160Ma of sporadic basaltic activity on the Languedoc volcanic … · 2011. 1. 26. · 160 Ma of sporadic basaltic activity on the Languedoc volcanic line (Southern France): A peculiar

Lithos 120 (2010) 202–222

Contents lists available at ScienceDirect

Lithos

j ourna l homepage: www.e lsev ie r.com/ locate / l i thos

160 Ma of sporadic basaltic activity on the Languedoc volcanic line(Southern France): A peculiar case of lithosphere–asthenosphere interplay

Jean-Marie Dautria ⁎, Jean-Michel Liotard, Delphine Bosch, Olivier AlardGéosciences Montpellier, UMR 5243, CC 60,Université Montpellier 2 Place E. Bataillon, 34095 Montpellier cedex 5, France

⁎ Corresponding author. Tel.: +33 467143293; fax: +E-mail address: [email protected] (J.-M. D

0024-4937/$ – see front matter © 2010 Elsevier B.V. Adoi:10.1016/j.lithos.2010.04.009

a b s t r a c t
a r t i c l e i n f o
Article history:Received 21 July 2009Accepted 16 April 2010Available online 26 April 2010

Keywords:Alkali basaltsPeridotite xenolithsLithosphereAsthenosphereCarbonated metasomatism

The N–S Languedoc volcanic province between the French Massif Central and the Mediterranean Sea ischaracterized by sporadic, scattered, low volume (∼2 km3) and geochemically homogeneous alkali basalticactivity, spanning from 161 to 0.5 Ma. The existence of magmatic activity of such a long duration within sucha small area (∼140 km long and ∼60 km wide), in spite of the extensive shift to the East of the Europeanplate (about 2500 km during the last 160 Ma) is problematic. Trace-element abundances in lavas suggest lowdegrees of melting (1–5%) in the spinel–garnet transition zone of an enriched lherzolitic source. The lavasdisplay rather large ranges in Sr isotopic ratios (0.70307–0.70436). The 143Nd/144Nd ratio variations aresmaller (0.51268–0.51300) and these of 206Pb/204Pb, 208Pb/204Pb and 207Pb/204Pb are 18.745–19.515, 38.532–39.228 and 15.567–15.680 respectively. The Languedoc lithospheric mantle, as sampled by xenoliths, isglobally similar to the Pyrenees lithosphere. The xenoliths show also rather large Sr, Nd and Pb isotopicvariations (87Sr/86Sr: 0.70287–0.70578; 143Nd/144Nd: 0.51256–0.51414; 208Pb/204Pb: 37.772–39.041; 206Pb/204Pb: 17.901–19.353) except for 207Pb/204Pb (15.570–15.620). The 206Pb/204Pb and La/Sm ratios arepositively correlated both in xenoliths and lavas. The increase of the 206Pb/204Pb (which could be interpretedas participation of the European Asthenospheric Reservoir, EAR) is probably related to volatile-rich(carbonated?) fluid percolation. This is corroborated by LILE and HFSE patterns observed in several xenoliths.Therefore, our data on lavas and xenoliths suggest a lithospheric origin for this long-lived magmatism. Wepropose (1) that the role of the asthenosphere in the Languedoc volcanism was restricted to volatile-richfluid supplying and (2) that the fluid injection within the lithosphere may be related to the arrival of theCentral Atlantic Plume head beneath Western Europe about 70 Ma ago. In this model, the isotopic signatureof the oldest lavas (N 70 Ma) would be that of the mantle lithosphere, inherited from Hercynian processes.The signatures of the subsequent lavas would be driven by the metasomatic component stored within thelithosphere and preferentially mobilized during incipient melting. This metasomatised lower lithospherewas close to its solidus and small changes in P (or T) triggered incipient melting leading to basalticvolcanism. Successive local re-adjustments of the lithospheric blocks, which accompanied the Meso-Cenozoic evolution of the Thetys Ligurian margin towards the present Mediterranean margin, are theprobable cause of these changes and so the sporadic volcanic activity in Languedoc is unrelated to deepasthenospheric processes.

33 467143603.autria).

ll rights reserved.

© 2010 Elsevier B.V. All rights reserved.

1. Introduction

The origin of the Cenozoic alkali magmatism in Western Europe isstill widely debated and widely varying models include: Miocene “hotspot(s)” (e.g. Granet et al., 1995) or “wet spot(s)” (cf. Wilson, 2007);superposition of active (Oligocene) and passive (Miocene) rifting(Michon and Merle, 2001) and Eocene impingement of the “CentralAtlantic Plume” on the European continent related to the opening ofthe Atlantic Ocean (Piromallo et al., 2008). Inmost of thesemodels therespective contributions of lithosphere and asthenosphere in the

magma genesis are not clearly defined. Furthermore, the causes ofmelting, e.g. thermal anomaly (Sobolev et al., 1996) and/or injectionof volatile-rich melts (e.g. Downes, 2001), remain elusive. Further-more, the role of the Alpine orogenesis in determining the location ofthe volcanism and the magmatic development, are still questionable(Michon and Merle, 2001; Piromallo et al., 2008).

Despite the low volume of erupted lavas (∼2 km3 on the whole)and their relatively uniform basaltic compositions, the Languedocvolcanic district (Southern France, Fig. 1) is probably one of the bestplaces in Europe to shed new light on several of these issues, because—(1) the episodic volcanic activity spans from the Mid-Jurassic toQuaternary—this long-term activity is seen nowhere else in Europe;(2) its exceptional geological location (Fig. 1): half-way between thePyrenean and Alpine belts, on the Gulf of Lion rim and on the South

mailto:[email protected]://dx.doi.org/10.1016/j.lithos.2010.04.009http://www.sciencedirect.com/science/journal/00244937

Fig. 1. Geological setting of Languedoc volcanic districts with locations and ages of samples dated in this paper. ROU: sample name (see Table 2), (): new age data (Ma) (see Table 1)CD: Causses district; ELD: Escandorgue-Lodévois district; LHVD: LowHérault Valley District; FMC: FrenchMassif Central. All volcanic rocks dated since 1974 and all samples analyzedfor this paper are given in Fig. RM1.

203J.-M. Dautria et al. / Lithos 120 (2010) 202–222

side of the French Massif Central lithospheric swell, less than 200 kmsouthwards of its apex.

The detailed study of the Languedoc basalts should thus bring newresults that may answer many questions. Has the mantle sources of

European basalts evolved during the last 160 Ma? Has it been affectedby the compressive and distensive geodynamic events related to theMesozoic and Cenozoic evolution of the Ligurian Tethys margin? Hasit been modified by the opening of the Atlantic Ocean and the

204 J.-M. Dautria et al. / Lithos 120 (2010) 202–222

Pyrenean and Alpine orogenesis? Has it been affected by the mantleevents responsible for the Oligocene Mediterranean rifting and theMiocene uplifts?

In the present work, an extensive petrological and geochemicalsynthesis of the Languedoc basalts has been carried out, includingmajor- and trace-element data as well as Sr, Nd and Pb isotopes forrepresentative samples of all age groups. Further, new whole-rock K–Ar ages have been obtained for key samples in order to refine thetemporal history of this volcanism. Finally, peridotitic xenolithshosted in several of these lavas were analyzed, in order to provide apetrological and geochemical characterization of the Languedoc sub-continental upper lithosphere. These data are compared with thetheoretical mantle source of the basalts inferred from our geochemicalcalculations from lavas and with the well-known lithospheric mantlerocks from the Pyrenees and the Massif Central.

2. Geological setting

Languedoc is the region of France extending between the upliftedFrench Massif Central and the Mediterranean Sea coast (Fig. 1). Thecentral part of Languedoc is almost entirely covered with Mesozoicsediments (mostly carbonates) deposed on the Northern Tethyanpassive continental margin. Major NE–SW strike–slip faults ofHercynian age (re-activated during the Mesozoic and Cenozoic, inconnection with the Pyrenean orogeny) crosscut the region (Fig. 1).

The Languedoc volcanics only appear in the sedimentary basinsand are grouped inside a NS area ∼140 km long and ∼60 km wide(Fig. 1) which can be geographically considered as the southernextension of the French Massif Central Mio-Plio-Quaternary volca-nism. Inside this area, several volcanic alignments are distinguishable.They are not clearly superimposed with faults but they probablycorrespond to major lithospheric scale structural discontinuities ofpossible Hercynian age.

The area is usually subdivided into 3 districts.

i. The Causses District (CD), in the northern part of Languedoc(Fig. 1), comprises small basaltic outcrops distributed alongtwo axes, one WNW–ESE that approximately corresponds tothe current northern boundary of the Mesozoic basin ofWestern Causses; the second NNE–SSW, roughlycorresponding to the present Central Causses basin axis(Fig. 1). These outcrops indicated in Figs. 1 and RM1 correspondeither to small lava lakes filling ancient maars (e.g. AZ),phreatomagmatic breccia pipes injected with dykes (e.g. EG),isolated dykes or sills (e.g. NT) or more rarely flows (e.g. Vi).Most of them have Miocene ages ranging between 5.8 and7.5 Ma (Fig. RM1). Such ages are very common in WesternEurope and correspond to the paroxysmal volcanic activity inthe French Massif Central. For instance, the basaltic plateau ofAubrac that bounds the Languedoc province to the north(Fig. 1) and is considered bymany authors (Brousse and Bellon,1974; De Goër de Hervé et al., 1991) as belonging to the FrenchMassif Central magmatic province, displays such a Messinianage. However, previous work (Gillot, 1974) has shown thatolder volcanic edifices (i.e. Dogger, Palaeocene, and Serraval-lian) are also present in CD and correspond to the precursors ofthe Languedoc magmatic activity (Fig. RM1).

ii. The Escandorgue–Lodévois District (ELD), in the central partof Languedoc, corresponds to a N–S continuous and narrowvolcanic trail about 35 km long and ∼3 km wide (Fig. 1). Thevolcanic activity was here essentially phreatomagmatic tosurtseyan and, in minor part, strombolian. The age of theactivity in the northern and central parts of this district(Escandorgue plateau) is well documented (Figs. 1 and RM1):more than 20 ages are available with values between 2.5 and1.5 Ma, (Gillot, 1974; Gastaud et al., 1983; Brugal et al., 1990;

Ambert et al., 1990), whereas no age is available at present forits southern end and its eastern side. Lodévois corresponds to a∼12 km eastwards extension of the southern part of Escan-dorgue (Fig. 1) and its activity (between 1.5 and 1.2 Ma) isslightly younger (Gastaud et al., 1983).

iii. The Hérault Low Valley District (HLVD) south of Lodévoiscomprises about twenty small, well-preserved monogenicstrombolian cones and hydromagmatic tuff rings of Quaternaryage (Von Frechen and Lippolt, 1965; Gastaud et al., 1983). Theyare grouped on the western bank of the Hérault river andconstitute a N–S volcanic line about 35 km long; this linecontinues up to the Mediterranean coast and is offset 15 kmeastward from the Escandorgue alignment (Fig. 1). The south-ernmost volcano (CPA, 0.73 Ma, Fig. 1) is a striking surtseyan tuffring outcropping along the present seaside, but aeromagneticdata show that the HLVD line extends offshore under the sea.

Two volcanic complexes belonging geographically to SouthLanguedoc have ages that are anomalous with regard to the HLVDactivity:POU (Fig. RM1), 20 km East of the Hérault river, a small-sizedbreccia pipe is dated to 46 Ma by Liotard et al. (1991); MTF (Fig. RM1),40 km to the East, includes breccia pipes and dykes that intrudeEocene sediments and have ages between 23 and 25 Ma (Gastaudet al., 1983). These two complexes are particularly interesting becausethey are the only volcanoes of Lutetian and Chattian ages.

3. Sampling and analytical techniques

Fifty-two lava samples have been selected for this study on the basisof their ages and freshness. This selection is representativeof thedistinctage groups in the various districts.We also collected three samples fromthe Messinian basaltic plateau of Aubrac (Fig. 1) for comparison.

About one third of the exposed lavas, whatever their age, containperidotitic xenoliths. Ten xenoliths included in the studied lavas havebeen selected on textural and petrological criteria to represent eachtype observed in Languedoc.

The lava and peridotite samples were crushed and then powderedin an agate mill. Whole-rock major elements were analyzed by X-rayfluorescence (XRF, SARM, Nancy). Trace elements and REE abun-dances were analyzed using a VG Plasmaquad II ICP-MS at theUniversity of Montpellier II (Ionov et al., 1993).

Before undertaking the acid digestion for the Sr, Nd and Pb isotopicanalyses, all WRwere leached for 30 min with 6 N HCl at 80 °C. After theleaching steps, the residues were rinsed three times in purified milli-QH2O. The total blank contents for Pb, Sr and Nd were less than 35, 40and 10 pg, respectively, for a 100 mg sample. Pb and Nd isotopiccompositions were measured on the VG Plasma 54 and the Nu 500 MC-ICP-MS located at the Ecole Normale Supérieure in Lyon (France). ThePb isotopic compositions were measured with an external precisionbetter than 300 ppm for 206, 207, 208Pb/204Pb, using the Tl normalizationmethod described byWhite et al. (2000). Further details about analyticaltechniques, accuracy and reproducibility are available in Bosch et al.(2008). The NIST 981 standard was measured after every two samples(206Pb/204Pb=16.9380±0.0030 (2σ); 207Pb/204Pb=15.4919±0.0022(2σ); 208Pb/204Pb=36.6925±0.0055(2σ); n=20); the Nd isotopicmeasurementswere bracketed between the “Lyon in-house”Nd standardevery two samples with an average of 143Nd/144Nd=0.512132±17(2σ)(n=55). The Sr isotopic compositions were measured on a FinniganTriton TImass spectrometer at the Laboratoire deGéochimieGIS of Nîmes(France). The NBS 987 Sr standard yielded a mean value of 87Sr/86Sr=0.710254±09 (2σ) (n=16).

