Zircon UPb and Hf isotopic constraints on the genesis of a post-kinematic S-type Variscan tin granite : the Logrosán cupola ( Central Iberian Zone )

The Variscan Orogeny produced widespread granites along the European Variscan belt. In relation to crustal thickening, post-collisional multiple Sn-bearing highly fractionated S-type leucogranites were emplaced. The Logrosán granite represents one of those granitic bodies and is the focus of this study. The Logrosán granite is located in the Central Extremadura Batholith, within the Central Iberian Zone (CIZ) and was emplaced during post kinematic stages of the Variscan Orogeny at ca. 308 Ma, as determined by combined U-Pb ID-TIMS and LA-ICPMS geochronology. The granitic body intruded the metasedimentary Schist Greywacke Complex (SGC) of Neoproterozoic age. A moderately evolved mediumto coarse-grained two mica leucogranite (Main Unit) and several highly evolved aplitic or microporphyritic units (Evolved Units) have been distinguished based on their petrography and whole rock geochemistry. Initial 87Sr/86Sr ratios vary from 0.7125 to 0.7286, whereas initial εNd shows a restricted range from -4.3 to -4.0 and εHf(t) ranges from +5.7 to -10.5 for Variscan-age zircons. Inherited zircons exhibit mostly Neoproterozoic ages and juvenile Hf-isotope composition (εHf up to +14.7) analogous to zircons from the SGC metasediments. The available geological, geochronological, geochemical and isotopic data allow us to propose partial melting of heterogeneous Neoproterozoic metasediments, similar to the outcropping SGC materials for the genesis of the Logrosán granite.


Introduction
The Variscan Orogeny generated a huge volume of granitic melts.The Iberian Belt shows the largest concentration of felsic magmatism within western Europe, mainly in its inner parts.The Central Iberian Zone (CIZ) is composed by vast granite batholiths drawing sub-concordant major linear arrays (Fig. 1a) (e.g., López Plaza and Martínez Catalán, 1987).During the last two decades, U-Pb zircon geochronology has been applied to different granite plutons of these batholiths (e.g., Fernández-Suárez et al., 2000;Dias et al., 1998).Much more recently, combined U-Pb and Lu-Hf isotope systematics on zircon grains have been used to identify source components and to constrain the age of heterogeneous inheritances incorporated in the granite genesis.This integrated isotope information contributed significantly to the discussions on the origin of the Central Iberian granites (e.g., Villaseca et al., 2012;Teixera et al., 2011).
In this work we have studied a small felsic cupola of the southern part of the CIZ related to two types of mineralizations: i) an intra-granitic D cassiterite-bearing vein-complex, usually related to greisen-like alteration, ii) an exo-granitic Prich ore of hydrothermal apatite (dahlite)-quartz veins (Vindel et al, 2014;Locutura et al., 2007).Other batholiths of this southern part of the CIZ have been dated by U-Pb zircon geochronology (Castelo Branco: Antunes et al., 2008;Cabeza de Araya: Gutiérrez-Alonso et al., 2011;Nisa-Alburquerque: Solá et al., 2009; Montes de Toledo: Orejana et al., 2012), but none of them have been studied by the combined U-Pb and Hf-isotope approach.Therefore, we have used a combination of zircon U-Pb CA-ID-TIMS and LA-ICP-MS geochronology to better constrain the age of the granite emplacement and of the inherited zircons.

