Lithos 132-133 (2012) 70–81

Contents lists available at SciVerse ScienceDirect

Lithos

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

New evidence of mantle heterogeneity beneath the Hyblean Plateau (southeast Sicily,Italy) as inferred from noble gases and geochemistry of ultramafic xenoliths

A. Correale a,⁎, M. Martelli b, A. Paonita b, A. Rizzo b, L. Brusca b, V. Scribano c

a Dipartimento di Scienze della Terra e del Mare (DiSTeM), Università degli Studi di Palermo, Via Archirafi 36, Palermo 90123, Italyb Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Palermo, Via Ugo La Malfa 153, Palermo 90146, Italyc Dipartimento di Scienze Geologiche, Università degli Studi di Catania, Corso Italia 55, Catania 95129, Italy

⁎ Corresponding author at: Dipartimento di Scienze dUniversità degli Studi di Palermo, Via Archirafi 36, Pale6809273; fax: +39 91 6809449.

E-mail address: a.correale@pa.ingv.it (A. Correale).

0024-4937/$ – see front matter © 2011 Elsevier B.V. Alldoi:10.1016/j.lithos.2011.11.007

a b s t r a c t

a r t i c l e i n f o

Article history:Received 14 July 2011Accepted 4 November 2011Available online 11 November 2011

Keywords:XenolithMantleHyblean PlateauMetasomatismNoble gasFluid inclusion

We analyzed major and trace elements, Sr and Nd isotopes in ultramafic xenoliths in Miocenic age Hybleandiatremes, along with noble gases of CO2-rich fluid inclusions hosted in the same products. The xenolithsconsist of peridotites and pyroxenites, which are considered to be derived from the upper mantle. Althoughthe mineral assemblage of peridotites and their whole-rock abundance of major elements (e.g.,Al2O3=0.8–1.5 wt.%, TiO2=0.03–0.08 wt.%) suggest a residual character of the mantle, a moderate enrich-ment in some incompatible elements (e.g., LaN/YbN=9–14) highlights the presence of cryptic metasomaticevents. In this context a deep silicate liquid is considered the metasomatizing agent, which is consistentwith the occurrence of pyroxenites as veins in peridotites. Both the Zr/Nb and 143Nd/144Nd ratios of the inves-tigated samples reveal two distinct compositional groups: (1) peridotites with Zr/Nb≈4 and 143Nd/144Nd≈0.5129, and (2) pyroxenites with Zr/Nb≈20 and 143Nd/144Nd≈0.5130. The results of noble-gas an-alyses also highlight the difference between the peridotite and pyroxenite domains. Indeed, the 3He/4He and4He/40Ar* ratios measured in the fluid inclusions of peridotites (respectively 7.0–7.4±0.1 Ra and 0.5–8.2,where Ra is the atmospheric 3He/4He ratio of 1.38×10−6) were on average lower than those for the pyrox-enites (respectively 7.2–7.6 Ra and 0.62–15). This mantle heterogeneity is interpreted as resulting from amixing between two end-members: (1) a peridotitic layer with 3He/4He≈7 Ra and 4He/40Ar*≈0.4, whichis lower than the typical mantle ratio (~1–4) probably due to melt extraction events, and (2) metasomatizingmafic silicate melts that gave rise to pyroxenites characterized by 3He/4He≈7.6 Ra, with a variable 4He/40Ar*due to degassing processes connected with the ascent of magma at different levels in the peridotite wall rock.The complete geochemical data set also suggests two distinct mantle sources for the xenolithic groupshighlighted above: (1) a HIMU (high-μ)-type source for the peridotites and (2) a DM (depleted mantle)-type source for the pyroxenites.

© 2011 Elsevier B.V. All rights reserved.

1. Introduction

Mantle xenoliths from diatremes often exhibit a pristine characterdue to the relatively low eruptive temperature and high ascent veloc-ity of such volatile-rich volcanic systems. Thus, a careful investigationof these xenoliths can provide unique information on upper-mantlecomposition and the processes that may modify it (e.g., Beccaluva etal., 2004; Downes, 2007; Dunai and Baur, 1995; Gautheron et al.,2005; Vaselli et al., 1995; Zangana et al., 1999). The Central Mediter-ranean area provides an attractive example of such an approach in ageodynamically complex region, where the characteristics of the lith-ospheric mantle have mostly been inferred from geophysical data(e.g., Berry and Knopoff, 1967; Calcagnile et al., 1982; Finetti and

ella Terra e del Mare (DiSTeM),rmo 90123, Italy. Tel.: +39 91

rights reserved.

Morelli, 1973). Our work within this area focuses on investigatingthe mantle beneath the Hyblean Plateau (southeastern Sicily, Italy),which is one of the rare European volcanic regions where xenolithsoccur.

The Hyblean Plateau has been characterized by several distinct ep-isodes of magmatism, starting from Triassic and lasting until Quater-nary (Carbone and Lentini, 1981; Cristofolini, 1966; Rocchi et al.,1998). Some of the volcanic events brought to the surface a largenumber of mantle-derived xenoliths, mainly spinel-facies peridotitesand subordinate pyroxenites. These products have been widely inves-tigated by many authors (Bianchini et al., 2010; Perinelli et al., 2008;Sapienza and Scribano, 2000; Sapienza et al., 2005; Tonarini et al.,1996), who highlighted the occurrence of metasomatic events affect-ing the local mantle. Sr- and Nd-isotope data of Hyblean peridotites(87Sr/86Sr=0.70288–0.70309 and 143Nd/144Nd=0.51287–0.51292)reveal a HIMU (high-μ)-like affinity, while the data for pyroxenites(87Sr/86Sr=0.70305–0.70326 and 143Nd/144Nd=0.51292–0.51299),which differ slightly from those of peridotites, has isotope

71A. Correale et al. / Lithos 132-133 (2012) 70–81

characteristics overlapping the alkaline lavas, implying that the py-roxenite domain contributed to the genesis of the Hyblean magmas(Bianchini et al., 2010; Tonarini et al., 1996).

Geochemical investigation of CO2-rich fluid inclusions hosted inolivines and pyroxenes confirmed a mantle-derived origin for the ul-tramafic xenoliths (Sapienza et al., 2005). Microthermometric ana-lyses show entrapment pressures of fluid inclusions in the range0.75–0.95 GPa, corresponding to a depth of 27–35 km, where spinelperidotites would be stable (Sapienza et al., 2005). Also, He-isotopemeasurements of the fluid inclusions hosted in peridotite minerals(both olivine and pyroxenes) have values of 7.3±0.3 Ra (where Rais the 3He/4He ratio of 1.38×10−6 as measured in air) (Sapienza etal., 2005), similar to the isotope signature of Pantelleria Island(Martelli et al., 2008; Parello et al., 2000), and are consistent with adepleted-mantle (DM) signature of the local mantle. Such values arethe highest of Plio-Quaternary Italian magmatism (e.g., Martelli etal., 2008) and underline the importance of the Hyblean province inthe evolution of the Italian area.

This study performed a comprehensive investigation of the geo-chemistry of ultramafic xenoliths from the Hyblean area. Sapienzaet al. (2005) investigated helium in fluid inclusions of Hyblean peri-dotites, whereas in the present study we investigated both heliumand argon and not only on the peridotites samples but also on asuite of pyroxenites. The same samples have been analyzed for traceelements (in whole rocks and clinopyroxenes) and Sr and Nd iso-topes. These data give new insights into the different roles playedby peridotites and pyroxenites in determining the Hyblean mantlecharacteristics, as well on the contributions of metasomaticprocesses.

2. Geological setting

The investigated area is in a critical geodynamical setting charac-terized by the collision between the European and African plates(Fig. 1; Barberi et al., 1974), with the Hyblean Plateau located in theundeformed northern portion of the Pelagian Block, in the forelandarea (Lentini et al., 1996). However, there is a scientific debateabout the nature of the lithosphere beneath the Hyblean region. In-deed, the hypothesis supported by Vai (1994) about the possible oce-anic character of this crust contrasts with the more commongeological models that consider this lithospheric block to be in

Fig. 1.Map of Hyblean Plateau. The enlarged area shows the provenance of main xeno-lithic samples.

continuity with the African plate, thus suggesting a continental char-acter (Burollet et al., 1978).

Discontinuous volcanic activity characterized the Hyblean Plateaufrom Cretaceous to Pleistocene (Cristofolini, 1966). The products ofthe numerous eruptions interrupted the Meso-Cenozoic deep-watercarbonate deposits and the Neogene–Quaternary clastic sequences(Bianchi et al., 1987). Although Cristofolini (1966) detected a Triassicigneous layer via drill holes near Ragusa, the oldest eruptive rocksthat outcrop are Cretaceous alkali basalts and are located in the east-ern part of the area (Capo Passero, Siracusa, and Augusta; Amore etal., 1988). After the Cretaceous activity, the volcanism stopped forabout 50 Ma and then restarted during the Miocene age with alkalineaffinity lavas, which can be found in the central-northern area of theHyblean Plateau, the so-called volcanic plateau (e.g., Bianchi et al.,1987). The last eruptive episode, during Plio-Pleistocene, was charac-terized by the eruption of tholeiite and alkaline basalts and minornephelinites (Beccaluva et al., 1998). Some diatreme-related depositsof Miocene age (Carbone and Lentini, 1981) and some Quaternarybasanitic and nephelinitic lavas sampled a part of lithosphere, carry-ing a huge amount of ultramafic xenoliths to surface (Scribano,1987a, 1987b). Among these, the Miocene Valle Guffari diatreme ischaracterized by the greatest variety and quantity of deep xenoliths,and is also the area where most of the investigated samples were col-lected for the present study (Fig. 1).

3. Analytical techniques

The studied samples consist of peridotitic and pyroxenitic xeno-liths found in some Miocene diatremes of the Hyblean area, in partic-ular the Valle Guffari diatreme and Cozzo Molino pipe (Fig. 1). Thesamples were selected on the basis of their size (~5 cm for the perido-tites and ~10 cm for the pyroxenites) and freshness.

Major- and trace-element analyses were performed on bothwhole rocks and on selected olivine and pyroxene grains. Wholerocks were analyzed at the laboratory of SGS Canada using ICP-AESand ICP-MS technical procedures. Selected portions of samples werecrushed and powdered with an agate mortar, then a weighted aliquot(~0.10 g) is digested by fusion with sodium peroxide in graphite cru-cibles or dissolution by multi-acid digestion using a combination ofHCl, HNO3, HF and HClO4. During digestion each sample was splitinto two aliquots for ICP-OES and ICP-MS analyses. The accuracy ofthe method was determined by analyzing certified reference mate-rials, while its precision was determined with replicate analyses(and found to be generally better than 10%).