K–Ar analyses have been performed at LSCE, CEA-CNRS, Gif-sur-Yvette from phenocryst-free samples. Age calculations arebased on the decay and abundance constants of Steiger and Jäger(1977): lb ̣=4.962×10−10 a−1; le=0.581×10−10 a−1; 40 K/K=1.167 10−4 mol/mol.

Table 1K/Ar ages of selected samples. For sample locations, see Fig. 2. Age calculations are based on the following decay (Steiger and Jäger, 1977) and abundance constants: lb−=4.962×10−10 a−1; le=0.581×10−10 a−1; 40 K/K=1.167 10−4 mol

.mol−1.

Sample RQH1 AG AR CX BR VI

1 2 1 2 1 2 1 2 1 2 1 2 3

K (wt.%) 0.930 0.930 1.428 1.428 1.971 1.971 1.511 1.511 1.190 1.190 1.934 1.934 1.934±2s 0.009 0.009 0.014 0.014 0.02 0.02 0.015 0.015 0.012 0.012 0.020 0.020 0.020Weight molten (g) 0.97699 0.99675 1.07422 2.06563 1.01983 1.19612 0.79319 1.00750 0.99371 1.01408 1.12587 0.17263 0.1648040Ar* (%) 7.978 5.564 14.233 16.17 12.342 11.456 18.694 17.587 6.954 15.732 89.087 71.823 84.33740Ar* (10–12 mol/g) 9.009 9.129 1.714 1.698 4.164 4.154 3.854 3.806 3.118 3.135 574.1 555.0 567.240Ar mean weight 9.063 1.705 4.159 3.823 3.127 565.41s 0.083 0.007 0.007 0.018 0.012 9.7Age mean value (My) 0.562 0.688 1.22 1.46 1.51 161.17±2s 0.015 0.015 0.03 0.03 0.02 1.78

.

.

Sample CE2 819 742 ROL1 EG1 TS

1 2 1 2 1 2 1 2 1 2 1 2

K (wt.%) 0.843 0.843 1.330 1.330 1.978 1.978 1.610 1.610 1.777 1.777 0.917 0.917±2s 0.008 0.008 0.013 0.013 0.02 0.02 0.016 0.016 0.018 0.018 0.009 0.009Weight molten (g) 0.96414 0.97481 1.03807 1.04300 1.01402 1.11561 0.92130 0.47516 1.00979 0.51514 0.47808 0.3946840Ar* (%) 7.045 22.054 16.203 20.899 23.164 20.349 9.663 19.357 79.216 64.97 79.135 82.87940Ar* (10–12 mol/g) 2.267 2.182 4.241 4.151 6.607 6.884 6.301 6.245 43.384 43.430 95.28 94.7840Ar mean weight 2.245 4.196 6.746 6.274 43.407 95.031s 0.06 0.064 0.196 0.025 0.154 0.35Age mean value (My) 1.52 1.82 1.97 2.25 14.03 58.79±2s 0.03 0.03 0.03 0.05 0.3 0.83

Table 1 (continued)

205J.-M

.Dautria

etal./

Lithos120

(2010)202

–222


4. Results

4.1. Lavas

4.1.1. GeochronologyTwelve new K–Ar ages have been obtained on selected key-

samples (Table 1; locations in Fig. 1). Fig. RM1 shows the ages of alllavas dated since 1965.

The Vi volcanic outcrop, located in the northern part of CD (Fig. 1),is unique in Languedoc (and in Western Europe), both for its fieldstructure and its age. It is a lava delta, with massive basalt sheets andbrecciated pillow-lavas interbedded within a Dogger coastal carbon-ate series. Baubron et al. (1978a), using the K–Ar method, determinedan age of 155±6 Ma for this basalt. Our new K–Ar data (Table 1)indicate a slightly older age of 161.2±1.8 Ma, corroborating theexistence of a magmatic event in Languedoc at the Callovian–Oxfordian boundary.

TS (Fig. 1) is an intrusive complex located 20 km NW from Vi.A new K/Ar age of 58.8±0.8 Ma has been obtained for this samplein agreement with the age of 57 Ma previously measured by Baubronet al. (1978b). Palaeocene and Eocene lavas are very uncommonin western Europe and only six occurrences are known in France. Allare highly SiO2-undersaturated basalts (nephelinites), sometimescarbonated (melilitites) and they are distributed along the faultsbounding the future Oligocene rifts. According to Lenoir et al. (2000a,b), this magmatism is related to the initiation of the major mantlemelting event leading to the Miocene–Pliocene–Quaternary volca-nism of French Massif Central.

The volcanic complex of Eglazine (EG) (Fig. 1) is one of the two sitesforwhich a Serravalian agewasmeasured in Languedoc. It is a relativelywell-preservedbreccia pipe exposed at thebottomof the Tarn canyon.Adyke crosscutting these breccias yields an age of 14.0±0.3 Ma, almostsimilar to the age (13.0±0.4 Ma) previously estimated byGillot (1974).The EG volcanic complex is thus contemporaneous with the firstvolcanic activity phase in Cantal and Velay, the largest volcanic districtsof French Massif Central (Nehlig, 1999; Mergoil et al., 1993).

Five new dates have been obtained for ELD in order to fill theage gap in the southern and eastern parts of this district (Fig. RM1).A glassy cauliflower bomb (ROL1) from the southernmost tuff-ring(Fig. 1) yields an age of 2.25±0.05 Ma, suggesting that no agegradient exists along the Escandorgue NS axis. BR (lava lake), 819 (aflow) and 742 (dyke) are samples collected along the eastern side ofEscandorgue: their ages (1.51±0.02, 1.82±0.03 and 1.97±0.03 Marespectively) do not differ from those of the central part of the massif.Sample AR corresponds to an isolated phreato-volcanic complexbelonging to East Lodévois: its age (1.22±0.03 Ma) is one of the mostrecent for this district.

Finally, four flows from the Quaternary HLVD have been datedduring this study: two come from its northern part (CX, the longestflow of Languedoc, 11 km, and CE) and two from its southern end (AGand RQH) (Fig. 1). The CX and CE lavas yield similar ages (1.46±0.03 Ma and 1.52±0.03 Ma respectively) and are contemporaneouswith the Lodévois activity; AG and RQH are younger (0.69±0.015 Maand 0.56±0.015 Ma respectively). RQH represents the most recentvolcanic event known in Languedoc. As shown by Figs. 1 and RM1, theyoungest volcanoes (b0.75 Ma) are all situated close to the coast.

4.1.2. Major and trace elementsAll analyzed samples (Table 2) belong to the alkaline series and

most of them are alkali basalts or basanites according to theclassification of Cox et al. (1979) (Fig. 2). The degree of SiO2-undersaturation is globally high and variable inside each lava group:2b%(Ne+Lc)normb15 for the basalts, 15b%(Ne+Lc)normb26 for thebasanites. In the “basalt tetrahedron” of Yoder and Tilley (seeRingwood, 1975) all Languedoc lavas except samples RQH and AGwould plot either in the basanitic (%NenormN5) or in the nephelinitic

(%NenormN15) fields. Only three samples (ROM, CAB, AZ) plot in thehawaiite field (Fig. 2), but their high MgO contents (6.9, 8.7 and 8.3respectively, Table 2) contradict this classification (hawaiite MgOcontents are usually around 5%). Samples RQH and AG plot in the sub-alkaline domain (Fig. 2) and they can be considered as olivinetholeiites.

The lavas as a whole display [mg] numbers ranging from 0.56 to0.75, SiO2 and alkali contents between 41 and 52% and 3.7 and 7.6%respectively (Table 2). The lack of truly differentiated lavas constitutesone major difference with the French Massif Central volcanic district.This feature thus suggests both of the absence of magma chambersbeneath the Languedoc area and relatively fast ascent of the magmas.Both features are consistent with the very low volume of the emittedlavas and the common occurrence of abundant mantle xenoliths.Thus, the major-element variations of the Languedoc basalts can beexplained by different degrees of partial melting and to a lesser extentby the extraction or accumulation of olivine crystals during themagma ascent. This observation is corroborated by the trace-elementdata (see Section 5.3).

From a petrographic point of view, only two uncommon mag-matic rocks have been found: a basanite containing very large (up to3 cm) phlogopite megacrysts (recently dated at 1.88±0.02 Ma bythe Ar/Ar method, Monié, unpublished data) and Ti-rich magnetitemegacrysts (up to 5 cm in size) from Lodévois (LO, Table 2,Fig. RM1),and a camptonitic lamprophyre (POU3) occurring as clasts in the46 Ma-old breccia pipe POU (Liotard et al., 1991) (Table 2,Fig. RM1).

The Languedoc basalts display very variable K2O/Na2O ratios(between 0.16 and 1.40) and their trace-element contents are alsovery variable (e.g. 3.5bThb17; 21bLab108; 38bNbb144, Table 1).Only RQH shows the low trace-element contents typical of tholeiites(e.g. Th=3 ppm, La=21 ppm, Table 1). Sample AG, in spite of itstholeiite-like major-element chemistry, has trace-element contents(Th=7 ppm, La=39 ppm) similar to the less enriched basanites (i.e.TS, Th=6 ppm, La=40 ppm). As shown in Fig. 3, the K/Rb ratios ofsome Plio-Quaternary lavas belonging to Group 2 (Table 2) (VA, FO,RO, LC1, BA, ROL2, TAU, COL, SM, GR, MIC, BAS) are anomalously low(K/Rbb200) compared to the mean OIB value of Sun and McDonough(1989); K/Rb=400. This suggests a loss of K (and/or a Rb increase)that we tentatively relate to late-magmatic and/or weatheringprocesses. The possible impact of late alteration is corroborated bythe relatively high LOI contents (N2%) measured in most low-K/Rbsamples (Table 2). Moreover, this loss of K could explain why mostsamples with anomalously low K/Rb ratios plot in the alkali basaltfield instead of the basanitic field (Fig. 2), in spite of their high degreeof SiO2-undersaturation.

The trace-element patterns of the studied lavas are remarkablyparallel (Fig. 4) and typical of alkali basalts, implying enriched OIB-type mantle sources. This indicates that weathering has not sig-nificantly affected the amounts of the incompatible elements, exceptfor the most mobile ones such as Rb, Ba and K. As expected, thetransitional basalts (RQH and AG) display less enriched patterns [(La/Yb)N≤15], while the Ne normative-rich lavas show themost enrichedpatterns (e.g. FO, LO, 742 and VA with (La/Yb)N ratios up to 33,Table 2). However, the lavas with the highest trace-element contentsare not those with the highest K and Lc normative contents (N5%)(Table 2). The K enrichment of these lavas would result consequentlyfrom melting of a K-rich and Th-, U-, Nb-, and LREE-poor phase likephlogopite. Slight negative anomalies in Th and U (e.g. (U/Nb)N:POU=0.738; NDG=0.748; MTF=0.843; PP1=0.567) and Zr–Hf(e.g. (Zr/Sm)N: MRS=1.04; TS=0.925; NT=0.948) are observed inseveral samples (Fig. 4). Small positive spikes in Pb are shown by thetransitional and low alkali basalts [(Pb/Ce)N: NDG=1.38 ; Vi=1.86].

4.1.3. Isotope dataThe isotopic data are reported in Table 3. The initial 87Sr/86Sr ratios

display rather large variation, ranging between 0.70307 and 0.70436

Table 2Major- and trace-element compositions of selected Languedoc lavas. The analytical methods are given in the text. Ne, Lc: nepheline and leucite normative content respectively, for the samples with L.O.I.b4%; [mg]: Mg/(Mg+Fe2+), withFe3+=0.15Fe2+. The abbreviations correspond to selected samples located in Fig. 2. Legends are: (i): isotopic composition corrected for in-situ decay; *: data from this study; []: published ages (for references see text); ** interpolatedage according to field observations. Note that samples NDG and AG belong to the same flow.