Geological Setting
The Logrosán pluton is located in the Central Iberian Zone (CIZ) (Julivert et al., 1972) of the Iberian Massif, at the southeast of the Cáceres province (Spain) (Fig. 1a).Large volumes of granitoids were emplaced during the post-collisional stage of the Variscan Orogeny, mostly late-to postkinematically to the D 3 event (e.g., Dias et al., 1998).The Logrosán granite is one of the post-kinematic bodies of the Central Extremadura Batholith (Castro, 1985) that intrudes into epizonal domains of the CIZ.The Logrosán granite is a small body that represents a typical felsic cupola related to complex hydrothermal mineralizations.Tin-tungsten deposits associated with granites in the Iberian Variscan Belt occur mainly in the Central Iberian Zone and constitute one of the most important Sn-W metallogenic provinces of the European Variscan Belt.Sn mineralization befalls as cassiterite-quartz stockwork and greisen type ore in the Logrosán granite.The Logrosán granite intruded the Neoproterozoic metasedimentary series of the Schist Greywacke Complex (SGC) which is characterized in this area by a monotonous decimetric to centimetric alternation of greywackes and shales with minor presence of sandstones and conglomerates (Fig. 1b).A Variscan low-grade regional metamorphism (Chl-Bt zone) has affected these Neoproterozoic country rocks.Moreover, the emplacement of the Logrosán granite has produced a superimposed contact metamorphism characterized by an inner biotite-tourmaline hornfel zone, an intermediate zone of micaceous spotted slates, and a transition zone of recrystallized slates and some metaquartzite levels (Fig. 1c).
The dominant regional structural pattern in the area is Variscan in age.The main Variscan deformation phases defined in the CIZ are also recognized in the area.D1 and D2 deformation phases correspond to the collisional stage of the Variscan orogeny and the crustal thickening which is associated with partial melting and a restricted production of peraluminous granitoids (Dias et al., 1998).The metasediments were folded during the first Variscan phase (D1) which started at 360 Ma in the CIZ (Ábalos et al., 2002;Dallmeyer et al., 1997).D1 compressional structures are constituted by large, subvertical folds oriented N100/110E and associated with a schistosity (S1) parallel or transverse to the axial plane (Quesada et al., 1987).The emplacement of the Logrosán granite produced some deviations in the S1 orientation until it reaches E-W directions.A greenschist facies metamorphic grade is reached in the area.The second deformation phase occurred during Lower-Middle Carboniferous (330-310 Ma) (Martínez Catalán, 2011;and references therein).D2 event has an extensional character (Díez Balda et al., 1995) and it is characterized by recumbent folds associated with dextral shear zones (N40-N60) with an axial plane crenulation cleavage or schistosity (S2).The S2 schistosity is occasionally recognized in the studied area.The D3 phase took place between 320 and 312 Ma (López-Moro et al., 2012;Valle Aguado et al., 2005;Dias et al., 2002) and constitutes the last ductile regional deformation phase.It is characterized by the generation of open to tight vertical folds with subhorizontal axes and subvertical shear zones with sinistral (NW-SE) or dextral (NE-SW) wrench movement (Dias et al., 1998).In the studied area shear zones are dextral with an ENE-WSW direction.On a regional scale, it has been assumed that such D3 structures were developed in a distensive context related to gravitational collapse and generalized extension across the belt (Díez Balda et al., 1995;Rodríguez-Pevida et al., 1990).Granite plutons, such as the Logrosán granite, were intruded later or post-kinematically to the D3 phase.Finally, the D4 fragile deformation phase corresponds to extension tectonics that took place from Middle Carboniferous to Permian times (Dias et al., 1998).In the studied area, D4 structures are characterized by the reactivation of old D3 faults in the opposite direction.These faults provided active channels for the flow of mineralized fluids and are almost certainly related to the tin ore vein system of the Logrosán granite.

Field relations and petrography
The Logrosán granite is an evolved leucogranitic apophysis with an outcropping area of about 4 km 2 (Fig. 1c).This body shows a sub-ellipsoidal shape elongated in the NE-SW direction, and a hidden larger volume of granite rocks may be assumed by the extent of the contact-metamorphic area.The Logrosán pluton is mainly composed of a coarse-to medium-grained two-mica leucogranite (Main Unit) (Fig. 2a) which grades to fine-grained or porphyritic varieties (Fig. 2b).Where K-feldspar megacrysts appear, flow textures are visible with a N130 main trend, following the contact of the granitic pluton in the most external areas.The two summits of the hill defined by the Logrosán granite (the San Cristóbal hill) are composed of a microporphyritic (two-texture) granite and an aplitic granite (Evolved Units) (Fig. 2c).No sharp contacts between the Evolved Units and the Main Unit could be observed.Pegmatitic and fine-grained leucocratic bodies of variable size are frequent and usually showing sharp contacts with the Main Unit.Micaceous, mostly biotitic enclaves are occasionally found, as well as schlieren structures.The bulk of the stock is composed of quartz, plagioclase, K-feldspar, muscovite, biotite and accessory minerals (tourmaline, apatite, zircon, monazite, ilmenite, Nb-rich rutile).
Tourmaline appears as an accessory phase within most of the Logrosán granite units, even in some pegmatitic miaroles.Tourmaline is also associated with all types of quartz-veining systems that cross-cut the granite body and occurs within the granite-metasediment contact.The Logrosán granite was triggered by complex hydrothermal events which are identified by the presence of sectors with a dense network of quartzrich veining (cm-scale in thickness of barren-, tourmaline-, cassiterite-rich varieties).Therefore, greisen-type alteration zones are found in the areas with a high proportion of mineralized veins and veinlets (Chicharro et al., 2011).Greisen alteration is not only restricted to the selvages but also affects some parts of the granitic body (Fig. 2d).This alteration is characterized by high contents of muscovite and the presence of disseminated cassiterite.A "sandy granite" is very common adjacent to the Sn-(W) veins and greisen alteration and in most cases is a result of old mining activity (processing/ panning for mineral recovery).

