Author's personal copy

Primary origin of some trachytoid magmas: Inferences from naturally quenchedglasses in hydrothermally metasomatized gabbroic xenoliths (Hyblean area, Sicily)

Marco Viccaro a, Vittorio Scribano a,⁎, Renato Cristofolini a, Luisa Ottolini b, Fabio C. Manuella a

a Università di Catania, Dipartimento di Scienze Geologiche, Corso Italia 57, I-95129, Catania, Italyb CNR-Istituto di Geoscienze e Georisorse, Sezione di Pavia, Via A. Ferrata 1, I-27100, Pavia, Italy

a b s t r a c ta r t i c l e i n f o

Article history:Received 9 February 2009Accepted 30 June 2009Available online 30 July 2009

Keywords:SicilyXenolithHydrothermal alterationDaly-gapTrachyteSIMS

Hydrothermally-modified gabbroic xenoliths from the Hyblean tuff-breccia deposits (Sicily) consist ofalbitized plagioclase, Fe–Mg-rich clays, aegirine–augite, ±zeolites, titanite, apatite, magnetite, andhydrothermal zircon. Pockets of silicate glass with perlitic cracking occur in some samples forming 15–20% (by volume) of the rock modal assemblage. Electron microprobe analyses show the trachyticcomposition of the glass, with generally peralkaline sodic affinity [molar Al2O3/(Na2O+K2O)~0.8 (averagevalue); molar Al2O3/(Na2O+K2O+CaO)~0.7 (average value); Na2O/K2O (wt.%)=1.7–2.3]. The glass traceelement abundances, obtained by secondary ion mass spectrometry (SIMS) analyses are consistent withthose of world-wide trachytes (e.g. Zr/Ti=0.15–018; Nb/Y=0.73–1). Relatively high abundances of Cl (700–1600 ppm) and F (N500 ppm) were also detected in the glass.Careful macroscopic and microscopic observations exclude the possibility that external silicate meltinfiltrated the xenolith. The occurrence of glass pockets between the mafic clay assemblages and the feldspargrains, along with comparisons between chemical compositions of the glass and the surrounding minerals,suggest that the glass is due to the melting of a eutectoid system consisting of Na-rich alkali feldspar, Fe–Mg-rich clays and aegirine–augite. Halogens had probably played an important role in the partial melting processby decreasing the melting temperature of modal minerals, especially feldspar.The occurrence of these trachytic glasses lends support to petrologic models suggesting that partial meltingof a hydrothermally altered, brine-rich oceanic crust induced by shallow-seated basic intrusions can produceprimary trachytoid melts. This may explain the “Daly-gap” characterizing some oceanic within-platevolcanoes.

© 2009 Elsevier B.V. All rights reserved.

1. Introduction

A number of within-plate volcanoes from oceanic and continentalgeodynamic settings erupts modest amounts of alkaline-intermediateto felsic products, particularly trachytes. Since the pioneeringexperimental work by Bowen (1945), Hamilton and MacKenzie(1965), and Bailey and Schairer (1964, 1966), trachytes have beenconsidered the result of the fractional crystallization of alkaline-maficmagmas. This process has been proved to be consistent with mass–balance calculations and melt inclusion geochemistry (e.g., Tanguyet al., 1997; Fedele et al., 2003). On the other hand, in some oceanicislands, fractional crystallization fails to account for the oddvolumetric relationship between trachytes and their supposed parentbasalts, especially when taking into account the absence of inter-mediate members throughout the differentiation suite (i.e. the “Daly-

gap” introduced by Chayes, 1977 on the basis of an early geologicalreport on the Ascension Island by Daly, 1925). Apart from a fewdismissive accounts considering this compositional gap an artifact dueto limitations of subaerial sampling (Harris, 1963; Baker and MvReath, 1972) or misleading petrochemical calculations (Clague, 1978),petrologists have offered different explanations for the paucity ofintermediate differentiates. Some theories invoke complex physico-chemical perturbations in the differentiating magma reservoir whichinterrupt the liquid line of descent, leading to the formation of endmembers rather than the intermediate ones (e.g. Bonnefoi et al., 1995;White et al., 2009). Mixing between already differentiated liquids inan anomalously stratified magmatic chamber was also suggested(Ferla and Meli, 2006).

On the other hand, near-solidus differentiation due to (hydrous)partial melting ofmafic intrusive rocks has repeatedly been proposed asa possiblemechanism for generating felsic primarymelts in oceanic andcontinental geodynamic settings (e.g., Chayes, 1977; Bohrson and Reid,1995; Bohrson et al., 1996; Bohrson and Reid, 1997, 1998; Avanzinelliet al., 2004; Martin and Sigmarsson, 2007; Koepke et al., 2007).Furthermore, open-system processes, such as brine-induced partial

Lithos 113 (2009) 659–672

⁎ Corresponding author. Tel.: +390957195743; fax: +390957195760.E-mail addresses: m.viccaro@unict.it (M. Viccaro), scribano@unict.it (V. Scribano),

rcristo@unict.it (R. Cristofolini), ottolini@crystal.unipv.it (L. Ottolini), manuella@unict.it(F.C. Manuella).

0024-4937/$ – see front matter © 2009 Elsevier B.V. All rights reserved.doi:10.1016/j.lithos.2009.06.037

Contents lists available at ScienceDirect

Lithos

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

Author's personal copy

melting of hydrothermally-modified oceanic crust or assimilation ofaltered wall rocks by basaltic magmas, can account for the isotopesystematics and trace element distributions of some OI-type trachytes(e.g., Lassiter et al., 2002; Gagnevin et al., 2003; Legendre et al., 2005).

Since the first discovery (Scribano,1986) of deep-seated xenoliths inthe Hyblean area (Southeastern Sicily, Italy), investigations on themhave aimed at obtaining insights into the nature of the unexposedlithospheric basement. In addition to providing the principal informa-tion on regional geology, Hyblean xenoliths have sometimes suppliedindications regarding petrological problems of broad interest, such asmechanisms of metasomatic processes in the lithospheric mantle andlower crust, as well as their bearing on the origin of different magmacompositions (e.g. Tonarini et al., 1996; Scribano et al., 2009).Specifically, Scribano and Manuella (2007) mentioned the intriguingoccurrence of glasses with trachytic compositions found in somehydrothermally altered mafic xenoliths from Valle Guffari (VG inFig. 1). These authors briefly noted that such an occurrencemay pertainto the primary origin of some trachytoid magmas, including peralkalinetypes. The present paper attempts to substantiate this hypothesis on thegrounds of newly acquired ion microprobe data on the same glasscompared with published trachyte compositions from various oceanicislands and the neighboring sites of Pantelleria and Mt. Etna.

2. Geological setting

The Hyblean Plateau is an uplifted emerged portion of thePelagian–Ionian foreland area in southeastern Sicily, Italy (Fig. 1).This plateau is bounded to the east by the northernmost segment of a

steep and long submarine slope, the Hyblean–Malta Escarpment,which separates the Pelagian shelf from the Ionian abyssal plain(~3000 m b.s.l.). The Hyblean Plateau is down-faulted to northwest,forming the Gela–Catania foredeep, which is filled by silico-clasticsediments of the Apennine–Maghrebian thrust belt front (Butler et al.,1992). Two main fault systems cross-cut the Hyblean area. One,trending NE–SW, is extensional; the other, trending NNW–SSE, mostlyconsists of strike–slip faults (Grasso and Reuther, 1988). The exposedpart of the Hyblean stratigraphic succession consists of UpperCretaceous to Cenozoic deep-water carbonate rocks, Neogene toQuaternary open-shelf terrigenous rocks and various levels of basicvolcanic rocks (Bianchi et al., 1987). The deepest subsurface levelsinferred by deep wells drilled for hydrocarbon prospecting consist ofdolomitic limestones and basic volcanic rocks middle Triassic in age(e.g. Rocchi et al., 1998). No direct evidence on the nature of the pre-Triassic basement has been reported so far, except for xenolithsbrought to the surface by diatremic eruptions (e.g. Scribano, 1986;Scribano et al., 2006a).

Oil wells drilled in the Hyblean sedimentary succession recorded anearly continuous, dominantly effusive magmatism from MiddleTriassic to Late Cretaceous, with an estimated magma effusion rateranging from 10 to 100 km3 Ma−1. The Mesozoic subsurface volcanicrocks consist of basalts with OIB affinity (Rocchi et al., 1998). UpperCretaceous volcanic rocks crop out in the southernmost area of Sicily(Fig.1) and hence they form scattered, small outcrops along the Ioniancoast, from Siracusa to Augusta. Most of the Hyblean volcanic rockscrop out in the northeastern part of the Plateau occupying an area ofabout 350 km2. They are generally Pliocene to Pleistocene basalticvolcanics (lavas and hyaloclastites), often submarine, with both Na-alkaline and tholeiitic affinity (e.g., De Rosa et al., 1991; Tonarini et al.,1996; Schmincke et al., 1997; Trua et al., 1998). It is important to notethat the interpretation of the Sr–Nd–Pb isotopic data for the UpperMiocene and Plio-Pleistocene lavas from the Hyblean area hasexcluded the presence of continental crust contamination (Truaet al., 1998).

3. The Hyblean xenoliths and their bearing on the nature of theunexposed basement

Although lava flows fed by fissure activity represent the mainvolcanic feature of the Hyblean area, there are also some alkaline-mafic diatremes in the central-eastern part of the plateau (Carboneand Lentini, 1981). The diatremic activity occurred during the UpperMiocene (Tortonian) and bore to the surface an important suite ofdeep-seated xenoliths the study of which has greatly improved ourknowledge of the underlying, unexposed lithosphere. Specifically, thexenoliths consist of mantle-derived ultramafic rocks, minor gabbroicrocks representing the unexposed crustal basement, and varioussedimentary and volcanic rocks coming from the Meso-Cenozoicsuccession. To date, no typical continental crust rocks have been foundamong the Hyblean xenoliths.