Single pyroxene and olivine crystals were analyzed for major ele-ments using a LEO™ 440 Scanning Electron Microscope coupled to anOxford-Link Energy Dispersive Spectroscopy system hosted at theDiSTeM laboratory, University of Palermo. More details on the proce-dure of sample preparation and on the analytical technique can befound in Lopez et al. (2006). Trace-element analyses were performedusing the laser ablation ICP-MS technique at Istituto Nazionale diGeofisica e Vulcanologia (INGV), Palermo. Selected samples were in-corporated into an epoxy-resin puck that was polished before analy-sis. The analytical system consisted of an Agilent-7500 CX quadrupolemass spectrometer coupled with an ArF excimer laser ablation system(GeoLas Pro). During analysis, samples were maintained in a heliumatmosphere, with a laser output energy of 10 J/cm2, a repetition rateof 10 Hz, and a 130-μm-diameter circular spot. We used Ca, Si, andFe as internal standards and NIST 612 as an external standard. TheNIST 612 analyses were carried out at the start, middle, and end ofeach analytical session. The precision was determined during eachanalysis session from the variance of ~15 NIST 612 measurements,which gave a relative standard deviation of b5%. The accuracy, calcu-lated using the BCR-2 international standard, was b10% for most ofthe elements.

72 A. Correale et al. / Lithos 132-133 (2012) 70–81

Sr- and Nd-isotope compositions of separated clinopyroxene phe-nocrysts (typically 1–3 g) were determined at INGV, OsservatorioVesuviano Napoli, by thermal ionization MS. Mineral samples thathad been careful hand-picked were crushed to powder in an agatemortar in order to prepare them for isotope analysis. More detailedinformation on the samples preparation and analytical procedurescan be found in Arienzo et al. (2009).

Noble gases were analyzed at the INGV Palermo laboratory bysingle-step in-vacuo crushing at a pressure of about 20 MPa (so asto minimize the contribution of noble gases from the crystal lattice)coupled with MS. He and Ne were analyzed by GVI-Helix SFT MS,while Ar was analyzed by a GVI-Argus device. Each sample was ana-lyzed twice, and in each analysis we used about 2 g of olivines and0.5–1 g of pyroxenes. We followed the same sample preparationand analytical techniques reported in Nuccio et al. (2008) andMartelli et al. (2011).

4. Petrography, and bulk-rock and mineral chemistry

On the basis of their mineralogical modal composition (50–75% ol-ivine, 8–25% orthopyroxene, 1–8% clinopyroxene, and 1–3% spinel),peridotites are classified as anhydrous spinel-facies lherzolite (XIH-3 sample) and harzburgite (XIH-1 and XIH-2 samples; Fig. 2). Theyhave a variable texture, from protogranular to porphyroclastic(Fig. 3). Data on the major elements are reported in Table 1. Olivinesshow an average composition of Fo90, orthopyroxenes were Fs9.4–11.2,Wo0.9–2.9, and En85.9–89.8, whereas the clinopyroxenes (Cr-diopsides)comprise Fs4.3–7.2, Wo43.1–49.5, and En46.2–50.1. The results of our ana-lyses are in accordance with available data in the literature forHyblean nodules, confirming a general homogeneity in the majorchemistry of the investigated xenoliths (Atzori et al., 1999;Bianchini et al., 2010; Nimis, 1998; Perinelli et al., 2008; Tonarini etal., 1996). The peridotite samples are essentially fresh, but sometimesit was possible to observe a slight degree of serpentinization along thecracks within olivines or along grain boundaries. Cr–Al spinel wasalso present both as interstitial grains and as vermiform intergrowthswith the pyroxene. Kink banding was ubiquitous in the olivine crys-tals (Fig. 3). Several samples were characterized by local modal in-creases in pyroxene contents, or centimeter-sized websterite veins.

Ol

Opx

Har

zbur

gite

Olivine bearing

Lherz

Ortopyroxenite

XXIH-2

XIH-1

Webs

Fig. 2. Modal composition of t

Four samples could be characterized as pyroxenites based on theirolivine, orthopyroxene, and clinopyroxene percentages (Fig. 2). Inparticular, samples XIP-4 and XIP-14 were clinopyroxenites (≥75%clinopyroxene and ≤5% orthopyroxene), while samples XIP-28 andXIP-17 were websterites (62–70% clinopyroxene and 13–23% ortho-pyroxene). Clinopyroxene from samples XIP-14, XIP-4, and XIP-17was an Al-diopside (Al2O3=6.8–9.5 wt.%; see Table 1) characterizedby several exsolution lamellae of Ca-poor pyroxene and Al-spinel.On the other hand, clinopyroxene from XIP-28 and XIC-26 was a Cr-diopside (Cr=0.75 and 1.14 wt.%, respectively; Table 2). The ortho-pyroxene composition varies in the range Fs9.5–20, Wo0.7–2.9, andEn77.3–89.6.

Sample XIP-28 contains ~8 vol.% Fo90 olivine. Most of the pyrox-enite samples contain variable amounts of Al–Cr spinel, which isparticularly abundant in sample XIP-4. This explains the exception-ally low silica content (SiO2=25.4 wt.%) and high alumina content(Al2O3=23.8 wt.%) in this sample (cf. Table 1). However, it mustbe noted that the distribution of the spinel in these xenoliths, andhence its grain size, were quite irregular. Considering that xenolithsare fragments of deep rocks, the percentage values therein might notbe representative of the original rock, especially for those withcoarse grain size. In fact, the averaged contents of this particular py-roxenite type deduced previously were 85% clinopyroxene, 5%orthopyroxene, and 10 vol.% Al-spinel (Punturo and Scribano,1997). It is also noteworthy that spinel is generally rimmed by akeliphytized garnet in sample XIP-17 (Table 1). In addition, we con-sidered a first-size composite xenolith (sample XIC-26), consistingof a harzburgitic peridotite frame cross-cut by two irregular 0.5-cm-wide clinopyroxenite veins. The data from the whole-rock anal-ysis reported in Table 1 represent the average composition of thiscomposite xenolith.

Observations of both peridotites and pyroxenites under the opticalmicroscope identified array of secondary fluid inclusions in olivineand pyroxene crystals (Fig. 3). Fluid inclusions were not distributeduniformly among the different mineralogical phases, in accordancewith previous observations in peridotite paragenesis by Sapienza etal. (2005). The clinopyroxenes are systematically richer in fluid inclu-sions relative to coexisting olivines and orthopyroxenes, as also ob-served in mantle xenoliths from different areas (Porcelli et al., 1986).

Cpx

Wherlite

Websterite

olite

Clinopyroxenite

Dunite

IH-3

XIP-28

XIP-17

XIP-4 XIP-14

terite

he investigated xenoliths.

Fig. 3. Thin-section photomicrographs showing typical petrographic features of the pe-ridotite samples: (a) rock-forming minerals (OL, olivine; Opx, orthopyroxene; Cpx,clinopyroxene; Spl, spinel) and their textural relations (sample XIH-3, crossed polars);(b) part of a kink-banded olivine grain cross-cut by composite serpentine and carbon-ate veins (sample XIH-3, crossed polars); and (c) fluid inclusions array within an oliv-ine grain (plane-polarized light).

73A. Correale et al. / Lithos 132-133 (2012) 70–81

5. Trace-element and Sr- and Nd-isotope geochemistry

Trace-element data for the peridotite and pyroxenite samples arelisted in Table 2. Fig. 4 shows the chondrite-normalized REE distribu-tion for both whole rocks and clinopyroxenes of peridotites. The REEpatterns are similar in Fig. 4a and b, although the clinopyroxenesshow clear REE enrichments relative to whole rocks, which are dueto the high affinity of REE for the pyroxene structure (Eggins et al.,1998).

The plots show consistent patterns of both whole rocks and clino-pyroxenes among different samples, suggesting a homogeneoussource composition. Compared to chondrite, all samples show evi-dent LREE enrichment (Lan/Ybn≈20) while HREE is slightly depleted(Fig. 4a). Evidence for this can also be found in previously publisheddata for other Hyblean peridotites and were attributed to a pervasiveor, more likely, cryptic metasomatism of a moderately depleted man-tle (Perinelli et al., 2008; Sapienza and Scribano, 2000; Sapienza et al.,

2005). The residual nature of the peridotites was also confirmed bythe depletion observed in HFSE relative to primordial mantle abun-dances (data not shown), similar to that reported by Sapienza andScribano (2000).

The chondrite-normalized REE pattern of pyroxenites is displayedin Fig. 5, both for whole rocks and clinopyroxenes. The pyroxenitesamples show a REE upward-convex pattern, characterized by aless-pronounced enrichment of the more incompatible elements(i.e., La, Ce, and Pr), and of HREE compared to MREE. Among the an-alyzed samples, only the composite peridotite–pyroxenite sample,XIC-26, show a different pattern, whose mineralogical compositionwas somewhat transitional between that of peridotites and pyroxe-nites. The enrichment of LREE in pyroxenites relative to chondritevaries among the studied samples (Lan/Ybn=2.4–11.3), opposite towhat was observed in peridotites (Fig. 5a and b). FollowingSapienza and Scribano (2000), the pyroxenites represent the crystal-lization product of deep magmatic liquids that intruded the perido-tites at different levels of the lithospheric mantle. In this framework,the differences in LREE enrichments among pyroxenite samplescould reflect varying degrees of metasomatism, depending on the ex-tent to which the metasomatizing melts interact with the surround-ing peridotite.

Fig. 6 plots Zr/Nb ratios versus Zr concentrations of the bulk rocks.The complete data set, comprising our data plus those in the litera-ture (Sapienza and Scribano, 2000), define two clearly distinguish-able compositional fields for pyroxenites and peridotites:pyroxenites are characterized by Zr concentrations of 26–40 ppmand a Zr/Nb ratio of ~20, while peridotites exhibit a much lower Zrcontent of ~8 ppm and a Zr/Nb ratio of ~4. The lower Zr concentrationof peridotites is related to their more refractory nature. A particularlynotable behavior is displayed by the XIC-26 pyroxenite sample, whichhas Zr and Zr/Nb values of 3.5 ppm and 5.3, respectively, which aremuch more similar to those of peridotites. Fig. 6 also shows the com-positions of erupted lavas having HIMU and DM signatures. Zr incom-patibility makes lavas obviously richer in this element than xenoliths;nevertheless, the Zr/Nb ratio is little affected by crystal-melt fraction-ation processes, so that lavas and xenoliths are directly comparable.Whereas pyroxenites exhibit the Zr/Nb ratio that is typical of DM, pe-ridotites clearly fall in the HIMU range.

