa.Subsequently, the western subduction segment initiated awest-directed roll-back to form the Alboran Basin.A general consensus on the geodynamic evolution of the

Alboran Domain does not as yet exist; however, each ofthe currently competing hypotheses involves large-scalemantle processes, and this feature motivates this study ofthe structure and petrology of the upper mantle in ques-tion. Within the Alboran Domain, this upper mantle isaccessible in two different ways; that is, in peridotite mas-sifs exposed in the Betic and Rif internal zones in the wes-ternmost part of the arc, and in the form of mantlexenoliths brought to the surface in southeastern Spain bylate Neogene alkaline as well as potassic to ultrapotassicvolcanic rocks.Large exposures of upper mantle rocks occur in the peri-

dotite massifs of the Sierra Bermeja, Alpujata andCarratraca near Ronda (western Betics), and in the BeniBousera massif (internal Rif, north Morocco). Structuraland geochemical studies of the Ronda peridotites of theSierra Bermeja (Van derWal, 1993; Van derWal & Vissers,1993, 1996; Van der Wal & Bodinier, 1996; Garrido &Bodinier, 1999; Lenoir et al., 2001) reveal a structural his-tory marked by early stage exhumation of the peridotitesfrom deep lithospheric levels (Davies et al., 1993), followedby a stage of progressive deformation leading to

JOURNAL OF PETROLOGY VOLUME 51 NUMBERS 1 & 2 JANUARY & FEBRUARY 2010

296

Dow

nloaded from https://academ

ic.oup.com/petrology/article/51/1-2/295/1457502 by guest on 22 June 2022

porphyroclastic and mylonitic microstructures. Thesedeformational structures in turn became overprinted by astage of intense annealing recrystallization in the presenceof percolating melts, prior to emplacement at crustallevels during the Early Miocene (�22 Ma, Priem et al.,1979) facilitated by plagioclase-facies ductile shear zones.In addition, a recent study of the Carratraca massif(Tub|¤a et al., 2004) has revealed the local preservationof early stage granular spinel-facies peridotites withprotogranular orthopyroxene^spinel clusters presumablyderived from precursor garnets. The emerging historyof early stage exhumation and garnet breakdown,followed by deformation and development of a porphyro-clastic microstructure prior to intense melt-assistedrecrystallization has been interpreted to reflect Mesozoicrifting and subsequent lithospheric thickening, prior tointense heating related to late orogenic extension (van derWal & Vissers, 1993, 1996; Vissers et al., 1995; Tub|¤a et al.,2004).

In this study, we present the results of a microstructuraland geochemical study of upper mantle xenoliths from thealkaline lavas of Cabezo Tallante, an eroded cinder coneNW of Cartagena (SE Spain) with Pliocene ArÂr agesranging from 2�93 to 2�29 Ma (Duggen et al., 2005). Thesexenoliths exceptionally preserve microstructural and geo-chemical evidence of a multi-stage history of deformation,recrystallization, melt^rock interaction and melt intrusiontracking the progressive uplift of this lithospheric mantlesector in response to lithosphere extension in the AlboranDomain. In terms of tectonic and magmatic processes, theevolution of the Tallante mantle shows marked similaritiesto that documented in the Ronda peridotites as outlinedabove. A major aim of this study, therefore, is to providefurther insights into the complex geodynamic and mag-matic evolution of the Alboran Domain through a mantleperspective. Specific aims concern (1) the origin and chem-ical affinity of melts migrating through and intruding theperidotites at different lithospheric depths, and (2) the

Fig. 1. Tectonic sketch map of the Betic^Rif arc and Alboran Sea, modified after Meijninger (2006), showing the location of the CabezoTallantevolcanic centre NWof Cartagena. The peridotite massifs amidst rocks of the Alboran Domain in the western part of the arc near Ronda andin northern Morocco should be noted. Ocean Drilling Program (ODP) Site 976 in the western Alboran Sea is indicated.

RAMPONE et al. EVOLUTION OFALBORAN DOMAIN

297

Dow



nature and conditions of these melt^rock interactionsduring progressive mantle uplift. The results of this studyare discussed in the context of current geodynamic scenar-ios for the evolution of the western Mediterranean region.

SAMPLES AND ANALYT ICALPROCEDURESThe investigated xenoliths have decimeter-scale sizes andgenerally do not exceed 10^15 cm. They comprise (1) clino-pyroxene-poor spinel lherzolites (cpx 510 vol. %), (2)plagioclase-bearing spinel peridotites, and (3) compositexenoliths made up of spinel and/or plagioclase peridotitesintruded by olivineâmphibole-bearing pyroxenites. Themost relevant microstructural relations are shown in Figs2^5. Xenoliths show little microstructural evidence ofinteraction with the host basalts. This is confined to rareand very thin (550 mm) glassy fractures mostly occurringalong grain boundaries or eventually crosscutting mantleminerals. When in contact with spinel, they cause tinydark reaction rims. These sites have, however, been care-fully avoided, and all the studied minerals were cleangrains, with no evidence of interaction with the hostbasalt. On the other hand, we specifically investigated theinteraction between peridotite and intruded amphibole-pyroxenite veins in composite xenoliths.Clinopyroxene-poor spinel lherzolites display variable

microstructures, mostly represented by porphyroclastic toequigranular types. The microstructure is: (1) dominantlyporphyroclastic with large to medium grain sizes in spinelperidotites 92T4, 92T6, 92T19 (Fig. 2a^d), 92T17 (Fig. 3f),and 92T21; (2) porphyroclastic to granular in samples92T3, 92T7 andT30; (3) fine-grained equigranular in sam-ples 92T18 and 92T20 (Figs 2e and 3b, c) and rathercoarse-grained equigranular in sampleT18.Plagioclase-bearing peridotites are clinopyroxene-poor

spinel lherzolites with variable amounts of plagioclase,mostly crystallized in interstitial aggregates, in associationwith olivine and/or orthopyroxene (Fig. 4a and b). Likethe spinel peridotites, the plagioclase-bearing peridotitesexhibit variable microstructure: (1) dominantly porphyro-clastic in samples 92T1, 92T9A,T11,T13 andT31A; (2) por-phyroclastic to granular in sample T32A; (3) tabularequigranular in samples 92T14 and 92T23. In mostsamples, the plagioclase modal abundance is lower than5 vol. %, whereas it occurs in higher amounts (up to10 vol. %) in samples 92T1 and 92T23. In peridotite92T14, plagioclase occurs in very low modal amounts(53 vol. %), mostly as thin rims around spinel.Spinel peridotites T18 and 92T20, and plagioclase-

bearing peridotites T11 and T13 contain thin (51cm thick)gabbronoritic to noritic veins, mostly consisting of ortho-pyroxene, plagioclase, and subordinate clinopyroxene.They crosscut the peridotite fabric and show a reaction

margin made of fine-grained orthopyroxene towards thehost peridotite (Fig. 4c). Similar gabbronoritic lithotypeshave been documented in previous studies on the Tallantexenoliths (Arai et al., 2003; Beccaluva et al., 2004; Shimizuet al., 2004, 2008) and ascribed to the intrusion of slab-derived melts. Here we describe the main microstructuralfeatures of the gabbronorites, which allow us to positionthis intrusion event in the context of the multi-stage meltinteractionîntrusion history recorded by the studied peri-dotite xenoliths. Detailed chemical investigations of thegabbronorite veins will be subject of a separate paper(Rampone et al., in preparation).Plagioclase-bearing peridotites 92T9A,T31A,T32A and

T13 are parts of composite xenoliths, consisting of countryperidotite and intruded olivineâmphibole-bearing pyrox-enite (samples 92T9B, T31B and T32B, respectively). Thepyroxenites display a rather sharp contact against the hostperidotites, and cut at high angles across the peridotite foli-ation defined by spinel trails (Fig. 5a and b).Major and trace element mineral chemistry data have

been obtained from a selected number of samples (spinelperidotites 92T20, T30, 92T6, 92T7, 92T19; plagioclase-bearing peridotite 92T1, composite xenoliths 92T9A-B,T31A-B), representative of all microstructural types(Tables 1^6). Mineral major element compositions wereanalysed using: (1) a Philips SEM 515 equipped with anX-ray dispersive analyser (accelerating potential 15 kV,beam current 20 nA), at the Dipartimento per lo Studiodel Territorio e delle sue Risorse, University of Genova,and (2) a JEOL JXA 8200 Superprobe equipped with fivewavelength-dispersive (WDS) spectrometers, an energy-dispersive (EDS) spectrometer, and a cathodolumines-cence detector (accelerating potential 15 kV, beam current15 nA), at the Dipartimento di Scienze della Terra,University of Milano. In situ trace element mineral analyseswere carried out by laser ablation inductively coupledplasma mass spectrometry (LA-ICP-MS) techniques atIGG-CNR in Pavia. Detailed descriptions of these analyti-cal procedures have been given by Tiepolo et al. (2003),Miller et al. (2007) and Rampone et al. (2008a).

MICROSTRUCTURESSpinel peridotitesPorphyroclastic spinel peridotites consist of large (milli-meter-size) olivine and orthopyroxene, and smallerclinopyroxene and spinel grains. Accessory amounts(51 vol. %) of phlogopite and amphibole have beenobserved in a few samples. Pyroxenes often display a pre-ferred orientation of the cleavage planes and exsolutionlamellae or blebs, whereas olivines are kinked, thus indi-cating internal plastic deformation of the primary mineralassemblage (Fig. 2a and b). Spinel occurs both as singlegrains and in rounded clusters, in association with ortho-pyroxene and subordinate clinopyroxene and olivine


298

Dow



Fig. 2. Microstructures in spinel peridotites. (a) Porphyroclastic spinel peridotite 92T6. Large kinked olivine porphyroclast (OLp) (cross-polarized light). (b) Porphyroclastic spinel peridotite 92T19. Large, exolved orthopyroxene porphyroclast partly replaced by new olivine(cross-polarized light). (c) Porphyroclastic spinel peridotite 92T19. Close-up view of (b), showing new underformed olivine grains (OLn)partly replacing, with lobate contacts, pyroxene (OPXp, CPXp) and olivine (OLp) porphyroclasts (cross-polarized light).(d) Porphyroclastic spinel peridotite 92T4. New unstrained poikilitic orthopyroxene (OPXn) partly replacing both new (OLn) and kinked por-phyroclast (OLp) olivine grains (cross-polarized light). (e) Equigranular spinel peridotite 92T18. Equigranular olivine, orthopyroxene, clino-pyroxene grains showing triple-point junctions (cross-polarized light). (f) Equigranular spinel peridotite 92T14. Orthopyroxene (opx)þ spinel(sp) rounded cluster surrounded by finer-grain equigranular matrix (plane-polarized light).


299

Dow



(e.g. 92T14; Figs 2f and 3a). In a recent study by Shimizuet al. (2008), similar clusters in theTallante peridotite xeno-liths have been interpreted to result from garnet break-down. Spinel^pyroxene clusters show variable shapes,ranging from approximately spherical and ellipsoidal insome xenoliths (92T14) to strongly ellipsoidal and elongate(e.g. 92T18); many xenoliths show conspicuous trails ofspinel suggesting that the clusters have been deformed tovery high strains (e.g. 92T18, 92T17, Fig. 3c and d; 92T3).This interpretation is supported by a distinct preferred ori-entation of the orthopyroxene [100] cleavage planes inmany of the stretched clusters. The highly deformed clus-ters contribute to defining a foliation in the rock, which ischaracterized by a spinel lineation in the plane of the folia-tion and is often accentuated by the elongation of olivineand orthopyroxene.To investigate the deformation associated with the spher-

ical to ellipsoidal clusters, we measured the olivine latticepreferred orientation (LPO) in plagioclase-bearing spinelperidotite 92T14; this is characterized by the occurrenceof such clusters and a tabular equigranular microstructure.

The result is shown in Fig. 6 (top panel). The LPO can beclassified as a typical A-type pattern (terminology accord-ing to Jung & Karato, 2001; Karato et al., 2008), with apoint maximum of the [010] axes at high angles to the foli-ation, a maximum of the [100] axes close to the lineation,and a maximum of the [001] axes in the plane of the folia-tion normal to the lineation. Similar fabrics have been pro-duced in simple shear deformation experiments (Zhang &Karato, 1995; Katayama et al., 2004) and are commonlyinterpreted to reflect dominant olivine (010)[100] slipexpected for lithospheric conditions of elevated tempera-tures (around 11008C), low stress (5300MPa) and lowwater content (Karato et al., 2008).The evidence in sample 92T14 for simple shear deforma-

tion led us to select a set of additional xenoliths showingincreasing degrees of deformation of the spinel^pyroxeneclusters, to explore if there are any trends in the LPOsand microstructure associated with this deformation.Figure 6 illustrates the microstructure and olivine LPOsin six xenoliths, shown in qualitative order of increasingflattening of the clusters. The upper one (92T14) described

Fig. 3. Microstructures in spinel peridotites showing progressive deformation of pyroxene^spinel clusters. (a) Equigranular spinel peridotite92T14 (plane-polarized light). (b) Equigranular spinel peridotite 92T20 (plane-polarized light). (c) Equigranular spinel peridotite 92T18(plane-polarized light). (d) Porphyroclastic spinel peridotite 92T17 (plane-polarized light).


300

Dow



Fig. 5. Meso- and micro-structures in composite olâmph-pyroxe-nite^peridotite xenoliths 92T9 and T31. (a) Decimeter-size xenolith;pyroxenite crosscutting peridotite foliation (marked by spinel trails).(b) Reaction zone at the contact between pyroxenite and hostperidotite, characterized by crystallization of new clinopyroxene(plane-polarized light). (c) Euhedral olivine (ol) and poikiliticbrown amphibole (amph) in pyroxenite (cross-polarized light).(d) Subhedral clinopyroxene (cpx) and interstitial plagioclase (plag)in pyroxenite (cross-polarized light).

Fig. 4. Microstructures in plagioclase-bearing peridotites and gab-bronorite veins. (a) Porphyroclastic peridotite T11. (ol þ plag þ opx)granoblastic pocket interstitial between kinked mantle olivines(OLp) (cross-polarized light). (b) Porphyroclasticêquigranular peri-dotite T32. Interstitial orthopyroxene (opx) replacing and lobatingtriple point junction between equigranular olivines (OLeq) (cross-polarized light). (c) Porphyroclastic peridotite T11: (orthopyroxene^plagioclase) gabbronoritic vein crosscutting large kinked olivine por-phyroclast (OLp). Noteworthy features are the fine-grained orthopyr-oxene reaction rim towards the host peridotite, and spinel-faciesmelt^rock interaction textures in the host peridotite [replacement ofolivine porphyroclast (OLp) by new orthopyroxene (Opxn)] (cross-polarized light).


301

Dow



92T18

92T14

92T21

92T17

T18

92T3

[100] [010] [001]

Fig. 6. Microstructures and olivine lattice preferred orientations (LPO) of six selected xenoliths. Microstructure sketches were made fromphotomicrographs, and show olivine without shading with fine dashed lines indicating subgrain boundaries, orthopyroxene with dashed linepattern, clinopyroxene with dashed-point ornamentation, and spinel in black. Ubiquitous olivine 1208 triple junctions in all samples should benoted. LPO patterns all shown with the foliation defined by stretched spinel aggregates oriented east^west and the spinel stretching fabric hori-zontal. Samples are shown in qualitative order of increasing strain, from relatively low strain in sample 92T14 to high strain and very intensestretching of the orthopyroxene^spinel aggregates in sample 92T3. The angle between [100] maxima and foliation in the upper four samplesshould be noted. All diagrams show 100 measurements contoured at 1, 3, 5, 7, etc. times uniform distribution. (For further explanation see text.)