Mantle xenoliths consist of spinel-facies, four-phase peridotites anddifferent pyroxenite types. The former are composed of spinel-harzburgite exhibiting both protogranular and porphyroclastic texture.The Mg/(Mg+Fe2+) values (Mg#) of the harzburgite olivine vary from0.90 to 0.92 (0.91 is the most common value), with NiO=0.2–0.5 wt.%.The Cr/(Cr+Al) ratio (Cr#) of spinel ranges from 0.25 to 0.45 and Mg#from0.7 to 0.8. TheMg#values of orthopyroxene vary between0.87 and0.91 with CaOb0.9 wt.%. Minor Cr-diopside exhibits a narrow range ofcompositional variations (En52Wo47Fs1–En50Wo43Fs7). Major elementcompositions in peridotites also suggest a moderate to high degree ofdepletion (whole rockAl2O3b2wt.%, TiO2b0.2wt.%). Their initial 143Nd/144Nd (0.51256–0.51277) is similar to that of the average Late Paleozoic–Early Mesozoic MORB (Tonarini et al., 1996). This time interval is thatestimated for the formation of the adjoining Ionian lithosphere (Vai,2003).

Fig. 1. Geological sketch map of the Hyblean area (after Lentini et al., 1994, mod.) andlocation of the main xenolith occurrence (VG-Valle Guffari; CM-Cozzo Molino; CA-Carlentini; VI-Vallone Iuso; SO-Sortino. Legend: a) Upper Cretaceous limestones;b) Upper Cretaceous OIB-type lava and hyaloclastite; c) Cenozoic limestones; d)Xenolith-bearing Miocenic diatremes and/or related tuff-breccia deposits; e) Plio-Pleistocenic OIB-type and E-MORB-type lavas and hyaloclastite; f) Neogene–Quaternaryopen-shelf clastics; g) Allochtonous sediments of the Maghrebian units; h) Main faults;i) Limit of the Maghrebian thrust belt.

660 M. Viccaro et al. / Lithos 113 (2009) 659–672

Author's personal copy

Different pyroxenite types represent both silicate melts trapped inthe depleted lithospheric mantle and products of metasomaticreactions (Scribano, 1987; Nimis and Vannucci, 1995; Punturo andScribano, 1998; Atzori et al., 1999). Pyroxenites are consideredeffective re-fertilizing media at the regional scale, with importantimplications for the origin of a wide spectrum of alkaline primarymagmas in the lithospheric mantle (Scribano et al., 2009).

Crustal xenoliths are composed of oxide-gabbro cumulatesexhibiting tholeiitic affinity: Zr/Nb=5–26; Nb/Y=0.008–0.7 (e.g.Winchester and Floyd, 1977); Th/Yb=0.14–047; Nb/Yb=1.43–8.3.The latter two ratios plot in the oceanic basalt array of Pearce's (2008)diagram, clustering around the E-MORB representative composition.In addition, these rocks exhibit important protoclastic and cataclasticdeformations similar to the sheared gabbros exposed at seafloor alongtransform segments of modern slow-spreading ridges (e.g. Cannatet al., 1997; Dick et al., 2000). Mafic granulites, with clear gabbroicparentage and MORB affinity are also abundant among the Hybleancrustal xenoliths (Scribano, 1988; Mazzoleni and Scribano, 1994;Tonarini et al., 1996; Scribano et al., 2006a).

The main results of the study on the Hyblean xenolith suite can besummarized as follows: 1) the dominant peridotites exhibit chemicalcomposition and isotopic signatures consistentwith a depletedmantleorigin; 2) typical rocks of the continental crust are absent; 3) shearedoxide-gabbros, with an E-MORB geochemical affinity, resemble thosefrom modern and fossil oceanic fracture-zones associated with slow-spreading ridges. On these grounds, Scribano et al. (2006a,b, 2009)introduced the hypothesis that the HybleanMeso-Cenozoic carbonaticand volcanic succession lies upon an oceanic core-complex, tectoni-cally exposed at the seafloor of an ancient ocean, most probably theOman–Levantine–Sicily seaway which connected the Panthalassa andwestern Tethys Oceans in the early Permian (Vai, 2003). This is analternative hypothesis to the earlier, still dominant, opinion that theSicily basement represents the northern promontory of the Africancontinental plate (e.g. Finetti, 1984).

3.1. Xenolith evidence for “fossil” abyssal-type hydrothermal systems

Various Hyblean xenoliths provide evidence for complex miner-alogical and textural metasomatic transformations related to long-lasting seawater circulation throughout the tectonized ultramafic/mafic basement (Scribano et al., 2006b). Accordingly, stronglyserpentinized peridotites are very common in all Hyblean xenolithoccurrences. Cryptocrystalline lizardite and platy crysotile representthe main serpentine polytypes, which are sometimes transformedinto retrograde clay minerals (sepiolite and palygorskite). RoundedNi–Fe sulfide micrograins (godlevskite and millerite), rarely sulfideveinlets, are often associated with the serpentinite veins, and aretherefore regarded as one of the products of the serpentine-formingreactions (e.g. Alt and Shanks, 2003).

The Hyblean gabbroic xenoliths exhibit various degrees ofhydrothermal alteration. Incipient albitization of the calcic plagioclaseand pervasive infiltration along brittle microfractures of a turbidclayey mat indicate the increasing degree of alteration. Albitization ofthe plagioclase generally does not erase its primary texture whereasigneous clinopyroxene is entirely replaced by a turbid mass of clayminerals, which also pervasively fills fractures cross-cutting thefeldspar framework. Diverse secondary minerals of likely hydrother-mal origin are immersed in this clayey mat: bright green aegirine–augite, gismondine group zeolites, Fe–Ti oxides, titanite, apatite,zircon, Ca-poor pyroxene, pyrite, sphalerite, gypsum et cetera.Aegirine–augite is ubiquitous, exhibiting either textural equilibriumor incipient to strong disequilibrium with the clayey matrix.

In some cases, both primary modal mineralogy and primarytexture are erased leaving a fully metasomatic rock dominated byalbite and Mg–Fe-rich clay minerals. These rocks generally exhibitporphyroclastic texture. The aegirine–augite observed above occurs

here as small, ameboid relics surrounded by a relatively large band ofclay minerals with sparse magnetite micrograins. Scribano et al.(2006b) also reported on relics of Fo91 olivine, likely deriving frommantle peridotites, giving these rocks a hybrid character. All thesetextural and mineralogical features closely recall the “blackwalls”occurring between serpentinized mantle peridotites and rodingitizedgabbro fromworld-wide ophiolite massifs (e.g. Dubińska et al., 2004).

Several metasomatized gabbroic xenoliths bear high molecularweight aliphatic hydrocarbons generally in the form of tan-coloredmicelles immersed in a clayey felt, probably originating from Fischer–Tropsch-type synthesis in a serpentinite-hosted hydrothermal system(Ciliberto et al., 2009). This suggestion is consistent with the abovehypothesis regarding the early seafloor exposure of an ultramaficcore-complex which pre-dates and holds up the Hyblean carbonaticsuccession.

4. Sample selection and analytical techniques

Silicate glass pockets are common in all types of metasomatizedgabbroic xenoliths, especially in the most altered ones, as shown byoptical microscope observations on 25 samples collected from ValleGuffari (VG), Melilli (CM) and Carlentini diatremes (Fig. 1). Pre-liminary microprobe analyses on eight of these xenoliths showed thatthe glass major elements abundances are surprisingly close to those oftrachytoid magmatic rocks (Scribano and Manuella, 2007). Unfortu-nately, obtaining a reliable glass composition is a problematic issue,since this is mostly replaced by a slightly to mildly birefringentcryptocrystalline felt of secondary minerals. In this respect, twosamples from Valle Guffari represent a favorable exception, exhibitingrelatively large zones with very fresh, perfectly isotropic glass. One ofthese samples (SBZ) has been used up for organic geochemistryinvestigation (Ciliberto et al., 2009), the second (FBV) was selected forthe present study, which focused on the glass pockets. This samplemeasures about 7×5×4 cm, and lacks host lava rind. It is trulyrepresentative of the “blackwall” type xenoliths described by Scribanoet al. (2006b).

4.1. Methods

Preliminary petrographic examination was conducted, as is usual,by polarizing microscope on 30 μm thin sections. Further observationsat micrometer scale were performed by TESCAN-VEGA\\LMU scanningelectron microscope (SEM) at the Department of Geological Sciences,Catania University (Italy). Observations were made in backscatteredelectron mode, generally under high vacuum conditions at accelerat-ing voltage 20 kV and beam current 0.2 nA. SEMwas equippedwith anEDAX Neptune XM4 60 microanalysis working in energy dispersivespectrometry (EDS) which provided preliminary information onmineral chemistry. Qualitative analyses on fluid inclusions, whereorganic compounds may occur, were performed on uncoated thinsections.

Major element abundances on mineral phases and glass, whichconstitute the sample, were acquired by means of a Cameca SX-50electron microprobe equipped with four wavelength dispersivespectrometers (WDS) and one energy dispersive spectrometer(EDS) at the CNR-IGG (Padua Section). Operating conditions wereset at 15 kV accelerating potential, 10 nA beam current, peak countingtimes of 10 s, Fe, Al, K and Na counted first. The electron beam wasdefocused to 6–7 μm and the stage slightly moved while acquiringsignals to avoid the loss of sodium. The accuracy is ~0.5% forabundances N15 wt.%, ~1% for abundances around 5 wt.% and ~20%for abundances around 0.5 wt.%. The precision is given better than 1%for SiO2, Al2O3, FeO, MgO and CaO and better than 3% for TiO2, MnO,Na2O, K2O and P2O5.

X-ray diffraction patterns of powdered whole rock samples wereacquired using a Siemens D5000 diffractometer equipped with Cu Kα

661M. Viccaro et al. / Lithos 113 (2009) 659–672

Author's personal copy

662 M. Viccaro et al. / Lithos 113 (2009) 659–672

Author's personal copy

anode, filter Ni, ΔV=40 kV, I=30 mA, windows of 2 mm, 1 mm,0.2 mm, speed scan 1°/min.