Group 1b0.8 My 1.2bGroup 2b2.3 My

RQH bNDG AGN MRS CPA E12 VA PA RO LC1 LC2 CE1 CE2 MCL FO CX AR SM

Long. E 3°22 11 3°27 44 3°28 17 3°24 29 3°31 07 3°21 52 3°21 40 3°21 32 3°17 47 3°24 47 3°24 34 3°23 48 3°23 39 3°24 13 3°23 23 3°21 58 3°29 33 3°23 49

Lat. N 43°1801 43°17 31 43°17 52 43°22 52 43°16 28 43°25 33 43°25 36 43°31 22 43°30 30 43°35 21 43°34 47 43°32 53 43°32 49 43°33 02 43°33 08 43°31 53 43°43 39 43°50 28

age (My) 0.56 0.69 0.69 0.68 0.73 1.4 1.4 2* 2* 1,5* 1,5* 1,5* 1.52 1,5* 1.46 1.22 2*

SiO2 51.06 51.06 49.68 45.78 47.33 44.34 42.74 43.18 43.87 44.04 41.58 44.59 45.67 43.91 40.82 44.66 42.43 43.90Al2O3 12.37 13.25 13.99 12.97 13.96 13.75 13.02 11.89 13.44 12.77 11.65 12.05 12.26 11.71 10.88 12.66 14.07 12.66Fe2O3* 11.23 10.67 10.92 12.81 11.89 12.36 12.03 13.52 11.85 12.18 13.32 11.54 12.01 11.90 13.10 12.96 12.79 12.30MnO 0.14 0.14 0.15 0.19 0.17 0.19 0.17 0.18 0.16 0.17 0.19 0.16 0.16 0.16 0.23 0.20 0.15 0.16MgO 9.00 7.92 8.49 10.67 8.34 6.93 8.04 9.84 8.69 8.67 9.71 11.41 10.30 10.39 10.50 9.26 9.07 10.22CaO 8.84 8.36 8.90 10.09 9.22 10.42 11.15 10.09 10.06 9.87 11.86 10.08 10.08 10.94 12.84 9.89 10.57 10.47Na2O 3.28 3.53 3.53 3.67 3.92 4.82 4.84 4.06 4.35 4.39 4.08 3.42 3.61 3.37 4.66 3.68 3.74 3.67K2O 0.90 1.56 1.54 1.53 1.66 0.83 0.85 0.75 0.94 0.97 0.98 1.55 0.90 1.47 0.87 1.73 2.29 0.78TiO2 2.06 2.05 2.10 2.25 2.33 2.90 2.77 2.60 2.97 2.99 2.70 2.34 2.32 2.54 2.63 2.46 3.53 2.65P2O5 0.42 0.59 0.61 0.76 0.76 0.91 1.12 1.21 0.83 0.91 1.39 0.93 0.92 0.83 1.25 1.01 0.75 0.84L.O.I. 0.02 0.33 −0.30 −0.57 −0.26 2.42 3.32 2.23 3.14 2.40 2.34 1.82 0.95 3.22 2.45 1.37 0.67 2.00Total 99.32 99.46 99.62 100.15 99.31 99.87 100.05 99.55 100.30 99.36 99.80 99.89 99.18 100.44 100.23 99.87 100.06 99.65[mg] 0.65 0.63 0.64 0.65 0.61 0.56 0.60 0.62 0.63 0.62 0.62 0.69 0.66 0.67 0.65 0.62 0.62 0.65Lc 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 4.2 0.0 0.0 0.0Ne 0.0 0.0 0.7 10.3 6.9 13.5 17.4 10.8 12.3 11.8 16.6 9.8 6.2 11.0 22.1 10.3 17.4 9.3Rb 19.36 19.46 33.11 33.33 39.83 38.69 46.85 64.92 20.76 43.68 53.58 27.44 43.72 22.69 14.01 57 46.71 57.5 74.23Sr 480.0 494.9 646.5 688.6 781.9 816.4 1270.30 1244.20 1078.90 1054.60 964.25 1193.50 903.28 967.40 969.3 1078 1020 938 963.04Y 21.21 21.68 23.35 23.62 26.06 25.31 26.03 30.95Zr 146.7 151.8 202.8 205.7 214.2 234.9 375.86 394.24 349.10 366.83 342.75 381.87 299.44 307.88 268.5 343 282.8 258 313.57Nb 38.13 57.10 58.17 72.46 73.41 143.73 151.52 110.86 112.25 109.50 120.19 95.95 98.08 79.98 113 82.36 94.8 94.41Cs 0.099 0.141 0.179 0.259 0.650 0.558 1.00 1.17 0.93 1.14 0.89 0.92 0.72 0.76 1.320 0.93 0.843 0.65 0.97Ba 293.5 304.2 514.3 510.0 606.1 598.7 988.31 1020.10 724.79 915.82 850.79 886.62 710.54 743.98 744.8 689 703.2 816 740.09La 21.05 21.57 37.30 39.29 51.36 48.61 91.97 108.01 78.36 73.38 64.01 87.62 67.32 72.29 53.07 79.87 68.63 47.1 60.12Ce 43.11 43.81 70.09 74.78 98.48 91.92 168.85 185.75 146.25 134.71 123.85 161.95 122.46 130.23 103.7 154 130.0 92.7 111.29Pr 4.985 5.075 7.693 8.189 10.73 10.05 18.21 19.45 15.98 14.51 13.80 17.95 13.42 14.23 11.55 17.53 14.03 10.5 12.38Nd 21.18 21.57 30.97 32.51 41.93 39.70 70.40 73.03 62.98 57.98 54.51 71.29 52.33 54.73 47.35 69.09 56.55 43.9 50.54Sm 5.088 5.208 6.455 6.536 8.120 7.828 12.06 12.10 11.01 9.94 9.68 12.44 9.03 9.65 8.646 11.87 10.05 8.36 9.04Eu 1.829 1.877 2.109 2.134 2.620 2.578 3.71 3.86 3.68 3.36 3.17 4.09 3.04 3.19 2.725 3.75 3.151 2.73 3.08Gd 5.535 5.584 6.236 6.415 7.482 7.405 10.71 11.00 10.58 9.60 8.74 11.24 9.18 9.32 7.601 10.28 8.957 8.06 8.77Tb 0.789 0.789 0.861 0.889 1.020 0.999 1.33 1.37 1.32 1.23 1.12 1.40 1.17 1.18 1.016 1.44 1.186 1.02 1.16Dy 4.793 4.849 5.290 5.227 6.082 5.837 7.23 7.73 7.35 6.96 6.28 7.73 6.66 6.79 5.492 7.09 6.472 5.38 6.61Ho 0.849 0.842 0.935 0.944 1.059 1.015 1.23 1.31 1.21 1.20 1.05 1.27 1.16 1.15 0.961 1.21 1.114 0.90 1.12Er 2.065 2.048 2.316 2.353 2.573 2.482 3.01 3.11 2.81 2.93 2.56 2.86 2.78 2.75 2.363 2.85 2.731 2.17 2.66Tm 0.270 0.267 0.313 0.312 0.344 0.325 0.39 0.39 0.34 0.37 0.32 0.36 0.36 0.36 0.303 0.36 0.369 0.265 0.34Yb 1.562 1.545 1.853 1.883 1.985 1.905 2.28 2.30 2.01 2.20 1.98 2.04 2.11 2.14 1.655 2.11 2.014 1.54 1.99Lu 0.231 0.227 0.282 0.282 0.303 0.285 0.34 0.32 0.28 0.33 0.27 0.29 0.31 0.31 0.253 0.31 0.299 0.216 0.29Hf 3.593 3.563 4.547 4.604 4.705 5.112 8.41 7.78 7.01 7.80 7.43 7.80 6.05 6.21 5.936 7.42 5.910 5.70 6.66Ta 1.838 1.825 2.731 2.762 3.577 3.608 7.65 7.64 6.03 6.89 6.26 7.12 5.02 5.29 4.537 6.25 4.259 4.89 5.34Pb 2.280 2.056 3.859 4.078 3.239 3.513 5.55 5.37 3.79 4.27 3.59 4.25 4.43 4.33 3.223 4.08 4.150 3.20 2.77Th 3.537 3.442 6.674 6.957 7.705 7.419 15.67 17.19 9.87 10.18 8.52 11.29 9.82 9.80 7.608 10.21 9.460 6.27 8.73U 0.683 0.729 1.256 1.294 1.721 1.482 3.68 3.90 2.37 2.52 2.11 2.73 2.27 2.33 2.089 2.55 2.493 1.80 2.29

207J.-M

.Dautria

etal./

Lithos120

(2010)202

–222

.

.

1.2bGroup 2b2.3 My 1.2bGroup 2b2.3 My

LO 819 BR 742 BA GR 809 MA SVT ROL1 ROL2 TAU LR COL BGE SAL FES CAB

Long. E 3°19 51 3°17 04 3°21 12 3°16 26 3°21 29 3°17 32 3°16 35 3°17 22 3°20 31 3°16 10 3°16 10 3°15 27 3°15 10 3°14 09 3°23 59 3°19 49 3°26 02 3°14 03

Lat. N 43°44 22 43°48 52 43°44 19 43°19 29 43°44 22 43°45 26 43°49 14 43°44 48 43°48 30 43°34 14 43°34 14 43°36 31 43°38 39 43°45 53 43°39 30 43°39 05 43°39 29 43°45 45

age (My) 2 1.82 1.51 1.97 2.25 2 2* 2

SiO2 39.18 42.71 44.34 41.72 43.39 42.62 46.36 41.60 41.31 43.97 44.94 44.88 45.35 42.97 47.26 44.01 44.16 46.90Al2O3 12.03 11.51 12.81 11.90 12.48 11.37 13.12 13.49 12.62 14.52 14.24 14.73 14.82 13.50 13.87 13.09 13.37 14.54Fe2O3* 13.35 13.78 12.42 12.08 13.07 13.20 11.59 13.90 13.94 12.28 11.99 11.83 10.43 11.97 12.00 12.43 12.56 10.34MnO 0.17 0.19 0.21 0.19 0.20 0.20 0.17 0.22 0.18 0.19 0.19 0.19 0.18 0.20 0.18 0.19 0.19 0.18MgO 8.89 12.74 8.60 11.69 9.48 13.38 11.17 8.83 9.73 6.40 7.51 7.13 6.21 9.09 7.98 10.48 9.52 8.66CaO 11.30 9.31 10.29 11.90 10.31 9.88 8.70 9.54 10.75 9.70 9.77 9.86 8.92 11.01 9.25 8.77 8.96 8.50Na2O 2.50 3.46 3.29 2.34 4.05 3.06 3.25 3.56 3.74 3.05 4.94 4.25 4.30 4.90 3.92 3.69 3.20 4.73K2O 2.40 1.74 1.46 2.42 1.08 0.62 1.96 2.33 2.07 2.45 1.25 1.12 3.11 1.03 1.94 2.50 2.44 2.71TiO2 3.82 3.12 3.02 3.08 2.74 2.57 2.76 3.71 3.92 3.17 3.03 3.13 2.76 3.28 2.48 2.71 2.81 2.45P2O5 0.72 1.01 1.07 0.66 1.00 0.75 0.50 0.95 0.78 0.97 0.85 0.85 0.87 0.93 0.99 0.90 0.95 0.84L.O.I. 4.64 0.57 2.93 1.89 2.23 3.10 0.89 2.67 0.86 3.71 1.78 2.59 2.49 1.70 0.84 1.62 2.03 0.46Total 99.00 100.14 100.44 99.87 100.03 100.75 100.47 100.80 99.90 100.42 100.48 100.55 99.44 100.57 100.73 100.40 100.18 100.31[mg] 0.60 0.68 0.61 0.69 0.62 0.70 0.69 0.59 0.61 0.54 0.59 0.58 0.58 0.63 0.60 0.66 0.63 0.66Lc 9.2 0.0 0.0 10.1 0.0 0.0 0.0 0.0 3.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0Ne 12.3 12.8 6.7 11.1 12.5 8.1 6.6 15.1 17.5 7.5 14.6 9.6 15.0 19.0 7.2 13.5 9.4 15.2Rb 54 42 57 61 62 83 49 55 74.5 69.00 79.57 141.01 83.79 94.22 50.21 56.17 65.64 80.82Sr 902 817 1142 1011 1011 929 687 1033 996 2391 1043 1141 1352 1243 958.9 892.5 1017 1058.50Y 33.54 28.76 31.79 30.10 32.59 30.88 26.11 30.11 28.80Zr 256 345 377 354 300 284 301 317 413.6 401.4 385.6 422.0 408.3 287.4 269.0 309.7 389.21Nb 84 106 110 110 95 82 85 113 123.6 105.1 109.8 124.3 125.8 90.26 72.76 99.76 115.45Cs 0.54 0.98 0.84 1.03 0.88 0.94 1.13 0.76 0.991 1.021 1.627 1.183 1.049 1.225 0.918 1.004 1.31Ba 767 568 863 808 843 684 599 723 936 994.5 985.9 961.1 1011 1081 727.7 759.9 884.5 907.34La 41.4 52.48 62 62.56 71.31 54.16 41.54 53.73 60.3 75.70 65.73 72.18 77.82 81.90 66.75 54.90 64.10 76.03Ce 83.9 102.5 120.6 122.8 138.5 105.8 78.07 104.8 117 144.3 124.8 138.4 144.0 157.5 123.3 105.8 125.8 139.77Pr 11.94 13.75 14.15 15.68 12.24 8.89 12.08 13.3 15.60 13.60 14.85 15.14 16.98 13.41 11.87 13.67 14.37Nd 41.1 48.62 54.94 56.64 62.49 49.19 35.63 48.41 54.4 62.19 54.12 59.53 58.93 66.82 53.50 47.91 55.70 54.99Sm 8.3 8.97 10.05 10.38 11.01 8.93 6.99 8.76 9.95 11.01 9.631 10.56 10.00 11.42 9.551 8.712 9.846 9.21Eu 2.27 2.89 3.22 3.3 3.47 2.86 2.28 2.8 3.22 3.346 2.902 3.192 3.121 3.512 3.027 2.695 3.166 2.90Gd 8.26 9.1 9.35 9.66 7.98 6.67 8 9.32 9.264 8.293 9.047 8.437 9.711 8.556 7.471 8.535 7.85Tb 0.93 1.15 1.29 1.34 1.36 1.14 0.97 1.14 1.18 1.229 1.078 1.200 1.120 1.255 1.156 0.993 1.133 1.05Dy 5.8 6.59 6.83 6.96 5.82 5.12 5.76 6.30 6.722 5.910 6.593 6.168 6.836 6.389 5.541 6.346 5.83Ho 0.99 1.15 1.18 1.2 1 0.9 1 1.07 1.164 1.024 1.159 1.064 1.182 1.116 0.971 1.088 1.02Er 2.38 2.77 2.95 2.9 2.38 2.25 2.4 2.64 2.980 2.570 2.926 2.727 3.002 2.850 2.428 2.794 2.64Tm 0.31 0.36 0.38 0.38 0.32 0.31 0.32 0.338 0.399 0.347 0.405 0.360 0.388 0.371 0.324 0.356 0.35Yb 1.47 1.8 2.2 2.25 2.25 1.87 1.83 1.91 1.91 2.308 1.928 2.293 2.156 2.328 2.179 1.836 2.094 2.12Lu 0.25 0.27 0.32 0.34 0.34 0.27 0.28 0.28 0.285 0.346 0.312 0.350 0.335 0.346 0.319 0.277 0.311 0.33Hf 5.2 5.78 7.59 8.08 8.18 6.78 6.2 6.47 7.16 8.511 8.480 8.872 8.589 8.739 6.346 6.108 7.011 8.16Ta 4.68 6 6.51 6.75 5.77 4.74 5.08 6.46 6.881 6.086 7.055 7.122 7.350 4.756 3.924 5.752 6.84Pb 3.6 3.32 3.92 4.43 5.47 3.5 3.72 4.28 3.51 4.547 5.251 4.981 5.873 4.637 5.385 3.586 3.969 5.34Th 5.2 6.7 8.34 8.69 9.79 7.48 7.44 7.62 8.21 10.54 9.847 10.73 12.51 12.79 10.07 7.708 8.533 12.56U 2.27 1.85 2.3 2.43 2.51 1.85 2.03 2.08 2.20 2.502 2.481 2.399 3.520 3.293 2.732 2.057 2.215 3.28