302

Dow



above contains plagioclase, four xenoliths are spinel lher-zolites, and one sample (T18) is transected by a gabbro-noritic dikelet. All of these samples were cutperpendicular to the macroscopically visible foliation andparallel to the spinel lineation. As in sample 92T14, mostLPO patterns are A-type fabrics, but samples 92T18 andto a lesser extent 92T21 show a tendency of the [010] and[001] axes to form girdles. These two latter LPOs are closeto D-type fabrics (Jung & Karato, 2001) and suggest atransition to olivine {0kl}[100] slip, again at elevated(although possibly somewhat lower) temperatures andlow water contents, but at higher stresses (4300MPa;Karato et al., 2008).Close inspection of the LPOs reveals a clear tendency of

the [010] axes to be more perpendicular to the plane ofthe foliation at inferred high strains, whereas the [100]axes at high strains make smaller angles or become paral-lel to the lineation. It follows that, at higher strains, thefoliation defined by the deformed spinel^pyroxene aggre-gates tends to make decreasing angles with the flow plane,a phenomenon to be expected when sampling along straingradients in heterogeneous simple shear zones. It shouldbe noted that all LPOs in Fig. 6 are shown with theinferred shear sense dextral.Although the orientations of the LPO patterns seem to

systematically vary with strain as inferred from thedeformed shapes of the clusters, there is no clear relation-ship between the increasing strain denoted by the elonga-tion of clusters, LPO patterns and the microstructure ofthe xenoliths. As shown in Fig. 6, in the six selected samplesrepresenting increasing amount of strain, the microstruc-tures vary from porphyroclasticêquigranular (92T14), toequigranular (92T18), to porphyroclasticêquigranular(92T21, 92T17), to again strongly equigranular types (T18,92T3). The ubiquitous presence in all samples of 1208 oli-vine triple junctions and relatively straight and slightlycurved olivineôlivine grain boundaries indicates that themicrostructure predominantly resulted from surfaceenergy driven migration recrystallization. We thereforeinfer that the microstructure of the xenoliths predomi-nantly reflects a stage of migration recrystallization thatto a variable extent obliterated the precursor, presumablyporphyroclastic microstructure associated with the high-temperature shearing.There is distinct microstructural evidence that the

strong equigranular recrystallization seen in many of theTallante xenoliths is in some way related to melt^rockinteraction processes. The most porphyroclastic types ofspinel peridotite exhibit peculiar microstructures indica-tive of multiple stages of melt^rock interaction: (1) crystal-lization of undeformed lobate olivine rims partlyreplacing exolved pyroxenes and kinked olivine porphyro-clasts (Fig. 2c) and (2) later diffuse crystallization ofunstrained poikilitic orthopyroxene grains that partly

corrode both the porphyroclastic minerals and the new oli-vine of the previous stage (Fig. 2d). Small unstrained clino-pyroxene grains are sometimes associated with thepoikilitic orthopyroxene. In peridotites with an intermedi-ate porphyroclastic^granular microstructure, both theporphyroclastic minerals and the melt^rock reaction struc-tures are increasingly replaced by the development of amedium- to fine-grained granoblastic assemblage of olivineþ pyroxeneþ spinel grains, typically showing sharpboundaries and triple junctions. The porphyroclasticfabric is almost completely obliterated in the equigranularperidotites, which consist of a recrystallized (olivineþpyroxeneþ spinel) and sometimes tabular, fine-grainedequigranular matrix (Fig. 2e). In these peridotites, (pyr-oxeneþ spinel) clusters are still recognizable, althoughthey also consist of equigranular recrystallized grains(Fig. 3b), but the melt^rock interaction textures are onlyrarely preserved. The equigranular recrystallization thusoccurs at the expense of variably deformed porphyroclasticperidotites, as also evidenced by the different degree ofcluster deformation and flattening in the equigranularxenoliths. For example, equigranular peridotite 92T20and tabular equigranular peridotite 92T14 exhibit roundedto ellipsoidal clusters, whereas porphyroclastic to granularperidotites 92T17 and T30 (not shown) are characterizedby the occurrence of aligned spinel trails (Fig. 3a^c). Weemphasize again that the olivines in the strongly recrystal-lized equigranular peridotites preserve their LPO relatedto the development of the spinel trails and partially pre-served porphyroclastic microstructure. Microstructuralevidence and the LPO results thus indicate that the finalstages of equigranular recrystallization occurred subse-quent to both the deformation and melt^peridotite interac-tion events.

Plagioclase-bearing spinel peridotitesIn the plagioclase-bearing peridotites the spinel-bearingmicrostructure ranges from porphyroclastic to equigranu-lar, thus covering the whole variability described above.In the less recrystallized porphyroclastic samples, themelt^rock reaction microstructures (olivine and/or ortho-pyroxene replacement of spinel-facies porphyroclasts) arefrequently preserved. The olivine LPOs in plagioclase-bearing equigranular peridotites 92T14 (Fig. 6) and92T23 (not shown) are again high-temperature, low-stressA-type fabrics as seen in the spinel peridotites.Plagioclase occurs in highly variable modal amounts in

the samples (from55% to about 10 vol. %) and mostlycrystallizes in small-grained granoblastic aggregates(together with olivine and/or orthopyroxene), interstitialbetween the spinel-facies minerals. In samples with lowplagioclase modal abundance (e.g. T11, 92T14), plagioclasecrystallizes both as rims replacing spinel (in associationwith fine-grained olivine), and in (plagioclaseþ olivine)


303

Dow



interstitial aggregates (Fig. 4a). Plagioclase-rich(410 vol. %) peridotites 92T1 and 92T23 exhibit the dif-fuse occurrence of gabbroic pockets, made of plagioclaseþorthopyroxene� olivine, that crystallized interstitiallyand partly replace porphyroclastic and equigranularspinel-facies minerals. In the equigranular peridotites,thin orthopyroxene and/or plagioclase rims are oftenobserved corroding triple junctions between olivine grano-blastic grains (Fig. 4b), indicating that crystallization ofthe plagioclase-bearing aggregates occurred in anadvanced stage of migration recrystallization, or evenafter the development of the equigranular microstructure.

Intrusive rocks: gabbronorites andolivineâmphibole pyroxenitesThin (51cm) gabbronoritic to noritic veins have beenobserved both in spinel (sample 92T20) and in plagio-clase-bearing (T11, T13, T18) peridotites, indicating thatthe melt impregnation process causing diffuse plagioclasecrystallization in the peridotites and the intrusion of gab-bronoritic veins are not directly related events. This is alsosupported by the distinct geochemical signature of therelated melts (see discussion below). The veins show clearintrusive relationships relative to the wall-rock peridotite,and frequently transect large olivine and orthopyroxeneporphyroclasts. The inner parts of the veins consist of upto millimeter-size orthopyroxene and plagioclase grains,whereas the vein margins against the host peridotites aremade by palisade-type, oriented, fine-grained orthopyrox-ene (Fig. 4c). Small euhedral apatite crystals are observedas inclusions in both vein-forming plagioclase and ortho-pyroxene. In plagioclase-bearing peridotite T13, the gab-bronoritic vein is cross-cut at a high angle by an olivineâmphibole-bearing pyroxenite, which indicates that thepyroxenite intrusion was subsequent to the gabbronoriteintrusion stage.

Pyroxenites display a clearly magmatic hypidiomorphictexture, defined by large (millimeter-size) euhedral olivineand subhedral clinopyroxene, poikilitic brown amphibole,interstitial plagioclase and minor phlogopite (Fig. 5c andd). Large poikilitic amphibole grains often include botholivine and clinopyroxene. Small euhedral Fe-oxide grainsfrequently occur as inclusions in clinopyroxene and amphi-bole. Fine-grained allotriomorphic aggregates of clinopyr-oxene þ olivine þ amphibole � phlogopite � plagioclasegrains have crystallized between the matrix (olivine, clino-pyroxene) minerals.The microstructure therefore indicatesthe following crystallization order: olivine, Fe-oxides, clin-opyroxene, amphibole, phlogopite and plagioclase.In all composite xenoliths, the contact between

pyroxenite and the host peridotite is marked by a centi-meter-thick reaction zone (Fig. 5b), characterized by anenrichment in clinopyroxene and by the occurrenceof small interstitial grains of brown amphibole andphlogopite. The newly formed clinopyroxene is clearlydistinguishable from mantle clinopyroxene because it isinterstitial, rich in fluid inclusions, and tends to replacemantle minerals (olivine and spinel).

MAJOR AND TRACE ELEMENTMINERAL CHEMISTRYDespite of the observed microstructural variability, min-eral compositions in the studied spinel peridotites arerather homogeneous and consistent with previously pub-lished data on Tallante cpx-poor lherzolites (Beccaluvaet al., 2004). Olivines show limited chemical heterogeneity(Table 1): Mg-numbers range between 90�5 and 91�6, withthe lowest values observed in equigranular peridotite92T20 and porphyroclastic^granular peridotite T30.Within a single sample, no appreciable chemical changesare observed in olivine pertaining to different

Table 1: Major (wt %) element compositions of olivines in peridotites and pyroxenites

Sample: 92T20 T30 T30 92T6 92T6 92T7 92T7 92T7 92T19 92T19 92T1 92T1 T31A T31A 92T9A 92T9B 92T9B 92T9B

grain new porph porph new new porph new new porph aggr porph aggr new porph idiom small small

SiO2 40�79 41�12 40�95 41�21 40�75 41�35 40�96 41�27 41�01 40�91 41�36 41�12 41�21 40�74 40�78 39�76 39�27 39�13

FeO 8�52 9�01 9�00 9�18 9�06 8�28 8�22 8�37 8�74 8�53 8�72 8�76 8�91 9�59 10�71 18�87 19�48 20�06

MgO 49�66 49�78 49�64 49�20 49�43 50�47 50�33 50�11 50�09 50�07 49�81 49�71 49�96 48�97 48�60 41�52 40�66 40�37

MnO 0�16 0�14 0�16 0�12 0�15 0�13 0�07 0�08 0�08 0�10 0�14 0�15 0�13 0�12 0�14 0�22 0�16 0�21

CaO 0�05 0�05 0�03 0�02 0�04 0�02 0�01 0�03 0�02 0�02 0�04 0�01 0�03 0�03 0�03 0�04 0�04 0�07

Total 99�18 100�10 99�78 99�73 99�43 100�25 99�59 99�86 99�94 99�63 100�07 99�75 100�24 99�45 100�26 100�41 99�61 99�84

Mg-no. 91�2 90�8 90�8 90�5 90�7 91�6 91�6 91�4 91�1 91�3 91�1 91�1 90�9 90�1 89�0 79�7 78�8 78�2

Grain: olivine in the equigranular recrystallized matrix. Porph: olivine porphyroclast. New: unstrained olivine replacingkinked olivine porphyroclasts. Aggr: olivine grain in interstitial granoblastic aggregates. Idiom: idiomorphic grain. Small:small anhedral crystal.


304

Dow



microstructures (i.e. porphyroclastic olivine, new olivinerims or recrystallized equigranular olivine grains).Clinopyroxenes show rather narrow Mg-number variation(91�2^93�2), consistent with that observed inolivine (Table 2; Fig. 7). They have slightly variable Ti andAl contents (TiO2¼0�24^0�44wt %; Al2O3¼5�42^6�64wt %), not correlated with the Mg-number; withineach sample, the highest Al contents are generallyobserved in clinopyroxene porphyroclasts, relative torecrystallized grains. In both porphyroclastic and equigra-nular peridotites, clinopyroxenes exhibit rather homoge-neous contents in moderately incompatible trace elements[i.e. rare earth elements (REE), Zr, Hf, Ti, Y], and arecharacterized by slight light REE (LREE) depletion andalmost flat middle to heavy REE (MREE^HREE) spectra(LaN/YbN¼ 0�37^0�62; SmN/YbN¼ 0�89^1�23). On theother hand, they display highly variable Th, U and Nbabundances, and such variations are observed within asingle sample (e.g. peridotite 92T20;Table 2, Fig. 8); despitehighly variable absolute contents, all clinopyroxenes showU and Th enrichment, and constant negative Nb anoma-lies. Similar trace element spectra in clinopyroxene, char-acterized by large U and Th heterogeneities coupledwith more constant REE contents, have already been doc-umented in previous studies of the Tallante spinel perido-tites (Beccaluva et al., 2004; Shimizu et al., 2008).Orthopyroxenes have Mg-numbers ranging from 90�1 to92�0, consistent with the compositions of olivine and clino-pyroxene; the lowest Mg-number values are recorded insamples 92T20 and T30. In the porphyroclastic perido-tites, no appreciable chemical differences are observedbetween orthopyroxene porphyroclasts and newpoikilitic grains crystallizing at the expense of olivine,except for slightly lower Al and Cr contents in the latter(Al2O3¼4�68^5�85wt % and 4�15^5�03wt %, Cr2O3¼

0�50^0�63wt % and 0�30^0�48wt %, in porphyroclasticand poikilitic grains, respectively). Few orthopyroxeneporphyroclasts have been analysed for trace elements:they have LREE-depleted spectra (LaN/SmN¼ 0�20^0�36), and display U and Th enrichment, similar toclinopyroxene.In plagioclase peridotite 92T1, the major element com-

positions of primary spinel facies minerals (olivine, ortho-pyroxene, clinopyroxene), as well as the chemicalvariations in a specific mineral pertaining to differentmicrostructures (e.g. orthopyroxene porphyroclasts andnew poikilitic crystals) are generally similar to those docu-mented in spinel peridotites (see Tables 1, 3 and 4). Interms of minor and trace elements, clinopyroxenes displayslightly higher Ti concentrations relative to clinopyroxenesin spinel peridotites, at constant high Mg-numbers(0�924^0�93; Fig. 7). They also exhibit overall enrichmentin REE, negative SrN anomalies, sometimes coupled todevelopment of negative EuN, although preserving LREE

depletion (LaN/SmN¼ 0�16^0�28; Fig. 8c). Similar chemi-cal effects in clinopyroxene (i.e. Ti, REE enrichment) arewidely documented in plagioclase-bearing impregnatedperidotites (Rampone et al., 1997, 2008a; Dijkstra et al.,2003; Piccardo et al. 2007). Clinopyroxenes moreover dis-play variable U andTh enrichment relative to Nb andTa,although not reaching the high values observed in somespinel peridotites. Plagioclases have anorthite (An) con-tents ranging from 57�6 to 58�6, rather high Sr abundances(220^265 ppm) and LREE-enriched REE spectra (LaN/SmN¼ 4�4^5�4).Minerals in the pyroxenites show rather evolved major

element compositions. Olivines have Mg-numbers rangingfrom 78�2 to 79�7, with the lowest values occurring in thefine-grained aggregates (Table 1). Clinopyroxenes alsoshow rather low Mg-number values (80�4^84�7) and ageneral Mg-number decrease from large euhedral tofine-grained anhedral crystals. Both clinopyroxene andamphibole have highTi contents (TiO2 is 0�73^0�99wt %and 2�70^4�40wt % in cpx and amph, respectively;Table 5). According to Leake’s (1997) classification, theamphiboles are pargasites to pargasitic hornblendes, withMg-numbers of 77�0^82�0. Plagioclases have low An con-tents (35�8^44�6; Table 6). The trace element compositionsof clinopyroxene and amphibole in pyroxenites are shownin Fig. 9. Clinopyroxenes display convex-upward MREEspectra and low HREE abundances (LaN/SmN¼ 0�99^1�6; SmN/YbN¼ 3�93^7�05), similar to clinopyroxenes inpyroxenites crystallized from alkaline melts (Fabries et al.,1989; Bodinier et al., 1990; Downes et al., 1991; Downes,2001). They moreover exhibit marked negative NbN^TaNanomalies, and strong U and Th enrichment [up to 70�Primitive Mantle (PM)]. The REE spectra of amphiboleare similar to those of clinopyroxene (LaN/SmN¼ 0�88^2�57; SmN/YbN¼ 6�27^7�05), but shifted to higher REEabsolute concentrations; in contrast to clinopyroxene, theydisplay NbN^TaN enrichment, and more variable U andTh contents. Higher trace element contents in amphiboleare coupled to lower Mg-number values. Phlogopite hashigh Ba (4865 ppm) and appreciable Sr (61ppm) and Nb(20�3 ppm) abundances.Clinopyroxene porphyroclasts in the host peridotites

T31A and 92T9A have more variable and, on average,lower Mg-number (89�8^92�6) relative to clinopyroxenein spinel and plagioclase peridotites (see Fig. 7). In termsof trace elements, clinopyroxenes preserve moderateincompatible element spectra similar to those in plagio-clase peridotites, but are selectively enriched in LREE(LaN/YbN¼ 0�76^0�81; SmN/YbN¼ 0�75^0�93) and haveU and Th abundances similar to those of clinopyroxenesin the pyroxenites. New interstitial clinopyroxenes, crystal-lized in the centimeter-thick reaction zone at the contactwith pyroxenite, have intermediate Mg-number (87�7^88�6)and TiO2 compositions with respect to clinopyroxenes in


305

Dow



Table 2: Major (wt %) and trace (ppm) element compositions of clino- and orthopyroxenes in spinel peridotites