Polished thin sections were Pt-coated and analyzed for trace elementsby Secondary Ion Mass Spectrometry (SIMS) with a Cameca IMS 4f ionmicroprobe installed at the CNR-IGG (Pavia Section). Samples werebombardedwith a 16O−primary beam, 6–15nAcurrent intensity, focusedonto a micro-spot area of ~8–20 μm in diameter. Thewidth of the energyslit was 50 eV and the voltage offset applied to the sample acceleratingvoltage (+4500 V) was −100 V. Positive filtered secondary ions wereextracted and focused under an ion image field of 25 μm. The largestcontrast diaphragm and field aperture (400 and 1800 μmØ, respectively)were used, giving amass resolving power of ~900 (M/ΔM). La, Ce, Nd, Sm,Er andYbwere analyzedbymeasuring the signals of one isotope (i.e.139La,140Ce, 146Nd, 149Sm, 167Er and 174Yb). For Eu, a deconvolution procedurewasused to eliminate BaO interferences fromEu isotopes. Ion signalsweremonitored at mass 154, 160 and 162 in order to distinguish the Gd (andDy) signal from CeO and NdO interferences. Dy was alsomonitored usingisotope 163. Moreover, in the case of Er, the reliability of the analysiswas improved by evaluating the 151Eu16O contribution to mass 167 fromthe Eu+ signal and from the oxide ratio (i.e., EuO+/Eu+) as determinedoncalibration standards. A similar procedurewas followed to correct Yb fromGdO interference at mass 174 (Bottazzi et al., 1994). With reference to theother trace elements, the following isotopes 7Li, 9Be, 11B, 19F, 39K, 45Sc, 47Ti,51V, 52Cr, 85Rb, 88Sr, 89Y, 90Zr, 93Nb, 133Cs, 138Ba, 232Th and 238U weremeasured. All were monitored during the same analysis along with theREE ion signals. 30Si+was selected as the isotope of the reference element(Si) for these matrixes. The duration of the analysis was 35 min, after apreliminary period of about 8 min required to obtain steady-statesputtering conditions for all the isotopes. Conversion from ion intensitiesto concentrations was empirically accomplished by measuring the ionyield relative to Si of each element in reference samples. Concentrations ofLi, Be, B and Fwere obtained bymeans of the calibration curve for silicates(Ottolini et al., 1993, 1994). International reference materials, i.e., NISTSRM-610 and R315 (Rautenschlein et al., 1985), in addition to CNR-IGGinner standards kaersutite Soda Springs KSS, and alkali–olivine–basalt BB(which is an in-house standard of the Geochemical Institute, GöttingenUniversity) were used during the calibration procedures. 45Sc+ and 52Cr+

were further corrected for 29Si16O+ and 24Mg28Si+ interferences at mass45 and 52 amu, respectively. The presence of residual Fe-basedinterferences at mass number 85 (Rb) was evaluated using silicatestandards with variable FeO contents. Accuracy and precision for REE andmost of the quantified trace elements are in the range 5–10% rel. at ppmlevel. Accuracy is typically ~15% for F, U and Th, and ~20% for Cs and Rb.

Bulk major element abundances were obtained at ActivationLaboratories Ltd. (Toronto, Canada) by inductively coupled plasmaoptical emission spectroscopy (ICP-OES) after sample fusion bylithium metaborate/tetraborate and digestion in a weak nitric acidsolution. Instrumental neutron activation analysis (INAA) andinductively coupled plasma mass spectrometry (ICP-MS) were usedto determine trace element concentrations. Additional information onmethods, as well as the standards used, sample preparation anddetection limits are reported in the ACTLABS official website (www.actlabs.com).

5. Sample description

Xenolith “FBV” consists of round edged dark clayey patchesimmersed in a porphyroclastic, feldspatic matrix. Na-rich alkali feldspar

(~40 vol.%), mafic clay (~30 vol.%), aegirine–augite (~8 vol.%), glasspockets (~15 vol.%), magnetite, zeolites, titanite and apatite form thexenolith mineral assemblage. A large number of glassy pockets appeargenerally associated with the clayey clumps and hence randomlydistributed in the rock (Fig. 2a; b). Thorough macroscopic andmicroscopic observations excluded the infiltration of an externallyderived melt.

Looking intomore details, the dark clumps exhibit ovoid or ameboidcross sections and an approximately concentric structure, with a core ofsieve-textured aegirine–augite relic surrounded by an inner shell ofcoalescing spherulites of Mg–Fe-rich clays, an intermediate shell of Mg-poor clay minerals and a glassy outer band lying between the clayeyassemblages and the feldspar matrix (Fig. 2c; d; e). Glass issuing fromthe outer shells alsofills irregular cracks in the adjoining feldsparmatrix(Fig. 2b; c) and pervasively infiltrates along the grain edges of thecataclastic domains. The clayey bands, especially the intermediate one,often display irregular thickness (Fig. 2c; e). Isolated glassy patchesapparently without associated clayey layers or aegirine–augite coresmay represent superficial slices of the glass shells. In addition, relativelywide, irregular glassy pockets are due to the coalescence of two ormoreglassy shells.

Major element compositions of minerals from the sample (Table 1)range within those of the hydrothermally metasomatized xenolithsfrom the Hyblean area (Scribano et al., 2006b; Ciliberto et al., 2009).Specifically, alkali feldspar consists of coarse albite porphyroclasts(Ab95–98–Or1–4–An0–3) with mortar textured crystal edges. Thesemortar textured zones exhibit an increased Or component, neverexceeding 25 mol%. The feldspar grains, especially in the shearedzones, display sieve textures and a cloudy aspect due to the close arrayof fluid and glass inclusions, the latter are generally replaced by clayminerals. The feldspar cell parameters, calculated on the basis ofwhole rock powder XRD peaks (JCPDS card 9-466), are the following:a=8.1382±0.0114 Å, b=12.8119±0.0201 Å, c=7.1518±0.0097 Å,α=94.106±0.075, β=116.541±0.077, γ=87.905±0.104,V=665.382 Å3. These values, compared with data in the literature(Ferguson et al., 1958; Martin,1970;Winter et al., 1977), are consistentwith a low-temperature, hydrothermal albite.

Compositions of the clay minerals conform to mixed layers ofchlorite/smectite (C/S) and smectite/illite (S/I). The number of atomsper formula unit (a.p.f.u.) of Si, on the basis of 28 oxygen equivalentatoms, varies from 7.17 to 7.50 and is always greater than the highestvalue for chlorites (=6.25: Bettison and Schiffman, 1988); Ca and Krepresent thedominant interlayer elements (Table 1; Fig. 3);Mg#variesfrom 0.4 to 0.6. XRD patterns display a wide, uneven peak of the basalreflection of clay minerals in the 4.200–7.240° 2θ range, centered on5.962° 2θ corresponding to d001 14.8121 Å (FWHM 1.035° 2θ Cu Kα),confirming the occurrence of random interlayers. Conversely, no micapeaks were detected. EMP analyses on the aegirine–augite relics(Table 1) evidence about 33 mol% of the aegirine component and lowalumina (Al2O3~0.7wt.%). Trace element abundances, obtained by SIMSmicrospots, show an elevated content in Zr (up to 1200 ppm) and REE,especially LREE (La~230× Chondrite; Ce and Nd~300× Chondrite:Table 2, Fig. 4). Euhedral titanitemicrocrysts, generally embedded in theclayey pools, exhibit a quite uniform composition, with SiO2~30 wt.%,TiO2~37%, CaO~27%; low-totals suggest the occurrence of someintracrystalline OH (Table 1). Fe-oxide grains generally consist ofmagnetite. Scant amounts of Ca–K-zeolites were also detected byEMPA and XRD. Specifically, zeolites belonging to the gismondine group

Fig. 2. Textural and mineralogical aspects of the glass-bearing xenolith FBV; a) low magnification photograph of part of the sample cut surface. The dark patches consist of aegirine–augite relics surrounded by mafic clays, the clear zone consists of albite and glass; b) magnified photograph of part of a polished thin-section of the sample. Arrows point to someglassy pockets showing a lustrous aspect; c) BSE image of a cross section of a typical clayey patch. Agt = aegirine–augite; C/S = chlorite/smectite interlayered clays; S/I = smectite/illite interlayered clays; Gls = glass; Ab= albite; d) Close view of the area enclosed in the rectangle depicted in c. Note the rough aspect of the glass near the clayey layer due to fluidinclusions; e) BSE microphotograph of a clayey patch with a very thin and irregular S/I shell (Full explanation is given in the text.); f) BSE image of a glassy zone between albite andclays. A small titanite grain (Ttn), probably an unmelted relict, is imbedded in the glass; g) microphotograph (plane pol. light) of part of a typical pocket of trachytoid glass showingstraight and curvilinear cracking and a fewmagnetite (mt) micrograins; h) Detail of a concentric fracture set in one of the glassy pockets (plane pol. light), with a fluid inclusion arrayarranged orthogonally to the perlitic cracking.

663M. Viccaro et al. / Lithos 113 (2009) 659–672

Author's personal copy

(garronite and merlinoite) were identified on the basis of XRD peaksakin to data reported by Nawaz andMalone (1982) andMohapatra andSahoo (1987), respectively. It is important to note that such zeolite typesgenerally crystallize from hydrothermal solutions in the temperaturerange 60–250 °C (Walker, 1962; Bayliss and Levinson, 1971).

5.1. Texture and composition of the glassy pockets

The interface between the glass and the feldspar grains is generallysharp (Fig. 2c), although a glassy film spreads through the grain

interstices near the contact surface, especially in the sheared zones.Conversely, the contact surface between the clayey layers and theglass appears quite irregular at a micrometer scale. Here dispersedclay minerals and a lot of fluid inclusions give the glass a dusty aspect(Fig. 2d). The relatively wide glass pockets display both straight andperlitic cracking (Figs. 2g; h and 5a; b). The glass is generally free ofminerals, except rare magnetite micrograins (Fig. 2g) and very rarerelics of titanite (Fig. 2f).