Table 2 (continued)

208J.-M

.Dautria

etal./

Lithos120

(2010)202

–222

.

.

1.2bGroup 2b2.3 My 5bGroup 3b7.5 My 13bGroup 4b161 My AUBRAC Group

BAG BAS GUI MIC ROM AZ SAU PP1 EG MTF POU1 POU3 TS bNT(1) NT(2)N VI 02 AU2 AU3 AU4

Long. E 3°13 59 3°17 03 3°15 15 3°12 44 3°14 10 2°59 46 3°21 47 3°23 15 3°13 10 3°51 38 3°38 56 3°38 56 2°54 43 2°52 26 3°13 44 3°12 11 3°01 16 2°55 13

Lat. N 43°44 11 43°39 06 43°44 35 43°43 14 43°49 03 44°08 48 43°58 32 44°03 24 44°12 22 43°40 43°31 34 43°31 34 44°24 25 44°28 29 44°16 16 44°44 05 44°35 53 44°31 01

age (My) 2 2 1.26 2 1.64 5.75 7.1 6.4 14 23.6 46 46 58.8 66.9 161.2 6.5 6.5 6.5

SiO2 44.73 44.79 43.75 42.31 46.24 46.98 44.20 43.98 45.52 45.26 40.88 41.90 45.08 39.53 47.69 43.68 47.31 44.88Al2O3 14.7 13.06 13.9 13.06 15.13 14.36 13.46 13.88 13.40 14.64 12.90 12.87 10.63 10.37 14.77 13.76 16.49 14.73Fe2O3* 12.85 11.63 12.84 12.96 11.94 11.35 12.62 12.55 11.09 12.30 12.51 12.87 11.11 10.55 10.80 13.03 12.48 11.83MnO 0.2 0.19 0.2 0.21 0.19 0.18 0.18 0.16 0.18 0.18 0.15 0.17 0.19 0.15 0.14 0.19 0.18 0.19MgO 7.51 9.83 8.15 9.18 6.89 8.29 10.52 10.02 11.81 8.36 9.21 9.38 14.34 11.79 7.61 9.27 5.73 7.13CaO 9.39 9.97 9.78 11.08 8.7 7.36 8.97 9.00 9.14 8.26 11.20 10.95 10.57 13.16 7.41 10.18 8.36 9.86Na2O 4.42 4.08 4.16 4.18 4.62 3.86 3.14 4.34 2.57 3.70 1.61 2.56 2.72 2.68 3.66 3.72 3.87 4.25K2O 2.69 1.28 2.64 0.66 2.85 2.67 1.91 1.73 1.96 2.25 2.26 2.25 0.91 1.75 2.06 2.07 1.76 1.12TiO2 3.21 2.83 3.29 3.19 3.16 2.22 2.77 2.56 1.89 2.59 3.54 3.22 2.11 2.58 2.80 3.19 3.24 3.13P2O5 0.9 0.83 0.94 1.08 0.84 1.00 0.77 0.69 0.56 0.63 0.65 0.59 0.53 0.64 0.68 0.87 0.72 0.85L.O.I. 0.07 1.59 0.31 2.32 −0.08 2.50 2.37 1.07 1.37 1.47 4.10 2.32 2.23 6.60 2.31 0.08 −0.08 2.59Total 100.67 100.08 99.96 100.23 100.48 100.78 100.91 99.98 99.49 99.64 99.01 99.03 100.42 99.80 99.93 100.03 100.05 100.55[mg] 0.57 0.66 0.59 0.62 0.57 0.62 0.66 0.65 0.71 0.61 0.63 0.62 0.75 0.72 0.62 0.58Lc 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 8.8 0.0 0.0 0.0 0.0Ne 16.8 11.4 17.2 13.8 15.1 7.0 8.1 15.3 4.9 9.3 6.9 11.8 6.0 13.3 2.2 14.0 3.8 9.6Rb 85.49 56.53 68.06 146.8 84.55 72.79 56.8 32.41 48.39 62.84 40.75 25.01 26.2 44.6 38.88 56.12 65.85 36.85 34.32Sr 1047 1097 1007 1045 1062 1197 2498 889.55 753.6 791.5 786.0 935.3 744 1124 1006 703 871.6 803.6 674.5Y 32.16 28.86 32.20 35.17 31.32 26.91 24.3 23.22 25.63 26.3 24.44 26.49 23.28 30.44 28.97 25.78Zr 411.7 396.0 413.6 417.9 458.3 407.1 333.7 313.46 216.8 315.2 248.4 270.9 177 227 241 253.8 309.4 300.8 249.7Nb 107.7 110.0 112.8 118.3 118.6 105.5 115.7 101.75 70.64 67.04 59.35 64.23 60.6 90.6 73.76 63.16 94.77 74.39 65.68Cs 1.05 1.13 0.99 3.88 1.03 1.41 1.03 0.76 1.26 0.762 2.633 1.305 0.81 3.37 3.33 1.746 1.147 0.366 0.484Ba 901.2 809.3 817.1 889.1 881.6 2060 720.8 599.00 603.9 538.2 738.5 599.3 503 980 943 270.7 660.3 451.3 432.5La 64.85 67.61 65.61 82.79 66.91 68.80 52.56 48.78 43.26 42.51 39.96 43.17 39.7 60.4 58.95 35.44 60.75 45.07 51.20Ce 127.3 128.0 129.6 158.6 130.3 133.3 101.7 94.21 84.27 88.45 85.99 92.89 75.3 115 112.8 75.76 122.3 96.64 101.3Pr 13.90 13.90 14.09 16.52 13.80 14.76 11.29 10.47 9.34 10.27 10.25 10.86 8.6 12.9 12.59 9.19 13.45 11.20 11.01Nd 56.66 54.44 56.82 64.55 55.19 56.28 44.43 41.32 38.23 42.25 41.74 46.10 36.2 52.5 52.49 40.44 54.53 47.48 44.42Sm 10.30 9.69 10.31 11.39 10.06 10.42 8.55 7.79 7.45 8.344 8.451 9.380 7.60 9.49 9.25 8.82 9.866 9.130 7.881Eu 3.19 2.96 3.20 3.46 3.11 3.24 2.74 2.57 2.44 2.735 2.770 2.978 2.49 2.95 2.87 2.89 2.900 2.742 2.360Gd 8.92 8.07 8.98 9.41 8.37 8.43 7.36 6.98 7.06 7.624 7.785 8.428 7.47 8.79 7.51 8.14 8.522 7.921 6.971Tb 1.22 1.11 1.21 1.31 1.17 1.14 0.98 0.92 0.97 1.036 1.018 1.124 0.99 1.05 1.00 1.06 1.127 1.046 0.927Dy 6.83 6.25 6.75 7.35 6.47 6.07 5.44 5.11 5.73 6.053 5.785 6.454 5.38 5.49 4.87 5.87 6.222 5.888 5.252Ho 1.16 1.06 1.17 1.26 1.13 1.03 0.92 0.87 1.05 1.076 0.968 1.079 0.93 0.93 0.89 0.96 1.092 1.039 0.928Er 2.97 2.66 2.93 3.13 2.84 2.50 2.26 2.07 2.81 2.593 2.282 2.471 2.18 2.19 2.16 2.29 2.712 2.676 2.400Tm 0.39 0.35 0.39 0.41 0.38 0.32 0.28 0.26 0.37 0.356 0.295 0.315 0.287 0.266 0.27 0.28 0.366 0.356 0.321Yb 2.37 2.11 2.31 2.46 2.29 1.90 1.66 1.48 2.27 2.043 1.649 1.809 1.67 1.53 1.53 1.6 2.078 2.057 1.896Lu 0.37 0.32 0.36 0.37 0.35 0.28 0.24 0.22 0.34 0.307 0.228 0.262 0.238 0.231 0.23 0.226 0.315 0.323 0.296Hf 7.29 6.89 7.13 7.19 7.95 9.10 7.40 6.78 5.18 6.741 5.618 6.206 4.03 5.41 5.40 5.95 6.927 6.568 5.453Ta 6.19 6.28 6.69 5.91 7.15 6.35 6.92 6.17 3.99 3.698 3.170 3.420 3.01 4.91 4.76 3.42 5.493 4.159 3.682Pb 3.93 4.89 3.98 4.19 4.41 5.34 3.54 3.44 3.51 3.104 4.105 3.076 3.23 4.93 4.50 5.65 4.056 3.232 2.924Th 9.14 10.21 9.26 11.51 9.88 9.87 7.68 6.70 6.37 5.970 4.544 4.985 6.04 8.36 8.06 6.08 8.199 4.891 6.513U 2.38 2.31 2.42 2.44 2.56 2.96 2.07 1.70 1.64 1.661 1.29 1.367 1.43 1.99 1.86 1.65 2.109 1.303 1.696

Table 2 (continued)

209J.-M

.Dautria

etal./

Lithos120

(2010)202

–222

Fig. 4. Extended trace-element patterns of selected Languedoc lavas. The normalizingvalues of Primitive Mantle (N) and incompatibility sequence are from Sun andMcDonough (1989). For samples location, see Fig. 1,RM1 and Table 2.

Fig. 2. %(K2O+Na2O) vs %SiO2 diagram for the Languedoc lavas. Boundaries are fromCox et al. (1979). Four groups of samples have been defined according to their age (seeTable 2): Group 1 b0.8 Ma; Group 2 1.2–2.3 Ma; Group 3 5–7.5 Ma; Group 4 13–161 Ma; AUB: Aubrac. (a): line separating alkaline and subalkaline domains is fromIrvine and Baragar (1971). Anhydrous recalculated values.


(except for sample SAU, which has 87Sr/86Sr=0.7072). The initial143Nd/144Nd ratios are more homogeneous (0.51269–0.51298). Theinitial Pb-isotope ratios also show significant variation (206Pb/204Pb18.735–19.658; 208Pb/204Pb 38.422–39.434; and 207Pb/204Pb 15.594–15.680). In Figs. 5a,b and 6a,b, the Languedoc basalts define relativelylarge fields included within the fields of Western and Central Europelavas (Piromallo et al., 2008) and superimposed on the French MassifCentral and Catalunya fields (Downes, 1984; Chauvel and Bor-Ming,1984;Wilson and Downes, 1991; Briot et al., 1991; Lenoir et al., 2000a,b; Dautria et al., 2004; Cebria et al., 2000).