92T20 92T20 92T20 92T20 92T20 T30 T30 T30 92T6 92T6

cpx cpx cpx cpx opx cpx cpx opx cpx cpx

cluster grain grain cluster cluster grain grain porph porph porph

wt %

SiO2 51�87 51�73 51�41 51�60 56�09 52�43 51�56 54�35 52�23 52�00

TiO2 0�43 0�42 0�40 0�47 0�07 0�44 0�44 0�09 0�32 0�30

Al2O3 6�10 6�29 6�39 6�07 3�24 5�42 6�56 5�85 6�41 6�64

Cr2O3 1�00 1�08 1�11 1�09 0�31 0�87 1�14 0�54 1�04 1�06

FeO 2�37 2�38 2�45 2�38 6�17 2�42 2�53 5�70 2�48 2�49

MgO 15�60 15�49 15�40 15�45 33�55 16�36 15�68 33�14 15�31 15�20

MnO 0�11 0�08 0�10 0�09 0�15 0�13 0�12 0�14 0�08 0�09

CaO 21�81 21�87 22�19 22�39 0�66 22�18 21�58 0�86 21�62 21�4

Na2O 0�88 0�91 0�82 0�73 0�03 0�48 0�62 0�00 1�11 1�14

Total 100�17 100�25 100�27 100�27 100�22 100�73 100�23 100�67 100�59 100�32

Mg-no. 92�1 92�1 91�8 92�0 90�6 92�3 91�7 91�2 91�7 91�6

ppm

Sc 52 52 50 52 17 61 53 19 64 56

V 262 260 271 233 84 244 250 97 262 255

Sr 62 58 60 71 0�47 27 34 0�23 50 45

Y 13�4 13�4 12�5 15�2 0�76 14�1 14�5 0�98 15�4 12�7

Zr 19 19 19 24 1�18 23 20 1�49 18 14

Nb 0�64 0�54 0�62 0�73 0�01 0�07 0�12 0�01 0�15 0�18

La 1�19 1�23 1�17 1�32 0�01 0�78 0�89 0�01 0�75 0�80

Ce 3�63 3�67 3�68 3�95 0�07 2�67 3�36 0�04 3�01 2�89

Pr 0�63 0�67 0�62 0�72 0�01 0�46 0�58 0�01 0�55 0�47

Nd 3�79 3�63 3�57 4�06 0�05 2�91 3�48 0�07 3�26 2�76

Sm 1�56 1�62 1�36 1�52 0�02 1�30 1�56 0�03 1�52 1�05

Eu 0�63 0�66 0�63 0�72 0�01 0�56 0�58 0�01 0�73 0�51

Gd 1�87 2�01 1�80 1�81 0�05 1�93 1�98 0�05 2�05 1�53

Tb 0�35 0�34 0�31 0�43 0�01 0�35 0�39 0�01 0�41 0�31

Dy 2�39 2�55 2�13 2�61 0�11 2�35 2�62 0�12 3�17 2�36

Ho 0�48 0�48 0�43 0�66 0�03 0�55 0�65 0�04 0�63 0�53

Er 1�59 1�44 1�45 1�42 0�10 1�51 1�41 0�13 1�87 1�61

Tm 0�19 0�20 0�20 0�24 0�02 0�27 0�23 0�03 0�27 0�24

Yb 1�36 1�49 1�38 1�58 0�18 1�58 1�61 0�28 1�88 1�58

Lu 0�19 0�19 0�19 0�20 0�04 0�18 0�27 0�05 0�25 0�21

Hf 0�68 0�59 0�52 0�77 0�03 0�75 0�62 0�06 0�56 0�37

Ta 0�080 0�054 0�039 0�047 50�004 0�014 0�028 0�001 0�041 0�028

Pb 50�02 0�04 0�03 0�07 0�02 0�07 0�06 — 0�037 0�031

Th 0�57 0�53 0�28 0�098 — 0�046 0�015 0�005 0�18 0�082

U 0�21 0�17 0�12 0�044 50�0025 0�041 0�024 0�002 0�070 0�028

Cr 7708 8371 8394 6664 2259 5955 6762 2818 7744 8445

Co 21 23 23 21 53 21 21 51 22 23

Ti 2251 2155 2205 2360 540 2570 2274 551 1840 1538

LaN/SmN 0�48 0�48 0�54 0�54 0�35 0�37 0�36 0�16 0�31 0�48

SmN/YbN 1�25 1�18 1�07 1�04 0�13 0�89 1�05 0�13 0�88 0�72

(continued)


306

Dow



the host peridotite and pyroxenite (Fig. 10a); their traceelement contents are variable, mostly similar to those ofclinopyroxenes in pyroxenites (Fig. 10b).

TRACK ING THE UPL I FTH ISTORY OF THE TALLANTEMANTLEThe Tallante xenoliths record a multi-stage history ofdeformation, recrystallization, melt migration and meltintrusion. Together with the specific geochemical finger-prints in the various lithotypes, the overprinting relation-ships between the different microstructural features allowus to reconstruct a relative chronology of tectonic andmagmatic events as follows.

Early decompression anddeformation eventsThe earliest microstructures, observed in both the moreporphyroclastic and the intensely recrystallized spinelperidotites, are rounded to ellipsoidal clusters made oforthopyroxeneþ spinelþminor clinopyroxene and oli-vine. A previous study of the Tallante xenoliths (Shimizuet al., 2008) that focused on symplectitic spinel^pyroxeneaggregates has shown textural and chemical characteristicssimilar to the spinel^pyroxene clusters of this study. Onthe basis of mass-balance calculations, Shimizu et al.(2008) demonstrated that the spinel^pyroxeneaggregates were derived from garnet breakdown, accord-ing to the subsolidus reaction garnetþolivine! spinelþorthopyroxeneþ clinopyroxene. This indicates re-equilibration of the Tallante peridotites from garnet- tospinel-facies conditions. In principle, the garnet^spinel

Fig. 7. Variation of Mg-number vs Ti (�1000) (atoms per six oxy-gens) in clinopyroxenes from spinel peridotites, plagioclase-bearingperidotite 92T1, and host peridotites to amphibole pyroxenites. Thefield refers to the compositions of clinopyroxenes in cpx-poor lherzo-lites from Beccaluva et al. (2004).

Table 2: Continued

92T7 92T7 92T19 92T19 92T19

cpx cpx cpx cpx cpx

porph porph porph porph porph

wt %

SiO2 52�70 52�33 52�40 51�72 53�94

TiO2 0�28 0�24 0�29 0�41 0�07

Al2O3 5�77 5�73 6�40 6�20 5�78

Cr2O3 1�27 1�18 0�99 1�08 0�63

FeO 2�25 2�15 2�10 2�19 5�47

MgO 16�12 16�42 15�97 15�94 33�39

MnO 0�09 0�06 0�06 0�04 0�15

CaO 20�86 21�16 21�64 21�20 0�78

Na2O 1�00 1�36 0�82 1�21 0�04

Total 100�34 100�63 100�67 99�99 100�25

Mg-no. 92�7 93�2 93�1 92�8 91�6

ppm

Sc 68 61 50 45 18

V 272 270 260 228 96

Sr 48 57 57 46 0�80

Y 9�69 10�6 11�1 10�1 1�02

Zr 16 14 18 17 1�81

Nb 0�11 0�52 0�67 0�31 0�02

La 0�90 1�01 1�41 1�22 0�01

Ce 3�01 3�61 5�00 4�15 0�08

Pr 0�55 0�58 0�84 0�69 0�01

Nd 3�21 3�84 4�62 3�79 0�06

Sm 1�18 1�36 1�41 1�30 0�04

Eu 0�54 0�56 0�69 0�56 0�01

Gd 1�63 1�67 1�50 1�48 0�05

Tb 0�28 0�31 0�30 0�27 0�02

Dy 2�08 2�15 2�20 1�89 0�16

Ho 0�44 0�47 0�43 0�41 0�04

Er 1�26 1�29 1�20 0�96 0�14

Tm 0�19 0�18 0�17 0�17 0�02

Yb 1�23 1�22 1�14 1�14 0�18

Lu 0�18 0�18 0�15 0�13 0�04

Hf 0�50 0�59 0�56 0�49 0�05

Ta 0�048 0�049 0�043 0�036 0�002

Pb 0�034 50�02 0�05 0�02 0�02

Th 0�095 0�111 0�078 0�070 0�013

U 0�074 0�068 0�026 0�025 0�005

Cr 9150 11050 8704 7277 3763

Co 20 25 25 20 49

Ti 1696 1597 2124 1932 539

LaN/SmN 0�48 0�46 0�62 0�59 0�20

SmN/YbN 1�04 1�21 1�35 1�24 0�25

Cluster: clinopyroxene grain in (pyroxenesþ spinel) clus-ters. Grain: clinopyroxene in the equigranular recrystallizedmatrix. Porph: clinopyroxene porphyroclast.


307

Dow



transition could be related either to isobaric heating or todecompression. However, equilibration temperatures inthe porphyroclastic peridotites discussed below do notallow us to make any inference on a possible heating event

related to spinel^pyroxene cluster development. On theother hand, a transition from deep lithospheric (garnet-facies) to spinel-facies conditions has been documented inthe Ronda and Beni Bousera massifs, both containinggraphite pseudomorphs after diamond (Davies et al., 1993);the inferred garnet breakdown and development of thespinel^pyroxene clusters in theTallante xenoliths may wellreflect this early decompressional evolution. Consistentwith previous work by Shimizu et al. (2008), we thereforefavour the same interpretation for theTallante mantle.The spinel^pyroxene clusters in the various xenoliths

show variable degrees of flattening, from moderately ellip-soidal to intensely stretched spinel^pyroxene aggregates.Irrespective of the recrystallization microstructure, the oli-vine LPOs in a selected set of samples with differentdegrees of flattening of the clusters are mostly of theA-type, common in upper mantle peridotite massifs aswell as in basalt-hosted peridotite xenoliths, suggestinghigh-temperature crystal^plastic ductile flow (e.g. Ave¤Lallemant & Carter, 1970; Karato et al., 2008, and refer-ences therein). In addition, there is a marked trend in thesamples studied for the angle between the stretching direc-tion (marked by the elongate spinel^pyroxene clusters)and the flow plane to decrease with increasing strain. Thisstrongly suggests that the strain in the samples and theassociated LPOs reflect different degrees of shearing(Zhang & Karato, 1995; Karato et al., 2008). The develop-ment of the LPOs was probably related to the simultaneousdevelopment of a coarse-grained porphyroclastic micro-structure, represented by a spinel-bearing mineral assem-blage with evidence of internal plastic deformation suchas large kinked olivines and coarse, partially exolvedorthopyroxenes. The remnants of this porphyroclasticmicrostructure are best preserved in some of the spinelperidotites.A critical issue, relevant to the tectonic history of the

upper mantle in this area, concerns the timing of the for-mation of the porphyroclastic microstructure and thedevelopment of the LPOs with respect to the inferredgarnet breakdown and development of the clusters. Itcould be argued that intense shear deformation of a pre-cursor garnet-bearing peridotite might have led tostrongly stretched garnets, known to occur in garnet-peridotite massifs (e.g. van Roermund et al., 2001) whichthen became transformed to spinel^pyroxene aggregatesduring later decompression. Alternatively, ductile shearflow may have affected a spinel peridotite containingspinel^pyroxene clusters resulting from a previous stage ofgarnet breakdown (i.e. ductile flow subsequent to earlydecompression). This latter interpretation is strongly sup-ported by a distinct preferred orientation of the orthopyr-oxene [100] cleavage planes in many of the stretchedclusters: hence we infer that the development of the por-phyroclastic microstructure and related LPOs postdates

(a)

(b)

(c)

Fig. 8. Primitive mantle normalized trace element abundances ofrepresentative clinopyroxenes, orthopyroxenes and plagioclases inspinel- and plagioclase-bearing peridotites. (a) Porphyroclastic peri-dotites 92T6, 92T7 and 92T19. The light grey and dark grey fieldsrefer to the compositions of clinopyroxenes inTallante spinel perido-tites from Beccaluva et al. (2004) and Shimizu et al. (2008), respec-tively. (b) Porphyroclasticêquigranular peridotite T30 andequigranular peridotite 92T20. Compositional fields as in (a).(c) Plagioclase-bearing peridotite 92T1. The dark grey field refers tothe compositions of clinopyroxenes in spinel peridotites from thisstudy. Normalizing values from McDonough & Sun (1995).


308

Dow



an earlier phase of decompression and garnet breakdown.At this stage, we note that the development of the spinelclusters probably involved an early stage of static recrystal-lization leading to grain coarsening before deformation,as breakdown of garnet in mantle rocks in the firstinstance typically leads to garnets surrounded by symplec-tites and/or kelyphites (e.g. van der Wal, 1993; Kaeseret al., 2006).In the majority of the spinel peridotites (i.e. in those

with a porphyroclastic^granular microstructure) theprimary porphyroclastic assemblage is clearly overprintedby largely surface energy driven annealing (migrationrecrystallization) leading to a commonly finer-grained,sometimes tabular, equigranular microstructure (e.g. sam-ples 92T18 and 92T20).The degree of equigranular recrys-tallization, however, is highly variable between xenoliths.Strikingly, there is no correlation between the strains,inferred qualitatively from the degree of stretching of thespinel^pyroxene clusters, and the recrystallization micro-structure. Highly recrystallized peridotites (e.g. samples92T14, 92T18 and 92T20) may preserve rounded to ellip-soidal clusters, whereas highly elongate spinel trails occurin porphyroclastic^granular peridotites (e.g. samples92T17 and T30). We therefore infer that the recrystallizedgranular microstructure of the xenoliths reflects a stage ofmigration recrystallization that largely obliterated a pre-cursor porphyroclastic microstructure associated withspinel-facies ductile shearing. This interpretation seemsconsistent with another aspect of the microstructure,namely that the grain sizes of the xenoliths do not showany systematic change with strain either. An unpublishedpilot study (de Boom, 1994) including four xenoliths fromour collection (92T18, 92T3, T11 and T18) suggested thatboth the maximum and average olivine grain sizesdepend on the grain sizes and volume fraction of theother phases, hence that the presence of the other phasescontrolled the recrystallized olivine grain sizes, again con-sistent with surface energy driven migration recrystalliza-tion. The intense recrystallization did not obliterate theLPO patterns, however, a phenomenon already knownfrom structural and microstructural studies in the Rondamassif (van der Wal, 1993; van der Wal & Vissers, 1993,1996; Vauchez & Garrido, 2001), but also from experimen-tal studies in quartz aggregates (Heilbronner & Tullis,2002).It can be concluded that the microstructures and LPO

data for the spinel peridotites point to the early decompres-sion of the Tallante mantle from garnet- to spinel-faciesconditions. This was followed by ductile shear flow in thespinel facies, presumably at lithospheric conditions. Thislatter notion seems consistent with the occurrence of fewLPO patterns showing a transition to D-type fabrics,known from experiments to develop at higher stressesexpected at lithospheric temperature conditions

Table 3: Major (wt %) and trace element (ppm)

compositions of clinopyroxenes in plagioclase peridotites

92T1 92T1 92T1 92T1

cpx cpx cpx cpx

porph grain porph grain

wt %

SiO2 51�83 52�14 51�61 51�67

TiO2 0�54 0�56 0�68 0�71

Al2O3 6�51 5�69 6�64 6�47

Cr2O3 0�96 1�05 1�13 1�11

FeO 2�16 2�23 2�16 2�23

MgO 16�06 16�31 15�86 15�26

MnO 0�04 0�07 0�07 0�12

CaO 21�76 21�66 21�86 22�24

Na2O 0�63 0�67 0�54 0�83

K2O 0�00 0�00 0�00 0�00

Total 100�49 100�38 100�55 100�64

Mg-no. 93�0 92�9 92�9 92�4

ppm

Sc 59 62 75 64

V 275 320 337 273

Sr 5�2 9�9 12 6�6

Y 22�6 21�4 29�4 24�5

Zr 30 25 31 28

Nb 0�07 0�10 0�48 —

La 0�63 0�64 0�97 0�62

Ce 3�61 3�83 4�50 3�14

Pr 0�83 0�78 0�94 0�60

Nd 5�53 4�88 7�37 5�36

Sm 2�42 2�04 2�16 2�32

Eu 0�75 0�72 1�15 0�94

Gd 3�22 3�10 3�91 3�18

Tb 0�65 0�53 0�78 0�64

Dy 4�24 4�02 4�88 4�30

Ho 1�00 0�82 1�46 0�97

Er 2�50 2�34 3�33 2�88

Tm 0�32 0�35 0�61 0�48

Yb 2�30 2�21 3�45 2�52

Lu 0�29 0�28 0�54 0�38

Hf 0�99 0�86 1�30 1�04

Ta 0�008 0�008 50�058 —

Pb 0�02 0�03 50�08 —

Th 0�054 0�020 0�043 0�037

U 0�018 0�008 50�008 0�009

Cr 6460 7821 7958 6410

Co 20 23 24 17

Ti 3831 3583 3401 3357

LaN/SmN 0�16 0�20 0�28 0�17

SmN/YbN 1�14 1�00 0�68 1�00

Grain: clinopyroxene in the equigranular recrystallizedmatrix. Porph: clinopyroxene porphyroclast. Aggr: plagio-clase grains in interstitial granoblastic aggregates.


309

Dow



(T511008C).The resulting porphyroclastic microstructurebecame in turn overprinted by variable degrees of migra-tion recrystallization leading to intermediate (porphyro-clastic^granular) as well as intensely recrystallizedequigranular microstructures.We envisage that this migra-tion recrystallization may well be related to the melt^rockinteraction processes discussed below.