Arrays of elongated fluid inclusions are arranged either radially orparallel to the concentric fractures (Figs. 2h and 5c). Examined ingreater detail, these inclusions sometimes display a mushroom shape(Fig. 5d; e) with a dark brownish “cap” and a thin, lighter colored“stalk”. The latter generally encases one ormore colorless bubbles (gasphase?). Ovoid shaped fluid–solid inclusions, ranging in size from~10 μm up to 50 μm, are also dispersed throughout the glass.Qualitative EDS analyses on uncoated thin sections revealed thepresence of reduced carbon and sulfur in the fluid fraction of theseinclusion which generally host sylvite daughter mineral (Fig. 5f).

About 60 EMP analyses were performed on four distinct glasspockets from two thin sections of the xenolith. Some representativedata are reported in Table 2. Considering the analyses recalculated onanhydrous basis, SiO2 varies from 60.2 to 62.7 wt.% (average=61.7,with a standard deviation σ=1.1); Al2O3 varies from ~13 to ~16 wt.%(average=14.5, σ=1.1); Na2O varies from 7. 2 to 7.8 wt.% (aver-age=7.3, σ=0.4); K2O varies from 3 to 4 wt.% (average 3.5,σ=0.24); FeO (total), varies from 7.6 to 9.7 (average 8.1, σ=0.9);MgO varies from 1.1 to 1.8, CaO from 1.8 to 2.4, Cl varies from 0.05 to0.2 wt.%. Totals, ranging from ~96% to ~100%, may suggest fluidconcentrations up to 4%. This variation appears unrelated to thelocation of the spots with respect to the concentric rings of themicroperlitic texture. Furthermore, there is no clear relation betweenthe color and the chemical composition of the glass as revealed bymicroprobe analyses.

Six micro-areas in two glass pockets from the sample studied werealso analyzed by SIMS to obtain trace element data (Table 3). Therelatively high abundances of Zr (~740 ppm on average), Ba(~390 ppm), F (~500 ppm) and LREE are noteworthy. The latterrange between 40× Chondrite and 100× Chondrite, whereas La(n)/Yb(n) is about 5.6 (Table 3, Fig. 4). Primordial Mantle normalized(McDonough and Sun, 1995) multi element pattern (Fig. 6) shows

Table 1Representative EMP data of main minerals forming the glass-bearing xenolith (sampleFBV).

wt.% a b1 b2 b3 b4 c1 c2 ⁎

Clay minerals

SiO2 35.77 43.50 44.35 44.50 44.1 47.90 46.55 45.30TiO2 0.08 0.76 0.91 0.98 1.48 0.00 0.00 0.47Al2O3 11.50 9.90 9.80 9.80 10.05 18.18 18.85 13.30FeO⁎ 23.11 23.6 23.10 22.90 23.20 3.19 0.43 17.00MnO 0.26 0.00 0.05 0.12 0.00 0.00 0.00 0.06MgO 17.60 7.15 6.60 6.55 6.15 1.13 0.00 4.11CaO 0.74 0.90 1.30 1.15 0.90 5.30 6.34 1.90Na2O 0.15 0.41 0.05 0.00 0.38 0.25 0.36 0.23K2O 0.21 3.35 3.30 3.30 3.35 4.81 4.69 3.80BaO nd nd nd nd nd 1.32 1.34 –

SrO nd nd nd nd nd 0.97 0.60 –

Total 89.42 89.57 88.46 89.30 89.61 83.05 79.16 –

Alkali feldspar Aegirine-augite Titanite

SiO2 69.50 68.70 67.10 69.35 51.80 52.21 52.4 31.15TiO2 0.00 0.01 0.00 0.00 0.40 0.40 0.42 37.15Al2O3 19.85 19.10 21.05 19.00 0.75 0.77 0.70 0.56FeO⁎ 0.20 0.20 0.28 0.12 20.05 19.00 18.90 1.98MnO 0.03 0.02 0.00 0.00 0.84 0.99 1.00 0.12MgO 0.00 0.00 0.00 0.00 5.80 6.25 6.66 0.00CaO 0.10 0.05 0.35 0.02 16.00 16.30 16.25 27.1Na2O 11.00 11.70 7.87 11.50 4.45 4.40 4.43 0.40K2O 0.28 0.35 3.23 0.25 0.00 0.00 0.00 0.00Total 99.96 100.13 99.88 100.24 100.09 100.32 100.76 98.46

Analyses of clayminerals: a: chlorite/smectitemixed layers; b1–4: smectite–illitemixedlayers; c1–2: illite/smectite mixed layers. The asterisk (⁎) represents the approximateaverage composition of the FBV clay minerals used for graphic comparison in Fig. 11(further explanation is given in the text).

Fig. 3. Classification diagrams (following D'Antonio and Kristensen, 2005; Fulignati et al., 1997) for clay minerals from the Hyblean metasomatized gabbroic xenoliths (shaded areas:data from Scribano et al., 2006b, this paper, and unpublished). a) Sum of major non-interlayer cations vs. Al (a.p.f.u.-atoms per formula unit calculated on 28 oxygen-equivalentbasis); b) sum of major interlayer cations vs. Si a.p.f.u. (28 oxygen-equivalent); c) K vs. Al(iv) a.p.f.u. (22 oxygen-equivalent).

664 M. Viccaro et al. / Lithos 113 (2009) 659–672

Author's personal copy

negative peaks for Sr and B in contrast to positive peaks for Zr and Ba(~1000× PM).

6. Discussion

Major element abundances of the glass pockets from sample FBV,obtained bymicroprobe analyses and recalculated on anhydrous basis,were plotted in the TAS diagram (Le Bas et al., 1986): the spot analysesform a cluster in the field for trachyte and trachydacite. Conversely,the host rock plots in the trachyandesite field (Fig. 7). The glass FeOtot/Al2O3 (wt.%) ratios range between the pantelleritic-trachyte andcomenditic-trachyte fields, clustering across the boundary linebetween the two fields (Fig. 8a) and generally display a mildperalkaline, rarely slight metaluminous, character [0.7bmolar Al2O3/(Na2O+K2O)b1.1; average=0.85: Fig. 8b]. Na2O/K2O ratios varybetween 1.9 and 2.3 (Fig. 8c).

Trace element ratios of the glass are also compatible with thosefrom trachytes. Specifically, the reported glass analyses plot in thealkaline trachyte field of the Nb/Y vs. Zr/Ti diagram (Pearce, 1996:Fig. 9) and the Ba/La ratio (=5.5) falls in the range of OIB-FOZO endmember (4.3–11.5 e.g. De Astis et al., 2006 and references therein).Furthermore, some trace element abundances of the glass werecompared with trachytic suites from several modern oceanic contexts,the greatest part showing compositional gap: Ascension (Harris,1983;

Nielson and Sibbet, 1996; Kar et al., 1998), Canary Islands (Chayes,1977; Rodiguez-Losada and Martinez-Frias, 2004), French-Polynesia(Legendre et al., 2005), Kerguelen (Gagnevin et al., 2003) and SocorroIsland (Bohrson and Reid, 1995; Bohrson et al., 1996; Bohrson andReid,1997,1998). Data for other oceanic trachytic suites as indicated inFig. 10 were desumed from the published GEOROC database, availableat http://georoc.mpch-mainz.gwdg.de/georoc/). The relatively nearsites of Pantelleria (Esperança and Crisci, 1995; Civetta et al., 1998;Avanzinelli et al., 2004) and Mt. Etna (Tanguy et al., 1997) were alsoconsidered, though the latter exhibits a very small proportion oftrachytes among its lavas and pyroclastic deposits and no Daly-gap hasbeen recognized (Viccaro and Cristofolini, 2008). Specifically, Fig. 10shows selected trace elements (Rb, Ba, La, Th, Nb) plotted against Zr,with the fields of each of the considered suites. The abundances of theselected elements in the glass, as reported in Table 3, are compatiblewith those of oceanic trachytes, except for the relatively low values ofRb and Nb. On the other hand, looking at the oceanic trachytes data setchosen for comparison, the diagram highlights the very wide range ofvariation in the selected elements, even within a particular suite. Forexample, Zr varies from about 50 ppm to about 2300 ppm; Ba variesfrom a few ppm up to about 2500 ppm. Since common petrogeneticprocesses, as well as fractionation crystallization or fractionationpartial melting may hardly account for such large variations, anadditional element exchange due to open-system processes mightalso be invoked in some cases (e.g., Legendre et al., 2005).

Considering the origin of the glass pockets in the sample, it isopportune to recall the lack of textural evidence for melt infiltrationfrom an external source. On the other hand, the presence in theHyblean crust of trachytic magma able to infiltrate wall rocks is veryunlikely, since igneous rocks with trachytic composition have neverbeen found in this region. The relatively homogeneous composition ofthe glass, in terms of major elements and most trace elements, withinthe same and different pockets suggests that it represents a minimumor a eutectoid composition. Given the absence of microlites (except formagnetite) in the glass, which could be responsible for local dif-ferentiation, the major element variations, shown by the standarddeviation values reported in Table 2, may be explained by composi-tional heterogeneity in the source minerals which influences theeutectoid composition.

The occurrence of the glass pockets between albite grains andclayey patches (Fig. 2) suggests that the eutectoid composition largelyconsists of unknown percentages of albite, mafic clays and aegirine–augite. This hypothesis is consistent with the chemical composition of

Table 2Major element abundances of FBV whole rock (ICP-OES) and its glass (selected EMP spot analyses).