Sr- and Nd-isotope data measured in clinopyroxenes from bothperidotites and pyroxenites are listed in Table 2 and plotted inFig. 7. The peridotites show almost homogeneous 87Sr/86Sr and143Nd/144Nd values of ~0.7029 and ~0.5129, respectively, while thepyroxenites exhibit variable 87Sr/86Sr (0.7028–0.7031) and 143Nd/144Nd≈0.5130. As shown in Fig. 7, our values are consistent withthose reported by Bianchini et al. (2010) and Tonarini et al. (1996).Consideration of the complete data set indicates the absence of anyappreciable differences in the 87Sr/86Sr ratios among peridotites andpyroxenites, whereas their 143Nd/144Nd ratios differed slightly. Inparticular, the isotope ratios were slightly higher in the pyroxenites(between 0.5129 and 0.5130) than in the peridotites (between0.5128 and 0.5129), highlighting the presence of two distinct compo-sitional groups. In accordance with inferences from the Zr/Nb ratio,inspection of Fig. 7 also suggests that the pyroxenites formed bymelts coming from a deep mantle (probably DM-type) source that in-truded into the shallower peridotite mantle level (with signaturessimilar to a HIMU-type source).

6. Chemical and isotope compositions of noble gases fromfluid inclusions

As already noted, fluid inclusions occur inside olivine and pyrox-ene crystals of Hyblean ultramafic xenoliths. These fluid inclusionsrepresent a primary gaseous phase (dominated by CO2) coexistingwith growingminerals at mantle depths, as demonstrated by thermo-barometric and microthermometric studies carried out in the same

Table 1Whole rock and mineral phases major element compositions of studied peridotites and pyroxenites xenoliths.

Sample wt.% SiO2 TiO2 Al2O3 FeO MnO MgO CaO Na2O K2O P2O5 Cr2O3

Peridotites XIH-1 WR 41.93 0.06 0.93 9.11 0.11 39.46 3.85 0.04 0.02 0.06 0.05Opx 56.66 0.11 2.76 6.25 0.15 32.67 0.57 0.49 b.d.l b.d.l 0.47Cpx 53.27 0.27 3.66 2.84 b.d.l 0.00 22.68 1.11 b.d.l b.d.l 0.85

XIH-2 WR 44.07 0.03 0.77 8.74 0.11 40.29 1.59 0.03 0.02 0.03 0.13WR 42.36 0.08 1.53 9.11 0.11 36.15 2.99 0.12 0.02 0.16 0.35Ol 41.00 b.d.l 0.03 10.39 0.26 47.58 0.12 0.70 0.01 b.d.l 0.08

XIH-3 Opx 55.24 0.16 4.05 6.72 0.08 31.00 1.39 0.64 0.03 b.d.l 0.70Cpx 52.23 0.32 5.50 3.88 0.15 16.09 19.59 1.35 b.d.l b.d.l 1.05WR 25.46 1.13 23.81 12.92 0.11 13.55 8.89 0.73 0.02 0.03 0.08

Pyroxenites XIP-4 Cpx 48.33 1.73 9.53 7.43 0.14 12.63 18.59 1.74 b.d.l b.d.l 0.05WR 44.18 1.50 11.17 7.75 0.11 14.41 16.07 0.76 0.09 0.11 0.02Opx 54.93 0.11 4.14 8.06 0.20 32.01 0.72 0.00 n.a n.a n.a.

XIP-17 Cpx 50.03 1.19 6.25 3.95 0.11 15.32 21.29 0.79 n.a n.a n.a.Spl n.a 0.45 65.14 11.72 0.09 22.14 n.a n.a n.a n.a 0.52Grn 40.36 0.11 22.62 16.16 0.46 14.12 6.75 0.03 n.a n.a n.a.WR 47.63 1.24 9.24 8.88 0.15 14.77 15.87 1.20 0.07 0.08 0.03Opx 51.23 0.54 6.89 12.53 0.19 27.20 1.35 0.18 n.a n.a n.a.

XIP-14 Cpx 48.48 1.59 9.52 4.26 0.10 13.54 17.81 1.41 n.a n.a n.a.Spl 0.09 0.56 59.78 20.35 0.11 18.02 n.a n.a n.a n.a 0.11Grn 41.60 0.38 23.22 13.50 0.35 17.32 5.57 n.a n.a n.a n.a.WR 42.95 1.07 8.21 10.57 0.13 17.95 13.66 0.80 b.d.l. 0.07 0.11Oliv 40.81 n.a n.a 9.80 0.15 49.80 n.a n.a n.a n.a n.a.

XIP-28 Opx 55.42 0.19 3.62 6.26 0.15 32.94 0.47 n.a n.a n.a 0.35Cpx 52.80 0.31 2.75 2.25 0.08 16.70 22.98 0.51 n.a n.a 0.65Spl n.a 0.11 55.80 11.61 0.09 20.80 n.a n.a n.a n.a 11.40WR 50.40 0.20 3.90 6.70 0.10 26.70 9.20 0.30 b.d.l. 0.10 0.68Oliv 40.61 n.a n.a 9.48 0.15 50.16 n.a n.a n.a n.a n.a.

XIC-26 Opx 55.80 0.12 2.85 6.65 0.16 34.30 0.35 0.01 n.a n.a 0.59Cpx 51.64 0.67 5.59 3.44 0.13 16.86 19.16 1.36 n.a n.a 1.14Spl n.a 0.42 33.10 19.60 0.20 16.15 n.a n.a n.a n.a 29.15

Table 2Trace elements abundance of whole rock and mineral phases and Sr–Nd isotopic compositions of handpicked clinopyroxenes from studied Hyblean enclaves.

Peridotites Pyroxenites

Sample XIH-1 XIH-1 XIH-1 XIH-2 XIH-3 XIH-3 XIH-3 XIH-3 XIP-4 XIP-4 XIP-17 XIP-17 XIP-14 XIP-28

WR OPX CPX WR WR OLIV OPX CPX WR CPX WR CPX WR WR

ppmBa 5.65 0.24 0.43 2.60 4.90 0.02 0.07 0.59 17.90 0.44 181 1.51 26.00 3.20Sc 6.35 18.41 94.86 8.45 8.75 3.20 21.22 58.45 23.90 55.70 42.20 57.87 37.70 n.a.Sr 202 0.57 276 186 216 0.02 2.27 250 60.10 94.61 648 57.69 161 150V 35.50 93.96 272.99 35.25 60.50 5.75 117 239 386 332 401 380 284 231Ce 5.25 0.21 41.22 4.01 4.69 0.01 0.36 32.61 6.42 12.46 18.00 7.47 16.00 10.60Co 129 58.91 25.57 121.50 123.50 134.45 71.49 29.58 100 41.82 n.a. 31.04 n.a. n.a.Cs 0.80 0.07 0.01 0.15 0.10 0.01 0.01 0.07 0.20 0.02 n.a. 0.02 n.a. n.a.Dy 0.26 0.12 1.91 0.19 0.31 0.01 0.19 2.17 2.13 4.27 n.a. 2.85 n.a. 2.28Er 0.10 0.12 0.88 0.09 0.15 0.01 0.13 0.96 1.00 2.00 n.a. 0.93 n.a. 1.11Eu 0.11 0.08 0.83 0.07 0.14 b.d.l. 0.05 1.11 0.77 1.43 1.26 1.18 1.00 0.96Gd 0.36 0.08 2.40 0.22 0.43 b.d.l. 0.14 3.01 2.46 4.63 n.a. 3.90 n.a. 2.75Hf 0.14 0.07 0.61 0.03 0.16 b.d.l. 0.09 0.80 0.90 2.28 1.90 2.11 1.30 1.54Ho b.d.l. 0.05 0.34 b.d.l. 0.06 b.d.l. 0.04 0.38 0.38 0.78 n.a. 0.44 n.a. 0.42La 3.05 0.11 20.11 2.15 2.50 b.d.l. 0.09 12.71 2.10 3.18 7.80 1.71 5.60 3.67Lu 0.02 0.05 0.11 0.01 0.02 b.d.l. 0.03 0.12 0.10 0.22 0.18 0.06 0.16 0.11Nb 2.95 0.18 1.97 1.00 1.90 0.01 0.24 1.85 1.00 0.59 n.a. 0.10 n.a. 2.06Nd 1.90 0.14 15.35 1.35 2.20 0.01 0.34 17.14 6.30 13.00 12.00 9.79 11.00 10.00Pb 0.70 0.09 0.93 b.d.l. 0.70 0.01 0.02 0.20 b.d.l. 0.06 n.a. 0.25 n.a. n.a.Pr 0.58 0.07 4.17 0.47 0.60 0.00 0.06 4.08 1.26 2.26 n.a. 1.52 n.a. 1.74Rb 1.15 0.09 0.04 0.68 0.60 0.02 0.03 0.15 0.45 0.08 9.00 0.12 9.00 0.99Sm 0.40 0.10 2.73 0.30 0.50 0.01 0.11 3.52 2.20 4.11 3.29 3.58 2.87 2.79Ta 0.07 0.04 0.22 0.60 0.05 b.d.l. 0.02 0.30 0.05 0.12 0.10 0.02 n.a. 0.18Tb 0.05 n.a. n.a. b.d.l. 0.06 n.a. n.a. n.a. 0.33 n.a. 0.60 n.a. 0.50 0.41Th 0.35 0.07 1.61 0.33 0.20 b.d.l. 0.01 0.36 0.10 0.06 0.50 0.02 0.30 0.23Tm b.d.l. 0.05 0.12 b.d.l. b.d.l. b.d.l. 0.02 0.14 0.12 0.26 n.a. 0.10 n.a. 0.13U 0.36 0.08 0.34 0.07 0.12 0.01 0.01 0.16 0.08 0.02 0.20 0.01 0.10 0.08Y 1.10 0.57 8.88 0.75 1.40 0.05 1.10 9.51 8.00 18.93 18.00 10.14 17.00 11.70Yb 0.10 0.18 0.79 b.d.l. 0.10 0.02 0.18 0.81 0.60 1.63 1.20 0.52 1.16 0.78Zr 7.00 0.94 17.07 1.88 9.40 0.11 5.50 46.29 26.20 48.73 45.00 41.45 40.00 39.3087Sr/86Sr n.a. n.a. 0.702956±5 n.a. n.a. n.a. n.a. 0.7030315±6 n.a. 0.7031435±5 n.a. 0.702859±7 n.a. n.a.143Nd/144Nd n.a. n.a. 0.512917±6 n.a. n.a. n.a. n.a. 0.512919±7 n.a. 0.512947±8 n.a. 0.512994±6 n.a. n.a.