Melt migration by reactive porous flowand annealing recrystallizationThe replacement of porphyroclastic minerals byunstrained olivine rims, and the subsequent crystallizationof undeformed poikilitic orthopyroxene at the expense ofboth porphyroclasts and newly crystallized olivine (seeFig. 2c and d) indicate that the porphyroclastic peridotiteswere affected by melt migration and melt^rock interaction.Similar microstructures have been described in spinel peri-dotites from ophiolitic and oceanic settings (Dijkstra et al.,2003; Piccardo & Vissers, 2007; Piccardo et al., 2007;Seyler et al., 2007; Rampone & Borghini, 2008; Ramponeet al., 2008a). They have been related to open-system reac-tive porous flow of olivine-saturated tholeiitic melts thatprogressively shift towards orthopyroxene saturationduring percolation and interaction with the host perido-tites. A series of studies on the mechanisms of melt migra-tion in the lithospheric mantle (Quick, 1981; Kelemen,1990; Kelemen et al., 1992, 1995, 1997; Kelemen & Dick,1995) have pointed out that melts rising adiabatically aresaturated in olivine; they will therefore crystallize olivineand dissolve pyroxenes in the host peridotites until,through continuous ascent and interaction with the hostperidotites, they reach pyroxene saturation and start to

crystallize ortho- and clinopyroxene. The diffuse crystalli-zation in the Tallante peridotites of orthopyroxene at theexpense of previous olivine-replacement textures may thusindicate various stages of reactive porous flow and interac-tion with increasingly modified melts.Another remarkable feature in variably deformed por-

phyroclastic peridotites is the partial obliteration of theporphyroclastic and melt^rock reaction microstructuresby the development of a granular texture. This again indi-cates that the allied migration recrystallization was notinduced by the shearing deformation as outlined above,but that the equigranular peridotites developed as theresult of extensive static, possibly melt-assisted annealingrecrystallization. As textural evidence clearly indicatesreactive porous melt flow, we suggest that the annealingrecrystallization was largely related to pervasive meltpercolation.Inferences on the origin and chemical affinity of the per-

colating melts at spinel facies conditions can be made con-sidering the trace element composition of clinopyroxenesin the spinel peridotites. Overall, clinopyroxenes showhighly variable U and Th enrichment, even within asingle thin section. In peridotite 92T20, clinopyroxeneswith the highest U and Th contents have been analysed ina (pyroxene^spinel) cluster close to a noritic vein, thislatter being significantly U^Th enriched (Rampone et al.,2008b, 2009). Based on this evidence, we believe that heter-ogeneity in theTh and U contents in clinopyroxenes prob-ably resulted from late-stage percolation of small meltfractions (Bedini et al., 1997; Ionov et al., 2002; Raffoneet al., 2009), and we did not consider these elements to con-strain the compositions of the migrating melts.

Table 4: Major element (wt %) compositions of orthopyroxenes in spinel- and plagioclase-bearing peridotites

92T20 92T20 92T20 T30 T30 92T6 92T6 92T7 92T7 92T7 92T19 92T19 92T1 92T9A 92T9A T31A T31A

cluster new porph porph new porph new porph new new porph new new porph new porph new

SiO2 56�09 55�68 55�05 54�35 54�66 54�13 54�69 54�91 54�97 54�95 54�37 55�20 55�11 53�77 55�00 54�47 54�87

TiO2 0�07 0�12 0�143 0�09 0�08 0�12 0�13 0�11 0�11 0�14 0�10 0�12 0�16 0�12 0�05 0�07 0�11

Al2O3 3�24 4�13 4�95 5�85 5�03 5�38 4�83 4�68 4�51 4�15 5�28 4�03 4�54 5�53 4�04 4�44 4�22

Cr2O3 0�31 0�49 0�59 0�54 0�30 0�53 0�38 0�50 0�52 0�48 0�55 0�43 0�51 0�56 0�55 0�48 0�52

FeO 6�17 6�44 6�38 5�70 5�72 5�85 6�05 5�44 5�29 5�33 5�51 5�55 5�66 5�82 6�10 5�73 5�86

MgO 33�55 33�11 32�42 33�14 34�04 33�10 33�53 33�83 34�12 33�86 33�21 34�22 32�87 33�28 33�34 33�94 34�05

MnO 0�15 0�139 0�17 0�14 0�16 0�23 0�14 0�05 0�07 0�18 0�15 0�17 0�20 0�13 0�11 0�10 0�17

CaO 0�66 0�67 1�09 0�86 0�75 0�81 0�86 0�78 0�79 0�73 0�81 0�78 0�86 0�86 0�79 0�76 0�81

Na2O 0�03 0�03 0�04 0�00 0�00 0�05 0�04 0�06 0�03 0�03 0�00 0�00 0�02 0�03 0�02 0�02 0�00

Total 100�27 100�81 100�83 100�67 100�74 100�20 100�65 100�36 100�41 99�85 99�98 100�50 99�93 100�10 100�00 100�01 100�61

Mg-no. 90�6 90�2 90�1 91�2 91�4 91�0 90�8 91�7 92�0 91�9 91�5 91�7 91�2 91�1 90�7 91�3 91�2

Cluster: orthopyroxene grain in (pyroxeneþ spinel) clusters. New: unstrained poikilitic orthopyroxene replacing olivineporphyroclasts and new olivine. Porph: orthopyroxene porphyroclast.


310

Dow



Table 5: Major (wt %) and trace (ppm) element compositions of clinopyroxenes, amphiboles and phlogopites in composite

xenoliths 92T9 andT31

Host peridotite Reaction zone Pyroxenite

T31A T31A 92T9A 92T9A 92T9A 92T9A 92T9A T31B T31B 92T9B 92T9B

cpx cpx cpx cpx cpx cpx cpx cpx cpx cpx cpx

porph porph porph porph react react react idiom idiom idiom idiom

wt %

SiO2 52�12 51�04 52�45 52�30 51�66 53�14 53�61 51�04 52�00 51�66 50�60

TiO2 0�62 0�43 0�32 0�32 0�76 0�44 0�30 0�73 0�83 0�99 0�98

Al2O3 4�68 5�80 4�42 6�01 5�16 3�21 3�00 5�25 4�30 5�10 5�85

Cr2O3 1�11 1�08 1�16 1�01 1�02 0�90 1�32 0�27 0�36 0�66 0�66

FeO 3�26 2�77 3�35 2�26 3�96 4�17 3�77 6�31 5�54 4�76 5�30

MgO 16�02 15�59 16�70 15�85 15�86 16�79 16�50 14�50 15�59 15�68 14�74

MnO 0�12 0�08 0�08 0�07 0�10 0�09 0�08 0�17 0�10 0�09 0�16

CaO 21�55 21�83 21�33 21�96 21�28 21�42 21�06 20�97 21�35 21�13 21�77

Na2O 0�44 0�83 0�82 0�56 0�56 0�73 0�84 1�11 0�18 0�46 0�21

K2O 0�03 0�01 0�03 0�00 0�02 0�00 0�02 0�00 0�01 0�06 0�02

Total 99�95 99�46 100�66 100�34 100�38 100�89 100�50 100�35 100�26 100�59 100�29

Mg-no. 89�8 90�9 89�9 92�6 87�7 87�8 88�6 80�4 83�4 82�6 83�2

ppm

Sc 64 63 49 70 54 44 57 57 76

V 257 287 336 241 306 204 338 307 344

Sr 10 18 87 78 116 77 78 94 96

Y 19�6 16�4 12�1 11�3 11�1 10�9 11�4 10�2 13�0

Zr 24 23 67 24 21 70 27 31 53

Nb 0�17 0�57 0�51 0�28 0�83 0�11 0�25 0�40 0�68

Ba — — — 1�1 — 0�1 0�7 — —

La 2�19 1�72 10�5 7�66 6�36 8�8 7�04 7�00 6�97

Ce 5�14 4�42 27�6 16�2 15�4 27�8 18�9 20�0 21�2

Pr 0�79 0�72 3�49 1�69 2�14 3�62 2�54 2�68 3�37

Nd 4�35 3�81 15�8 6�31 10�2 16�4 12�7 12�9 17�6

Sm 1�81 1�32 3�74 1�67 3�10 3�39 3�42 3�14 4�43

Eu 0�52 0�55 1�16 0�60 0�92 1�11 1�10 1�00 1�40

Gd 2�39 2�02 3�32 1�73 2�55 3�13 3�45 3�15 4�47

Tb 0�46 0�36 0�47 0�29 0�36 0�40 0�45 0�41 0�56

Dy 3�66 2�95 2�62 1�97 2�37 2�42 2�61 2�31 3�19

Ho 0�74 0�58 0�43 0�41 0�42 0�36 0�45 0�42 0�55

Er 2�22 1�91 1�13 1�09 1�04 0�92 1�18 1�01 1�26

Tm 0�31 0�27 0�13 0�15 0�15 0�14 0�16 0�12 0�15

Yb 2�11 1�90 1�09 1�17 0�91 0�72 0�95 0�82 0�92

Lu 0�28 0�26 0�12 0�15 0�12 0�12 0�11 0�09 0�14

Hf 0�69 0�63 1�77 0�64 0�69 2�10 1�11 1�30 2�44

Ta 0�05 0�06 0�06 0�03 0�06 0�02 0�03 0�04 0�09

Pb 0�39 0�72 3�3 1�8 3�6 2�1 2�2 1�9 1�8

Th 2�35 2�3 5�5 3�1 6�3 2�4 5�0 3�1 3�3

U 1�13 0�82 1�3 0�83 1�5 0�55 1�2 0�79 0�83

Cr 8514 9159 7436 8231 9627 689 2890 5553 5745

Co 19 22 24 22 29 35 37 29 30

Ti 2737 2028 8272 1632 4130 1910 6338 5793 8426

LaN/SmN 0�76 0�81 1�75 2�88 1�28 1�63 1�29 1�40 0�99

SmN/YbN 0�93 0�75 3�72 1�55 3�72 5�14 3�93 4�15 5�25

(continued)


311

Dow



Despite variable U and Th enrichment, clinopyroxenesdisplay smooth trace element patterns characterized byslight LREE depletion. Such REE compositions (inwhole-rock and clinopyroxene) are widely documented inlithospheric peridotites from both orogenic massifs andxenoliths, and can reflect either a pristine fertile mantlecomposition or a previously depleted lithospheric mantlemodified and refertilized by interaction with percolatingmelts (Vannucci et al., 1991; Bianchini et al., 2007; LeRouxet al., 2007; Rivalenti et al., 2007; Raffone et al., 2009).In the Tallante xenoliths, significant geochemical varia-

tions are not observed between the porphyroclastic andequigranular spinel peridotites. Nevertheless, it is notewor-thy that the clinopyroxenes in the equigranular peridotitesare slightly more homogeneous in terms of REE contents,and exhibit less pronounced Zr, Hf and Ti anomalies. Thissuggests, in agreement with microstructural evidence, alarger time-integrated melt^rock ratio in the equigranularperidotites and, in turn, indicates that clinopyroxenechemistry probably reflects refertilization by melt^rockinteraction. We therefore used the REE compositions ofclinopyroxenes in the equigranular peridotites 92T20 andT30, inferred to represent the peridotites most equilibratedwith the percolating melts, to derive the compositions ofthe equilibrium melts. Two sets of cpx/meltREE distributioncoefficients were used, for Si-undersaturated andSi-saturated melts (Vannucci et al., 1998; Ionov et al.,2002), to account for the chemical modification of percolat-ing melts from olivine to orthopyroxene saturation(inferred from microstructural evidence) during ascentand interaction with the host peridotites. The results areshown in Fig. 11a, together with the compositional fieldsfor most primitive basaltic andesites and alkaline magmasof the Alboran region (Duggen et al., 2004, 2005, 2008).Despite variable absolute REE concentrations, dependingon the choice of cpx/melt distribution coefficients,computed melts show moderate LREE enrichment andalmost flat MREE to HREE patterns, consistent withthe sub-alkaline magmatism of the Alboran Domain.We thus infer that the melt percolation events documentedin the Tallante spinel peridotites were related to thesub-alkaline magmatism that affected the Alboran region-since the latest Oligocene, presumably in response to pro-cesses in the upper mantle driving late-orogenicextension. It should be noted that a similar scenario hasbeen envisaged for the Ronda peridotites (Van derWal &Bodinier, 1996).

Melt impregnation: emplacement atshallow lithospheric depthExhumation of the Tallante mantle to shallower litho-spheric levels is documented by the occurrence of plagio-clase-bearing spinel peridotites showing peculiarmicrostructures (e.g. gabbroic pockets interstitial betweenporphyroclastic and equigranular grains) and anomalous

Table 5: Continued

Pyroxenite

92T9B T31B 92T9B 92T9B T31B

cpx amph amph amph phlog

idiom poikil poikil poikil idiom

wt %

SiO2 52�18 41�85 41�70 43�16 37�12

TiO2 0�74 3�81 4�39 2�69 4�17

Al2O3 4�36 12�97 13�03 13�14 15�14

Cr2O3 0�60 0�13 0�08 0�11 0�00

FeO 5�00 9�17 9�32 7�68 9�43

MgO 15�56 14�86 14�15 15�95 18�67

MnO 0�11 0�16 0�19 0�11 0�08

CaO 21�66 11�51 11�40 11�78 0�08

Na2O 0�36 2�59 2�40 2�44 1�84

K2O 0�01 1�29 1�41 1�31 8�81

Total 100�58 98�34 98�07 98�37 95�34

Mg-no. 84�7 78�0 77�0 82�0 78�0

ppm

Sc 63 39 33 40 5�34

V 366 428 492 420 296

Sr 84 574 778 514 62

Y 12�1 21�9 18�7 16�0 0�05

Zr 39 88 55 136 14�9

Nb 0�61 50 28 72 20�4

Ba — 543 710 534 4865

La 6�1 20�0 11�4 26�6 —

Ce 16�8 58�8 40�0 68�9 —

Pr 2�39 8�40 6�39 8�20 —

Nd 12�1 40�6 33�8 32�5 —

Sm 3�40 8�98 8�10 6�48 —

Eu 1�21 2�60 2�43 2�03 —

Gd 3�66 7�48 7�11 5�44 —

Tb 0�48 0�91 0�88 0�63 —

Dy 2�91 5�30 4�60 3�63 —

Ho 0�50 0�85 0�75 0�56 —

Er 1�16 2�14 1�85 1�49 —

Tm 0�13 0�26 0�22 0�19 —

Yb 0�93 1�75 1�25 1�12 —

Lu 0�11 0�23 0�16 0�15 —

Hf 1�78 3�09 1�99 3�47 0�34

Ta 0�08 2�89 1�66 3�58 1�40

Pb 2�9 12 6�2 9�4 10�6

Th 3�1 8�8 0�64 4�2 0�08

U 0�59 1�8 0�29 1�3 0�21

Cr 4145 605 231 579 723

Co 33 60 55 55 77

Ti 8653 27712 36055 18360 21818

LaN/SmN 1�12 1�40 0�88 2�57

SmN/YbN 3�97 5�57 7�05 6�27

Porph: mantle porphyroclast. React: new clinopyroxenecrystallized in the reaction zone. Idiom: idiomorphic grain.Poikil: poikilitic grain.


312

Dow



(410 vol. %) plagioclase modal enrichment. Similarmicrostructures have been widely documented in ophioliticand oceanic peridotites, and are commonly ascribed toentrapment and crystallization of migrating melts (Dick,1989; Cannat et al., 1990; Rampone et al., 1997, 2008a;Dijkstra et al., 2001; Tartarotti et al., 2002; Piccardo et al.,2004, 2007; Chazot et al., 2005).In plagioclase-bearing peridotite 92T1, clinopyroxenes

exhibit significant Sr depletion and overall enrichment inREE,Ti and Zr for a similar LREE depletion as comparedwith clinopyroxenes in the spinel peridotites (Figs 7 and8c). Comparable chemical features, specifically the REEincrease, have previously been documented in clinopyrox-enes from impregnated peridotites (Rampone et al., 1997,2008a; Barth et al., 2003; Dijkstra et al., 2003; Piccardoet al., 2004, 2007; Rampone & Borghini, 2008), andascribed to one (or a variable combination) of the follow-ing effects: (1) entrapment within the peridotites of small(53%) melt fractions; (2) melt^rock reaction at decreasingmelt mass; (3) increase of the solid/liquid trace elementpartition coefficients as a result of the chemical evolution

Table 6: Major (wt %) and trace (ppm) element compo-

sitions of plagioclases

T31A 92T9B T31B 92T1 92T1 92T1

wt %

SiO2 52�69 57�00 58�5 54�62 54�63 54�37

TiO2 0�00 0�07 0�12 0�02 0�00 0�01

Al2O3 30�55 27�49 26�06 28�73 28�76 28�96

FeO 0�34 0�22 0�23 0�06 0�07 0�09

MgO 0�00 0�00 0�00 0�03 0�03 0�05

MnO 0�00 0�00 0�00 0�00 0�01 0�02

CaO 11�93 8�99 7�00 11�90 11�96 11�93

Na2O 4�65 6�18 6�94 4�84 4�81 4�83

K2O 0�15 0�23 0�53 0�07 0�10 0�10

Total 100�31 100�18 99�38 100�28 100�37 100�35

An 58�6 44�6 35�8 57�6 57�9 57�7

ppm

Sr 189 223 256 265

Y 0�11 0�12 0�12 0�41

La 1�37 0�70 1�13 0�99

Ce 1�64 1�44 1�83 1�70

Pr 0�11 0�12 0�18 0�22

Nd 0�38 0�44 0�74 0�62

Sm 0�051 0�09 0�13 0�14

Eu 0�17 0�60 0�83 0�84

Gd 0�045 0�35 0�12 0�14

Ti 51 108 157 142

Fig. 9. Primitive mantle normalized trace element abundances ofclinopyroxenes and amphiboles in pyroxenites T31B and 92T9B.Normalizing values as in Fig. 8.