FBV Glass Glass

wt.% (w.r.) a b c d e f g h i J k l Avg σ

SiO2 53.75 58.76 60.69 60.38 60.09 59.47 61.41 60.88 62.30 60.75 62.24 61.56 61.60 61.51 1.11TiO2 0.58 1.13 1.34 1.28 1.16 1.19 1.06 1.09 0.44 0.90 0.70 0.50 1.06 1.00 0.24Al2O3 15.77 12.16 14.41 14.85 14.46 15.03 12.86 14.67 15.60 12.91 15.95 13.94 14.91 14.85 1.07FeO(t) 6.36 9.37 7.98 7.90 8.05 6.83 8.82 8.43 5.51 8.39 6.16 7.49 8.10 7.5 1.24MnO 0.14 nd nd nd nd nd nd nd nd nd Nd nd nd – –

MgO 4.44 1.62 1.45 1.34 1.34 1.22 1.36 1.31 0.84 1.22 0.82 1.62 1.18 1.36 0.34CaO 1.92 2.35 2.41 2.37 2.17 2.14 2.74 2.26 1.76 2.52 1.91 2.10 2.70 2.24 0.32Na2O 4.13 7.15 7.47 7.07 7.14 7.44 6.90 7.59 7.70 7.32 6.51 7.11 6.75 7.30 0.32K2O 3.29 3.40 3.54 3.47 3.63 3.47 3.71 3.15 3.75 3.27 3.67 3.47 3.33 3.50 0.22P2O5 0.04 0.41 0.25 0.34 0.21 0.28 0.39 0.18 0.05 0.44 0.15 0.19 0.36 0.23 0.12BaO nd 0.14 0.14 0.00 0.00 0.00 0.07 0.14 0.00 0.00 0.00 0.14 0.00 0.05 0.06Cl nd 0.16 0.09 0.09 0.11 0.10 0.10 0.07 0.09 0.11 0.07 0.13 0.09 0.11 0.02L.O.I 9.30 nd nd nd nd nd nd nd nd nd Nd nd nd – –

Total 99.72 96.76 99.81 99.14 98.45 97.25 99.46 99.81 98.10 97.89 98.22 98.30 100.10A/N+K 1.51 0.82 0.70 0.96 0.92 0.93 0.86 0.92 0.93 0.82 1.08 0.90 1.01 0.94A/N+K+C 1.32 0.64 0.58 0.75 0.73 0.75 0.64 0.73 0.78 0.63 0.87 0.72 0.76 0.75Mg# 66.5 23.6 27.6 23.2 22.9 24.2 21.6 21.7 21.4 20.6 19.2 27.8 20.06 24.4D.I. 72.1 66.4 75.2 77.5 75.5 77.3 70.5 76.4 81.5 70.3 83.3 73.6 78.3 77.8

Avg represents average of 63EMPanalyses ondifferent glassypockets recalculated on anhydrousbasis.σ: standard deviation.A=Al2O3mol%;N=Na2Omol%; K=K2Omol%; C=CaOmol%.Mg#=Mg/(Mg+Fe2+ total); D.I.-differentiation index = normative Qz+Ab+An+Or+Ne+Le (wt.%).

Fig. 4. Chondrite normalized (Boynton, 1984) REE patterns for the averaged glass, thewhole rock (sample FBV) and the coexisting aegirine–augite (data reported in Table 3).

665M. Viccaro et al. / Lithos 113 (2009) 659–672

Author's personal copy

the mineral phases noted above, as shown graphically in Fig. 11 formajor elements and in Fig. 6 for selected trace elements. Unfortu-nately, no reliable mass–balance calculations are possible due to thelarge uncertainty inherent in averaging the claymineral compositions.The relatively wide variations in Zr, as shown in Fig. 10, may betentatively attributed to the variability of Zr to micro-heterogeneity atthe melting scale, where variable proportions of aegirine–augiteparticipate in partial melting together with albite and clay minerals,bearing in mind the high Zr concentration found in aegirine–augite(Zr up to 1200 ppm; Table 3). Furthermore, the rather flat trend

observed for the other trace elements, which is difficult to ascribe to asimple liquid line of descent, may be evidence that trachytic glassesare unaffected by differentiation processes such as fractional crystal-lization. Given the origin of the glass through partial melting, itslocalized infiltration between sheared feldspar grains andmicrocracks(Fig. 2b; e) may be explained by textural adaptation to variations involume related to partial melting.

The abundance of fluid inclusions in the glass is consistent with therole played by clay minerals in the melting. The reduced carboncompounds in somefluid inclusions (Fig. 5d), inferredbyqualitativeEDS

Fig. 5. Microphotographs showing some textural elements of the glassy pockets from the sample FBV; a) BSE image of typical concentric cracking of the glass (Gls). S/I represents asmectite/illite spherulite post-dating the glass formation; FI are fluid inclusions; b) BSE microphotograph showing clay minerals (S/I) post-dating the perlitic cracking;c) microphotograph (Plane pol. light) representing a glassy zone with a close array of elongated fluid inclusions in an approximately concentric pattern; d) microphotograph (planepol. light) of “mushroom” shaped fluid inclusion in the glass (more explanation is given in text); e) BSE image of a long-side cut, empty “mushroom” shaped fluid inclusion; f) BSEimage of a fluid–solid inclusion (Fi), approximately indicated by the dark halos, occurring in the glass. KCl represents sylvite daughter mineral.

666 M. Viccaro et al. / Lithos 113 (2009) 659–672

Author's personal copy

analyses in uncoated thin section, is consistent with the occurrence ofaliphatic and minor polycyclic aromatic hydrocarbon compounds,probably originating in a serpentinite-hosted hydrothermal system,confined in the clayey fraction of several hydrothermally alteredgabbroic xenoliths from the study area (Ciliberto et al., 2009). Inparticular, these authors detected hydrocarbons in the xenolith SBZ,which has almost identical composition and texture to the sampleselected for the present study, including trachytic glass pockets.

The partial melting process was probably triggered during ascentof the diatremic system and stopped (with quenching of the melt)during eruption. Although the temperature range of the volatile-richbasaltic diatreme systems is certainly lower than that pertaining tonormal effusive eruptions (e.g., Dawson, 1980), no inference on theeruptive temperatures of the HybleanMiocenic diatremes is available.However it is opportune to recall published experimental dataindicating that the melting point of the very refractory KAlSiO4

drops from about 1700 °C down to about 650 °C under high-F fugacity,since the substitution of one O2− by two F− destabilizes the crystallinestructure (Veksler et al., 1988 and references therein). These authorsalso noted that, despite the high fluorine content of the KAlSiO3F2component, no liquid immiscibility between fluorine- and silica-richdomains has been observed and the melt quenches as a perfectly

homogeneous glass. Assuming, reasonably enough, that albitebehaves like kalsilite, it seems probable that halogens have drama-tically lowered the temperature of the eutectoid. Thus melting tem-peratures as low as 600–650 °C are not inconceivable.

Since “the use of the very small to understand the very large” is awidely acknowledged approach in the Earth sciences, the occurrenceof this microscopic-scale partial melting event may pertain to theworld-wide debate on the primary origin of some trachytic magmas.In this respect it must be noted that the Hyblean area is not directlyinvolved in this debate, since igneous rocks with trachytic composi-tion have never been found in this region. To our knowledge, there islittle tangible evidence of alkali-rich silicate melts produced bynatural partial melting of hydrothermally altered mafic rocks. One ofthe rare cases, reported by Wood and Browne (1995), concerns some“paralava” bombs with Cl-rich peralkaline composition, from WhiteIsland volcano, New Zealand. Such paralava derives from pockets ofshallow-seated melt due to the pyrometamorphic interaction ofandesitic basaltic magma with hydrothermally altered vent breccias.

Table 3Abundances (ppm) of some trace elements of the FBV glass, aegirine–augite (Agt)obtained by SIMS spot analyses and the whole rock (FBV) using ICP-MS.

Ppm a b c d e f Average Agt FBV

B 4.0 3.8 3.8 3.9 4.0 4.6 4 0.3 ndF 540 559 518 541 583 560 550 38 ndCr b.d.l. b.d.l. b.d.l. b.d.l. b.d.l. b.d.l. – 11 5.2Ni n.d. n.d. n.d. n.d. n.d. n.d. – n.d. 44Rb 30 32 26 27 28 29 28 n.c. 30Sr 95 82 92 89 91 90 90 25 1100Y 43.3 29.6 46 39 40 37 39 101 11.1Zr 687 468 920 703 839 814 738 1222 615Nb 36 22 46 35 38 19 32 4.3 70Ba 399 373 398 391 399 391 392 1.4 752La 82 60 75 67 68 63 69 74 17.4Ce 159 116 157 140 139 128 140 247 34.3Nd 71 46 66 58 60 56 59 182 13.1Sm 12.8 8.4 11.7 10.9 9.7 9.4 10.5 41 2.4Eu 3.2 1.8 2.9 2.4 2.2 2.1 2.4 8.6 0.6Gd 8.8 5.8 9.6 7.8 9.0 7.8 8.1 32 1.8Dy 7.7 5.6 8.0 6.7 6.8 6.1 6.8 25 2.1Er 4.5 3.4 4.9 4.3 4.2 4.1 4.9 11.9 1.2Yb 9.5 5.9 9.1 8.0 7.9 8.6 8.1 19.9 2.3Th 5.0 3.4 5.8 5.2 5.0 5.8 5.0 0.1 6.1U 0.8 0.6 1.1 1.0 1.2 1.0 0.9 0.7 0.9

a–f: individual spot analyses. n.d.: not detected; b.d.l.: below detection limits; n.c.: notcalibrated.

Fig. 6. Primordial Mantle normalized (McDonough and Sun, 1995) multi-elementdiagram for the averaged glass composition, the whole rock (sample FBV) and thecoexisting aegirine–augite averaged composition (data from Table 3).

Fig. 7. Total alkalis vs. silica diagram (Le Bas et al., 1986) for the glass and its hostxenolith. a) Gray area includes the glass compositions, square represents the glass-bearing whole-xenolith (FBV), striped area comprises the Hyblean gabbro xenolithsexcept the oxide-rich types (data source: Scribano et al., 2006a); b) close view of thebroken-line edged rectangular area shown in (a) enclosing distinct EMP analyses of theglass (circles).