74 A. Correale et al. / Lithos 132-133 (2012) 70–81

Fig. 4. C1-normalized REE patterns of a) whole rock (this work and Sapienza and Scri-bano, 2000) and b) cpx (this work and Perinelli et al., 2008) from peridotites.Normalization to C1 is after Anders and Grevesse (1989).

0.1

1

10

100

La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu

rock

/C1

Cpx XIP-4

Cpx XIP-17

Cpx XIC-26

b

0.1

1

10

100

La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu

rock

/C1

XIP-4 XIP-17 XIP-14XIP-28 XIC-26

a

Fig. 5. C1-normalized REE patterns of a) whole rock from this work (symbols) andSapienza and Scribano (2000) (shaded area) and b) cpx (this work) from pyroxenites.Normalization to C1 as in Fig. 4.

1

10

100

1 10 100 1000

Zr (ppm)

Zr/

Nb

peridotites this work pyroxenites this work peridotites Sapienza & Scribano (2000) pyroxenites Sapienza & Scribano (2000)

DM

HIMU

DM

HIMU

Fig. 6. Zr/Nb vs Zr diagram for whole rock from peridotites and pyroxenites. The Zr/Nbratio of peridotites approaches slightly those of a HIMU-type source whereas those DMare more similar to a HIMU source.Reference fields: DM source from Sun and McDonough (1989), Hofmann (1988), Are-valo and McDonough (2010) while HIMU source from Chauvel et al. (1992).

75A. Correale et al. / Lithos 132-133 (2012) 70–81

area (Sapienza et al., 2005; Tonarini et al., 1996). Fluid inclusions alsoretain noble gases, which can be used as powerful tracers of the man-tle source.

The concentrations of noble gases in the studied xenoliths arelisted in Table 3. The He content varies from 7.3×10−14 to2.6×10−11 mol/g in mineral separates of peridotite nodules, andfrom 5.1×10−13 to 3.4×10−11 mol/g in those of pyroxenite (Fig. 8).The He abundance in peridotites overlapped the range of datareported for Hyblean samples by Sapienza et al. (2005). The Ar con-centration was measured for the first time in both peridotites and py-roxenites, varying from 1.89×10−13 to 6.64×10−12 mol/g and from8.2×10−13 to 3.2×10−11 mol/g, respectively. Samples that are richerin He are generally also richer in Ar; this behavior was observed in allof the investigated samples, although the He/Ar ratios did differ be-tween the samples.

The He and Ar concentrations differ systematically among thecogenetic minerals (olivines, orthopyroxenes, and clinopyroxenes)of each sample (see Table 3), with them being slightly higher in clin-opyroxenes and orthopyroxenes than in olivines. The 3He/4He ratioswere also the highest in clinopyroxenes and orthopyroxenes. Thispartially agrees with the findings of Sapienza et al. (2005), whoreported generally comparable values in olivine and orthopyroxenebut lower values than in cogenetic clinopyroxenes.

It is well known that fluid inclusions can be contaminated by air.In order to evaluate the air contribution in our samples, we plotted4He/20Ne versus 40Ar/36Ar ratios, as shown in Fig. 9. The 4He/20Neratio varies between 61 and 4740, while that of 40Ar/36Ar varies

0.5127

0.5128

0.5129

0.5130

0.5131

0.5132

0.7026 0.7027 0.7028 0.7029 0.7030 0.7031 0.7032 0.703387Sr/86Sr

143 N

d/14

4 Nd

peridotites this work pyroxenites this work peridotites Bianchini et al. (2010); Tonarini et al. (1996) pyroxenites Bianchini et al. (2010); Tonarini et al. (1996)

HIMU

DM

HIMU

Fig. 7. Sr–Nd isotopic composition of Hyblean xenoliths from this work and from liter-ature (Bianchini et al., 2010; Tonarini et al., 1996). Nd isotopes of peridotites fall fully inthe range of HIMU source, while higher ratio in pyroxenites may testify a certain ten-dency toward a DM end member.The DM and HIMU values are from Zindler and Hart (1986).

Table 3Noble gases analyses of mineral phases from peridotitic and pyroxenitic Hyblean xenoliths.

Sample Mineralphase

Weight 4He 10−13 20Ne 10−15 40Ar 10−12 4He/20Ne 3He/4He Err +/− 40Ar/36Ar Err 4He/40Ar*(g) (mol/g) (mol/g) (mol/g) (R/Ra) (%)

PeridotitesXIH-1 Ol 3.0 1.2 5.6 1.4 211.9 7.0 0.19 314.8 0.05 1.3

1.7 0.7 0.4 1.5 180.2 7.0 0.25 312.2 0.05 0.9XIH-1 Opx 2.2 4.2 0.7 2.2 637.9 7.0 0.11 342.7 0.06 1.4

1.4 3.4 0.5 2.0 685.3 7.0 0.12 357.1 0.07 1.0XIH-1 Cpx 1.6 4.2 0.3 1.9 1236.5 7.0 0.13 445.7 0.07 0.7

1.0 4.3 0.4 1.3 1034.8 7.0 0.10 414.9 0.06 1.1XIH-2 Ol 3.2 2.1 0.3 1.1 596.6 7.1 0.11 323.4 0.03 2.2

3.3 2.1 0.3 0.2 762.5 7.1 0.08 513.7 0.02 2.6XIH-2 Opx 2.0 10.0 2.8 1.1 359.2 7.2 0.06 479.6 0.04 2.3

1.6 5.9 6.5 1.0 90.7 7.3 0.07 461.8 0.05 1.71.7 11.4 3.7 1.9 310.3 7.2 0.06 611.1 0.02 1.10.7 10.6 2.2 3.6 489.2 7.2 0.07 804.8 0.27 0.5

XIH-3 Ol 2.0 66.6 3.3 2.7 2033.0 7.2 0.06 426.8 0.03 8.21.9 45.4 2.2 2.3 2036.1 7.2 0.06 414.1 0.07 6.9

XIH-3 Opx 2.1 168.6 7.5 5.7 2245.8 7.2 0.06 502.1 0.05 7.21.2 102.7 4.2 4.1 2453.8 7.2 0.06 553.5 0.06 5.3

XIH-3 Cpx 1.1 263.0 5.5 6.6 4739.9 7.4 0.05 716.4 0.04 6.70.7 188.5 5.1 4.3 3668.6 7.4 0.05 641.8 0.05 8.1

PyroxenitesXIH-4 Cpx 0.8 302.1 35.4 25.6 853.9 7.3 0.06 620.1 0.10 2.3

0.6 337.3 14.7 13.6 2294.1 7.3 0.05 900.9 0.04 3.7XIH-14 Cpx 0.5 13.9 22.7 2.4 61.4 7.6 0.10 315.9 0.03 8.8

0.5 30.4 8.4 1.6 360.8 7.6 0.10 338.3 0.07 15.40.4 11.5 6.3 2.7 183.0 7.4 0.11 314.9 0.42 6.80.6 10.5 3.9 0.8 269.7 7.2 0.10 342.7 0.07 9.00.6 5.1 0.3 0.8 1833.5 7.1 0.13 550.4 0.04 1.3

XIC-26 Cpx 0.5 116.2 107.9 32.3 107.7 7.4 0.10 323.3 0.07 4.20.2 83.6 6.2 6.9 1342.8 7.3 0.05 649.2 0.27 2.2

XIP-28 Cpx 0.5 170.4 4.7 3.7 3660.9 7.4 0.06 554.2 0.04 9.80.3 136.2 5.3 3.9 2580.4 7.4 0.06 477.1 0.04 9.1

XIH-17 Cpx 0.3 45.8 9.3 4.5 490.3 7.4 0.08 431.4 0.04 3.20.5 88.8 9.8 20.7 904.1 7.4 0.05 938.6 0.02 0.60.5 81.4 13.0 23.9 624.8 7.3 0.05 556.6 0.04 0.7

Air 0.3 295.5

76 A. Correale et al. / Lithos 132-133 (2012) 70–81

between 316 and 939. All of the samples plot close to a computedcurve of the binary mixing between an atmospheric term and a hypo-thetical MORB source (Graham, 2002; Marty et al., 1983), therebyconfirming atmospheric contamination of the gases released fromfluid inclusions (Fig. 9; see caption for further details). It is notewor-thy that the highest 4He/20Ne and 40Ar/36Ar ratios (indicating oursamples with the lowest air contamination) were generally measuredin samples with the highest gas contents released from fluid inclu-sions. The most likely causes of the atmospheric signature are (1)air contamination in the mantle due to subduction of atmosphericcomponents (e.g., Sarda, 2004) and (2) air entrapment in microcracksof minerals during or after the eruptive activity (e.g., Nuccio et al.,2008). Regarding the first cause, we recall that gas emissions at

6.5

7.0

7.5

8.0

1.E-14 1.E-13 1.E-12 1.E-11 1.E-10

He (mol/g)

3 He/

4 He(

R/R

a)

Fig. 8. 3He/4He (expressed as R/Ra) ratio vs He concentration in the investigatedsamples.