(a)

(b)

Fig. 10. (a) Variation of Mg-number vs Ti (�1000) (atoms per sixoxygens) in clinopyroxenes from pyroxenites, the reaction zone at thehost peridotite^pyroxenite contact, and host peridotites. The darkgrey field refers to the compositions of clinopyroxenes in the spinelperidotites (this study). (b) Primitive mantle normalized trace ele-ment abundances in clinopyroxenes as in (a); the dark and light greyfields refer, respectively, to the compositions of clinopyroxenes inspinel and plagioclase peridotites (this study). Normalizing values asin Fig. 8.


313

Dow



of the percolating melts. On the other hand, the similarityof LREE fractionation in clinopyroxenes from spinel andplagioclase peridotites suggests that the impregnatingmelts probably had a similar (sub-alkaline) chemical affin-ity to that of the percolating melts of the previous reactiveporous flow stages. This is also consistent with the diffuseoccurrence of orthopyroxene (�olivine) in the plagioclase-bearing gabbroic pockets, indicative of an opx-saturatedsignature of the impregnating melts. As inferred in ophioli-tic peridotites from the AlpsÂpennine system and theOthris Massif (Dijkstra et al., 2003; Piccardo et al., 2007;Rampone & Borghini, 2008; Rampone et al., 2008a),open-system reactive porous flow at spinel-facies condi-tions and melt impregnation at plagioclase-facies condi-tions could represent different stages of the same meltpercolation event at different lithospheric levels.

Specifically, the impregnated plagioclase-bearing perido-tites would correspond to the stage when upward migrat-ing melts reached rather shallow and colder lithosphericenvironments where conductive cooling caused melt crys-tallization and possibly entrapment.In this context it is important to note the occurrence, in

a few peridotites (e.g. 92T14, T11), of fine-grained rims ofplagioclaseþ olivine around spinel, indicating that someplagioclase formation can be related to subsolidus recrys-tallization from spinel- to plagioclase-facies conditionsaccording to the reaction orthopyroxene þ spinel! plagi-oclase þ olivine þ new pyroxenes. According to a recentexperimental study of the spinel^plagioclase transition(Borghini et al., 2009), crystallization in fertile anddepleted lherzolites of such a plagioclase-bearing assem-blage indicates re-equilibration at P5 0�9 GPa at about1000^11008C. In addition, the microstructures of the plagi-oclase-bearing peridotites clearly indicate that plagioclasecrystallization occurred in mantle rocks that previouslyexperienced the deformation, melt^rock interaction andannealing recrystallization history documented in thespinel peridotites. This constitutes strong evidence that theplagioclase-bearing xenoliths are not simply samples fromthe shallower levels of the Tallante lithosphere, but thatthey record the progressive exhumation of the Tallantemantle, presumably in response to progressive extensionin the Alboran region.

The melt intrusion events: from diffusedto focused melt migrationVeins and dykes of different compositions (gabbronoritesand olivineâmphibole pyroxenites) with clear cross-cutting relationships intrude both spinel and plagioclase-bearing peridotites. Veins and dykes provide furtherevidence for the exhumation of the Tallante mantletowards a shallower and colder lithospheric environment,reflected by the transition from porous melt flow tomagma emplacement in fractures, presumably related to achange in the rheology of the lithospheric mantle duringextension-related exhumation and cooling.Gabbronoritic xenoliths and veins comparable with

those described in this study have been reported in previ-ous studies of the Tallante xenoliths (Arai et al., 2003;Shimizu et al., 2004, 2008; Beccaluva et al., 2007). On thebasis of their Si-oversaturated composition, trace element(Th, U, REE, Sr) enrichment and radiogenic Sr isotopecompositions, it has been inferred that they representsubduction-related melts, either approaching MioceneBetic lamproites in composition (Beccaluva et al., 2007) oradakitic melts produced by partial melting of a sinkingslab beneath the Alboran Domain (Shimizu et al., 2004).Detailed microstructural and geochemical investigationsof the gabbronorites will be the subject of a separate paper(Rampone et al., in preparation), because they constituteone of the few documented occurrences of Si-oversaturated

(a)

(b)

Fig. 11. (a) Primitive mantle normalized trace element abundancesin computed melts in equilibrium with average clinopyroxene fromequigranular peridotite 92T20 and porphyroclastic^granular perido-tite T30. Symbols refer to different sets of partition coefficients usedfor the calculation (filled triangle, Ionov et al., 1992; filled square,Vannucci et al., 1998). The grey fields refer to the compositions of themost primitive basaltic andesites from the Alboran Domain (lightgrey; data from Duggen et al., 2004, 2008) and alkali basaltsfrom Cabezo Tallante (dark grey; data from Duggen et al., 2005).(b) Primitive mantle normalized trace element abundances of com-puted melts in equilibrium with averaged clinopyroxenes from pyrox-enitesT31B and 92T9B. Dark grey field as in (a).


314

Dow



large ion lithophile element enriched melts intruded andcrystallized within the shallow lithospheric mantle (Baliet al., 2008; Mazzucchelli et al., 2004), which justifies spe-cific attention. Preliminary geochemical investigations(Rampone et al., 2009) have confirmed that the parentalmelts to the gabbronorites are significantly enriched inLREE, Th, U and volatile (Cl) components: this providesfurther support for their slab-derived origin, and points toa continental crust (or sediment) source component,rather than oceanic crust. In the context of this study,which aims to explore the whole geodynamic evolutionrecorded in the Tallante mantle, the gabbronorites repre-sent an important stage, as they provide striking evidencefor the existence of a subducting slab beneath the region.The gabbronorite^norite veins are crosscut by olivine^

amphibole pyroxenite dykes, which represent the latestmagmatic event affecting the Tallante mantle. Accordingto phase relations in the H2O-bearing picrobasalticsystem (Helz, 1973; Allen & Boettcher, 1983; Ulmer,1989; Grove et al., 2003), the observed crystallizationorder in pyroxenites (olivine^clinopyoxeneâmphibole^plagioclase) is indicative of intrusion pressures above 0�5GPa, probably in the range 0�6^0�9 GPa. This is consistentwith the evidence that the pyroxenites are intruded withinplagioclase-bearing peridotites (equilibrated at P50�8GPa). The inferred pressures are further supported by the0�7^0�9 GPa estimates achieved using the Nimis & Ulmer(1998) single-clinopyroxene geobarometer.Major element mineral compositions in the amphibole

pyroxenites indicate that they crystallized from ratherevolved melts (e.g. Mg-number in olivine and clinopyrox-ene 78�2^79�7 and 80�4^84�7, respectively). Nevertheless,inferences on the chemical affinity of the parental meltscan be made using the trace element compositions of clino-pyroxene, which is an early crystallizing phase and thusbetter reflects the chemical composition of the migratingmelt. In terms of REE patterns, clinopyroxenes in the pyr-oxenites closely resemble clinopyroxenes crystallized fromalkaline melts (Fabries et al., 1989; Bodinier et al., 1990;Downes et al., 1991; Downes, 2001; see Fig. 9). This is con-firmed by the REE compositions of computed equilibriummelts (using cpx/meltREE distribution coefficients fromIonov et al., 2002; Fig. 11b); they are similar to the composi-tions of the host alkali basalts (Duggen et al., 2005) interms of LREE/HREE fractionation, although shifted tohigher absolute concentrations because of their moreevolved chemical signature. Parental melts to the amphi-bole pyroxenites were therefore alkaline magmas similarto theTallante host alkali basalts.It is noteworthy that the clinopyroxenes exhibit more

pronounced depletion in Nb, and to a minor extent Ta, rel-ative to clinopyroxenes in equilibrium with alkaline melts.In principle, this could be an effect of the early precipita-tion of Fe-oxides (similar to those occurring as inclusions

in clinopyroxene and amphibole), which can fractionatethese elements relative to the REE (Bodinier et al., 1996;Gregoire et al., 2000; Rivalenti et al., 2004). At the bulk-rock scale, the low Nb and Ta contents in clinopyroxeneare then primarily compensated by high Nb and Ta con-centrations in pargasitic amphibole, as well as moderateNb abundances in phlogopite.Both clinopyroxene and amphibole in the pyroxenites

also show remarkable Th and U enrichment, at least oneorder of magnitude higher than expected on the basis ofexperimentally determined mineral/melt partition coeffi-cients (Hauri et al., 1994; Lundstrom et al., 1994; Tiepoloet al., 2007) and ocean island basalt (OIB) compositions.Large Th and U concentrations can result from late-stageentrapment of small volumes of alkaline melts (Raffoneet al., 2009). This could be especially true for poikiliticminerals such as amphibole, as suggested by its larger Thand U variability (see Fig. 9). On the other hand, new clin-opyroxenes crystallized in the wall^peridotite reactionzone are similar to clinopyroxenes in pyroxenites in termsof U, Th and LREE (Fig. 10b): this argues against asimple trapped melt effect, and rather suggests that theobserved Th and U enrichment in clinopyroxene repre-sents the chemical signature of the parental melts.Chromatographic percolation of small, highly incompati-ble element enriched melt volumes through the host peri-dotites is then documented by the compositions ofclinopyroxene porphyroclasts in the host peridotites (i.e.outside the reaction zone), these latter being selectivelyenriched in Th, U and LREE relative to clinopyroxenesin porphyroclastic spinel peridotites (see the comparisonfield in Fig. 10b). This process has been widely documentedin metasomatized peridotites (Bedini et al., 1997; Ionovet al., 2002; Bodinier & Godard, 2003; Bodinier et al., 2004;Rivalenti et al., 2007; Raffone et al., 2009, and referencestherein), particularly in the host peridotites of amphibole-bearing veins crystallized from alkaline melts (Bodinieret al., 1990, 2004; Zanetti et al., 1996).In summary, we infer that the observed Th and U

enrichment in pyroxenite minerals possibly reflects thechemical signature of parental alkaline melts that couldhave been acquired by interaction with lithosphericmantle material enriched in these elements during previ-ous U^Th-rich metasomatic stages (such as gabbronorite^norite intrusion). In spite of this, the petrographic and goe-chemical characteristics of the amphibole pyroxenitesclearly indicate that their parental melts had alkalineaffinity, consistent with theTallante host alkali basalts.

Geothermobarometry data and theirsignificanceOur microstructural and geochemical study of theTallantexenoliths allows us to identify a multi-stage history ofrecrystallization, deformation, melt migration and intru-sion as outlined above. The P^T evolution of the Tallante


315

Dow



mantle associated with this history may help to elucidateits geodynamic significance; however, the thermobaro-metric data summarized below and their interpretationare surrounded with uncertainties, mostly related to late-stage cooling of the studied xenoliths at shallow litho-spheric depths.Early uplift of the Tallante mantle rocks and develop-

ment of spinel^pyroxene clusters (qualitatively shown inFig. 12 as field Ta1) was followed by ductile shear flow atspinel-facies conditions leading to a porphyroclastic micro-structure. Geothermometric estimates on the porphyro-clastic minerals using different methods (Brey & Ko« hler,1990; Taylor, 1998; Witt-Eickschen & O’Neill, 2005) haveyielded variable results, as has been frequently observedin the case of significantly exolved pyroxene grains. Moreconsistent results were obtained with the Brey & Ko« hler(1990) Ca-in-Opx and the Witt-Eickschen & O’Neill(2005) geothermometers. Most temperature estimates arein the range 960^10108C, with a few values up to 1050^11008C (shown as fieldTa2 in Fig. 12). In a few orthopyrox-ene porphyroclasts we analysed core^rim traverses; Caprofiles are generally smooth, and the resulting geother-mometric estimates, using the Brey & Ko« hler (1990)method, mostly range from 950 to 10008C, with a fewhigher temperatures (to 11008C) rarely preserved in thecores. SimilarTestimates (960^10208C) were also obtainedin spinel-facies porphyroclastic minerals from plagioclase-bearing peridotites. Overall this indicates that any earlyhigh-Tequilibration (predating exolutions in pyroxenes) ispoorly preserved. The data are consistent, however, withthe development of the porphyroclastic microstructure ina lithospheric environment.These early microstructures are variably overprinted by

recrystallization, which in all likelihood occurred in thepresence of percolating melts, as suggested by the replace-ment of porphyroclastic minerals by unstrained olivinerims and the subsequent crystallization of undeformed poi-kilitic orthopyroxene at the expense of porphyroclasts andnewly crystallized olivines. Temperature estimates on thesepoikilitic, replacive orthopyroxenes using the Brey &Ko« hler (1990) Ca-in-Opx method are in the range 960^10508C, similar to theT estimates on exsolved porphyro-clasts; but pressures (within the spinel stability field) arepoorly constrained (field Ta3 in Fig. 12).In several spinel peridotites recrystallization led to the

development of a granular microstructure. Despite themicrostructural evidence for high-temperature annealing,geothermometric estimates on these granular assemblagesagain yield rather low equilibration temperatures of�950^10008C. Progressive melt^rock interaction andmigration (annealing) recrystallization was probablyaccompanied by exhumation of the mantle rocks to shal-lower levels, where melts began to crystallize interstitialplagioclase and gabbroic pockets, causing the anomalous

(410 vol. %) plagioclase modal enrichment observed inthe plagioclase-bearing peridotite xenoliths. The lateststages of the Tallante mantle history are marked by theintrusion in spinel- as well as plagioclase-bearing perido-tites of gabbronorite veins followed by olivineâmphibolepyroxenites, which occurred at 0�7^0�9 GPa (tentativelyshown as fieldTa4 in Fig. 12).The sequence of P^T conditions obtained from the

Tallante xenoliths is characterized by markedly low tem-peratures, too low to account for the observed reactiveporous flow and melt impregnation events. In all docu-mented stages,Testimates vary in a narrow range, despitedifferent microstructural sites and mineral assemblages.This suggests that all of the geothermometric results maylargely reflect the effects of late-stage cooling before thexenoliths were sampled by the ascending host magmas.Such late-stage cooling in theTallante mantle seems consis-tent with documented geochronological results from thecrustal metamorphic rocks of the Alboran Domain, in par-ticular in the western Betics, in which U^Pb ion micro-probe dating of zircon, Ar/Ar dating of hornblende, Ar/Arlaserprobe dating of muscovite and biotite, and Ession-track analysis of zircon and apatite all reveal that coolingwas extremely rapid in the interval 21�2^20�4 Ma (e.g.Platt & Whitehouse, 1999; Platt et al., 2003); that is, wellbefore xenolith entrainment in the uprising host lavas,during the Pliocene.Unlike the uncertainty on the geothermometry results,

the reconstructed multi-stage history of deformation,recrystallization, melt migration and intrusion in theTallante xenoliths clearly points to about 30 km of exhuma-tion of the Tallante mantle, from P42 GPa (as indicatedby the orthopyroxene^spinel clusters after garnet), to0�7^0�9 GPa (marked by partial peridotite re-equilibrationat plagioclase-facies conditions and intrusion of alkalinepyroxenites). This implies exhumation to a shallow levelin response to Neogene lithosphere extension. In this con-text we emphasize that equilibration at 0�7^0�9 GPa of theTallante xenoliths is entirely consistent with current esti-mates of the present-day crustal thickness of 15^20 km inthe Alboran Sea and about 22 km in the Cartagena area(Torne et al., 2000).