667M. Viccaro et al. / Lithos 113 (2009) 659–672

Author's personal copy

The authors also suggested that the fusion probably occurred at atemperature between 800 and 1000 °C in a regime where silicatemelts were buffered by NaCl-saturated aqueous vapor and hydrosalinemelt. Since these paralava bombs exhibit an odd and highly variablechemical composition (e.g. their agpaitic indexes vary from 1 to 22:Wood and Browne, 1995), their relevance for hypotheses regardingthe origin of magmas is limited, except for highlighting the crucial roleof saline brines in partial melting processes.

Contrary to the above reported case, the Hyblean glass shows arelatively narrow variation in its chemical composition which closelymatches that of world-wide trachytes. It is therefore a plausiblecandidate to substantiate petrogenetic models invoking the partialmelting of hydrothermally altered mafic rocks (e.g., Lassiter et al.,2002) to produce primary trachytic magmas. Given the relatively

elevated halogens and water concentration necessary to decrease theeutectoid temperature and hence to produce the melts, additionalconditions that can allow these melts to be mobilized and thenerupted are: 1) the extent of partial melting and 2) the melt viscosity.With respect to the former, it should be noted that the chemicalcomposition of the glass, and hence of the early eutectoid assemblage,is notably different from that of its host whole rock (e.g. Fig. 7),yielding no more than 20% (by volume) of trachytic melt. Althoughsuch a percentage may be sufficient to determine the coalescence ofdifferent melt pockets and hence the melt percolation through crustalfractures, it is possible that elsewhere large portions of ancient,hydrothermally altered, oceanic deep crust may exhibit, as a whole, aeutectoid-like composition capable of yielding even much largervolumes of trachytic melt.

Considering the potential mobility of such a trachytoid melt, thecoexistence of water and halogens certainly reduces the degree ofpolymerization of the melt structure, reducing in turn the viscosity ofthe system and allowing it to flow (Baker and Vaillancourt, 1995;Giordano et al., 2004, 2008). In addition, the effect of the peralkalinityof the melt on its rheology is most important: it is one of the reasonsfor the occurrence of kilometers-long and relatively thin peralkalinetrachyte/rhyolite lava flows in some oceanic island volcanoes (e.g.Bohrson et al., 1996).

An attempt was made to calculate the viscosity of the trachyticmelt with the composition reported in Table 2 considering theexperimental model of Vetere et al., (2007 Eq. 5), which predictsviscosity variation as a function of melt water content and tempera-ture. Volatile concentration was assumed as 4 wt.%, as suggested bylow-totals of some microprobe spots (see previous section). Thevalues of viscosity obtained at T=1000 °C are close to the Newtonianviscosities (~102.2 Pa s) for tholeiitic and alkali basaltic melts atT=1100 °C (cf. Giordano and Dingwell, 2003). On the other hand, aviscosity as high as ~104.7 Pa s is obtained at T=750 °C, which is areliable temperature value for trachytic systems (Scaillet andMacdonald, 2006; Stevenson and Wilson, 1997). Nevertheless theVetere et al. (2007) model considers H2O as the only volatilecomponent dissolved in the melt.

From a hydrodynamic standpoint, the segregation of such arelatively high-silica melt from the source and its upward migrationmay be therefore possible, provided that the system has temperaturesN750 °C, high halogen content and high alkaline oxides/aluminaratios. Such a temperature value in the lower/middle level of the

Fig. 8. Tentative characterization of the FBV glass using classic diagrams for trachytoidrocks. a) FeO vs. Al2O3 diagram shows the pantelleritic-trachyte affinity of the glass(circles); b) Al2O3/(Na2O+K2O) vs. Al2O3/(CaO+Na2O+K2O) molar ratios show theperalkaline and metaluminous affinity of the glass. Note the peraluminous character ofthe whole rock (FBV); c) Na2O vs. K2O (wt.%) diagram showing the sodic affinity of theFBV glass (black area). Fields for trachytic rocks from different volcanic areas are alsoindicated [Data sources: Ascension Island-Harris (1983); Canary Islands-Chayes (1977)and Rodiguez-Losada andMartinez-Frias (2004); Mt. Etna-Tanguy et al. (1997); French-Polynesia-Legendre et al. (2005); Rallier-du-Baty, Kerguelen-Gagnevin et al. (2003);Pantelleria-Civetta et al. (1998), Avanzinelli et al. (2004); Socorro Island-Bohrson andReid (1998)].

Fig. 9. Averaged glass composition (open circle) plotted on the Nb/Y vs. Zr/Ti diagram(Pearce, 1996). Inset in the upper-right corner represents the enlarged area of thediagram enclosed in the broken-line edged rectangle, where the cluster of distinct spotanalyses (full circles) is shown.

668 M. Viccaro et al. / Lithos 113 (2009) 659–672

Author's personal copy

oceanic crust would imply the occurrence of basaltic intrusion(s) (e.g.a dikes network). In this respect, the actual formation of importantvolumes of trachytic magmas depends on the convergence of several

favorable factors, primarily the volume relations between the basaltand the altered gabbroic rocks and their respective distance in thecrust. Obviously, basalt must furnish only the heat to trigger partial

Fig. 10. Concentrations (ppm) of Zr versus Rb, Ba, La, Th and Nb for the FVB glasses of trachytic composition (filled circles). FVB glasses are plotted within the context of selectedoceanic trachytic suites desumed from the published GEOROC database, available at http://georoc.mpch-mainz.gwdg.de/georoc/. Arrows indicating a particular trachytic suite areplaced just at the edge of the respective field. See text for discussion. Differences in the Zr compositional range for a given suite are artifacts since no coupled trace element data areavailable in the considered database.

669M. Viccaro et al. / Lithos 113 (2009) 659–672

Author's personal copy

melting and no chemical interaction between the two magma typesshould occur to preserve the compositional identity of the anatecticmelts. Such conditions may be fully achieved in geological settingswith starving magmatic regimes. Accordingly, slow-spreading ridgesand fracture-zones in off-ridge locations should represent the mostsuitable setting for primary trachyte magma formation. The possibilitythat part of the supposed basalt and trachytic magmas canindependently reach the surface as extrusive eruptions may explainthe typical bimodal volcanism of some ocean island volcanoes.

A final comment addresses the eruptive styles of these trachyticmagmas which, owing to their high volatile concentration, canproduce powerful explosive eruptions, not very common in within-plate settings. In this regard, it must be noted that magma ascentdynamics represent a crucial issue: in case of slowascending, large gasbubbles might rise buoyantly through the magma and escape quietlyfrom the system yielding lava flows. On the other hand, if magmaascends rapidly to the surface, disequilibrium bubble nucleation andstrongly explosive fragmentation will probably occur (e.g. Fisher andSchmincke, 1984).

7. Conclusions

Although our results do not provide inferences at the regionalscale, since no igneous rocks with trachytic compositionwere found inthe Hyblean area, they can provide some indications regarding thegeneral, long-lasting debate on the origin of the bimodal trachyte–basalt association occurring in diverse anorogenic settings, especiallyin within-plate oceanic islands.

The results obtained here lead to the followingmain conclusions: 1)the glass pockets found in some hydrothermally-modified, gabbroicxenoliths from the Hyblean area are due to the melting of eutectoidsystems chiefly consisting of albite and mafic clays and hydrothermalaegirine–augite. The clays derived from an open-system breakdown ofigneous mafic minerals; 2) major element abundances of the glassgenerally matches a peralkaline trachyte composition; 3) the distribu-tion of trace elements in the glass is consistent with that of trachyticvolcanic rocks, especially those from oceanic island settings; 4) anatexisof a brine-rich, altered oceanic crust triggered by shallow-seatedintrusions of basaltic magma can produce primary trachytic melts,accounting for the “Daly-gap”which characterizes someoceanicwithin-plate volcanoes; 5) segregation from the source and upward migrationof trachytic primary melts is favored by high content in alkalis and highconcentration of dissolved volatiles, including halogens, which depoly-merize the melt structure and hence decrease its viscosity; 6) such avolatile-rich magma can account for unusual explosive volcanic activityin within-plate environments.

Acknowledgments

Editorial guidance from A. Kerr and constructive criticism fromW.Bohrson and an anonymous reviewer are gratefully acknowledged. M.Wilkinson supplied assistance with the English language. This paperrepresents one of the results of an ongoing research program onHyblean xenoliths conducted by V. Scribano. The Università di Catania(Ricerche di Ateneo, grants to V.S and R.C.) provided partial financialsupport for EMP and SIMS analyses.

References

Alt, J.C., Shanks, W.C., 2003. Serpentinization of abyssal peridotites from theMARK area,Mid-Atlantic Ridge: sulfur geochemistry and reaction modeling. Geochimica etCosmochimica Acta 67 (4), 641–653.

Atzori, P., Mazzoleni, P., Punturo, R., Scribano, V., 1999. Garnet–spinel–pyroxenite xenolithsfromHybleanPlateau(South-easternSicily, Italy).MineralogyandPetrology66,215–226.

Avanzinelli, R., Bindi, L., Menchetti, S., Conticelli, S., 2004. Crystallisation and genesis ofperalkaline magmas from Pantelleria Volcano, Italy: an integrated petrological andcrystal–chemical study. Lithos 73, 41–69.

Bailey, D.K., Schairer, J.F., 1964. Feldspar–liquid equilibria in peralkaline liquids: theorthoclase effect. American Journal of Science 262, 1198–1206.

Bailey, D.K., Schairer, J.F., 1966. The system Na2O–Al2O3–Fe2O3 at 1 Atm and thepetrogenesis of alkaline rocks. Journal of Petrology 7, 114–170.

Baker, P.E., Mv Reath, I., 1972. The acid rocks of the ocean basins. Bulletin of Volcanology36, 328–341.

Baker, D.R., Vaillancourt, J., 1995. The low viscosity of F+H2O-bearing granitic meltsand implications for melt extraction and transport. Earth and Planetary ScienceLetters 132, 199–211.