In the 143Nd/144Nd vs 87Sr/86Sr diagram (Fig. 5a,b),most Languedoclavas plot along the Mantle Array. From this diagram, two importantobservations can be made: (1) some sub-contemporaneous sam-ples from adjacent locations show very different isotopic signatures(e.g. LR and SAL that are 3 km from each other, Table 3, Fig. RM1);(2) conversely, some samples with significantly different locationsand ages have similar isotopic compositions [e.g. AG (0.69 Ma) and Vi(161 Ma), separated by 150 km, (Table 3, Fig. 1). Thus, at first sight,there is no regional or age control on the isotopic signatures of theLanguedoc lavas. Nevertheless, a large number of Group 2 (Table 2)samples are slightly more enriched in radiogenic Nd. Most of thesesamples plot within or close to the Low Velocity Component (LVC)field as defined by Hoernle et al. (1995). According to Wilson and

Fig. 3. K/Rb vs. (Na2O+K2O) wt.% diagram for the Languedoc lavas. Dashed square:lavas with anomalously low K/Rb ratios. For legend, see Fig. 2.

Downes (1991), Granet et al. (1995) and Downes (2001), such asignature corresponds to the asthenospheric component of theEuropean plume.

In the 206Pb/204Pb vs 207Pb/204Pb diagram (Fig. 6a), most samplesplot off the Northern Hemisphere Reference Line (NHRL). The Pre-Miocene lavas (from 161 to 25 Ma) define a horizontal trendcharacterized by a decrease in 206Pb/204Pb with increasing age,while the 207Pb/204Pb ratio remains constant. In the 206Pb/204Pb vs.208Pb/204Pb diagram (Fig. 6b), the lavas plot within a narrow areaalong the NHRL, but no samples plot in the LVC field. As previouslyshown for Nd and Sr isotopes, significantly different Pb isotopicsignatures can be found in samples close in age (e.g. 819 and 742,Table 3) while some samples of very different ages and/or fromdistant locations have similar signatures [e.g. AG (0.69 Ma) and Vi(161 Ma), Table 3]. However, as shown by the 206Pb/204Pb vs agediagram (Fig. 7), lavas with ages between 66.9 and 23.6 Ma display Pbisotopic heterogeneities of same order of magnitude as those of theMiocene–Pliocene–Quaternary lavas. Thus, these observations showthat the Languedoc lava sources are isotopically variable and suggestthat (1) the length-scale of variation is very small and (2) the isotopicheterogeneities were probably entirely acquired before the Miocene.As suggested by Fig. 7, the first influence of the EAR componentappeared around 70–60 Ma, which is in agreement with observationsmade in the post-collisional Cenozoic volcanic districts of the Adriaticdomain (Bianchini et al., 2008). In this hypothesis, only the oldestmagma (Vi, 161.2 Ma) sources could be considered as being free ofEAR influence.

4.2. The peridotite xenoliths

Studies of the petrology and geochemistry of the peridotitexenoliths from the Languedoc basalts have shown that this part ofthe French sub-continental mantle lithosphere is rather heteroge-neous (e.g. Albert et al., 1967; Brousse and Ildefonse, 1970; Coisy,1977; Berger, 1981; Fabries et al., 1987; Cabanes and Mercier, 1988;Jakni et al., 1996; Dautria et al., 2006). Harzburgites and lherzolites arefound in almost similar proportions and rare wehrlites have been

Table 3Sr, Nd, Pb isotopic compositions of selected Languedoc lavas. Absolute ages are from this study and from literature data (Gillot, 1974; Baubron et al., 1978a). *: data from this study; []: published ages (for references see text); ** interpolatedage according to field observations; (i), in situ decay corrected. For lead-isotope ratios isotopic composition uncertainties are better than 300 ppm.

Sample RQH1 AG MRS CPA MCL FO CX LO 819 BR 742

Age (My) 0.56* 0.69* [0.68] [0.73] 1.5** 1.5** 1.46* 2* 1.82* 1.51* 1.97*

87Sr/86Sr 0.703452±08 0.703917±08 0.703433±07 0.703574±11 0.703581±05 0.70335±50 0.703253±02 0.70399857±06 0.70325±60 0.703370±40 0.703080±3087Sr/86Sr(i) 0.70345 0.70391 0.703432 0.70357 0.70358 0.70335 0.70325 0.70399 0.70325 0.70337 0.70307143Nd/144Nd 0.512803±10 0.512789±08 0.512727±09 0.512847±09 0.512920±06 0.512910±30 0.512894±04 0.512912±05 0.512860±30 0.513000±20 0.512900±20143Nd/144Nd(i) 0.512802 0.512788 0.512726 0.512846 0.512919 0.512909 0.512893 0.512910 0.512858 0.512999 0.512898208Pb/204Pb 38.8690±09 38.9800±03 38.8810±08 38.8790±06 38.9400±15 38.9760±12 38.9480±18 38.9780±37 38.6290±15 38.9880±15 39.1680±57208Pb/204Pb(i) 38.866 38.976 38.876 38.874 38.929 38.964 38.937 38.969 38.617 38.977 39.155207Pb/204Pb 15.6340±08 15.6390±03 15.6250±07 15.6250±05 15.6221±06 15.6310±40 15.6227±05 15.6157±08 15.5670±60 15.6361±60 15.6810±22207Pb/204Pb(i) 15.634 15.639 15.625 15.625 15.622 15.631 15.622 15.615 15.566 15.636 15.68206Pb/204Pb 18.9500±07 18.9000±03 19.0490±06 19.0250±04 19.2682±06 19.3150±50 19.2718±13 19.3321±07 18.7450±70 19.1230±70 19.3790±270206Pb/204Pb(i) 18.948 18.898 19.045 19.022 19.259 19.306 19.263 19.32 18.735 19.114 19.368

.

.

Sample BA GR 809 MA ROL1 ROL2 TAU LR COL BGE SAL

Age (My) 1.5** 1.5** 1.5** 2** 2.25* 2.25* 2.25** [1.3] [2] 1.5** 1.5**

87Sr/86Sr 0.703360±30 0.703170±40 0.703210±40 0.704360±40 0.703614±03 0.703331±03 0.703274±03 0.703815±03 0.703224±03 0.703551±05 0.703363±0287Sr/86Sr(i) 0.70336 0.70316 0.70321 0.70436 0.70361 0.70332 0.70326 0.70381 0.70322 0.70355 0.70336143Nd/144Nd 0.512920±30 0.512920±10 0.512910±20 0.512930±40 0.512896±03 0.512877±08 0.512894±05 0.512914±05 0.512901±04 0.512877±06 0.512905±04143Nd/144Nd(i) 0.51292 0.51292 0.512909 0.51293 0.51289 0.51288 0.51289 0.51291 0.5129 0.51288 0.5129208Pb/204Pb 39.0720±22 38.9600±10 38.9040±16 38.9840±24 39.0323±08 38.5317±15 39.0355±13 38.9959±13 38.8937±25 38.9375±14 38.8703±12208Pb/204Pb(i) 39.063 38.949 38.894 38.972 39.015 38.518 39.0196 38.987 38.876 38.928 38.86207Pb/204Pb 15.6670±08 15.6040±04 15.5940±06 15.6550±10 15.6236±03 15.5990±05 15.6349±04 15.6171±04 15.6178±08 15.6332±04 15.6210±04207Pb/204Pb(i) 15.667 15.604 15.594 15.655 15.623 15.599 15.634 15.617 15.617 15.633 15.621206Pb/204Pb 19.1870±±09 19.2620±05 19.1590±08 19.0920±12 19.3771±03 18.8164±05 19.3201±04 19.3416±05 19.2247±08 19.1577±05 19.1747±04206Pb/204Pb(i) 19.180 19.254 19.151 19.082 19.365 18.806 19.309 19.334 19.211 19.15 19.166

Table 3 (continued)

211J.-M

.Dautria

etal./

Lithos120

(2010)202

–222

.

.

Sample FES CAB BAG BAS GUI MIC ROM AZ SAU PP1 MTF

Age (My) 1.5** 2** 2** 2** [1.26] 2** [1.64] [5.75] [7.1] [6.35] [23.6]

87Sr/86Sr 0.703223±02 0.703226±02 0.703585±16 0.703536±09 0.703187±09 0.703529±08 0.703221±03 0.704002±08 0.707214±12 0.703417±11 0.703325±1087Sr/86Sr(i) 0.70322 0.70322 0.70358 0.70353 0.70318 0.70352 0.70321 0.70398 0.70721 0.70341 0.70325143Nd/144Nd 0.512906±03 0.512887±06 0.512910±03 0.512912±04 0.512912±03 0.512892±03 0.512941±06 0.512796±12 0.512989±09 0.512949±08 0.512695±08143Nd/144Nd(i) 0.5129 0.51289 0.51291 0.51291 0.51291 0.512892 0.51294 0.51279 0.51298 0.51294 0.51268208Pb/204Pb 38.8668±12 39.0157±18 38.9972±12 38.9338±28 38.9978±12 39.0019±23 38.9980±12 38.9530±07 38.9600±08 38.9682±46 39.0230±05208Pb/204Pb(i) 38.856 39.000 38.982 38.92 38.988 38.984 38.986 38.918 38.909 38.927 38.871207Pb/204Pb 15.6172±04 15.6153±06 15.6190±04 15.6215±08 15.6177±04 15.6239±07 15.6214±04 15.6090±06 15.6050±08 15.6230±04 15.6160±05207Pb/204Pb(i) 15.617 15.615 15.618 15.621 15.617 15.623 15.621 15.607 15.603 15.621 15.61206Pb/204Pb 19.1960±04 19.3621±06 19.3281±04 19.2523±03 19.3804±04 19.2898±08 19.3281±03 19.2030±06 19.2790±07 19.1568±04 19.5150±05206Pb/204Pb(i) 19.188 19.349 19.316 19.243 19.373 19.278 19.319 19.171 19.238 19.126 19.387

.

.

Sample POU3 EG TS NT VI AU2 AU3 AU4

Age (My) [46] 14.0* 58.8* [66.9] 161.2* [6.5] [6.5] [6.5]

87Sr/86Sr 0.703799±10 0.703854±08 0.703582±09 0.704317±09 0.704271±07 0.703678±03 0.703589±02 0.703542±0587Sr/86Sr(i) 0.70375 0.70382 0.7035 0.70421 0.70374 0.70366 0.70358 0.70353143Nd/144Nd 0.512821±07 0.512872±12 0.512939±10 0.512855±07 0.512834±07 0.512864±05 0.512836±04 0.512871±05143Nd/144Nd(i) 0.51278 0.51286 0.51289 0.51281 0.51269 0.51286 0.51283 0.51287208Pb/204Pb 39.1720±04 39.2275±07 39.3010±09 39.0960±08 39.0045±06 39.4330±12 39.1528±10 39.4819±14208Pb/204Pb(i) 38.925 39.1503 38.9364 38.7226 38.4222 39.3894 39.1204 39.4337207Pb/204Pb 15.6260±04 15.6480±05 15.6290±08 15.6280±08 15.6430±05 15.6494±04 15.6449±04 15.6420±05207Pb/204Pb(i) 15.616 15.645 15.617 15.615 15.618 15.648 15.644 15.64206Pb/204Pb 19.5090±04 19.2120±06 19.3050±07 19.0780±05 18.7790±04 19.5691±05 19.2546±08 19.6957±05206Pb/204Pb(i) 19.305 19.151 19.044 18.809 18.309 19.535 19.229 19.658

Table 3 (continued)

Table 3 (continued)

212J.-M

.Dautria

etal./

Lithos120

(2010)202

–222

Fig. 5. (a, b): 143Nd/144Nd(i) versus 87Sr/86Sr(i) for Languedoc peridotitic xenoliths and host lavas. Lavas groups (1 to 4) are those defined in Table 2; AUB: Aubrac district;EG: Eglazines spinel–garnet peridotite; PYR Lith: Pyrenean Lithosphere (Mukasa et al., 1991); FMC xeno: French Massif Central peridotite xenoliths (for references, see text); Olotxeno (stippled area): peridotite xenoliths form Olot (Iberian plate, NE Spain; Bianchini et al., 2007); Sardinia xeno (dotted area): Peridotite xenoliths from Sardinia (Beccaluva et al.,2001); EUR Lavas: field of European lavas from Piromallo et al. (2008); FMC lavas: lavas of the French Massif Central (for references, see text); CAT lavas: Catalunya lavas (includingOlot) from Cebria et al. (2000); LVC: Low Velocity Component from Hoernle et al. (1995). Sr and Nd isotopic compositons of the analyzed xenoliths have been reported bothuncorrected and corrected for in situ decay at 161 Ma (age of the oldest lava).


described. All are equilibrated in the spinel domain, except somelherzolitic samples (EG9 and 11, Table 4) from a single CD breccia pipe(EG, Fig. 1) which contain spinel and garnet in textural equilibrium(Berger, 1981). Except for the coarse equant texture, all the classicaltextural types are present (porphyroclastic, e.g. PP-2, granoblastic, e.g.SOU, mylonitic: Pg-0, Table 4). Several poikilitic samples have beencollected in Lodévois (e.g. PCV-9, Table 4). Most xenoliths displaygeochemical evidence of metasomatism, and in several localitiessamples contain secondary phases such as amphibole (pargasiteto kaersutite, e.g. Pg), Al-poor diopside (e.g. SOU), and more rarelyphlogopite and/or alkali feldspar. Calcite of secondary origin hasalso been described in some lherzolitic xenoliths from NE Languedoc(PP-2).