THE TALLANTE MANTLE IN THEALBORAN CONTEXTComparison with the Ronda massifThe Alboran Domain is an exceptional region for uppermantle studies, in that both mantle xenoliths (Tallante)and orogenic peridotites (Ronda) are clearly associated inspace and time with one and the same large-scale orogenicprocess. This invites comparison of our xenolith resultswith the Ronda history as inferred by other studies (vander Wal, 1993; van der Wal & Vissers, 1993, 1996; van der


316

Dow



Wal & Bodinier, 1996; Garrido & Bodinier, 1999; Lenoiret al., 2001); with this aim we briefly review the main char-acteristics of that history.The Ronda peridotites contain evidence of an early stage

uplift from the diamond stability field to the deeper partof the spinel-facies field (Davies et al., 1993). This was fol-lowed by intense deformation and development of

porphyroclastic spinel-facies tectonites, representing theoldest deformational structure preserved (e.g. van derWal, 1993). Along the outer margin of the Ronda body,adjacent to the high-grade metamorphic crustal envelope,these tectonites pass into garnet^spinel mylonites in amylonitic shear zone of a few hundred meters width (seealso Precigout et al., 2007). Towards the core of the massif,

Fig. 12. P^Tdata and inferred P^T paths for theTallante mantle and Ronda peridotites. Peridotite solidus, garnet^spinel (g-s) reaction curveand ariegite^seiland subfacies boundary (a-s) redrawn after van der Wal & Vissers (1993); spinel^plagioclase peridotite boundary for fertilelherzolite (s-p) and depleted lherzolite (s-p’) according to Borghini et al. (2009). GRT, SPL and PLAG denote garnet, spinel and plagioclaseperidotite facies, respectively. Fields Ta1^Ta4 show qualitative estimates (circled fields) and thermobarometry results (rectangles) for theTallante mantle, withTa1 garnet breakdown to spinel^pyroxene clusters, Ta2 geothermometry results obtained in porphyroclastic microstruc-tures, Ta3 results for poikilitic replacive orthopyroxenes formed by melt^rock interaction and recrystallization, Ta4 development of plagio-clase-bearing assemblages, fine granoblastic recrystallization, and intrusion of gabbronoritic and olivineâmphibole-pyroxenite veins. A(dashed black line), plausible P^T trajectory for the Tallante mantle inferred from extensive evidence for melt^rock interaction, as well asfrom comparison with Ronda peridotites. (Note inferred cooling stage towards field Ta4 at shallow levels.) B (light grey), P^T path for theRonda peridotites according toVan derWal & Vissers (1993), with R1 early spinel-facies equilibration, R2 development of spinel tectonites, R3garnet^spinel mylonites, R4 conditions in spinel tectonites close to recrystallization front, R5 granular spinel peridotites, R6 emplacement-related plagioclase-facies shear zones. C (dark grey), P^T path inferred by Platt et al. (2003) for rocks at the Ronda recrystallization front.Shaded field near R2 and R3 represents equilibrium conditions during development of spinel tectonites and garnet^spinel mylonites, shadedfield labelled R5L shows conditions for the granular peridotites as estimated by Lenoir et al. (2001), shaded field at the s-p boundary betweenwet and dry solidus represents plausible conditions for the plagioclase-facies shear zones. Depth scale assumes average crustal density of2800 kg/m3. (For further explanation see text.)


317

Dow



the porphyroclastic microstructures, as well as occasionalnarrow mylonite zones, are overprinted by intense recrys-tallization and development of granular peridotites. Theboundary between the spinel tectonites and the granularrocks, currently known as the Ronda recrystallizationfront (van der Wal & Vissers, 1993, 1996) has been shownto result from extensive melt^rock interaction and melt-assisted recrystallization (van der Wal & Bodinier, 1996)affecting the foliated spinel tectonites and occasional mylo-nites at spinel-facies conditions. In a more recent study,Lenoir et al. (2001) have interpreted the Ronda recrystalli-zation front as a lithospheric melting front. The meltingand melt^rock interaction processes were followed by peri-dotite emplacement into the crust along high-temperatureplagioclase-facies shear zones developed in the deeperparts of the spinel-facies granular peridotites.Van derWal & Vissers (1993) ascribed the early uplift of

the Ronda peridotites from diamond facies across thegarnet^spinel facies boundary (stage R1 in Fig. 12) to theJurassic phase of extension and breakup in the Neotethys(Dercourt et al., 1986). They attributed the subsequentdeformation in the porphyroclastic spinel peridotites(spinel tectonites) and garnet^spinel mylonites to progres-sive ductile shearing in a subduction-zone hanging-wallenvironment during Cretaceous^Paleogene collision. Asubduction-zone setting for the spinel tectonites andgarnet^spinel mylonites was principally suggested on thebasis of the higher inferred pressures associated with thegarnet^spinel mylonites in combination with low tempera-tures obtained from geothermometry on syntectonicassemblages (810^9008C and 830^8808C for the spinel tec-tonites and garnet^spinel mylonites, respectively, denotedin Fig. 12 as stages R2 and R3). A subduction zone setting(i.e. with the Ronda mantle in the subduction zone hang-ing wall) was also proposed by Davies et al. (1993) on thebasis of the geochemistry of the garnet pyroxenites. Upliftof the Ronda rocks in the margin of the body proceededunder relatively cool conditions between 800 and 9008C(shown as a dashed trajectory in Fig. 12), whereas thedeeper part of the body became heated and extensivelyaffected by melting and melt^rock interaction processes(shown in Fig. 12 as stage R5), probably in response to con-vective removal of overthickened lithosphere or detach-ment of a subducting slab, causing ascent of asthenosphereand extensional thinning of the remaining overlying litho-sphere (Van der Wal & Vissers, 1993; Van der Wal &Bodinier et al., 1996; Garrido & Bodinier, 1999). This led toextensional exhumation of the peridotites along plagio-clase-facies ductile shear zones (stage R6 in Fig. 12) esti-mated at �22 Ma (Priem et al., 1979).Prior to any comparison of the Tallante xenolith results

with the Ronda peridotite history, three aspects of that his-tory need to be considered that have been discussed inrecent studies. First, as emphasized by Platt et al. (2003),

the P^T path suggested by van der Wal & Vissers (1993)is in fact a composite P^T trajectory for the Ronda perido-tite as a whole, rather than for single material pointswithin the massif. Second, van der Wal & Vissers (1993)obtained rather low temperatures (770^8808C) in thespinel tectonites close to the recrystallization front(labelled in Fig. 12 as R4), but renewed thermobarometrystudy by Lenoir et al. (2001) of similar samples has yieldedmuch higher temperatures of around 1050^11008C. Inaddition, Lenoir et al. (2001) concluded that the develop-ment of the recrystallization front occurred at tempera-tures in the range 1180^12258C and pressures near those ofthe ariegite^seiland subfacies boundary (field R5L inFig. 12); that is, at higher temperatures but also higher pres-sures than inferred by van der Wal & Vissers (1993).Third, with the aim of reconciling the structural andgeothermometric data from the peridotites with thoseseen in the crustal envelope, Platt et al. (2003) have recon-sidered the significance of the spinel tectonites and in par-ticular the garnet^spinel mylonites. Those workers haveelegantly shown that, instead of (Cretaceous toPaleogene) subduction-related lithosphere thickening, theinferred low-temperature conditions during developmentof the garnet^spinel mylonites may equally be consistentwith the onset of lithospheric extension, and that the mylo-nites may represent an extensional ductile shear zone thatdeformed at relatively low temperatures as a result of con-tinuous cooling against a hanging wall of previously thick-ened crust. Recent structural work (Precigout et al., 2007)in addition suggests that the spinel tectonites and garnet^spinel mylonites may form one heterogeneous shear zonesystem, such that the spinel tectonites may equally repre-sent extensional deformation in the upper mantle. On thebasis of this alternative interpretation, and using the vander Wal & Vissers (1993) estimates for the garnet^spinelmylonites in combination with the new Lenoir et al. (2001)thermobarometric results for the granular peridotites,Platt et al. (2003) proposed an alternative P^T trajectoryfor the Ronda peridotites, also shown in Fig. 12.Consistent with van der Wal & Vissers (1993, 1996) andVissers et al. (1995), the mantle uplift associated with thisevolution largely results from continuous extension andthinning during the early Miocene of the overlying crust,initially thickened to values of up to 55 km as a resultof Alpine collision, and thinned in response to a lateOligocene delamination of gravitationally unstable litho-sphere and consequent ascent of asthenosphere and ther-mal erosion of the remaining lithospheric mantle.The sequence documented in the Tallante xenoliths of

early uplift, followed by ductile flow in porphyrocasticspinel peridotites, in turn overprinted by extensive melt-assisted recrystallization, clearly recalls the earlier part ofthe Ronda history. The early uplift in the Tallante mantleevidenced by the spinel^pyroxene clusters after former


318

Dow



garnet could be the equivalent of the early stage uplift ofthe Ronda peridotite body. The porphyroclastic micro-structures and associated LPOs in the Tallante xenolithsmay represent a same stage of upper mantle deformationas the Ronda spinel tectonites and garnet^spinel mylonites.It is plausible that these ductile flow structures mark theonset of extensional deformation at upper mantle levels asoutlined above. The subsequent melt^rock interaction pro-cesses and associated annealing recrystallization in theTallante xenoliths and the melting and melt^rock interac-tion processes documented in the Ronda massif probablyreflect the same orogen-scale heating event.Accommodated by plagioclase-facies shear zones, the

Ronda massif became emplaced in a rapidly thinning pileof HP^LTcrustal rocks at 22 Ma (Priem et al.,1979), shortlyafter heating since around 27 (Platt & Vissers, 1989) or 25Ma (Platt et al., 2003). During the stages of melt^rockinteraction and melt impregnation, the Tallante mantlerocks record a similar uplift, from pressures around 2GPa to those of the plagioclase stability field (0�7^0�9GPa), and we suggest that this uplift likewise reflects rapidcrustal thinning. However, in contrast to the Ronda peri-dotites, the Tallante mantle rocks continued to reside inthe upper mantle until they were sampled by ascendingalkaline melts at 2�93^2�29 Ma (Duggen et al., 2005); thatis, for a time span of almost 20 Myr during which a signif-icant amount of the transient heat was probably lost byconductive cooling. This notion may have important con-sequences for comparisons between the two suites ofmantle rocks, and for the inferred P^T evolution of theTallante mantle.Although the similarity of the tectonic sequences in the

Tallante xenoliths and Ronda peridotites is obvious, corre-lation between rock microstructure and associated P^Tconditions in the two mantle domains is surrounded withuncertainties. In particular, there are significant micro-structural differences between the coarse-granular perido-tites of Ronda and the much finer-grained equigranularperidotites of Tallante. In addition, the Ronda coarse-granular peridotites differ geochemically from the spineltectonites and have been interpreted as the solid residuesafter partial melting of a spinel tectonite protolith (Lenoiret al., 2001). In Tallante, the equigranular peridotites donot show a marked geochemical difference from the por-phyroclastic xenoliths. As all recent studies converge onthe idea that the Ronda recrystallization^melting frontwas thermally controlled and associated with a significantthermal gradient through the spinel tectonites, one wouldexpect ‘Tallante-type’ equigranular peridotites ahead ofthe front (i.e. in the spinel tectonite domain). Such fine-granular, recystallized microstructures, however, are notobserved in Ronda. Following a suggestion by J. L.Bodinier (personal communication, 2009) we hypothesizethat the equigranular microstructures did not develop in

Ronda because of the rapid cooling of the massif, whicheffectively ‘quenched’ the deformation and melt^rock reac-tion microstructures.If our inferences on the analogy of the Tallante micro-

structural record and the Ronda structural and thermalhistory are essentially correct, it is likely that the Tallantemantle underwent re-equilibration during Neogene cool-ing over a time span of up to 20 Myr. This cooling stageexplains, in our view, the lack of a geothermometricrecord in the Tallante xenoliths of elevated temperaturesassociated with melt percolation and annealing recrystalli-zation, as any mineral equilibria reached during melt-assisted high-temperature annealing may well have beenreset by diffusion. In addition, and as a consequence ofthis younger part of the thermal history, it is also impossi-ble to ascertain whether the Tallante mantle ever experi-enced cooling before the development of (equi)granularperidotites; that is, before pervasive porous melt percola-tion, such as documented in the Ronda rocks for the transi-tion from spinel tectonites to garnet^spinel mylonites(with temperatures as low as 8508C). Early mineral equi-libria attained at such relatively low temperatures can beexpected to have been affected by re-equilibration duringsubsequent melt^rock interaction, and the analyticalresults from the porphyroclastic microstructures maysimply reflect the cumulative effects of reheating and sub-sequent late-stage cooling.

P^T history of the Tallante mantleIn view of the above considerations on the thermobaro-metric results, inferences on the P^T evolution of theTallante mantle are necessarily qualitative. However,given the marked similarity to the Ronda peridotiterecord, from the early uplift to spinel-facies conditions, fol-lowed by ductile flow, in turn overprinted by melt-assistedrecrystallization and exhumation from pressures around2 GPa to those of the plagioclase field, we suggest that theTallante mantle must have gone through an analogousP^T evolution, but probably at lower temperatures thanrecorded in Ronda, and ending in a distinct cooling stageat plagioclase-facies conditions (Fig. 12, black dashedpath) prior to sampling in Pliocene times. Lower climaxtemperatures during the stage of melt^rock interactionand melt impregnation in theTallante xenoliths are largelybased on the lack of evidence for partial melting related tothe development of equigranular peridotites, whereasthroughout the granular domain the Ronda massifwas affected by a partial melting event implying thatsuper-solidus (at least some ‘lithospheric’ hydrous solidus)temperatures were reached. More depleted, harzburgiticperidotites have been documented at Tallante byBeccaluva et al. (2004) but they were ascribed to pre-Paleozoic melting events. In addition, we recall that theRonda P^T path as proposed by Platt et al. (2003) and


319

Dow



shown in Fig. 12 is in part based on thermobarometric esti-mates for the coarse-granular rocks at the recrystallizationfront by Lenoir et al. (2001), who noted that these tempera-ture estimates should be considered as minimum values.On the other hand, microstructural evidence for melt per-colation and melt^rock interaction in theTallante xenolithsdoes indicate that ambient temperatures should have beenhigh enough for such processes to occur, and it seemslikely that temperatures exceeded 11008C during thatstage (Fig. 12).

Implications for west MediterraneangeodynamicsThere are three aspects of the Tallante mantle record withimportant consequences for the geodynamic evolution ofthe Alboran Domain. First, spinel-facies shearing was fol-lowed by a stage of reactive porous melt flow, melt impreg-nation and melt intrusion, and despite the lack of reliablegeothermometric evidence this calls for elevated tempera-tures during that stage. Second, theTallante mantle under-went some 30 km of exhumation, from P �2 GPa to 0�7^0�9 GPa, indicating an equivalent thinning of the overlyingcrustal and mantle rocks. Third, reactive porous flow andmelt impregnation in the Tallante spinel- and plagioclase-bearing peridotites were probably related to migration ofsub-alkaline melts, and this was followed by intrusion ofmelts of alkaline affinity.Consistent with interpretations of the upper mantle

structure in the Alboran region by Gutscher et al. (2002),Duggen et al. (2005) inferred that west-directed roll-backand steepening of a subducting oceanic plate inducedextension and thinning in the Betic^Rif belt, leading tothe formation of the Alboran Basin. We emphasize, how-ever, that slab roll-back by itself cannot account for theboth extreme and rapid heating at shallow depths docu-mented in the metamorphic crustal envelope of theRonda peridotites (Platt et al., 2003) and in the metamor-phic rocks at Site 976 in the Alboran Sea (Comas et al.,1999), nor for the high temperatures implied by the melt^rock interaction and melt impregnation processes in theTallante mantle and similar melt^rock interaction anddevelopment of a lithospheric melting front in the Rondamassif. Both the extreme heating and rapid crustal thin-ning and mantle exhumation lend support to some formof removal of overthickened lithosphere and ascent of hotasthenosphere (Platt & Vissers, 1989; Vissers et al., 1995).In this context we recall that Paleogene subduction of theLigurian ocean in the western Mediterranean regionoccurred in a northwesterly direction underneath theBalearic Islands, Sardinia and Corsica, leading to a colli-sional ridge between Iberia and North Africa (Fig. 13).The roll-back process is inferred to have started duringthe latest Oligocene to early Miocene (i.e. around 25Ma), when this subduction^collision system became

divided into two segments by some form of lithospheretearing (Spakman & Wortel, 2004) leading to separationof the Alboran and Calabrian parts of the collisionaldomain and subsequent west-directed roll-back of theAlboran segment to form the Alboran Basin (Fig. 13; seealso Gutscher et al., 2002; Booth-Rea et al., 2007). LateOligocene to early Miocene extensional collapse of the col-lisional ridge as envisaged by Platt & Vissers (1989) wouldhave coincided with this process of lithosphere rupture.We therefore hypothesize that the retreating subductionsystem, now reflected by a curved slab observed in seismictomography studies at the western end of the Betic^Rifarc (Fig. 13, shaded domain), evolved from Late Oligoceneremoval of gravitationally unstable lithosphere and ascentof hot asthenosphere, triggering extension and rapid tran-sient heating of the overlying remaining lithosphericcolumn. As a result, this remaining lithosphere deprivedof its lithospheric root started to extend and spread over aLigurian oceanic remnant located WSW of the collisionaldomain, which effectively led to subduction and roll-backof that oceanic lithosphere underneath the extending andthinning collisional ridge. This scenario is consistent withand supported by the notion that the Flysch Trough unitsin the western Betics were most probably floored by oce-anic crust (e.g. Booth-Rea et al., 2007, and referencestherein).The interaction with melts of different sources in the

Tallante spinel and plagioclase-bearing peridotites lendsstrong support to the above scenario. Uplift of theTallantemantle was accompanied by reactive porous flow andmelt impregnation probably related to migration of sub-alkaline melts, and this is consistent with the sub-alkalinemagmatism affecting the Alboran region since theOligocene. This magmatism has been largely ascribed tomelting of mantle sources contaminated by hydrous fluidsor melts derived from dehydration and/or melting of sub-ducting oceanic lithosphere, induced by slab roll-back andsteepening (Duggen et al., 2004, 2005, 2008; see, however,Turner et al., 1999).Duggen et al. (2005) also proposed a geodynamic sce-

nario to explain the Neogene transition from subduction-related to intraplate-type alkaline magmatism in theAlboran region, which can be observed both on a regionalscale and in single volcanic systems. Consistent with theinterpretations of the upper mantle structure in theAlboran region by Gutscher et al. (2002) they inferredthat, close to the Miocene^Pliocene boundary, continuingslab roll-back triggered delamination of bands of subconti-nental lithosphere (continental edge delamination) at theedges of the subducting plate (i.e. at the southern Iberiaand northern African plate margins), and that this causedupwelling of plume-contaminated sub-lithospheric mantle,which generated the alkaline magmatism as seen in theTallante volcanic center. Turner et al. (1999) instead


320

Dow



explained the underlying process in terms of an advancedstage of convective removal leading to the loss of theremaining thick lithosphere underneath the Iberian (andalso African) margin, whereas Spakman & Wortel (2004)described this stage of the upper mantle history in a differ-ent way; namely, as a laterally propagating detachment ofthe slab. In this context we note that a diachronouschange from marine to continental conditions in theNeogene basins (Meijninger, 2006) of the eastern Betics,from LateTortonian in the east to Early Messinian furtherwest, lends independent support to this latter hypothesis.The consequences in terms of mantle upwelling, however,should be the same in each of these interpretations, andwe emphasize that Cabezo Tallante is indeed locatedroughly above the slab edge as envisaged by Duggen et al.(2005); that is, above the zone of slab detachment asobserved in seismic tomography images by Spakman &Wortel (2004).Remarkably, the melt migration and intrusion stages

recognized in the Tallante peridotite xenoliths record the

same transition from sub-alkaline to alkaline magmatism,and this magmatic evolution occurred during progressiveexhumation probably accommodated by large-scaleextension-related crustal thinning as outlined above. Wetherefore propose that lithosphere extension, initiated bysome detachment process in a position near the BalearicIslands and followed by west-directed slab roll-back,led to uplift and migration of lithospheric mantle sectors(such as the mantle sampled at present at Tallante)from an inner part of the mantle wedge, where they expe-rienced deep (spinel-facies) pervasive porous melt flow,towards a position above a westward propagatingslab edge or slab detachment zone developing along thesouthern Iberian continental margin since the earlyTortonian. This allowed upwelling of plume-contaminatedsub-lithospheric mantle, which generated the alkalinemagmatism reflected in the shallow intrusion of olivineâmphibole pyroxenites at 0�7^0�9 GPa, followed by furtherascent of alkaline melts and sampling of the Tallantemantle domain.