Mofeta dei Palici – which is located in the northern Hyblean areaand close to Quaternary volcanic systems – showed 40Ar/36Ar valuesin the range 1600–2000, which are consistent with a mixing betweenair and a MORB mantle (Nakai et al., 1997; INGV-PA database). Theseratios are much higher than those measured in our fluid inclusions.Given that the Hyblean mantle surely has 40Ar/36Ar ratios above2000, the low 40Ar/36Ar ratios of fluid inclusions cannot be inheritedfrom themantle but instead are probably caused by air contaminationthat occurs at shallow levels or after the entrapment of fluid inclu-sions. Similar conclusions have been previously drawn by fluid inclu-sions studies from other areas (Graham, 2002; Martelli et al., 2011;Nuccio et al., 2008; Porcelli and Ballentine, 2002). At the present

0.1

1

10

100

1000

10000

250 350 450 550 650 750 850 950

40Ar/36Ar

4 He/

20N

e

peridotites

pyroxenites AIR

MORB

Fig. 9. 4He/20Ne vs 40Ar/36Ar ratios of fluid inclusions from Hyblean xenoliths. Thecurve defines a mixing trend between two end-members: 1) MORB, having 4He/20Ne~10,000 (Marty et al., 1983), 40Ar/36Ar ~40,000 (Graham, 2002); 2) Air, having 4He/20Ne=0.318, 40Ar/36Ar=295.5.

77A. Correale et al. / Lithos 132-133 (2012) 70–81

state of knowledge, we therefore believe that air componentsentrapped in microcracks of minerals during or after their eruptionprovide the most likely explanation of air contamination in our fluidinclusions (Ballentine and Barfod, 2000).

We tested if air contamination could affect the 3He/4He ratios bymeans of the formula of Giggenbach et al. (1993) that uses the 4He/20Ne ratio of the atmospheric end-member to evaluate the degree ofcontamination of a sample. The results demonstrate that these cor-rections have practically negligible effects. The He contents in fluidinclusions were in fact practically unmodified by air contaminationdue to the low concentration of He (5.2 ppm) in air. For the same rea-son, even when considering a fractionated air (e.g. air saturated wateror air saturated sea water) as a contaminant, the 3He/4He valueswould remain unchanged after correction. The 40Ar concentration influid inclusions was corrected by assuming that all of the 36Ar foundin the samples was of atmospheric origin, according to the followingreported equation:

40Ar � ¼40Armeasured−40Ar=36Ar

� �air�36ArmeasuredÞ

�

where 40Ar* represents the corrected 40Ar. This equation allowed usto also compute a corrected-for-air 4He/40Ar ratio, hereafter referredas 4He/40Ar* (see Table 3).

The 3He/4He values vary between 7.0 and 7.4 Ra in the peridotites,in accordance with those observed by Sapienza et al. (2005), whilethey vary between 7.2 and 7.6 Ra in the pyroxenites. The 4He/40Ar*ratios range between 0.4 and 8 in the peridotites and between 0.6and 15 in the pyroxenites, indicating partial overlap in the values.The 3He/4He ratios are plotted versus the 4He/40Ar* ratios in Fig. 10.In general, the average 3He/4He and 4He/40Ar* ratios were lower inperidotite samples than in pyroxenite samples.

7. Discussion

7.1. Noble gases as geochemical tracers of mantle processes

In order to account for the observed variations of 3He/4He and4He/40Ar* ratios, we need to consider the main processes, both post-eruptive and mantle-related, that can affect the noble-gas signature.

7.1.1. Post-eruptive processesPost-eruptive processes that could affect the variability of 3He/4He

ratios in fluid inclusions are cosmogenic 3He production and

6.6

6.8

7.0

7.2

7.4

7.6

7.8

8.0

0.1 1.0

4He/

3 He/

4 He

(R/R

a)

peridotites this workpyroxenites this workmixing curve

Magma degassing

HIMU

Fig. 10. Plot of 3He/4He (R/Ra) corrected for air contamination vs 4He/40Ar* ratios of peridotimetasomatic processes, between a DM and a HIMU end-member. The dark thick arrow indicinformation about DM (light gray area) and HIMU (deep gray area) sources. The meaning o

radiogenic 4He. Samples were collected from surfaces in rapid ero-sions, and the XIH1 and XIH3 samples were from road cuts, whichshould have made massive ingrowth of cosmogenic helium highlyunlikely. In principle, the crushing procedure should release onlygas retained in the bubbles and not matrix-sited components suchas post-eruptive 3He and 4He, further preventing both cosmogenicand radiogenic contributions implanted in the crystal matrix. Also,our data on peridotites display 3He/4He values (7.0–7.4 Ra) that over-lap those of Sapienza et al. (2005; 7.0–7.6 Ra) for similar samples, de-spite us using a single-step crushing while Sapienza et al. (2005) usedprolonged crushing (strokes for 2.5 min). It is reasonable to assumethat if the samples were rich in a post-magmatic component thatcould be released by crushing, very different crushing techniquesshould give different results.

In addition, considering that the two principal parameters thatcontrol the post-eruptive production of 3He and 4He (age of the sam-ple and exposure at the surface) are similar for pyroxenites and peri-dotites, and that for the same He concentration the 3He/4He ratio is inmost cases higher for pyroxenites than for peridotites (Fig. 8), we at-tribute this isotopic difference to genetic processes rather than topost-eruptive processes. Therefore, even if we cannot definitely ex-clude slightly alteration of the original 3He/4He ratio of individualsamples by post-eruptive processes, the mean difference between pe-ridotites and pyroxenites should be largely attributable to mantleprocesses.

7.1.2. Mantle processesIn their study of worldwide mantle xenoliths, Yamamoto et al.

(2009) observed that 4He/40Ar* decreased from a typical mantlevalue of 1–4 (Graham, 2002; Ozima and Podosek, 1983) down to0.1, paralleled by a decrease in 3He/4He from 7 to 3 Ra. They attribut-ed this to kinetic fractionation among noble-gas atomic species due totheir diffusion through the mineral assembly of mantle, towardsmagma channels crossing the mantle itself. Because of their high in-compatibilities, noble gases would be preferentially partitioned inthe magma relative to mantle minerals so as to diffuse from mantleto magma flowing through the channels. Under such conditions, thedifferent diffusivities of 3He, 4He, and 40Ar would induce a kineticfractionation of these isotopes, and so the 3He/4He and 4He/40Ar* ra-tios would decrease in the mantle source. Following the approach ofYamamoto et al. (2009), we calculated how the noble-gas ratiosshould vary as a result of the diffusive fractionation (see Yamamotoet al., 2009, for further details on boundary conditions). The pro-cess produces a dramatic decrease in 3He/4He compared to that

10.0 100.0

40Ar*

diffusive fractionation

DM

tes and pyroxenites. The dashed curves result from a mixing, which is a consequence ofates the variations of 4He/40Ar* ratio during degassing processes. See the text for moref the diffusive fractionation curve (thin curve) is exposed in the text (Section 7.1.2).

6.9

7.0

7.1

7.2

7.3

7.4

7.5

Sm (ppm)

Peridotites Pyroxenites

20%40%

60% 80%

6.9

7.0

7.1

7.2

7.3

7.4

7.5

0 0.5 1 1.5 2 2.5 3 3.5 4

0 2 4 6 8 10 12 14

Nd (ppm)

Peridotites Pyroxenites

20%40%

60% 80%

3 He/

4 He

(R/R

a)3 H

e/4 H

e (R

/Ra)

a

b

Fig. 12. Plot of 3He/4He (R/Ra) vs Sm (a) and Nd (b) concentrations of whole rock fromHyblean peridotites and pyroxenites. The curves describe a mixing trend between twohypothetical end-members associated respectively with peridotitic and pyroxeniticsources. Ticks indicate percentages of the pyroxenitic end-member. The He concentra-tions of the two end-members, used to calculate the mixing path, were the highestmeasured in pyroxenites and the lowest in peridotites. Accordingly, the ratio betweenthe He contents of these two end-members (which determines the convexity of thecurve) is ~460.

0.1

1.0

10.0

100.0

0 200 400 600 800 1000 1200

Pressure (MPa)

4 He/

40A

r

vesicles

melt

Fig. 11. Variations of 4He/40Ar* ratio in melt (gray curves) and vesicles (black curve)during closed (dashed curve) and open (continuous curve) system degassing due tothe decompressive ascent of magma. The shaded areas show the variation range ofthe 4He/40Ar* ratio for the vesicles during a hypothetical magmatic depressurizationfrom ~900 to 700 MPa (see the text for details). The equilibrium degassing modeland volatile solubilities were the same as reported by Paonita and Martelli (2007) fora basalt melt at 1200 °C. The initial conditions were H2O and CO2 contents of 0.3 wt.%and 1 wt.%, respectively, and an initial He/Ar* ratio of 3, in accordance with a pristinebasaltic magma from the upper mantle (Paonita and Martelli, 2007).

78 A. Correale et al. / Lithos 132-133 (2012) 70–81

of 4He/40Ar*, showing a conflicting behavior with respect to the trendobserved in the data set (Fig. 10).

In order to explain the 4He/40Ar* variations observed in our data-set we focused on the process of magma degassing, given that the dif-ferent solubility of noble gases in silicate melts could generate largechanges of 4He/40Ar* (e.g., Burnard, 2004; Moreira and Sarda, 2000;Paonita and Martelli, 2006, 2007, and references therein). Indeed asmagma ascends throughout the mantle as a consequence of depres-surization, the noble gases leave the magma in proportion to theirsolubilities. Specifically, due to the solubility of He being higherthan that of Ar (e.g., Iacono-Marziano et al., 2010; Nuccio andPaonita, 2000), the degassing process would increase the 4He/40Ar*ratio of the residual magma. Based on the equilibrium degassing cal-culations of Paonita and Martelli (2007) for a typical basalt comingfrom the upper mantle and exsolving CO2-dominated fluids, the firstvesicles that separate from a melt with 4He/40Ar≈3 have a 4He/40Ar ratio of ~0.2 (see Fig. 11). This ratio would increase as themagma ascends. Fluid inclusions entrapped at different pressures inthe forming minerals can hence record 4He/40Ar* ratios reflectingvariable extents of degassing. By assuming an open-system degassing,a hypothetical pressure decrease from ~900 to 700 MPa – which iswithin the range of the expected depths of the investigated products(27–35 km; Sapienza et al., 2005) – is readily able to explain the ob-served 4He/40Ar* variations (Fig. 11).

In accordance with Paonita and Martelli (2007), the kinetic frac-tionation of 3He and 4He during magma degassing of CO2-dominatedfluids (like our fluid inclusions) can be excluded, and hence a furtherprocess must occur in conjunction with degassing to explain the mea-sured 3He/4He variations. Starting from the mentioned petrologic andgeochemical evidences, we propose that the He-isotope variability re-sults from a mixing between heterogeneous mantle sources, wheretwo local end-members having different 3He/4He values can be iden-tified: (1) the peridotite domain having 3He/4He≈7 Ra and (2) thepyroxenite domain with 3He/4He≈7.6 Ra.