We have selected 10 mantle xenoliths out of 92 collected by Alardand Dautria, unpublished. These selected samples are representativeof the range of textures, equilibrium conditions and geochemicalfingerprints (see REE patterns, Fig. 8) identified in Languedoc. Theirmajor- and trace-element compositions as well as Sr, Nd and Pbisotopic ratios are given in Table 4, along with their equilibriumtemperatures. Several of these xenoliths were partially previouslydescribed by Lorand et al., 2003 and by Dautria et al., 2006.

4.2.1. Major and trace elementsOverall the fertility of the xenoliths from the ELD and the HLVD

districts is comparable to those from the Southern French MassifCentral (south of 41°S as defined by Lenoir et al., 2000a,b). For

instance the average whole-rock Al2O3 content is about 2.4±0.9 wt.%(n=49) for the HLVD-ELD, indistinguishable from the composition ofxenoliths from the Southern French Massif Central (Al2O3=2.5±1.0 wt.%; n=134). The Causses xenoliths have whole-rock Al2O3contents (2.9±0.6 wt.%; n=15) similar within error to those of theSouthern French Massif Central and the HLVD. Although we cannottotally rule out a sampling bias, the Montferrier xenoliths (n=17)appear to be more fertile with Al2O3 contents of 3.5±0.7 wt.%. Thenearby Pyrenees orogenic massifs have essentially a bi-modalharzbugite–Lherzolite composition volumetrically dominated by thelherzolite type with Al2O3 wt.% ranging between 2.5 and 4.5 (Le Rouxet al., 2007).

Although there are significant variations in term of REE contentsdue to various degrees of depletion (e.g. YbCI: 0.73–2.7; CI: CI-1chondrite nomalised), the REE patterns of most of the Languedocmantle xenoliths (except poikilitic samples, PCV-9) can be describedas a continuum between two types of REE patterns. The first,illustrated by the Gt-Sp sample Eg-9 (Fig. 8), shows a depletion inthe middle to light REE (MREE and LREE, respectively) relative to theheavy REE (HREE). Such a pattern is reminiscent of the so-called“DepletedMORBMantle (DMM) pattern” and is classically consideredas resulting from the extraction of partial melts. The second “end-member” pattern, illustrated by sample SOU-6, is characterized by arelatively flat HREE to MREE segment [(Eu/Yb)N≈1] and a markedenrichment of the LREE relative to MREE and HREE [(La/Sm)NN3].Most of the other Languedoc xenoliths have REE patterns

Fig. 6. (a, b): Diagrams of 207Pb/204Pb vs 206Pb/204Pb and 208Pb/204Pb vs 206Pb/204Pb forLanguedoc peridotitic xenoliths and host lavas. For identification of groups andabbreviations, see Fig. 4. The Group 4 lavas are plotted individually and identified bytheir ages (see Tables 1 and 2). MO-OR lavas: Monchique and Ormonde lavas fromBernard-Griffiths et al. (1997). EAR: European Asthenospheric Reservoir (Granet et al.,1995).

Fig. 7. ln (age×10) vs 206Pb/204Pb(i). Symbols and abbreviations as in Fig. 6.


intermediate between these two end-members (Fig. 8). A fractionalmelting model (source composition and partition coefficients as forlava modelling) suggests that the Languedoc lithosphere hasundergone small to moderate amounts of partial melting (≤4%fractional melting), except for samples CX2, PP-2 and PCV-9 for whichthe degree of melting reaches 10–15%. This slightly depleted“protolith” was subsequently metasomatized to a variable extent bypercolation of small-volume melts enriched in the most incompatible

elements (Navon and Stopler, 1987). PCV-9 does not belong to thiscontinuum, and shows an almost flat REE pattern consistent withprevious data reported for samples from the same locality (Lorandand Alard, 2001). This pattern is typical of poïkilitic samples anddenotes melt–rock interaction at high melt–rock ratios with an OIB-like melt (Alard et al., 1996; Xu et al., 1998; Lorand and Alard, 2001).

Abundances of Large Ion Lithophile Elements (LILE: Ba, Rb, Th, U,Pb, Sr) are more variable and may depend upon the occurrence ofmetasomatic phases such as amphibole. A detailed discussion of thebehaviour of these elements is beyond the scope of this contribution,but we note the distinct behaviour of U relative to Th and to a lesserextent of Sr and Pb relative to Ce. The depleted patterns (e.g. GRM-2,Eg-9) are marked by selective enrichment of U relative to Th, yieldingextremely high U/Th [(U/Th]PM as high as 27 [i.e. Pg5)]; Pb also showspronounced positive anomalies relative to Ce (Fig. 9). Such anomalieshave been commonly described in the FrenchMassif Central xenoliths(Alard et al., 1996; Lenoir et al., 2000a,b). In contrast, LREE enrichedsamples such as Pg-0 and SOU-6 show concomitant enrichment in Uand Th and have broadly chondritic U/Th ratios. In these samples Pband Sr do not define marked anomalies.

The distribution of High Field Strength Elements (HFSE: Nb, Ta, Zrand Hf) in these samples allows us to identify other differencesbetween the two end-members. LREE-enriched samples (SOU-6, PG-0, PP-2) have pronounced negative anomalies in Nb and Ta relative toTh and La [e.g. (Nb/La)PM=0.03–0.41]. Such negative anomalies areoften considered as symptomatic of carbonate metasomatism (e.g.,Dautria et al., 1992, 2006; Ionov et al., 1993). The HFSE abundance ofthe LREE-depleted end-member samples follows the general trend ofincompatible-element depletion except for the two garnet-bearingsamples [(Nb/La)PM=0.34–2.0]. The pattern of PCV-9 shows slightpositive anomalies inNbandTa relative to the LREE [i.e. (Ta/La)PM=2.3;(Nb/La)PM=2.6)].

4.2.2. Sr, Nd and Pb isotopesThe isotopic ratios of Sr, Nd and Pb display rather large ranges

(b87Sr/86Sr 0.70287–0.70578; 143Nd/144Nd 0.51256–0.51414; 208Pb/204Pb 37.772–39.041; 206Pb/204Pb 17.901–19.353) except 207Pb/204Pb,which shows little variation (15.570–15.620). Two garnet–spinelperidotites (i.e. EG samples) show strongly positive εNd as high as 29.Such positive values have only found in the North French MassifCentral and have been interpreted as the fingerprint of an old melt-depletion event (Downes et al., 2003). However, contrary to the N-MCF, the EG xenoliths have highly radiogenic Sr (87Sr/86SrN0.705),which could then be interpreted as the signature of a time-integrated

Table 4Major- and trace elements and Sr, Nd, Pb isotopic compositions of selected peridotite xenoliths from Languedoc. Sp, spinel; Gt, garnet; Lhz, Lherzolite; Hz, harzburgite; Granulo,granuloblastic; Porhyro., Porphyroclastic; T°B&K(1991), 2 pyroxenes equilibrium temperature after Brey and Kohler (1990); T°Wells (1977), 2 pyroxenes equilibrium temperatureafter Wells (1977). The uncertainties for Pb isotopic ratios are better than 300 ppm.

Samples CX-2 PCV-9 PP-2 SAM-6 GRM-2 SOU-6 Eg-9 Eg-11 Pg-0 Pg-5

Host lava ref. CX MCL PP1 Non analyzed 819 BR EG EG MTF MTF

Rock type Sp Lhz Sp Hz Sp Lhz Sp Lhz Sp Lhz Sp Lhz Gt-Sp Lhz Gt-Sp Lhz Sp Lhz Sp Lhz

Texture Granulo. Poikilitic Porphyro. Porphyro. Granulo. Porphyro. Porphyro. Porphyro. Mylonitic Porphyro.

T° B&K (1990) 800 1214 835 1010 925 950 1220 1230 641 8841σ 15 6 35 24 30 25 10 10 12 9T°Wells (1977) 832 1236 870 993 913 938 1165 1174 723 8661σ 22 18 20 25 32 25 14 16 15 15SiO2 (wt.%) 44.34 43.73 43.36 44.85 44.4 43.4 43.89 44.13 43.89 44.13TiO2 0.12 0.13 bLD 0.18 bLD 0.092 0.06 0.08 0.06 0.08Al2O3 1.92 1.8 2.01 3.57 3.4 3.1 2.92 3.63 2.92 3.63Cr2O3 bLD bLD bLD 0.45 0.13 0.33 bLD bLD bLD bLDFe2O3 8.4 7.82 9.14 8.78 8.67 9.36 8.75 8.81 8.75 8.81MnO 0.12 0.11 0.13 0.14 0.112 0.13 0.12 0.13 0.12 0.13MgO 42.58 43.7 42.4 38.15 39.4 39.7 40.42 38.58 40.42 38.58CaO 1.62 1.32 1.52 2.96 2.43 2.02 2.85 3.04 2.85 3.04Na2O 0.01 0.01 bLD 0.61 0.38 0.29 0.22 0.22 0.22 0.22K2O 0.04 0.01 bLD 0.02 0.00 0.04 bLD bLD bLD bLDP2O5 0.02 0.03 bLD 0.04 0.00 0.02 0.09 0.08 0.09 0.08LOI 1.11 1.47 0.3 0.84 0.36 0.35 0.23 0.98 bLD bLD

100.28 100.13 98.86 100.59 99.28 98.83 99.55 99.68 99.32 98.7Rb (ppm) 0.59 0.13 0.57 0.39 0.20 1.76 0.25 1.41 0.21 0.30Sr 7.39 5.08 15.00 11.68 8.88 15.22 8.87 9.20 20.98 13.51Y 1.32 0.83 1.04 3.46 2.92 2.20 2.50 2.90 3.06 3.42Zr 2.31 3.05 5.59 4.93 4.42 3.12 2.27 2.28 5.30 5.84Nb 0.28 0.61 0.47 0.19 0.05 0.20 0.08 0.65 0.05 0.10Cs 0.01 0.01 0.04 0.005 0.001 0.01 0.008 0.02 0.006 0.02Ba 6.72 4.32 3.29 0.82 2.16 2.90 1.09 0.94 1.48 1.86La 0.52 0.23 1.12 0.51 0.08 2.30 0.04 0.31 1.66 0.28Ce 0.57 0.69 1.62 0.61 0.40 4.53 0.13 0.61 2.64 0.54Pr 0.07 0.09 0.2 0.15 0.11 0.47 0.03 0.08 0.24 0.12Nd 0.36 0.47 0.86 0.81 0.68 1.73 0.22 0.39 0.92 0.72Sm 0.13 0.14 0.17 0.29 0.28 0.32 0.12 0.15 0.28 0.29Eu 0.05 0.05 0.05 0.13 0.11 0.10 0.055 0.07 0.12 0.13Gd 0.21 0.17 0.17 0.48 0.43 0.36 0.24 0.30 0.48 0.48Tb 0.04 0.03 0.03 0.09 0.08 0.06 0.05 0.06 0.09 0.09Dy 0.29 0.19 0.19 0.65 0.54 0.42 0.41 0.49 0.61 0.66Ho 0.06 0.04 0.04 0.15 0.12 0.09 0.09 0.11 0.13 0.15Er 0.19 0.12 0.13 0.44 0.35 0.28 0.28 0.34 0.38 0.42Tm 0.03 0.02 0.02 0.07 0.05 0.04 0.04 0.05 0.06 0.07Yb 0.20 0.12 0.14 0.43 0.35 0.29 0.29 0.34 0.38 0.42Lu 0.03 0.02 0.02 0.07 0.06 0.05 0.05 0.06 0.06 0.07Hf 0.08 0.10 0.21 0.19 0.16 0.13 0.08 0.09 0.20 0.20Ta 0.01 0.03 0.04 0.03 0.003 0.008 0.002 0.03 0.005 0.01Pb 0.63 0.58 0.2 0.28 0.76 0.82 0.04 0.08 0.55 0.27Th 0.06 0.04 0.45 0.01 0.007 0.20 0.002 0.05 0.185 0.006U 0.02 0.01 0.15 0.01 0.02 0.05 0.003 0.02 0.05 0.0487Sr/86Sr 0.703274±05 0.703619±16 0.704434±08 0.704199±08 nd 0.702867±09 0.705396±06 0.705777±11 0.703188±12 0.702972±08εSr −17.4 −12.5 −0.9 −4.3 −23.2 12.7 18.1 −18.6 −21.7143Nd/143Nd 0.513323±19 0.512922±25 0.512559±14 0.512890±04 nd 0.513191±09 0.514139±17 0.513499±14 0.513106±20 0.513336±11εNd 13.4 5.5 −1.5 4.9 10.8 29.3 16.8 9.1 13.6206Pb/2046Pb 18.5988±11 18.654* 19.3530±67 18.7551±09 nd 18.5897±18 17.9012±07 18.2673±08 18.4871±06 18.1341±07207Pb/204Pb 15.5705±11 15.594* 15.6115±85 15.6075±10 nd 15.5959±16 15.5755±06 15.5919±06 15.5977±07 15.6196±13208Pb/204Pb 38.3478±42 38.557* 39.0408±87 38.4988±18 nd 38.3603±36 37.7719±17 38.2098±18 38.3741±25 37.9633±27


metasomatic enrichment in Rb. With the exception of the EG samples,the Languedoc xenoliths do not show any peculiar characteristicsrelative to the FrenchMassif Central xenoliths when plotted in the Sr–Nd isotopic space (Fig. 5; Downes et al., 2003, references therein). Inthese diagrams, the Pyrenees domain encompasses the Massif Centralfield, and it is not possible to discriminate between the two domains.