23 Ma

Late Miocene slab detachment

Early Miocene lithosphere tearing

lithosphere tearing slab detachment in Middle Miocene

500 km

Ronda

Tallante

IBERIA

AFRICA

23 Ma

23 Ma

Late Oligocene

collisional domain

Calabriansegment

Balearic Islands

Alboransegment

Fig. 13. Kinematics of slab roll-back in the Betic^RifÂlboran region, slightly modified after Spakman &Wortel (2004). Grey shaded area indi-cates location of the BeticÂlboran slab at a depth of 200 km as observed in seismic tomography. Present-day coastlines shown as continuouslines; dashed lines indicate location of African margin and Balearic Islands 23 Myr ago. Dashed north^south-trending boundary in the eastindicates Late Oligocene separation of the Calabrian and Alboran segments. Set of curved dashed lines illustrates westward migration of thesubduction trench with time from near the Balearic Islands, and ending at the time of slab detachment under the Betics (Late Miocene).Initially, the trench retreated in a south to southwesterly direction while slab bending progressed. This was accommodated by Early MioceneWSW-directed lithosphere tearing along the Balearic margin, and simultaneous west-directed detachment evolving into lithosphere tearingalong the African margin. Along the Balearic margin, lithosphere tearing came to a halt during the Miocene when the trench became roughlyparallel to the Iberian margin, and laterally propagating slab detachment allowed the last stages of west-directed roll-back. (Note position ofCabezoTallante above zone of Late Miocene slab detachment.)


321

Dow



ACKNOWLEDGEMENTSWe acknowledge Laura Negretti (Genova) and AndreaRisplendente (Milano) for assistance with the EDS andWDS analyses. We are indebted for a helpful review byTomoaki Morishita, for challenging and helpful commentsby Arjan Dijkstra, and for a well-thought out and extre-mely constructive review by Jean-Louis Bodinier. Theirreviews have been of great help to better clarify the resultsof our work and to considerably improve the paper, notonly with respect to our study of the CabezoTallante xeno-liths but also regarding our comparison with a state-of-the-art structural and geochemical interpretation of theRonda massif.

REFERENCESAllen, J. C. & Boettcher, A. L. (1983). The stability of amphibole in

andesite and basalt at high pressures. American Mineralogist 68,307^314.

Arai, S., Shimizu, Y. & Gervilla, F. (2003). Quartz diorite veins in aperidotite xenolith fromTallante, Spain: implications for reactionand survival of slab-derived SiO2-oversaturated melts in theupper mantle. Proceedings of theJapan Academy, Series B 79, 145^150.

Ave¤ Lallemant, H. G. & Carter, N. L. (1970). Syntectonic recrystalli-zation of olivine and modes of flow in the upper mantle. GeologicalSociety of America Bulletin 81, 2203^2220.

Bali, E., Zajacz, Z., Kova¤ cs, I., Szabo¤ , C. S., Halter, W., Vaselli, O.,To« ro« k, K. & Bodnar, R. J. (2008). A quartz-bearing orthopyrox-ene-rich websterite xenolith from the Pannonian Basin, WesternHungary: evidence for release of quartz-saturated melts from asubducted slab. Journal of Petrology 49, 421^439.

Banks, C. J. & Warburton, J. (1991). Mid-crustal detachmentin theBetic system of southeast Spain.Tectonophysics 191, 275^289.

Barth, M. G., Mason, P. R. D., Davies, G. R., Djikstra, A. H. &Drury, M. R. (2003). Geochemistry of the Othris Ophiolite,Greece: evidence for refertilization? Journal of Petrology 44,1759^1785.

Beccaluva, L., Bianchini, G., Bonadiman, C., Siena, F. & Vaccaro, C.(2004). Coexisting anorogenic and subduction-related metasoma-tism in mantle xenoliths from the Betic Cordillera (southernSpain). Lithos 75, 67^87.

Bedini, R. M., Bodinier, J.-L., Dautria, J. M. & Morten, L. (1997).Evolution of LILE-enriched small melt fractions in the lithosphericmantle: a case study from the East African Rift. Earth and

Planetary Science Letters 153, 67^83.Bianchini, G., Beccaluva, L., Bonadiman, C., Nowell, G., Pearson, G.,

Siena, F. & Wilson, M. (2007). Evidence of diverse depletion andmetasomatic events in harzburgite^lherzolite mantle xenolithsfrom the Iberian plate (Olot, NE Spain): implications for litho-sphere accretionary processes. Lithos 94, 25^45.

Bodinier, J.-L. & Godard, M. (2003). Orogenic, ophiolitic and abyssalperidotites. In: Holland, H. D. & Turekian, K. K. (eds) Treatise onGeochemistry. Oxford: Elsevier Science, pp. 103^170.

Bodinier, J.-L., Vasseur, G., Vernieres, J., Dupuy, C. & Fabries, J.(1990). Mechanism of mantle metasomatism: geochemical evidencefrom Lherzorogenic peridotite. Journal of Petrology 31, 597^628.

Bodinier, J.-L., Merlet, C., Bedini, R. M., Simien, F. & Remaı̈di, M.(1996). Distribution of niobium, tantalum and other highly incom-patible trace elements in the lithospheric mantle: the spinel para-dox. Geochimica et Cosmochimica Acta 60, 545^550.

Bodinier, J.-L., Menzies, M. A., Shimizu, N., Frey, F. A. &McPherson, E. (2004). Silicate, hydrous and carbonate metasoma-tism at Lherz, France: contemporaneous derivatives of silicatemelt^harzburgite reaction. Journal of Petrology 45, 299^320.

Booth-Rea, G., Ranero, C. R., Mart|¤nez-Mart|¤nez, J. M. &Grevemeyer, I. (2007). Crustal types andTertiary tectonic evolutionof the Albora¤ n sea, western Mediterranean. Geochemistry, Geophysics,Geosystems 8, 1^25, doi:10.1029/2007GC001639.

Borghini, G., Fumagalli, P. & Rampone, E. (2009).The spinel^plagio-clase transition in the upper mantle: subsolidus experiments on fer-tile and depleted lherzolite. Journal of Petrology (submitted).

Brey, G. P. & Ko« hler, Y. (1990). Geothermobarometry in four-phaselherzolites II. New thermobarometers, and practical assesment ofexisting thermobarometers. Journal of Petrology 31, 1353^1378.

Cannat, M., Bideau, D. & Hebert, R. (1990). Plastic deformation andmagmatic impregnation in serpentinized ultramafic rocks fromthe Garrett transform fault (East Pacific Rise). Earth and PlanetaryScience Letters 101, 216^232.

Chazot, C., Charpentier, S., Kornprobst, J.,Vannucci, R. & Luais, B.(2005). Lithospheric mantle evolution during continental break-up:The West Iberia non-volcanic passive margin. Journal of Petrology46, 2527^2568.

Comas, M. C., Platt, J. P., Soto, J. L. & Watts, A. B. (1999).The originand tectonic history of the Alboran Basin: insights from Leg 161Results. In: Zahn, R., Comas, M. C. & Klaus, A. (eds) Proceedingsof the Ocean Drilling Program, Scientific Results, 161. College Station,TX: Ocean Drilling Program, pp. 555^580.

Coulon, C., Megartsi, M., Fourcade, S., Maury, R. C., Bellon, H.,Louni-Hacini, A., Cotten, J., Coutelle, A. & Hermitte, D. (2002).Post-collisional transition from calc-alkaline to alkaline volcanismduring the Neogene in Oranie (Algeria): magmatic expression ofa slab breakoff. Lithos 62, 87^110.

Davies, G. R., Nixon, P. H., Pearson, D. G. & Obata, M. (1993).Tectonic implications of graphitized diamonds from the Rondaperidotite massif, southern Spain. Geology 21, 471^474.

de Boom, J. (1994). The influence of secondary phases on olivine grainsize and deformation mechanisms in upper mantle rocks. M.Sc.thesis, Utrecht University, 38 pp.

Dercourt, J., Zonenshain, L. P., Ricou, L.-E., Kazmin, V. G., LePichon, X., Knipper, A. L., Grandjacquet, C., Sbortshikov, I. M.,Geyssant, J., Lepvrier, C., Pechersky, D. H., Boulin, J., Sibuet, J.-C., Savostin, L. A., Sorokhtin, O.,Westphal, M., Bazhenov, M. L.,Lauer, J. P. & Biju-Duval, B. (1986). Geological evolution of theTethys belt from the Atlantic to the Pamirs since the Lias.Tectonophysics 123, 241^315.

Dick, H. J. B. (1989). Abyssal peridotites, very low spreading ridgesand ocean ridge magmatism. In: Saunders, A. d. & Norry, M. J.(eds) Magmatism in the Ocean Basins. Geological Society, London, Special

Publications 42, 71^105.Dijkstra, A. H., Drury, M. R. & Vissers, R. M. L. (2001). Structural

petrology of plagioclase lherzolites in the West Othris Mountains(Greece): melt impregnation in mantle lithosphere. Journal of

Petrology 42, 5^24.Dijkstra, A. H., Barth, M. G., Drury, M. R., Mason, P. R. D. &

Vissers, R. M. L. (2003). Diffuse porous melt flow and melt^rockreaction in the mantle lithosphere at the slow-spreading ridge: astructural petrology amd LA-ICP-MS study of the OthrisPeridotite Massif (Greece). Geochemistry, Geophysics, Geosystems 4,1^24.

Doblas, M., Lopez-Ruiz, J. & Cebria, J.-M. (2007). Cenozoic evolu-tion of the Alboran Domain: a review of the tectonomagmaticmodels. In: Beccaluva, L., Bianchini, G. & Wilson, D. S. (eds)


322

Dow



CenozoicVolcanism in the Mediterranean Area. Geological Society of America,

Special Papers 418, 303^320.Docherty, C. & Banda, E. (1995). Evidence for the eastward migration

of the Alboran Sea based on a regional subsidence analysis: a casefor basin formation by delamination of the subcrustal lithosphere?Tectonics 14, 804^818.

Downes, H. (2001). Formation and modification of the shallow litho-spheric mantle: a review of geochemical evidence from ultramaficxenolith suites and tectonically emplaced ultramafic massifs ofWestern and Central Europe. Journal of Petrology 42, 233^250.

Downes, H., Bodinier, J.-L., Thirlwall, M. F., Lorand, J.-P. &Fabries, J. (1991). REE and Sr^Nd isotopic geochemistry ofEastern Pyrenees peridotite massifs: sub-continental lithosphericmantle modified by continental metasomatism. Journal of Petrology(Special Lherzolite Issue) 97^115.

Duggen, S., Hoernle, K., van den Bogaard, P., Ru« pke, L. & PhippsMorgan, J. (2003). Deep roots of the Messinian salinity crisis.Nature 422, 602^606.

Duggen, S., Hoernle, K., van den Bogaard, P. & Harris, C. (2004).Magmatic evolution of the Alboran Region: the role of subductionin forming the western Mediterranean and causing the MessinianSalinity Crisis. Earth and Planetary Science Letters 218, 91^108.

Duggen, S., Hoernle, K., van den Bogaard, P. & Garbe-Scho« nberg, D.(2005). Post-collisional transition from subduction- to intraplate-type magmatism in the westernmost Mediterranean: evidence forcontinental-edge delamination of subcontinental lithosphere.Journal of Petrology 46, 1155^1201.

Duggen, S., Hoernle, K., Klugel, A., Geldmacher, J., Thirlwall, M.,Hauff, F., Lowry, D. & Oates, N. (2008). Geochemical zonation ofthe Miocene Alboran Basin volcanism (westernmostMediterranean): geodynamic implications. Contributions to

Mineralogy and Petrology 156, 577^593.Fabries, J., Bodinier, J.-L., Dupuy, C., Lorand, J.-P. & Benkerrou, C.

(1989). Evidence for modal metasomatism in the orogenic spinellherzolite body from Cassou (North-eastern Pyrenees, France).Journal of Petrology 30, 199^228.

Garc|¤a-Duen‹ as,V., Balanya¤ , J. C. & Mart|¤nez-Mart|¤nez, J. M. (1992).Miocene extensional detachments in the outcropping basement ofthe northern Alboran basin (Betics) and their tectonic implica-tions. Geo-Marine Letters 12, 88^95.

Garc|¤a-Herna¤ ndez, M., Lo¤ pez-Garrido, A. C., Rivas, P., Sanz deGaldeano, C. & Vera, J. A. (1980). Mesozoic palaeogeographic evo-lution of the External zones of the Betic Cordillera. Geologie en

Mijnbouw 59, 155^168.Garrido, C. & Bodinier, J.-L. (1999). Diversity of mafic rocks in the

Ronda peridotite: evidence for pervasive melt/rock reactionduring heating of subcontinental lithosphere by upwelling astheno-sphere. Journal of Petrology 40, 729^754.

Gill, R. C. O., El Aparicio, A., El Azzouzi, M., Hernandez, J.,Thirlwall, M. F., Bourgois, J. & Marriner, G. F. (2004). Depletedarc volcanism in the Alboran Sea and shoshonitic volcanism inMorocco: geochemical and isotopic constraints on Neogene tec-tonic processes. Lithos 78, 363^388.

Gregoire, M., Lorand, J. P., O’Reilly, S. Y. & Cottin, J. Y. (2000).Armalcolite-bearing, Ti-rich metasomatic assemblages in harzbur-gitic xenoliths from the Kerguelen Archipelago: implications forthe oceanic mantle budget of high-field strength elements.Geochimica et Cosmochimica Acta 64, 673^694.

Grove, T. L., Elkins-Talton, L. T., Barman, S. W., Chatterjee, N.,Mu« ntener, O. & Gaetani, G. A. (2003). Fractional crystallizationand mantle-melting controls on calc-alkaline differentiationtrends. Contributions to Mineralogy and Petrology 145, 515^533.

Gutscher, M.-A., Malod, J., Rehault, J.-P., Contrucci, I.,Klingelhoefer, F., Mendes-Victor, L. & Spakman, W. (2002).Evidence for active subduction beneath Gibraltar. Geology 30,1071^1074.

Hauri, E. H.,Wagner, T. P. & Grove, T. L. (1994). Experimental andnatural partitioning of Th, U, Pb and other trace elements betweengarnet, clinopyroxene and basaltic melts. Chemical Geology 117,149^166.

Heilbronner, R. & Tullis, J. (2002). The effect of static annealing onmicrostructures and crystallographic preferred orientations ofquartzites experimentally deformed in axial compression andshear. In: De Meer, S., Drury, M. R., De Bresser, J. H. P. &Pennock, G. M. (eds) Deformation Mechanisms, Rheology andTectonics:

Current Status and Future Perspectives. Geological Society, London, Special

Publications 200, 191^218.Helz, R. T. (1973). Phase relations of basalts in their melting range at

PH2O¼ 5 kb as a function of oxygen fugacity. Journal of Petrology14, 249^302.