In this regard it is important to note that the values of peridotitesare closer to those of a HIMU-type mantle source (3He/4He=5–7 Ra;Hanyu and Kaneoka, 1998; Moreira and Kurz, 2001), while those ofpyroxenites approach those of a DM-type source (3He/4He=8±1Ra; Allègre et al., 1995). Accordingly, the deep high-3He/4He pyroxe-nite melt decompresses during ascent, reaching the low-3He/4He

peridotitic levels. Open-system degassing processes can easily in-crease the 4He/40Ar ratios of the pyroxenite from values starting ataround 0.4, such as for an early vapor separated from melt havingtypical mantle ratio (see above), up to 15 or even more (Fig. 11).The subsequent mixing process would occur between the high-3He/4He pyroxenite having variable 4He/40Ar ratios and a low-3He/4Heperidotite mantle with 4He/40Ar=0.4–1.0 (Fig. 10). This range,which is slightly lower than the typical mantle ratios of 1–4 (Ozimaand Podosek, 1983), could result from the extraction of liquids pro-duced by partial melting of the primordial peridotite. In fact, due tothe incompatibility being higher for He than for Ar(DHe=1.17×10−4 versus DAr=1.10×10−3; Heber et al., 2007),~1% of melting would account for the required decrease in He/Ar inthe residual peridotite. Fig. 10 sketches the process using a set oftwo end-member mixing curves between a fixed term and otherend-members with different 4He/40Ar ratios. The grid clearly showsthat the described process may easily explain the complete data set.

In a previous study of Hyblean peridotitic xenoliths, Sapienza et al.(2005) had already observed 3He/4He values in the range 7.3±0.3 Ra,although they did not investigate pyroxenitic xenoliths, and they pro-posed a deep metasomatizing source. Based on our results, theHyblean pyroxenites would therefore represent a metasomatizingagent (having 3He/4He≈7.6 Ra) that is located in a deeper portionof the local mantle, while the peridotites would correspond to a shal-lower layer (having 3He/4He≈7 Ra) that is occasionally crossed bypyroxenite melts ascending from depth (Perinelli et al., 2008;Sapienza and Scribano, 2000; Scribano et al., 2008). This would resultin the shallower portion of the Hyblean mantle being partially or to-tally refertilized by such metasomatizing melts.

It is noteworthy that the measured 3He/4He values of the Hybleanxenoliths (7.0–7.6 Ra) were slightly but distinguishably higher than

79A. Correale et al. / Lithos 132-133 (2012) 70–81

those measured in all of the other mantle xenoliths from Europe andNorth Africa (5.6–7.0 Ra; Beccaluva et al., 2007, 2008; Dunai and Baur,1995; Gautheron et al., 2005; Martelli et al., 2011). Also, such valuesare the highest measured in recent basaltic lavas of Italian volcanism(Martelli et al., 2008, and references therein; Marty et al., 1994), withonly free gases from Pantelleria Island reaching similar values(Parello et al., 2000).

7.2. Evidence of mantle metasomatism by coupling noble-gas and trace-element data

The geochemical investigations of Hyblean xenoliths carried out inthis study have suggested the existence of a vertical heterogeneity ofHyblean mantle, with it being characterized by a pyroxenitic deeplayer and a peridotitic shallow portion that occasionally is partiallyor totally metasomatized. We subsequently therefore investigatedthe quantitative relationships between trace elements and 3He/4Hedata.

Fig. 12 shows average 3He/4He versus Sm and Nd concentrationsof single nodules. Pyroxenites and peridotites defined two distinctend-members of Hyblean mantle, where the highest 3He/4He datawere found in the pyroxenites that also showed the highest Sm andNd concentrations, while the same correspondence at the lowestvalues was observed in the peridotites. The two mantle layers, char-acterized by extreme Sm, Nd, and 3He/4He values of the completedata set, would be (1) the peridotitic one, with an average 3He/4He≈7 Ra, Sm≈0.25 ppm, and Nd≈0.1 ppm, and (2) the pyroxenit-ic one, characterized by an average 3He/4He≈7.4 Ra, Sm≈3.5 ppm,and Nd≈13 ppm.

Different degrees of metasomatism were modeled by a hypothet-ical mixing between the two end-members, as already observed forthe noble gases and the REE data independently. Fig. 12 shows thatthe strongly convex shape of the resulting mixing curves is in goodagreement with the data, suggesting that the metasomatic processcontrols both trace-element and noble-gas geochemistry. WhereasSm and Nd mix linearly, the higher 3He/4He ratios in pyroxenites,coupled to the high He content, makes He extremely sensitive tometasomatic events. In fact, a very low contribution of fluids from py-roxenite dramatically changes the isotope ratio of peridotite mantle.The scarcity of both He and incompatible trace elements accountsfor a strongly depleted character of Hyblean peridotites, which prob-ably also suffered extensive degassing during melt extraction.

The He-isotope ratio gives some indications about the widely de-bated genesis of pyroxenites on a worldwide scale, with the twomain groups of interpretations being (e.g., Bodinier and Godard,2003; Downes, 2007) (1) crystal precipitation from deep-mantlemagmas in conduits passing through the lithosphere and (2) recy-cling and recrystallization of subducted components belonging toold oceanic crust in convecting mantle. Our data indicate the highHe-isotope ratio of Hyblean pyroxenites, which is close to the meanvalue of MORBs, intended as samples of the convecting upper mantle(Graham, 2002). Our data are thus consistent with the Hyblean py-roxenites originating from the first of the two hypothesized mecha-nisms, while processes connected to recycled components bysubducted crust should involve a lower He-isotope ratio due totime-integrated 4He production from U and Th radiogenic decay.

8. Inferences from the heterogeneity of the Hyblean mantle

The geochemistry of Zr/Nb, Sr and Nd isotopes, and noble gases inour mantle xenoliths highlights that Hyblean peridotites and pyroxe-nites resemble well-known HIMU and DM mantle sources, respec-tively. Furthermore, the petrographic evidence (peridotitic nodulesveined from pyroxenites) account for a deeper origin for the pyroxe-nites, so that the mantle layer having HIMU characteristics would beshallower than the DM-type one. Such inferences raise two main

questions: (1) can a recycled component be associated with a perido-tite mantle and a DM be associated with a pyroxenite mantle, and (2)what is the meaning of their peculiar vertical stratification?

With regard to the first question, Sobolev et al. (2008) showedthat the enriched component is frequently linked to reaction pyroxe-nite, whereas the depleted component is likely to be derived from aperidotitic source, which would contradict our results. However, bystudying mantle xenoliths from the Canary Islands, Gurenko et al.(2009) suggested that a recycledmantle component is not necessarilylinked to an eclogite–pyroxenite paragenesis. In fact, it could alsoexist in the form of peridotite when an old (>1 Ga) recycled compo-nent had sufficient time to be stirred back into the peridotite matrix.The results of our study are thus consistent with the second hypoth-esis, and suggest that the mineralogical assemblage of recycled man-tle component can range from pyroxenites, “hybrid melts”, up toperidotites.

Concerning the inferred vertical stratification, our results contrastwith the view that the Euro-Mediterranean HIMU is generated by up-welling of a deep plume (e.g., Hoernle et al., 1995). In fact, when link-ing the HIMU signature to a mantle plume of recycled material, wewould expect the HIMU to originate from deeper than the MORB.However, different models support our inferred location of HIMUbeing shallower than MORB.

Scribano et al. (2008) assumed the presence of a serpentinite-hosted hydrothermal system in the Hyblean lithosphere as a resultof tectonic uncovering and seafloor exposure of the uppermost man-tle since middle Triassic. This hypothesis could account for the HIMUmarker of peridotites through the hydrothermal addition of U to al-tered rocks, especially by serpentine formation (Michard andAlbarède, 1985). The homogeneous isotope marker of He in olivines,orthopyroxenes, and clinopyroxenes of the peridotite sampleswould support the radiogenic production of 4He from serpentineveins and its uniform diffusion toward the three mineralogicalphases. Since the serpentine veins were mainly found in microcrackswithin olivine or along grain boundaries, we expect the 3He/4He ra-tios to be more radiogenic in olivines than in pyroxenes. This wouldalso require migration of He from the lattice into fluid inclusions, al-though we showed that this process seems to be of minor importance(see Section 7.1.1). The more primitive He-isotope composition of py-roxenites would also imply either that the latter ones infiltrated theperidotite matrix subsequent to the hydrothermal circulation causingthe HIMU signature or that the analyzed pyroxenites sampled por-tions of pyroxenite veins that were physically distant from the con-tact area with the surrounding peridotite, so as to avoid the maininteraction with serpentinizing fluids.

Indeed, the most striking feature of the investigated HIMU signa-ture is its widespread occurrence in several Euro-Mediterraneanareas (Cebria and Wilson, 1995; Macera et al., 2003; Wilson andBianchini, 1999; Wilson and Downes, 1992), so that any explanationshould preferentially involve a regional scale. Lustrino and Wilson(2007) summarized some of the different models that support thepresence of a Euro-Mediterranean HIMU mantle without invoking amantle plume actively upwelling from a thermal boundary layer atthe core–mantle boundary. The signature of recycled componentmay be simply inherited by the shallow upper mantle in the form ofa metasomatized lithosphere or enriched asthenosphere, as sug-gested by the SUMA model (statistical upper mantle assemblage)(Meibom and Anderson, 2004). Based on this, Piromallo et al.(2008) explained the common HIMU-like character of erupted lavasin different tectonic environments of the Euro-Mediterranean regionby proposing a sublithospheric dragging of the plume head locatedin the Canary–Cape Verde zone as a result of a north-to-eastward mi-gration of the Eurasian and African plates away from the hot spot.This would have allowed a spreading of plumematerial in the shallowsublithospheric mantle so as to produce a geochemically anomalous(HIMU-like) level located above the depleted upper mantle. We

80 A. Correale et al. / Lithos 132-133 (2012) 70–81

therefore conclude that the somewhat anomalous mantle stratifica-tion that we have suggested here for Hyblean mantle can beexplained within the framework of the above model.

9. Conclusions

A comprehensive geochemical study of mantle-derived ultramaficxenoliths hosted in some Hyblean volcanic systems was performed tobetter characterize the lithospheric mantle below this area, therebyexpanding the existing knowledge. The investigated nodules consist ofperidotites and pyroxenites entrapped in someMiocene-age diatremes.