In the 206Pb/204Pb vs 207Pb/204Pb diagram (Fig. 6a) the Languedocxenoliths clearly plot within the Pyrenées field, off the NHRL. TheLanguedoc mantle lithosphere is characterized by a relativelyconstant 207Pb/204Pb (15.58±0.06) despite variable 206Pb/204Pbratios (17.48–19.35). However, 208Pb/204Pb is positively correlatedwith 206Pb/204Pb (Fig. 6b), and discrimination between the two do-mains is difficult.

5. Discussion

5.1. Age constraints

As clearly shown in Figs. 1 and RM1, and as previously noted byBrousse and Bellon (1974) and Ghristi (1985), a progressive re-juvenation of the magmatic activity towards the South is clear inLanguedoc: the lavas are essentially Miocene to the north, Pliocene–Lower Quaternary in the central part and Late Quaternary in the south.According to Brousse and Bellon (1974), this rejuvenation resultedfrom the activity of the mantle plume that was emplaced during theMiocene beneath the French Massif Central (∼150 km North of theLanguedoc) and subsequently expanded southward.

Fig. 8. Chondrite-normalized Rare Earth Element patterns of selected peridotiticxenoliths from Languedoc. The normalizing values are from Sun and McDonough(1989). For samples location, see Table 4 and Fig. 3.

Fig. 10. (a, b): Primitive Mantle-normalized La/Sm vs. Yb (a) and La/Sm vs Yb/Eu(b) diagrams for the selected peridotitic xenoliths from Languedoc. Symbols and datareferences as in Fig. 5a,b. CD: Causses district, ELD: Escandorgue–Lodévois district andHLVD: Hérault Low Valley district. Mtf: Montferrier xenoliths. Normalizing values forPrimitive Mantle as in Fig. 9. Pyr: Pyrenees peridotitic massifs; N-FMC and S-FMC:


Although the progressive rejuvenation of the magmatic activitytowards South is clear for the last 7 Ma, such a spatial and temporalevolution is not obvious in the pre-Miocene activity. The majorMessinian volcanic episode of North Languedoc (CD) has beenpreceded by several older magmatic events: Callovian–Oxfordianboundary (Vi), Palaeocene (TS, NT), Serravallian (EG). On the otherhand, there wasmagmatic activity in the South before the Quaternary,in the Eocene (POU) and in the Upper Oligocene (MTF). Except for thecentral part of the volcanic line (ELD) where the magmatic activitymostly occurred between 2.5 and 1.2 Ma, basaltic magmas wouldhave been generated periodically between 160 and 6 Ma beneathNorth Languedoc and 46 and 0.6 Ma beneath South Languedoc: thisobservation suggests that the Languedoc magmatic activity is notrelated to the Miocene French Massif Central plume.

peridotitic xenoliths from the Northern and the Southern French Massif Central,respectively.

5.2. Xenolith constraints

As shown in Fig. 10, there is a geographic zonation of the REEfractionation between the peridotite xenoliths from the different

Fig. 9. Extended trace-element patterns of selected peridotitic xenoliths from Languedoc. TheMcDonough (1989). For samples location, see Table 4 and Fig. 3.

French volcanic areas and the Pyrenean peridotitic massifs. Thissuggests the existence of regional differences in composition withinthe French mantle lithosphere. Four domains (North-French Massif

normalizing values for PrimitiveMantle and incompatibility sequence are from Sun and


Central (N-FMC), South-French Massif Central (S-FMC), Pyrenees(PYR) and Languedoc are distinguished in Fig. 10. The geochemicaldifferences could be either inherited from the protolith or acquiredduring subsequent metasomatic events. This zonationmay be due to adifference in the protolith and indeed Lenoir et al. (2000a,b) proposedthat the North-FrenchMassif Central protolithmantle lithosphere wassignificantly more depleted than the South-French Massif Central andthat this depletion occurred in Proterozoic time. However, thedifference in protolith is not as marked as Lenoir et al. (2000a,b)suggested: i.e. Al2O3 contents of xenoliths from North-French MassifCentral = 2.0±0.8 are within error of those from the South-FrenchMassif Central (2.5±1). This is also supported by Fig. 10a, where thethree domains N-FMC, S-FMC and Languedoc show comparable Yb-ranges. However, many of the South-French Massif Central xenolithsshow Yb contents twice as high as the Primitive Mantle value. Such“hyper”-fertility is relatively uncommon in the other domains, evenwithin the extremely fertile Pyrenees Lherzolites. A long-termdepletion of the North-French Massif Central mantle is supported bytheir highly positive εNd and εHf (Downes et al., 2003; Wittig et al.,2007), which has not been found in the South-French Massif Central.However, age zonation within the French Massif Central is notsupported by Os data, as relicts of 2.2±0.2 Ga age are found in allFrench Massif Central domains (Alard, 2000; Alard et al., 2002).Languedoc xenoliths (CD) like the North MCF xenoliths, show highlypositive εNd (i.e., +29.3; this study).

Os melt-depletion ages of ca 2.4±0.2 Ga (i.e. within the rangeof estimates for the French Massif Central melt-depletion age) havebeen obtained for the Pyrenees massifs (Reisberg and Lorand, 1995;Burnham et al., 1998). Thus, although it cannot be ruled out, a North–South variation in degree of melting and age zonation is not stronglysupported. Rather, Fig. 10 suggests that the differences observedbetween the four domains are more likely to be related to subsequentmelt–rock percolation reaction processes. Indeed the REE fractiona-tions reported here are not consistent with a simple melt-depletiontrend. Thus the differences between these lithospheric domainswould have been acquired during “metasomatic” processes (depend-ing on percolation–reaction characteristics and the nature of thepercolating fluid). A detailed discussion on the origin of this zonationis beyond the scope of this contribution. However the differencesbetween the French Massif Central and the Languedoc are statisticallysignificant, and consequently we suggest that the similarities with thePyrenees are meaningful. The Languedoc xenoliths clearly showstrong affinities with the Pyrenees orogenic massifs. Fabries et al.(1987), through a study of the Montferrier xenoliths, noted thesimilarity between these two lithospheric mantle domains in term offertility, deformation, equilibrium temperature and metasomatism(amphibole, sulphide; see also Alard et al., in press). The similarities inREE patterns indicate that the Languedoc domain and the Pyrenees

Table 5Modal composition of the source, melting proportions of the different phases and partition cclinopyroxene; Gt, garnet; Sp, spinel, Phlo, phlogopite; C0, source composition; PM, primitivfrom Halliday et al. (1995) except for spinel (Elkins et al., 2008); *, interpolated values.

Phases Ol Opx

Proportions in source 0.7 0.2Melting proportions 0 0.01

Elements Th

Ol 0.000006Opx 0.00002

Partitioning Cpx 0.0021Coefficients Gt 0.0021

Sp b0.0002Phlo 0

Co (×PM) 0.085Co (×1.6 PM) 0.136Co (×3.2 PM) 0.272

mantle lithosphere share a similar protolith and have undergonecoeval melt-percolation reaction. This is further attested by the factthat peridotites from both localities share the same 206Pb/204Pb–208Pb/204Pb space (Fig. 6b).

In Section 4.2.1, the trace-element patterns of the Languedocxenoliths have been described in terms of a continuum between (1) aMREE-LREE depleted pattern ascribed to partialmelting and (2) a LREE-LILE-enriched pattern without concomitant HFSE enrichment. Previousauthors (Dautria et al., 2006; Ionov et al., 1993) have ascribed this typeof incompatible-element fractionation as due to the percolation of a“carbonated” metasomatic fluid through a variably depleted mantlelithosphere. Petrographic evidence (i.e. carbonate, secondary clinopyr-oxe, sulphide, ±melt pockets) for such metasomatism has been foundin xenoliths from several localities throughout the Languedoc (Jakni etal., 1996; Dautria et al., 2006) and notably in sample PP-2. The 143Nd/144Nd ratio decreases as La/Sm increases, and inversely the 206Pb/204Pband 208Pb/204Pb become more radiogenic as La/Sm increases. Theserelationships strongly suggest that the isotopic composition of theLanguedoc xenoliths is significantly affected by metasomatism (exceptfor EG samples). As thefingerprint ofmetasomatism becomesmore andmore predominant, the isotopic composition is progressively shiftedtoward the European Asthenospheric Reservoir (EAR) composition.Thus, within the Languedoc mantle lithosphere, the carbonatedmetasomatism is associated with the EAR isotopic signature. We notethat PP-2, the only carbonate bearing xenolith, shows the mostradiogenic Pb composition of all the xenoliths studied here. Xenolithsfrom Montferrier display the two signatures: depleted [i.e. low (La/Sm)N, low 206Pb/204Pb (b18.5) and high εNd (N10)] and enriched [i.e.high (La/Sm)N, high 206Pb/204Pb (≥18.5) and low εNd (≤10)]suggesting that the two signatures coexist in the same area. Thisobservation precludes geographic variation. Furthermore, the occur-rence of the enriched signature in Montferrier xenoliths indicates thatthe carbonated and related metasomatism affected the Languedoclithosphere before 25 Ma. It is noteworthy that themostmetasomatisedxenoliths have Pb-isotope compositions that overlap those of the oldestLanguedoc lavas (≤45 Ma). This suggests that the isotopic signature ofthe oldest lavas was mainly driven by the mantle lithospherecomposition, and that the asthenospheric component (EAR) becameprogressively predominant.

5.3. Constraints from the lavas

In a previous paper devoted to the Lodévois basalts (Liotard et al,1999), we estimated, from a trace element modelling of non-modalbatch partial melting, that these lavas resulted from 1 to 2% of partialmelting of a lherzolitic source enriched in garnet (4%) and phlogopite(0.5%). In this modelling, the source enrichment factor was1.6×Primitive Mantle (PM; from Sun and McDonough, 1989) for

oefficients used for the batch melting modelling; Ol, olivine; Opx, orthopyroxene; Cpx,e mantle composition according to Sun and McDonough (1989); partition coefficients

Cpx Gt Sp Phlo

0.05 0.02 0.02 0.010.77 0 0.05 0.17

La Sr Sm Yb

0.0002 0.00004 0.0009 0.0240.0031 0.0007 0.0037 0.0380.054 0.091 0.27 0.430.0007 0.0007 0.22 6.40.0002 0.0047 0.0047* 0.0047*0.003 0.044 0.0059 0.030.687 21.1 0.444 0.4931.0992 21.1 0.444 0.4932.1984 21.1 0.444 0.493


the most incompatible elements (e.g. LREE, Th) and 1×PM for the lessincompatible elements (e.g. Sr, Sm, and Yb).

Does the modelled source of the Lodévois basalts account for thegeochemical characteristics of all Languedoc lavas? The same trace-element modelling as used for Lodévois shows that all Languedoclavas could be derived from the Lodévois source but with significantchanges in both the extent of partial melting and the proportions ofgarnet and phlogopite, either in the source or in the residue. However,this rather simple model does not account for the variations of severaltrace-element ratios (e.g. La/Sm, Sm/Yb, Th/Sr, and Sr/Yb). Theseratios are keys to distinguishing between the effects of source-enrichment processes and the degree of melting. For this paper, wehave calculated the theoretical melting curves for three enrichmentfactors (i.e. ×1PM, ×1.6 PM, ×3.2 PM) and for different aluminousphases (i.e. pure garnet, garnet+spinel, and pure spinel). This rangeof enrichment factors has been chosen because it is similar to the LREEenrichment observed in several metasomatised peridotitic xenolithsfrom Languedoc (e.g. CX2, PP2, SOU6, Table 4). For the HREE, weassumed an enrichment factor of 1×PM, as observed for instance inperidotite PP2 (Table 4). The modal composition of the source, themelting proportions of the different phases and the partitioncoefficients used for this modelling are given in Table 5. The best fitis obtained for themelting of a lherzolitic source containing 2% garnet,2% spinel and 1% phlogopite (Fig. 11a, b). This low phlogopite contentsuggests that the LILE enrichment of source is essentially cryptic andnot related to modal mineralogy. Fig. 11a,b shows that the Languedoclavas plot between the PM and 3.2×PM curves with degrees ofmelting ranging between 1 and 5% in the La/Sm vs. Sm/Yb diagram(Fig. 11a) or between 0.5 and 3% in the Th/Sr vs. Sr/Yb diagram(Fig. 11b). These degrees of melting are low in comparison to thoseobtained experimentally by Green and Falloon (2005) for alkalibasalts (10%). According to these authors, they rather correspond tothe degrees of melting that would produce nephelinites (2.5%) and

Fig. 11. (a, b): La/Sm vs Sm/Yb and [(Th/Sr)×1000] vs Sr/Yb diagrams for the studiedlavas. Curves for the partial melting of Primitive Mantle (PM) and two enriched sources(1.6 PM and 3.2 PM) are plotted. Partial melting degrees (0.5, 1, 3, 5, and 10%) areindicated by dotted lines. PM values are from Sun and McDonough (1989).

basanites (4.5%). As noted above, most Languedoc basalts display highlevels of SiO2 undersaturation [%(Ne+Lc)norm, up to 15 in basalts, upto 26 in basanites] and high LILE concentrations (e.g. Th up to17 ppm). Such high contents of normative feldspathoïds and LILE aregenerally observed in basanites and nephelinites. Thus, althoughmostLanguedoc basalts plot within the alkali basalt field of Cox et al.(1979); Fig. 2, our calculated low degrees of melting remainacceptable. These results confirm that the sources of the Languedoclavas are lithospheric and located at the spinel–garnet transition zone(70–90 km depth).