Hoernle, K., van den Bogaard, P., Duggen, S., Mocek, B & Garbe-Schonberg, D. (1999). Evidence for Miocene subduction beneaththe Alboran Sea: 40Ar/39Ar dating and geochemistry of vocanicricks from Holes 977A and 978A. In: Zahn, R., Comas, M. C. &Klaus, A. (eds) Proceedings of the Ocean Drilling Program, Scientific

Results, 161. College Station, TX: Ocean Drilling Program,pp. 357^373.

Ionov, D. A., Savoyant, L. & Dupuy, C. (1992). Application of theICP-MS technique to trace element analysis of peridotites andtheir minerals. Geostandards Newsletter 16, 311^315.

Ionov, D. A., Bodinier, J.-L., Mukasa, S. B. & Zanetti, A. (2002).Mechanisms and sources of mantle metasomatism: major andtrace element conditions of peridotite xenoliths from Spitzbergenin the context of numerical modelling. Journal of Petrology 43(12),2219^2259.

Jolivet, L. & Faccenna, C. (2000). Mediterranean extension and theAfricanÊurasian collision.Tectonics 19, 1095^1106.

Jung, H. & Karato, S. (2001). Water-induced fabric transitions in oli-vine. Science 293, 1460^1463.

Kaeser, B., Kalt, A. & Pettke, T. (2006). Evolution of the lithosphericmantle beneath the Marsabit Volcanic Field (northern Kenya):constraints from textural, P^T and geochemical studies on xeno-liths. Journal of Petrology 47(11), 2149^2184.

Karato, S., Jung, H., Katayama, I. & Skemer, P. (2008). Geodynamicsiginficance of seismic anisotropy of the upper mantle: new insightsfrom laboratory studies. Annual Review of Earth and Planetary Sciences

36, 59^95.Katayama, I., Jung, H. & Karato, S. (2004). New type of olivine fabric

at modest water content and low stress. Geology 32, 1045^1048.Kelemen, P. B. (1990). Reaction between ultramafic wall rock and

fractionating basaltic magma, I, Phase relations, the origin ofcalc-alkaline magma series, and the formation of discordantdunite. Journal of Petrology 31, 51^98.

Kelemen, P. B. & Dick, H. J. B. (1995). Focused melt flow and localizeddeformation in the upper mantle: juxtaposition of replacive dunitesand ductile shear zones in the Josephine peridotite, SW Oregon.Journal of Geophysical Research 100, 423^438.

Kelemen, P. B., Dick, H. J. B. & Quick, J. E. (1992). Formation of harz-burgites by pervasive melt^rock reaction in the upper mantle.Nature 358, 635^641.

Kelemen, P. B., Shimizu, N. & Salters, J. M. (1995). Extraction of mid-ocean-ridge basalt from the upwelling mantle by focused flow ofmelt in dunite channels. Nature 375, 747^753.

Kelemen, P. B., Hirth, G., Shimizu, N., Spiegelman, M. & Dick, H. J.B. (1997). A review of melt migration processes in the adiabatically


323

Dow



upwelling mantle beneath oceanic spreading ridges. PhilosophicalTransactions of the Royal Society of London, Series A 355, 283^318.

Leake, B. E. (1997). Nomenclature of amphiboles: report of thesubcommittee on amphiboles of the international mineralogicalassociation, commission on new minerals and mineral names.American Mineralogist 82, 1019^1037.

Lenoir, X., Garrido, C. J., Bodinier, J.-L., Dautria, J. M. &Gervilla, F. (2001). The recrystallization front of the Ronda perido-tite: evidence from melting and thermal erosion of subcontinentallithosphere mantle beneath the Alboran basin. Journal of Petrology42, 141^158.

LeRoux, V., Bodinier, J.-L., Tommasi, A., Alard, O., Dautria, J.-M.,Vauchez, A. & Riches, A. J. V. (2007). The Lherz spinel lherzolite:refertilized rather than pristine mantle. Earth and Planetary Science

Letters 259, 599^612.Lonergan, L. & White, N. (1997). Origin of the Betic^Rif mountain

belt.Tectonics 16, 504^522.Lundstrom, C. C., Shaw, H. F., Ryerson, F. J., Phinney, D. L., Gill, J.

B. & Williams, Q. (1994). Compositional controls on the partition-ing of U,Th, Ba, Pb, Sr and Zr between clinopyroxene and haplo-basaltic melts: implications for uranium series disequilibria inbasalts. Earth and Planetary Science Letters 128, 407^423.

Mazzucchelli, M., Correia, C. T., Rivalenti, G., Peressini, G.,Zanetti, A., Tassinari, C. C. G. & Girardi, V. A. V. (2004).Mesozoic igneous event into the Baldissero mantle peridotite(Ivrea^Verbano Zone, NW Italy). EGU04, Nice (France), 25^30April. Geophysical Research Abstracts 6, 03966 .

McDonough,W. F. & Sun, S.-s. (1995). The composition of the Earth.Chemical Geology 120, 223^253.

Meijninger, B. M. L. (2006). Late-orogenic extension and strike-slipdeformation in the Neogene of southeastern Spain. Geologica

Ultraiectina 269, 179 .Miller, C., Zanetti, A., Tho« ni, M. & Konzett, J. (2007). Trace element

mineral chemistry of the type locality (Koralpe, Saualpe) andPohorje eclogites (Eastern Alps): implications for behaviour offluid-mobile elements in a continental subduction zone, geochro-nology and geothermometry. Chemical Geology 239, 96^123.

Nimis, P. & Ulmer, P. (1998). Clinopyroxene geobarometry of mag-matic rocks. Part 1: An expanded structural geobarometer foranhydrous and hydrous, basic and ultrabasic systems. Contributionsto Mineralogy and Petrology 133, 122^135.

Piccardo, G. B. & Vissers, R. L. M. (2007). The pre-oceanic evolutionof the Erro^Tobbio peridotite (Voltri Massif, Ligurian Alps,Italy). Journal of Geodynamics 43, 417^449.

Piccardo, G. B., Muntener, O., Zanetti, A. & Pettke, T. (2004).Ophiolitic peridotites of the AlpineÂpennine system: mantle pro-cesses and geodynamic relevance. International Geological Review 46,1119^1159.

Piccardo, G. B., Zanetti, A. & Muntener, O. (2007). Melt/peridotiteinteraction in the Southern Lanzo peridotite: Field, textural andgeochemical evidence. Lithos 94, 181^209.

Platt, J. P. & Vissers, R. L. M. (1989). Extensional collapse of thick-ened continental lithosphere: a working hypothesis for theAlboran Sea and Gibraltar Arc. Geology 17, 540^543.

Platt, J. P. & Whitehouse, M. J. (1999). Early Miocene high-tempera-ture metamorphism and rapid exhumation in the Betic Cordillera(Spain): evidence from U^Pb zircon ages. Earth and Planetary

Science Letters 171(4), 591^605.Platt, J. P., Soto, J. I., Comas, N. C. & Leg 61 Shipboard Scientists.

(1996). Decompression and high-temperature low-pressure meta-morphism in the exhumed floor of an extensional basin, AlboranSea, western Mediterranean. Geology 24, 447^450.

Platt, J. P., Argles,T.W., Carter, A., Kelley, S. P.,Whitehouse, M. J. &Lonergan, L. (2003). Exhumation of the Ronda peridotite and itscrustal envelope: constraints from thermal modelling of a P^T^time array. Journal of the Geological Society, London 160, 655^676.

Precigout, J., Gueydan, F., Gapais, D., Garrido, C. J. & Essaifi, A.(2007). Strain localization in the subcontinental mantleça ductilealternative to the brittle mantle.Tectonophysics 445(3^4), 318^336.

Priem, H. N. A., Hebeda, E. H., Boelrijk, N. A. I. M.,Verdurmen, T.E. A. & Oen, I. S. (1979). Isotopic dating of the emplacement ofthe ultramafic masses in the Serrania de Ronda, southern Spain.Contributions to Mineralogy and Petrology 70, 103^109.

Quick, J. E. (1981). The origin and significance of large, tabulardunites bodies in the Trinity peridotite, Northern California.Contributions to Mineralogy and Petrology 78, 413^422.

Raffone, N., Chazot, G., Pin, C., Vannucci, R. & Zanetti, A. (2009).Metasomatism in the lithospheric mantle beneath Middle Atlas(Morocco) and the origin of Fe- and Mg-rich wehrlites. Journal ofPetrology 50, 197^249.

Rampone, E. & Borghini, G. (2008). The melt intrusion/interactionhistory of the Erro^Tobbio peridotites (Ligurian Alps, Italy):Insights on mantle processes at non-volcanic passive margins.EuropeanJournal of Mineralogy 20, 573^585.

Rampone, E., Piccardo, G. B., Vannucci, R. & Bottazzi, P. (1997).Chemistry and origin of trapped melts in ophiolitic peridotites.Geochimica et Cosmochimica Acta 61, 4557^4569.

Rampone, E., Piccardo, G. B. & Hofmann, A.W. (2008a). Multi-stagemelt^rock interaction in the Mt. Maggiore (Corsica, France)ophiolitic peridotites: microstructural and geochemical records.Contribution to Mineralogy and Petrology 156, 453^475.

Rampone, E., Vissers, R. L., Poggio, M., Scambelluri, M. &Zanetti, A. (2008b). Subduction- vs- intraplate-type melt migrationin the Alboran lithospheric mantle: insights from theTallante xeno-liths (Betic Cordillera, SE Spain). EOS Transctions, American

Geophysical Union 89(53), Fall Meeting Supplement, Abstract V33C-2234.

Rampone, E., Vissers, R. L. M., Poggio, M., Scambelluri, M. &Zanetti, A. (2009). Subduction-related melts in the Alboranmantle: the Tallante xenoliths (B. Cordillera, SE Spain).Goldschmidt Conference Abstracts, p. A1071.

Rivalenti, G., Zanetti, A., Mazzucchelli, M., Vannucci, R. &Cingolani, C. A. (2004). Equivocal carbonatite markers in themantle xenoliths of the Patagonia backarc: the GobernadorGregores case (Santa Cruz Province, Argentina). Contributions toMineralogy and Petrology 147, 647^670.

Rivalenti, G., Zanetti, A., Girardi, V. A. V., Mazzucchelli, M.,Tassinari, C. G. C. & Bertotto, G. W. (2007). The effect of theFernando de Noronha plume on the mantle lithosphere in north-eastern Brazil. Lithos 94, 111^131.

Royden, L. H. (1993). Evolution of the retreating subduction bound-aries formed during continental collision.Tectonics 12, 629^638.

Seyler, M., Lorand, J. P., Dick, H. J. B. & Drouin, M. (2007).Pervasive melt percolation reactions in ultra-depleted refractoryharzburgites at the Mid-Atlantic Ridge, 158200N: ODP Hole 1274A.Contributions to Mineralogy and Petrology 153, 303^319.

Shimizu, Y., Arai, S., Morishita, T., Yurimoto, H. & Gervilla, F.(2004). Petrochemical characteristics of felsic veins in mantle xeno-liths from Tallante (SE Spain): an insight into activity of silicicmelt within the mantle wedge. Transactions of the Royal Society of

Edinburgh: Earth Sciences 95, 265^276.Shimizu,Y., Arai, S., Morishita,T. & Ishida,Y. (2008). Origin and sig-

nificance of spinel^pyroxene symplectite in lherzolite xenolithsfromTallante, SE Spain. Mineralogy and Petrology 94, 27^43.


324

Dow



Spakman, W. & Wortel, R. (2004). A tomographic view on WesternMediterranean Geodynamics. In: Cavazza, W., Roure, F.,Spakman, W., Stampfli, G. M. & Ziegler, P. (eds) The Trasmed

Atlas, The Mediterranean Region from Crust to Mantle. Berling,Heidelberg, NewYork: Springer, pp. 31^52.

Tartarotti, P., Susini, S., Nimis, P. & Ottolini, L. (2002). Melt migra-tion in the upper mantle along the Romanche Fracture Zone(Equatorial Atlantic). Lithos 63, 125^149.

Taylor, W. R. (1998). An experimental test of some geothermometerand geobarometer formulations for upper mantle peridotites withapplication to the thermobarometry of fertile lherzolite and garnetwebsterite. NeuesJahrbuch fu« r Mineralogie, Abhandlungen 172, 381^408.

Tiepolo, A., Bottazzi, P., Palenzona, M. & Vannucci, R. (2003). Alaser probe coupled with ICP- double-focusing sector-field massspectrometer for in situ analysis of geological samples and U^Pbdating of zircon. Canadian Mineralogist 41, 259^272.

Tiepolo, M., Oberti, R., Zanetti, A.,Vannucci, R & Foley, S. F. (2007).Trace-element partitioning between amphibole and silicate melt.In: Hawthorne, F.C., Oberti, R., DellaVentura, G. & Mottana, A.(eds) Amphiboles: Crystal Chemistry, Occurrence, and Health Issues.

Mineralogical Society of America, Reviews in Mineralogy 67, 417^451.Torne, M., Fernandez, M., Comas, M. C. & Soto, J. I. (2000).

Lithospheric structure beneath the Alboran basin: results from 3Dgravity modeling and tectonic relevance. Journal of Geophysical

Research 105, 3209^3228.Tub|¤a, J. M., Cuevas, J. & Esteban, J. J. (2004). Tectonic evidence in

the Ronda peridotites, Spain, for mantle diapirism related to dela-mination. Geology 32, 941^944.

Turner, S. P., Platt, J. P., George, R. M. M., Kelley, S. P., Pearson, D.G. & Nowell, G. M. (1999). Magmatism associated with orogeniccollapse of the BeticÂlboran Domain, SE Spain. Journal of

Petrology 40, 1011^1036.Ulmer, P. (1989). High pressure phase equilibria of a calc-alkaline

picro-basalt: implications for the genesis of calc-alkaline magmas.Carnegie Institution ofWashingtonYearbook 88, 28^35.

Van der Wal, D. (1993). Deformation processes in mantle peridotites.Geologica Ultraiectina 102, 180 pp.

Van derWal, D. & Bodinier, J. L. (1996). Origin of the recrystallizationfrom the Ronda peridotite by km-scale pervasive porous melt flow.Contributions to Mineralogy and Petrology 122, 387^405.

Van derWal, D. & Vissers, R. L. M. (1993). Uplift and emplacement ofupper mantle rocks in theW Mediterranean. Geology 21, 1119^1122.

Van derWal, D. & Vissers, R. L. M. (1996). Structural petrology of theRonda peridotite, SW Spain: deformation history. Journal of

Petrology 37, 23^43.Vannucci, R., Shimizu, N., Bottazzi, P., Ottolini, L., Piccardo, G. B.

& Rampone, E. (1991). Rare earth and trace element geochemistryof clinopyroxenes from the Zabargad peridotite^pyroxenite asso-ciation. Journal of Petrology, Special Lherzolites Issue, 255^269.

Vannucci, R., Bottazzi, P., Wulff-Pedersen, E. & Neumann, E.-R.(1998). Partitioning of REE,Y, Sr, Zr andTi between clinopyroxeneand silicate melts in the mantle under La Palma (Canary Island):implications for the nature of the metasomatic agents. Earth and

Planetary Science Letters 158, 39^51.Van Roermund, H. L. M., Drury, M. R., Barnhoorn, A. & de

Ronde, A. (2001). Relict majoritic garnet microstructures fromultra-deep orogenic peridotites in western Norway. Journal of

Petrology 42, 117^130.Vauchez, A. & Garrido, C. J. (2001). Seismic properties of an astheno-

spherized lithospheric mantle: constraints from lattice preferredorientations in peridotite from the Ronda massif. Earth and

Planetary Science Letters 192, 235^249.Vissers, R. L. M., Platt, J. P. & Van derWal, D. (1995). Late orogenic

extension of the Betic Cordillera and the Alboran domain: A litho-spheric view.Tectonics 14, 786^803.

Witt-Eickschen, G. & O’Neill, H. St. C. (2005). The effect of tempera-ture on the equilibrium distribution of trace elements between clin-opyroxene, orthopyroxene, olivine and spinel in upper mantleperidotite. Chemical Geology 221, 65^101.

Wortel, M. J. R. & Spakman,W. (2000). Subduction and slab detach-ment in the Mediterranean^Carpathian region. Science 290,1910^1917.

Zanetti, A., Vannucci, R., Bottazzi, P., Oberti, R. & Ottolini, L.(1996). Infiltration metasomatism at Lherz as monitored by sys-tematic ion-microprobe investigations close to a hydrous dike.Chemical Geology 134, 113^133.

Zeck, H. P., Kristensen, A. B. & Williams, I. S. (1998). Post-collisionalvolcanism in a sinking slab settingçcrustal anatectic origin of pyr-oxene-andesite magma, Caldear volcanic province, southeasternSpain. Lithos 45, 499^522.

Zhang, S. & Karato, S. (1995). Lattice preferred orientation of olivineaggregates deformed in simple shear. Nature 375, 774^777.


325

Dow



the Tallante Mantle Xenolith Record (Betic Cordillera, SE Spain)

Documents