New data of major and trace elements coupled to Sr and Nd iso-topes are reported here and compared to those available from similarstudies. The present study integrates these geochemical data with aninvestigation of noble gases of fluid inclusions hosted in the same xe-noliths paragenesis.

The obtained data led us to the following conclusions:

1) The mantle below the Hyblean area is heterogeneous, featuring ashallower peridotitic layer with more evolved geochemical char-acteristics (3He/4He≈7 Ra, 143Nd/144Nd≈0.5129, and Zr/Nb≈4) relative to a deeper pyroxenite domain that shows a prim-itive character (3He/4He≈7.6 Ra, 143Nd/144Nd≈0.5130, and Zr/Nb≈30). Peridotites and pyroxenites seem to display HIMU andDM affinities, respectively.

2) Metasomatic processes occur in the lithosphere below this area.Particularly, deep pyroxenite melts were identified as a metaso-matizing agent. By ascending toward the surface, they intrudethe peridotite mantle at different levels by partially or totallyrefertilizing it.

3) The metasomatic processes control both trace-element and noble-gas geochemistry. In contrast, previous studies of mantle xenoliths(i.e., Matsumoto et al., 2000, and references therein) found adecoupling between noble gases in fluid inclusions and radiogenicisotopes and trace elements in the whole rock can very often beobserved in the same samples. The present study has revealedthat different geochemical tracers can display very different sensi-tivities to the effects of metasomatic mixing between two end-members, and hence this process should be carefully consideredwhen formulating hypotheses of the processes underlying decou-pling between noble-gas and trace-element geochemistry.

Acknowledgments

We thank Ilenia Arienzo for performing Sr and Nd isotope analysesat INGV-Osservatorio Vesuviano and Mariano Tantillo for help innoble-gas analyses. Sivio Rotolo is also thanked for assistance duringSEM-EDS analyses. This work is part of the PhD thesis of A.C. financiallysupported by the Università di Palermo. Comments by two anonymousreviewers and Editor-in-Chief A. Kerr greatly improved the article.

References

Allègre, C.J., Moreira, M., Staudacher, T., 1995. 4He/3He dispersion and mantle convec-tion. Geophysical Research Letters 22, 2325–2328.

Amore, C., Carveni, P., Scribano, V., Sturiale, C., 1988. Facies ed età del vulcanismo nellafascia sudorientale della Sicilia (Pachino-Capo Passero). Bollettino della SocietaGeologica Italiana 107, 3–4.

Anders, E., Grevesse, N.N., 1989. Abundances of the elements: meteoritic and solar.Geochimica et Cosmochimica Acta 53, 197–214.

Arevalo, R., McDonough, W.F., 2010. Chemical variations and regional diversity ob-served in MORB. Chemical Geology 271, 70–85.

Arienzo, I., Civetta, L., Heumann, A., Wörner, G., Orsi, G., 2009. Isotopic evidence foropen system processes within the Campanian Ignimbrite (Campi Flegrei Italy)magma chamber. Bulletin of Volcanology 71, 285–300.

Atzori, P., Mazzoleni, P., Punturo, R., Scribano, V., 1999. Garnet-spinel pyroxenite xeno-liths from Hyblean plateau (South-eastern Sicily, Italy). Mineralogy and Petrology66, 215–226.

Ballentine, C.J., Barfod, D.N., 2000. The origin of air-like noble gases in MORB and OIB.Earth and Planetary Science Letters 180, 39–48.

Barberi, F., Civetta, L., Gasparini, P., Innocenti, F., Scandone, R., Villari, L., 1974. Evolutionof a section of the Africa–Europa plate boundary: paleomagnetic and volcanologicevidence from Sicily. Earth and Planetary Science Letters 22, 123–132.

Beccaluva, L., Siena, F., Coltorti, M., Digrande, A., Lo Giudice, A., Macciotta, G., Tassinari,R., Vaccaro, C., 1998. Nephelinitic to tholeiitic magma generation in a transten-sional tectonic setting: an integrated model for the Iblean vulcanism, Sicily. Journalof Petrology 39, 1547–1576.

Beccaluva, L., Coltorti, M., Giunta, G., Siena, F., 2004. Tethyan vs. Cordilleran ophiolites:a reappraisal of distinctive tectono-magmatic features of supra-subduction com-plexes in relation to the subduction mode. Tectonophysics 393, 163–174.

Beccaluva, L., Azzouni-Sekkal, A., Benhallou, A., Bianchini, G., Ellam, R.M., Marzola, M.,Siena, F., Stuart, F.M., 2007. Intracratonic asthenosphere upwelling and lithosphererejuvenation beneath the Hoggar swell (Algeria): evidence from HIMU metasoma-tised lherzolite mantle xenoliths. Earth and Planetary Science Letters 260, 482–494.

Beccaluva, L., Bianchini, G., Ellam, R.M., Marzola, M., Oun, K.M., Siena, F., Stuart, F.M., 2008.The role of HIMU metasomatic components in the African lithospheric mantle:petrological evidence from the Gharyan peridotite xenoliths, NW Libya. In: Coltorti,M., Gregoire, M. (Eds.), Mantle Metasomatism in Intra-plate and Supra SubductionSettings: London Geological Society, Special Publication, 293, pp. 253–277.

Berry, M.J., Knopoff, L., 1967. The structure of the Upper Mantle under the WesternMediterranean Basin. Journal of Geophysical Research 72, 3613–3627.

Bianchi, F., Carbone, S., Grasso, M., Invernizzi, G., Lentini, F., Longaretti, G., Merlini, S.,Mostardini, F., 1987. Sicilia orientale: Profilo geologico Nebrodi-Iblei. Memoriedella Società Geologica Italiana 38, 429–458.

Bianchini, G., Yoshikawa, M., Sapienza, G.T., 2010. Comparative study of ultramafic xe-noliths and associated lavas from South-Eastern Sicily: nature of the lithosphericmantle and insights on magma genesis. Mineralogy and Petrology 98, 111–121.

Bodinier, J.L., Godard, M., 2003. Orogenic, ophiolitic, and abyssal peridotites. Treatise onGeochemistry, Part 2. Elsevier Ltd, pp. 103–170.

Burnard, P., 2004. Diffusive fractionation of noble gases and helium isotopes duringmantle melting. Earth and Planetary Science Letters 220, 287–295.

Burollet, F.P., Mugniot, J.M., Sweeney, P., 1978. The geology of the Pelagian Block: themargins and basins of Southern Tunisia and Tripolania. In: Nairm, A.E.M., Kanes,W.H., Stehli, F.G. (Eds.), The Ocean Basins and Margins. : The Western Mediterra-nean, vol. 4b. Plenum Press, NY, pp. 331–359.

Calcagnile, G., D'Ingeo, F., Farrugia, P., Panza, G.F., 1982. The lithosphere in the Central-Eastern Mediterranean area. Pure and Applied Geophysics 120, 389–406.

Carbone, S., Lentini, F., 1981. Caratteri deposizionali delle vulcanite del Miocene super-iore negli Iblei (Sicilia sud-orientale). Geologica Romana 20, 79–101.

Cebria, J.M., Wilson, M., 1995. Cenozoic mafic magmatism in Western/Central Europe:a common European asthenospheric reservoir. Terra Nova 7, 162.

Chauvel, C., Hofmann, A.W., Vidal, Ph., 1992. HIMU-EM, the French Polynesian connec-tion. Earth and Planetary Science Letters 110, 99–119.

Cristofolini, R., 1966. Le manifestazioni eruttive basiche del Trias superiore nel sotto-suolo di Ragusa (Sicilia Sud Orientale). Periodico di Mineralogia 35, 1–28.

Downes, H., 2007. Origin and significance of spinel and garnet pyroxenites in the shal-low lithospheric mantle: ultramafic massifs in orogenic belts in Western Europeand NW Africa. Lithos 99, 1–24.

Dunai, T.J., Baur, H., 1995. Helium, neon, and argon systematics of the European sub-continental mantle: implications for its geochemical evolution. Geochimica et Cos-mochimica Acta 59, 2767–2783.

Eggins, S.M., Rudnick, R.L., McDonough, W.F., 1998. The composition of peridotites andtheir minerals: a laser-ablation ICP-MS study. Earth and Planetary Science Letters154, 53–71.

Finetti, I., Morelli, C., 1973. Geophysical exploration of the Mediterranean Sea. Bollet-tino di Geofisica Teorica ed Applicata 15, 263–340.

Gautheron, C., Moreira, M., Allègre, C., 2005. He, Ne and Ar composition of the Europe-an lithospheric mantle. Chemical Geology 217, 97–112.

Giggenbach, W.F., Sano, Y., Wakita, H., 1993. Isotopic composition of helium, and CO2

and CH4 contents in gas produced along the New Zealand part of a convergentplate boundary. Geochimica et Cosmochimica Acta 57, 3427–3455.

Graham, D.W., 2002. Noble gas isotope geochemistry of mid-ocean ridge and ocean is-land basalts: characterization of mantle source reservoir. In: Porcelli, D.P., Ballen-tine, C.J., Wieler, R. (Eds.), Noble Gases in Geochemistry and Cosmochemistry:Rev. Mineral. Geochem., vol. 47, pp. 247–317. Washington, D. C.,.

Gurenko, A.A., Sobolev, A.V., Hoernle, K., Hauff, F., Schmincke, H.U., 2009. Enriched,HIMU-type peridotite and depleted recycled pyroxenite in the Canary plume: amixed-up mantle. Earth and Planetary Science Letters 277, 514–524.

Hanyu, T., Kaneoka, I., 1998. Open system behavior of helium in case of the HIMUsource area. Geophysical Research Letters 25, 687–690.

Heber, S.V., Brooker, R.A., Kelley, P.S., Wood, J.B., 2007. Crystal–melt partitioning ofnoble gases (helium, neon, argon, krypton, and xenon) for olivine and clinopyrox-ene. Geochimica et Cosmochimica Acta 71, 1041–1061.

Hoernle, K., Zhang, Y.S., Graham, D., 1995. Seismic and geochemical evidence for largescale mantle upwelling beneath the eastern Atlantic and western and central Eu-rope. Nature 374, 34–39.

Hofmann, A.W., 1988. Chemical differentiation of the earth: the relationship betweenmantle, continental crust, and oceanic crust. Earth and Planetary Science Letters90, 297–314.