Bayliss, P., Levinson, A.A., 1971. Low temperature hydrothermal synthesis from dolomiteor calcite, quartz and kaolinite. Clays and Clay Minerals 19, 109–114.

Bettison, L.A., Schiffman, P.,1988. Compositional and structural variations of phyllosilicatesfrom the Point Sal ophiolite, California. American Mineralogist 79, 671–695.

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

Bohrson, W.A., Reid, M.R., 1995. Petrogenesis of alkaline basalts from Socorro Island, Mexico:trace element evidence for contamination of ocean island basalt in the shallow oceancrust. Journal of Geophysical Research-Solid Earth 100 (B12), 24555–24576.

Bohrson, W.A., Reid, M.R., 1997. Genesis of silicic peralkaline volcanic rocks in an oceanisland setting by crustal melting and open-system processes: Socorro Island,Mexico. Journal of Petrology 38 (9), 1137–1166.

Bohrson, W.A., Reid, M.R., 1998. Genesis of evolved ocean island magmas by deep- andshallow-level basement recycling, Socorro Island, Mexico: constraints from Th andother isotope signatures. Journal of Petrology 39 (5), 995–1008.

Bohrson, W.A., Reid, M.R., Grunder, A.L., Heizler, M.T., Harrison, T.M., Lee, J., 1996.Prolonged history of silicic peralkaline volcanism in the eastern Pacific Ocean.Journal of Geophysical Research 101 (B5), 11.457–11.474.

Bonnefoi, C.C., Provost, A., Albarède, F., 1995. The “Daly Gap” as a magmatic catastrophe.Nature 378, 270–272.

Bottazzi, P., Ottolini, L., Vannucci, R., Zanetti, A., 1994. An accurate procedure for thequantification of rare earth elements in silicates. In: Benninghoven, A., Nihei, Y., Shimizu,N.,Werner,H.W.(Eds.), IX InternationalConferenceonSecondary IonMassSpectrometry,Yokohama (Japan). JohnWiley & Sons, Chichester (England), pp. 927–930.

Bowen, N.L., 1945. Phase equilibria bearing on the origin and differentiation of thealkaline rocks. American Journal of Science 243, 75–89.

Boynton, W.V., 1984. Geochemistry of the rare earth elements: meteorite studies. In:Henderson, P. (Ed.), Rare earth elements geochemistry. Elsevier, pp. 63–114.

Butler, R.W.H., Grasso, M., La Manna, F., 1992. Origin and deformation of the Neogene–Recent Maghrebian foredeep at the Gela Nappe, SE Sicily. Journal of the GeologicalSociety of London 149, 547–556.

Cannat, M., Chatin, F., Whitechurch, H., Ceuleneer, G., 1997. Gabbroic rocks trapped inthe upper mantle at the Mid-Atlantic ridge. In: Karson, J.A., Cannat, M., Miller, D.J.,Elthon, D. (Eds.), Proceedings O.D.P. Science Results, vol. 153 , pp. 243–264.

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

Chayes, F., 1977. The oceanic basalt–trachyte relation in general and in the CanaryIslands. American Mineralogist 62, 666–671.

Ciliberto, E., Crisafulli, C, Manuella, F.C., Samperi, F., Scirè, S., Scribano, V., Viccaro, M.,Viscuso, E., 2009. Aliphatic hydrocarbons in metasomatized gabbroic xenolithsfrom Hyblean diatremes (Sicily): genesis in a serpentinite hydrothermal system.Chemical Geology 258, 258–268.

Civetta, L., D Antonio, M., Orsi, G., Tilton, G.R., 1998. The geochemistry of volcanic rocksfrom Pantelleria Island, Sicily Channel: petrogenesis and characteristics of themantle source region. Journal of Petrology 39 (8), 1453–1491.

Clague, D.A., 1978. The oceanic basalt–trachyte association: an explanation of the Dalygap. Journal of Geology 86, 739–743.

Daly, R.A., 1925. The geology of Ascension Island. Proceedings of the AmericanAcademia of Arts and Sciences 60, 3–80.

D'Antonio, M., Kristensen, M.B., 2005. Hydrothermal alteration of oceanic crust in theWest Philippine Sea basin (Ocean Drilling Program leg 195, Site 1201): inferencesfrom a mineral chemistry investigation. Mineralogy and Petrology 83, 87–112.

Fig. 11. Major elements of the main minerals from FBV xenolith normalized to theaveraged glass composition reported in Table 2. Clay: average of clay mineralscompositions (Table 1); agt: aegirine–augite composition (analysis from Table 1). Ab:albite composition (analysis from Table 1).

670 M. Viccaro et al. / Lithos 113 (2009) 659–672

Author's personal copy

Dawson, J.B., 1980. Kimberlites and Their Xenoliths. Springer-Verlag, Berlin, p. 252.De Astis, G., Kempton, P.D., Peccerillo, A., Wu, T.W., 2006. Trace element and isotopic

variations fromMt Vulture to Campanian volcanoes: constraints for slab detachmentandmantle inflow beneath southern Italy. Contributions to Mineralogy and Petrology151, 331–351.

De Rosa, R., Mazzuoli, R., Scribano, V., Trua, T., 1991. Nuovi dati petrologici sulle vulcanitidei Monti Iblei (Sicilia sud-orientale): implicazioni genetiche e geotettoniche.Mineralogica et Petrographica Acta 34, 133–151.

Dick, H.J.B., Natland, J.H., Alt, J.C., Bach, W., Bideau, D., Gee, J.S., Haggas, S., Hertogen, J.G.H.,Hirth, G., Holm, P.M., Idelfonse, B., Iturrino, G.J., John, B.E., Kelley, D.S., Kikawa, E.,Kingdom, A., LeRoux, P., Maeda, J., Meyer, P.S., Miller, D.J., Naslund, H.R., Niu, Y.L.,Robinson, P.T., Snow, J., Stephen, R.A., Trimby, P.W.,Worm, H.U., Yoshinobu, A., 2000. Along in situ sectionof lowerocean crust: result ofODPLeg176drillingat theSouthwestIndian ridge. Earth and Planetary Science Letters 179, 31–51.

Dubińska, E., Bylina, P., Kozłowski, A., Dörr, W., Nejbert, K., Scastock, J., Kulicki, C., 2004.U–Pb dating of serpentinization: hydrothermal zircon from metosomatic rodingiteshell (Sudetic Ophiolite, SW Poland). Chemical Geology 203, 183–203.

Esperança, S., Crisci, G.M., 1995. The island of Pantelleria: a case for the development ofDMM-HIMU isotopic compositions in a long-lived extensional setting. Earth andPlanetary Science Letters 136, 167–182.

Fedele, L., Bodnar, R.J., DeVivo, B., Tracy, R., 2003. Melt inclusion geochemistry andcomputer modelling of trachyte petrogenesis at Ponza, Italy. Chemical Geology 194,81–104.

Ferguson, R.B., Traill, R.J., Taylor, W.H., 1958. The crystal structures of low-temperatureand high-temperature albites. Acta Crystallographica 11, 331–348.

Ferla, P., Meli, C., 2006. Evidence of magma mixing in the “Daly Gap” of alkaline suites: acase study from the enclaves of Pantelleria (Italy). Journal of Petrology 47,1467–1507.

Finetti, I., 1984. Geophysical study of the Sicily Channel Rift Zone. Bollettino di GeofisicaTeorica e Applicata 26, 3–28.

Fisher, R.V., Schmincke, H.-U., 1984. Pyroclastic Rocks. Springer-Verlag, Berlin, p. 472.Fulignati, P., Malfitano, G., Sbrana, A., 1997. The Pantelleria caldera geothermal system: data

from the hydrothermal minerals. Journal of Volcanology and Geothermal Research 75,251–270.

Gagnevin, D., Ethien, R., Bonin, B., Moine, B., Féraud, G., Gerbe, M.C., Cottin, J.Y., Michon, G.,Tourpin, S., Mamias, G., Perrache, C., Giret, A., 2003. Open-system processes in thegenesis of silica–oversaturated alkaline rocks of the Rallier-du-Baty Peninsula,Kerguelen Archipelago (Indian Ocean). Journal of Volcanology and GeothermalResearch 123, 267–300.

Giordano, D., Dingwell, D.B., 2003. Non-Arrhenian multicomponent melt viscosity: amodel. Earth and Planetary Science Letters 208, 337–349.

Giordano, D., Romano, C., Papale, P., Dingwell, D.B., 2004. The viscosity of trachytes, andcomparison with basalts, phonolites, and rhyolites. Chemical Geology 213, 46–61.

Giordano, D., Russell, J.K., Dingwell, D.B., 2008. Viscosity of magmatic liquids: a model.Earth and Planetary Science Letters 271, 123–134.

Grasso, M., Reuther, C.D., 1988. The western margin of the Hyblean Plateau: a neotectonictransform system on the SE Sicilian foreland. Annales Tectonicae 2, 107–120.

Hamilton, D.L., MacKenzie, W.S., 1965. Phase equilibria studies in the system NaAlSiO4–KAlSiO4–SiO2–H2O. Mineralogical Magazine 34, 214–231.

Harris, P.G., 1963. Comments on a paper by F. Chayes “Relative abundances ofintermediate members of the oceanic basalt–trachyte association”—a discussion.Journal of Geophysical Research 68, 5103–5107.

Harris, C., 1983. The petrology of lavas and associated plutonic inclusions of AscensionIslands. Journal of Petrology 24, 424–470.

Kar, A., Weaver, B., Davidson, J., Colucci, M., 1998. Origin of differentiated volcanic andplutonic rocks from Ascension Island, South Atlantic Ocean. Journal of Petrology 39,1009–1024.

Koepke, J., Berndt, J., Feig, S.T., 2007. The formation of SiO2-rich melts within the deepoceanic crust by hydrous partial melting of gabbros. Contributions to Mineralogyand Petrology 153, 67–84.