As shown by Fig. 11a,b, the oldest lavas (between 161 and 14 Ma)display exactly the same range of degrees of melting and sourceenrichment factors as the youngest lavas (Plio-Quaternary). Thesefigures also show that the sources of lava for adjacent volcanoes candiffer in both LILE enrichment and degree of melting. Such sourceheterogeneities and such variability in partial melting within a small(kilometre)-scale magmatic province strongly suggest a lithosphericsource for these lavas.

This modelling suggests that the transitional basalts RQH and AGwould result from 3 to 5% of melting. Such low degrees of melting donot agree with the high SiO2 contents of these samples (49.70% and51.06%, respectively), which are more akin to melts derived by higherdegrees of melting (≥10%). Thus, this SiO2 enrichment results eitherfrom contamination by SiO2-rich crustal materials during the ascent ofthe magma or from melting of a mantle source dominated byorthopyroxene (harzburgite). Considering that the Sr and Nd isotopiccompositions do not indicate any significant crustal contamination,we favour the second hypothesis. However, the relatively high LILEcontents of these lavas, (e.g. La=21–39, Nb=38–58, and Th 3.4–6.9),imply that their hypothetical harzburgitic source was LILE-enriched(between ×1 and ×1.6 PM, Fig. 11b). The harzburgite xenoliths aregenerally LILE-impoverished: for instance, the only studied harzbur-gite from our xenolith set (PCV9, Table 4) displays an enrichmentfactor of 0.3×PM. However, as observed in the Lherz massif, veins ofLILE-enriched websterite and amphibole-rich veins commonly cross-cut the harzburgitic bodies (Bodinier et al., 2004). The source of theLanguedoc transitional basalts may be such a veined and hydratedharzburgite.

If the Languedoc lavas were of asthenospheric origin, their isotopicheterogeneities would imply either contamination during magmaascent, or source heterogeneities, conflicting with a purely astheno-spheric origin. The Sr and Nd isotopic compositions of the Languedoclavas preclude significant crustal contamination. The variations in206Pb/204Pb ratios at nearly constant 207Pb/204Pb have been observedpreviously for the French Massif Central basalts and were interpretedas resulting from contamination by granulite-derived melts (Downes,1984), or more recently to mixing between asthenospheric andlithospheric melts (Wilson and Downes, 2006). Such a mixing modelcould be also applied to the Languedoc basalts and, in this case, thelithospheric melt would derive from melting of a lithosphere akin tothe Pyrenean lithosphere. In the 87Sr/86Sr vs. 143Nd/144Nd diagram(Fig. 5), all Languedoc basalts are included within the field of theLanguedoc xenoliths, and, at a larger scale, within the fields of boththe FrenchMassif Central and the Pyrenean lithosphere. This feature israther consistent with a purely lithospheric origin, but it can alsosimply indicate that the asthenospheric and the lithospheric compo-nents cannot be distinguished in such an isotopic space, that iscorroborated by the position of the LVC field in Fig. 5b.

In the 208Pb/204Pb vs. 206Pb/204Pb diagram (Fig. 6b), most of thestudied basalts plot away from the field of Languedoc xenoliths and inthe Pyrenean lithosphere field. In the 207Pb/204Pb vs. 206Pb/204Pbdiagram (Fig. 6a), all Languedoc basalts plot away from the two latterfields, except for the two oldest (NT and Vi), which plot within bothfields (Fig. 6a and b). This strongly suggests that the isotopiccharacteristics of the Vi and NT sources are ancient and similar tothe Pyrenean lithosphere. These source characteristics may have been


acquired before 161 Ma, perhaps during the Hercynian orogeny oreven before. This also shows that an unmodified Pyrenees-likelithosphere was present beneath Languedoc until at least 67 Ma.The basalts with intermediate ages (between 59 and 24 Ma, i.e. TS,MTF, Group 4) show higher 206Pb/204Pb ratios uncorrelated with207Pb/204Pb and plot between the Pyrenean field and the EAR domain.This shift can be assigned to the increasing participation of an EAR-like asthenospheric component in their source. The Miocene–Pliocene–Quaternary lavas plot within, or close to the Group 4 do-main (Table 2). Except for Vi, the mantle sources of all Languedoclavas can be therefore described in terms of variablemixing between aPyrenees-like lithosphere and an EAR-like component.

In the various diagrams showing trace-element contents versusisotopic ratios (e.g. La/Sm vs 206Pb/204Pb, Fig. 11), a weak correlationcan be observed for both lavas and xenoliths. This suggests that (1) themechanisms responsible for the incompatible-element enrichmentand the isotopic variations are probably correlated, implying that theEAR-like component must be LILE-enriched, (2) the upper lithosphere(sampled by the xenoliths) and the lower lithosphere correspondingto the sources of the basalts have been similarly affected. For thexenoliths, the “enriched” signature is clearly related to volatile-rich(carbonated?) metasomatism. Such metasomatic fluids are classicallyattributed to small-volume melts issuing from the underlyingasthenosphere. As suggested by the presence of metasomatic phasesin many xenoliths (amphibole, phlogopite, K feldspar, and carbonate),we propose that the EAR signature may be also partly stored inthe lower lithosphere within secondary metasomatic minerals. How-

Fig. 12. 206Pb/204Pb vs La/Sm diagram for Languedoc xenoliths and host lavas. Sym

ever, the small-volume melts can be also responsible for a crypticenrichment bearing the EAR signature.

The 206Pb/204Pb vs. age diagram(Fig. 7) shows that (1) the injectionand percolation of the small-volume melts through the lowerlithosphere started around 67 Ma ago, (2) the ranges of geochemicaland Pb isotopic heterogeneity observed in the lavas erupted between67 and 24 Ma (Group 4, Table 2) and in the Miocene–Pliocene–Quaternary lavas (Groups 1, 2, 3, Table 2) are identical. This wouldimply that the Miocene asthenospheric uprising beneath the FMC hasnot induced any significant geochemical (LILE and isotopes) modifi-cation in the sources of the Languedoc lavas.

5.4. A model for the 160 Ma-long episodic magmatic activityin Languedoc

The only significant asthenospheric upwelling event underWesternEurope between 160 Ma and 25 Ma is the arrival of the Central Atlanticplume head (Oyarzun et al., 1997 and Piromallo et al., 2008). This majormantle event occurred ∼70 Ma ago (Piromallo et al., 2008). Oyarzunet al. (1997) and Piromallo et al. (2008) proposed that, the CA plumehead was composed of asthenospheric material contaminated bylithospheric components entrained towards the East by the drift of theEuropean plate. At the European scale, Piromallo et al. (2008) considerthat the Upper Cretaceous–Eocene magmatism was derived entirelyfrom partial melting of the Central Atlantic Plume head while thesubsequent volcanic activity would be favoured by rifting and regional-scale convection related to the recent geodynamic evolution of Europe.

bols as in Fig. 5. Field of Cape Verde carbonatites is from Hoernle et al. (2002).


However, the geochemical and isotopic heterogeneities observedamong the European lavas would only reflect the heterogeneity of thisplume head.

in contrast at the Languedoc scale, our observations suggest thatthe alkali magmas result predominantly from partial melting of thelower lithosphere. The role of the Central Atlantic Plume head wouldbe limited to supplying the EAR-like small-volume melts thatmetasomatised the overlying lithosphere. However, the chemicalnature (silicated or carbonated?) of thesemelts remains questionable.At Cape Verde, the Central Atlantic Plume has regularly producedcarbonated magmas in the past (Gerlach et al., 1988). Many CapeVerde carbonatites have EAR-like 206Pb/204Pb (Hoernle et al., 2002)and high La/Sm ratios (Fig. 12) making them very good candidates forour Small Volume Melt component. Furthermore, carbonated meta-somatism has been directly and indirectly evidenced in the Languedoclithosphere following petrologic studies (Jakni et al., 1996; Dautria etal., 2006) and geochemical evidence (e.g. trace elements, this study).In this model, the increase in 206Pb/204Pb ratios observed in theLanguedoc lavas between 67 and 24 Ma (Fig. 12) would simply resultfrom an increasing participation, during partial melting, of thecarbonated small-volume melts stored in the lithosphere. Significant-ly, the two oldest lavas (Vi, and NT) have Pb-isotope compositions inthe range of the Pyrenean lithosphere, suggesting that they mayhave been produced without the addition of any small-volume melts.On the other hand, some younger Quaternary lavas (e.g. RQH, AG,ROL2, 819) display characteristics close to those of Vi and NT, sug-gesting that the Languedoc basalt source has been affected hetero-geneously by melt percolation. The 206Pb/204Pb variability observed inthe Cenozoic lavas (Fig. 7) is probably partly the consequence of thisheterogeneity.

The recurrence of partial melting episodes in the lower lithosphereover a 160 Ma timespan in the same restricted area (∼8000 km2),implies that the Languedoc lower lithosphere has been chemically andalso probably thermo-barometrically close to its solidus conditionssince at least the Mid-Jurassic. These peculiar conditions could bepartly inherited from the Hercynian orogeny. The ∼67 Ma oldmetasomatic event, which probably affected the whole lithosphere,would be responsible for the crystallization of volatile-rich secondaryphases in the lower lithosphere, that would enhance its meltingpropensity. In such conditions, subtle P and/or T changes could triggerlow-degree partial melting and generate small volumes of magma.Two heat sources can be considered to be operative in Languedoc: thepassive mantle uprising associated with the Oligocene rifting in theGulf of Lions, and the Miocene–Pliocene mantle upwelling of theFrench Massif Central. Local decompression events (associated withthe evolution of the Languedoc sedimentary basins and the re-adjustment of lithospheric blocks) also occurred during the Mesozoicand Cenozoic due to the progressive movement of the Thetys Ligurianmargin towards the present Mediterranean margin. These events arePaleocene uplift, Lutetian relaxation and Oligocene rifting (Séranne etal.,2002). Volcanic eruptions occur in the Languedoc area, during all ofthese periods. However, the arrival of the Central Atlantic Plume headnear 70 Ma may also have induced sufficient perturbation to triggerlocal partial melting within the lower lithosphere. The Paleocenevolcanic episode observed in North Languedoc (TS, NT) may beattributed to this event.

The 160 Ma volcanic episode (Vi) cannot be explained by the sameprocess. The Vi basalt would be the only lava derived from a Pyreneanlithosphere that was unmodified since the Hercynian. In this case, Viwould be the only Southern France magmatic event related to theextension that affected the Ligurian Tethys passive margin duringDogger.

The progressive rejuvenation of the volcanic activity towardsSouth during the last 7 Ma is also debatable. We propose that thismagmatism is related to the distension migration related to theSouth-Southeastwards roll-back of the subduction of the Thethyan

oceanic crust beneath Southern Europe and associated back-arc rifting(Seranne, 1999).

The succession of events that have affected Languedoc istentatively summarized in Fig. RM2a and b.

6. Conclusions

The sporadic volcanic activity of Languedoc spans the last 160 Maand consists of very small volumes of alkali basalts (b2 km3). Thesebasalts eruptedwithin the same small area (∼8000 km2), even thoughthe European plate shifted about 2500 km to the East during the sametime period. Such details suggest (1) a lithospheric origin for thismagmatism as proposed by Beccaluva et al. (2007) for the Adriaticvolcanism and (2) a relationship to regional tectonic events ratherthan to large-scale and deep mantle events. This would imply that theLanguedoc lithosphere has been chemically, and probably thermo-barometrically, close to its solidus conditions from at least 160 Ma.Themantle event responsible for these characteristicsmust be prior toMiddle Jurassic and it may be a Hercynian heritage.

Our new data and modelling confirm a lithospheric origin of theLanguedoc magmatism: the lava sources are located in the lowerlithosphere, at the transition between the garnet and spinel stabilityfields. The EAR signature observed in both the Cenozoic lava sourcesand overlying xenoliths are suggested to be associated withmetasomatism involving the percolation of volatile-rich small volumemelts (probably carbonated) and heterogeneously affecting thewholemantle lithosphere. Our isotopic data suggest that this metasomaticevent occurred ∼67 Ma ago: it may be related to the arrival of theCentral Atlantic plume head under southern France.

The lower lithosphere beneath Cenozoic Languedoc was at thesame time close to its solidus conditions and metasomatised (whichenhanced its melting propensity). In such conditions, subtle P and/or Tchanges resulting from the Mesozoic–Cenozoic tectonic evolution ofthe Thetys Ligurian margin towards the present Mediterraneanmargin, would be able to trigger local low-degree partial melting.

Finally, we suggest that the role of the asthenosphere in theLanguedoc volcanism was minor.

Documents

160Ma of sporadic basaltic activity on the Languedoc volcanic … · 2011. 1. 26. · 160 Ma of sporadic basaltic activity on the Languedoc volcanic line (Southern France): A peculiar