Iacono-Marziano, G., Paonita, A., Rizzo, A., Scaillet, B., Gaillard, F., 2010. Noble gas solu-bilities in silicate melts: new experimental results and a comprehensive model ofthe effects of liquid composition, temperature and pressure. Chemical Geology279, 145–157.

Lentini, F., Carbone, S., Catalano, S., Grasso, M., 1996. Elementi per la ricostruzione delquadro strutturale della Sicilia orientale. Memorie Società Geologica Italiana 51,179–195.

81A. Correale et al. / Lithos 132-133 (2012) 70–81

Lopez, M., Pompilio, M., Rotolo, S.G., 2006. Petrology of some amphibole-bearing volca-nics of the pre-Ellittico period (102–80 ka), Mt. Etna. Periodico di Mineralogia 75,151–166.

Lustrino, M., Wilson, M., 2007. The circum-Mediterranean anorogenic Cenozoic igne-ous province. Earth-Science Reviews 81, 1–65.

Macera, P., Gasperini, D., Piromallo, C., Blichert-Toft, J., Bosch, D., Del Moro, A., Martin,S., 2003. Geodynamic implications of deep mantle upwelling in the source of Ter-tiary volcanics from the Veneto region (South-Eastern Alps). Journal of Geody-namics 36, 563–590.

Martelli, M., Nuccio, P.M., Stuart, F.M., Di Liberto, V., Ellam, R.M., 2008. Constraints onmantle source and interactions from He–Sr isotope variation in Italian Plio-Quaternary volcanism. Geochemistry, Geophysics, Geosystems. doi:10.1029/2007GC001730.

Martelli, M., Bianchini, G., Beccaluva, L., Rizzo, A., 2011. Helium and argon isotopic com-positions of mantle xenoliths from Tallante and Calatrava, Spain. Journal of Volca-nology and Geothermal Research 185, 172–180.

Marty, B., Zashu, S., Ozima, M., 1983. Two noble gas components in a Mid-AtlanticRidge basalt. Nature 302, 238–240.

Marty, B., Trull, T., Lussiez, P., Basile, I., Tanguy, J.C., 1994. He, Ar, O, Sr and Nd isotopeconstraints on the origin and evolution of Mount Etna magmatism. Earth and Plan-etary Science Letters 126, 23–39.

Matsumoto, T., Masahiko, H., McDougall, I., O'Reilly, S.Y., Norman, M., Yaxley, G., 2000.Noble gases in pyroxenites and metasomatised peridotites from the Newer Volca-nics, southeastern Australia: implications for mantle metasomatism. Chemical Ge-ology 168, 49–73.

Meibom, A., Anderson, D.L., 2004. The statistical upper mantle assemblage. Earth andPlanetary Science Letters 217, 123–139.

Michard, A., Albarède, F., 1985. Hydrothermal uranium uptake at ridge crests. Nature317, 244–246.

Moreira, M., Kurz, M.D., 2001. Subducted oceanic lithosphere and the origin of the ‘highμ’ basalt helium isotopic signature. Earth and Planetary Science Letters 189, 49–57.

Moreira, M., Sarda, P., 2000. Noble gas constraints on degassing processes. Earth andPlanetary Science Letters 176, 375–386.

Nakai, S., Wakita, H., Nuccio, P.M., Italiano, F., 1997. MORB type neon in an enrichedmantle beneath Etna, Sicily. Earth and Planetary Science Letters 153, 57–66.

Nimis, P., 1998. Clinopyroxene geobarometrv of pyroxenitic xenoliths from Hybleanplateau (SE Sicity, Itaty). European Journal of Mineralogy 10, 521–533.

Nuccio, P.M., Paonita, A., 2000. Investigation of the noble gas solubility in H2O–CO2

bearing silicate liquids at moderate pressure II: the extended ionic porosity (EIP)model. Earth and Planetary Science Letters 183, 499–512.

Nuccio, P.M., Paonita, A., Rizzo, A., Rosciglione, A., 2008. Elemental and isotope covari-ation of noble gases in mineral phases from Etnean volcanics erupted during2001–2005, and genetic relation with peripheral gas discharges. Earth and Plane-tary Science Letters 272, 683–690.

Ozima, M., Podosek, F.A., 1983. Noble Gas Geochemistry. Cambridge University Press,New York.

Paonita, A., Martelli, M., 2006. Magma dynamics at mid-ocean ridges by noble gas ki-netic fractionation: assessment of magmatic ascent rates. Earth and Planetary Sci-ence Letters 241, 138–158.

Paonita, A., Martelli, M., 2007. A new view of the He–Ar–CO2 degassing at mid-oceanridges: homogeneous composition of magmas from the upper mantle. Geochimicaet Cosmochimica Acta 71, 1747–1763.

Parello, F., Allard, P., D'Alessandro, W., Federico, C., Jean-Baptiste, P., Catani, O., 2000.Isotope geochemistry of Pantelleria volcanic fluids, Sicily Channel rift: a mantlevolatile end-member for volcanism in southern Europe. Earth and Planetary Sci-ence Letters 180, 325–339.

Perinelli, C., Sapienza, G.T., Armienti, P., Morten, L., 2008. Metasomatism of the uppermantle beneath the Hyblean Plateau (Sicily): evidence from pyroxenes and glassin peridotite xenoliths. London Geological Society, Special Publication 293,197–221.

Piromallo, C., Gasperini, D., Macera, P., Faccenna, C., 2008. A late Cretaceous contamina-tion episode of the European–Mediterranean mantle. Earth and Planetary ScienceLetters 268, 15–27.

Porcelli, D., Ballentine, C., 2002. Models for the distribution of terrestrial noble gasesand evolution of the atmosphere. In: Porcelli, D.P., Ballentine, C.J., Wieler, R.(Eds.), Noble Gases in Geochemistry and Cosmochemistry: Review in Mineralogyand Geochemistry, vol. 47, pp. 411–480. Washington, D. C.,.

Porcelli, D.R., O'Nions, R.K., O'Reilly, S.Y., 1986. Helium and strontium isotopes in ultra-mafic xenoliths. Chemical Geology 54, 237–249.

Punturo, R., Scribano, V., 1997. Dati geochimici e petrologici preliminari su xenoliti diclinopirossenite a grana ultragrossa e websteriti nelle vulcanoclastiti miocenichedell'alta Valle Guffari (Monti Iblei, Sicilia). Mineralogica et Petrographica Acta 40,95–116.

Rocchi, S., Longaretti, G., Salvadori, M., 1998. Subsurface Mesozoic and Cenozoic mag-matism in south-eastern Sicily: distribution, volume and geochemistry of magmas.Acta Vulcanologica 10, 395–408.

Sapienza, G., Scribano, V., 2000. Distribution and representative whole rock chemistryof deep-seated xenoliths from the Iblean Plateau, South-Eastern Sicily, Italy. Peri-odico di Mineralogia 69, 185–204.

Sapienza, G., Hilton, D.R., Scribano, V., 2005. Helium isotopes in peridotite mineralphases from Hyblean Plateau xenoliths (southeastern Sicily, Italy). Chemical Geol-ogy 219, 115–129.

Sarda, P., 2004. Surface noble gas recycling to the terrestrial mantle. Earth and Plane-tary Science Letters 228, 49–63.

Scribano, V., 1987a. Origin of websterite nodules from some alkaline volcanic rocks ofHyblean Plateau (South-Eastern Sicily). Periodico di Mineralogia 56, 51–69.

Scribano, V., 1987b. The ultramafic and mafic nodule suite in a tuff-breccia pipe fromCozzo Molino (Hyblean Plateau SE Sicily). Rendiconti della Società Italiana diMineralogia e Petrologia 42, 203–217.

Scribano, V., Viccaro, M., Cristofolini, R., Ottolini, L., 2008. Metasomatic events recordedin ultramafic xenoliths from the Hyblean area (Southeastern Sicily, Italy). Mineral-ogy and Petrology 95, 232–250.

Sobolev, A.V., Hofmann, A.W., Brügmann, B., Batanova, V.G., Kuzmin, D.V., 2008. Aquantitative link between recycling and osmium isotopes. Science 321, 536.

Sun, S., McDonough, W.F., 1989. Chemical and isotopic systematics of oceanic basalts:implications for mantle composition and processes. In: Saunders, A.D., Norry, M.J.(Eds.), Magmatism in the Ocean Basins: Geological Society Special Publication,pp. 313–345.

Tonarini, S., D'orazio, M., Armienti, P., Innocenti, F., Scribano, V., 1996. Geochemical fea-tures of Eastern Sicily lithosphere as probed by Hyblean xenoliths and lavas. Euro-pean Journal of Mineralogy 5, 1153–1174.

Vai, G.B., 1994. Crustal evolution and basement elements in the Italian area: palaeogeo-graphy and characterization. Bollettino di Geofisica Teorica ed Applicata 36,411–434.

Vaselli, O., Downes, H., Thirlwall, M.F., Dobosi, G., Coradossi, N., Seghedi, I., Szakacs, A.,Vannucci, R., 1995. Ultramafic xenoliths in Plio-Pleistocene alkali basalts from theEastern Transylvanian Basin: depleted mantle enriched by vein metasomatism.Journal of Petrology 36, 23–53.

Wilson, M., Bianchini, G., 1999. Tertiary–Quaternary magmatism within the Mediterra-nean and surrounding regions. In: Durand, Jolivet, Horvath, Seranne (Eds.), TheMediterranean Basins: Tertiary Extension Within the Alpine Orogen: GeologicalSociety, London, vol. 156, pp. 141–168.

Wilson, M., Downes, H., 1992. Mafic alkaline magmatism associated with the EuropeanCenozoic Rift system. Tectonophysics 208, 173–182.

Yamamoto, J., Nishimura, K., Takeshi, S., Keiji, T., Naoto, T., Yuji, S., 2009. Diffusive frac-tionation of noble gases in mantle whit magma channels: origin of low He/Ar inmantle-derived rocks. Earth and Planetary Science Letters 280, 167–174.

Zangana, N., Downes, H., Thirlwall, M.F., Bea, F., Marriner, G.F., 1999. Geochemical var-iations in peridotite xenoliths and their constituent clinopyroxenes from Ray Pic(French Massif Central): implications for the composition of the shallow litho-spheric mantle. Chemical Geology 153, 11–35.

Zindler, A., Hart, S.R., 1986. Chemical geodynamics. Earth and Planetary Science Letters14, 493–571.