Lassiter, J.C., Hauri, E.H., Nikogasian, I.K., Barsxzus, H.G., 2002. Chlorine/potassiumvariations in melt inclusions from Raivavae and Rapa, Austral Islands: constraintson chlorine recycling in the mantle and evidence for brine-induced melting ofoceanic crust. Earth and Planetary Science Letters 202, 525–540.

Le Bas, M.J., Le Maitre, R.W., Streckeisen, A., Zanettin, B., 1986. A chemical classification ofvolcanic rocks based on total alkali–silica diagram. Journal of Petrology 27, 745–750.

Legendre, C., Maury, R.C., Savanier, D., Cotton, J., Chauvel, C., Hémond, C., Bollinger, C.,Guille, G., Blais, S., Rossi, P., 2005. The origin of intermediate and evolved lavas inthe Marquesas archipelago: an example from Nuku Hiva island (French Polynesia).Journal of Volcanology and Geothermal Research 143, 293–317.

Lentini, F., Carbone, S., Catalano, S., 1994. Main structural domains of the centralMediterranean region and their Neogene tectonic evolution. Bollettino di GeofisicaTeorica ed Applicata 36, 141–144.

Martin, R.F., 1970. Cell parameters and infrared absorption of synthetic high to lowalbites. Contributions to Mineralogy and Petrology 26, 62–74.

Martin, E., Sigmarsson, O., 2007. Crustal state and origin of silicic magma in Iceland: thecase of Torfajökull, Ljósufjöll and Snæfellsjökull volcanoes. Contributions toMineralogy and Petrology 153, 593–605.

Mazzoleni, P., Scribano, V., 1994. Preliminary geochemical information on selectedupper-mantle and lower-crust xenoliths from Hyblean Plateau (South EasternSicily). Mineralogica et Petrographica Acta 37, 295–305.

McDonough, W.F., Sun, S.S., 1995. The composition of the Earth. Chemical Geology 120,223–253.

Mohapatra, B.K., Sahoo, R.K., 1987. Merlinoite in manganese nodules from the IndianOcean. Mineralogical Magazine 51, 749–750.

Nawaz, R., Malone, J.F., 1982. Gobbinsite, a new zeolite mineral from Co. Antrim, N.Ireland. Mineralogical Magazine 46, 365–369.

Nielson, D.L., Sibbet, B.S., 1996. Geology of Ascension Island, South Atlantic Ocean.Geothermics 25, 427–448.

Nimis, P., Vannucci, R., 1995. An ion microprobe study of clinopyroxenes in websteriticand megacrystic pyroxenites from Hyblean Plateau (SE Sicily, Italy): constraints onHFSE/REE/Sr fractionation at mantle depth. Chemical Geology 124, 185–197.

Ottolini, L., Bottazzi, P., Vannucci, R., 1993. Quantification of lithium, beryllium andboron in silicates by secondary ion mass spectrometry using conventional energyfiltering. Analytical Chemistry 65, 1960–1968.

Ottolini, L., Bottazzi, P., Zanetti, A., 1994. Quantitative analysis of hydrogen, fluorine andchlorine in silicates using energy filtering, SIMS IX. In: Benninghoven, A., Nihei, Y.,Shimizu, N., Werner, H.W. (Eds.), Proceedings of the 9th International Conferenceon Secondary Ion Mass Spectrometry, Yokohama (Japan). John Wiley & Sons,Chichester (England), pp. 191–194.

Pearce, J.A., 1996. A user's guide to basalt discrimination diagrams. In: Wyman, D.A.(Ed.), Trace Element Geochemistry of Volcanic Rocks: Applications for MassiveSulphide Exploration. Short Course Notes, vol. 12. Geological Association of Canada,pp. 79–113.

Pearce, J.A., 2008. Geochemical fingerprint of oceanic basalts with applications toophiolite classification and the search for Archean oceanic crust. Lithos 100,14–48.

Punturo, R., Scribano, V., 1998. Ultracoarse clinopyroxenites, websterites and phlopite-bearing lherzolite xenoliths from Valle Guffari (Buccheri, Sicily): petrographicaland mineralogical data. Acta Vulcanologica 10, 77–81.

Rautenschlein, M., Jenner, G.A., Hertogen, J., Hofmann, A.W., Kerrich, R., Schmincke, H.U.,White,W.M.,1985. Isotopic and trace element compositionof volcanic glasses fromtheAkaki Canyon, Cyprus: implications for the origin of the Troodos ophiolite. Earth andPlanetary Science Letters 75, 369–383.

Rocchi, S., Longaretti, G., Salvadori, M., 1998. Subsurface Mesozoic and Cenozoicmagmatism in south-eastern Sicily: distribution, volume and geochemistry ofmagmas. Acta Vulcanologica 10, 395–408.

Rodiguez-Losada, J.A., Martinez-Frias, J., 2004. The felsic complex of the VallehermosoCaldera: interior of an ancient system (La Gomera, Canary Islands). Journal ofVolcanology and Geothermal Research 137, 261–284.

Scaillet, B., Macdonald, R., 2006. Experimental constraints on pre-eruptive conditions ofpantelleritic magmas: evidence from the Eburru complex, Kenya Rift. Lithos 91, 95–108.

Schmincke, H.U., Behncke, B., Grasso, M., Raffi, S., 1997. Evolution of the NorthwesternIblean Mountains, Sicily: uplift, Pliocene/Pleistocene sea-level changes, paleoenvi-ronment, and volcanism. Geologische Rundschau 86, 637–669.

Scribano, V., 1986. 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., 1987. Origin of websterite nodules from some alkaline volcanic rocks ofHyblean Plateau (South Eastern Sicily). Periodico di Mineralogia 56, 51–69.

Scribano, V., 1988. Petrological notes on lower-crustal nodules from Hyblean Plateau(Sicily). Periodico di Mineralogia 57, 41–52.

Scribano, V., Manuella, F.C., 2007. Xenolith evidence for partial melting of hydro-thermally modified mafic/ultramafic rocks in the Hyblean lithospheric basement(South-eastern Sicily): general implications in the origin of different magmacompositions with misleading geodynamic signatures. Epitome 2, 141–142.

Scribano, V., Sapienza, G.T., Braga, R., Morten, L., 2006a. Gabbroic xenoliths in tuff-brecciapipes from theHyblean Plateau: insights into the nature and composition of the lowercrust underneath South-Eastern Sicily, Italy. Mineralogy and Petrology 86, 63–88.

Scribano, V., Ioppolo, S., Censi, P., 2006b. Chlorite/smectite–alkali feldspar metasomaticxenoliths from Hyblean Miocenic diatremes (Sicily, Italy): evidence for earlyinteraction between hydrothermal brines and ultramafic/mafic rocks at crustallevels. Ofioliti 31, 161–171.

Scribano, V., Viccaro, M., Cristofolini, R., Ottolini, L., 2009. Metasomatic events recordedin ultramafic xenoliths from the Hyblean area (Southeastern Sicily, Italy).Mineralogy and Petrology 95, 235–250.

Stevenson, R.J., Wilson, L., 1997. Physical volcanology and eruption dynamics ofperalkaline agglutinates from Pantelleria. Journal of Volcanology and GeothermalResearch 79, 97–122.

Tanguy, J.C., Condomines, M., Kieffer, G., 1997. Evolution of the Mount Etna magma:constraints on the present feeding system and eruptive mechanism. Journal ofVolcanology and Geothermal Research 75, 221–250.

Tonarini, S., D Orazio, M., Armenti, P., Innocenti, F., Scribano, V., 1996. Geochemicalfeatures of Eastern Sicily lithosphere as probed by Hyblean xenoliths and lavas.European Journal of Mineralogy 8, 1153–1173.

Trua, T., Esperança, S., Mazzuoli, R., 1998. The evolution of the lithosperic mantle alongthe N. African plate: geochemical and isotopic evidence from the tholeiitic andalkaline volcanic rocks of the Hyblean Plateau, Italy. Contributions to Mineralogyand Petrology 131, 307–322.

Vai, G.B., 2003. Development of the palaeogeography of Pangea from Late Carboniferousto Early Permian. Palaeo 196, 125–155.

Veksler, I.V., Fedorchuk, Y.M., Nielsen, T.F.D., 1988. Phase equilibria in the silica-undersaturated part of the KAlSiO4–Mg2SiO4–Ca2SiO4–SiO2–F system at 1 atm andthe larnite–normative trend of melt evolution. Contributions to Mineralogy andPetrology 131, 347–363.

Vetere, F., Behrens, H., Misiti, V., Ventura, G., Holtz, F., De Rosa, R., Deubener, J., 2007. Theviscosity of shoshonitic melts (Vulcanello Peninsula, Aeolian Islands, Italy): insightson the magma ascent in dikes. Chemical Geology 245, 89–102.

Viccaro, M., Cristofolini, R., 2008. Nature of mantle heterogeneity and its role in thegeochemical and volcanological evolution of Mt. Etna (Italy). Lithos 105 (3–4),272–288.

Walker, G.P.L., 1962. Garronite, a new zeolite, from Ireland and Iceland. MineralogicalMagazine 33, 173–186.

671M. Viccaro et al. / Lithos 113 (2009) 659–672

Author's personal copy

White, J.C., Parker, D.F., Ren, M., 2009. The origin of trachyte and pantellerite fromPantelleria, Italy: insights frommajor element, trace elements, and thermodynamicmodeling. Journal of Volcanology and Geothermal Research 179, 33–55.

Winchester, J.A., Floyd, P.A., 1977. Geochemical discrimination of different magma seriesand their differentiation products using immobile elements. Chemical Geology 20,325–343.

Winter, J.K., Ghose, S., Okamura, F.P., 1977. A high-temperature study of the thermalexpansion and of the anisotropy of the sodium atom in low albite. AmericanMineralogist 62, 921–931.

Wood, C.P., Browne, P.R.L., 1995. Chlorine-rich pyrometamorphic magma at WhiteIsland volcano, New Zealand. Journal of Volcanology and Geothermal Research 72,21–35.

672 M. Viccaro et al. / Lithos 113 (2009) 659–672