Analytical methods
A total of 9 representative samples weighted between 3-5 kg were collected for whole-rock geochemistry (6 medium to coarse-grained granites from the Main Unit and 3 finegrained or microporphyritic granites from the Evolved Units) (Table 1).Fresh and less altered fractions of each sample were selected for crushing and powdering.Each sample was fused using a lithium metaborate-tetraborate mixture.The melt produced by this process was completely dissolved with 5% HNO 3 .Major analyses were carried out using fusion-inductively coupled plasma mass spectrometer (FUS-ICPMS) while trace elements were analyzed by fusion-inductively mass spectrometer (FUS-MS) at Activation Laboratories (ACTLABS, Canada).Uncertainties in major elements are bracketed between 1 and 3% relative, except for MnO (5-  Individual zircon and monazite crystals were separated for geochronology and Lu-Hf isotopic studies from a mixture of two representative granite samples (AQ1 and AQ2, see Table 1).About 5 kg granite of each sample were crushed and sieved with a steel jaw-crusher and a disk mill at the Complutense University of Madrid.Zircon and monazite were preconcentrated with a Wifley table using a modified version of the "water-based" separation technique of Söderlund and Johansson (2002) at the Spanish Geological Survey laboratories (IGME, Tres Cantos, Spain).Further separation based on paramagnetic properties was done using a Franz isodynamic separator.Finally the zircon and monazite selected for analyses were hand-picked under a microscope.Several monazite and zircon crystals were carefully selected by picking the most idiomorphic crystals to avoid all possibility of inheritance for ID-TIMS.A representative selection of zircons were strewn and mounted on epoxy resin for LA-ICPMS microanalysis.The mount was polished to expose the zircon central portions and studied with transmitted and reflected light on a petrographic microscope.The internal structure, inclusions, fractures and physical defects were analyzed using back scattered electron (BSE) imaging.
The zircon fractions selected for ID-TIMS were pre-treated with the chemical abrasion (CA) method of Mattinson (2005).Zircon annealing was carried at 900 o C for 48 hours and the chemical attack was done in Parrish-type minibombs inside Parr bombs at 180 o C for 12 hours.Final zircon dissolution was achieved after placing the bomb at 240 o C for 72 hours.The procedure for extraction and purification of Pb and U is a scale-down version of that of Krogh (1973).A 208 Pb-235 U spike was used to obtain the U/Pb ratios by isotope dilution (ID).Isotopic ratios were measured with a Triton TIMS multi-collector mass spectrometer equipped with an axial secondary electron multiplier (SEM) ion counter.The instrument is set up to do measurements both in static and peak-jumping mode using the SEM.For static measurements the 204 Pb was measured with the calibrated SEM (92-93% Yield calibration).The Pb measurements were done in the 1300-1460 o C range, and U was measured in the 1420-1500 o C 10%).The precision for Rb, Sr, Zr, Y, V, Hf and most of the REE range from 1 to 5%, and between 5 and 10% for the rest of trace elements.Some granite samples have concentrations in transition metals below detection limits (V: 5 ppm, Cr: 20 ppm, Sc: 1 ppm, Co: 10 ppm) and almost all the granites have Ni < 20 ppm and Mo < 2 ppm.More information on the procedure, precision and accuracy of ACTLABS ICP-MS analyses can be found at www.actlabs.com.
Fluorine and Lithium were determined in selected samples in the laboratory of the Spanish Geological and Mining Institute (IGME), where F was determined by spectrophotometric methods after its extraction by pyrohydrolysis and Li was extracted by digestion with HF-HNO 3 -HClO 4 and determined by atomic absorption spectrophotometry.An analytical error of ± 10% has been estimated.
Stable isotope data were obtained at the Stable Isotope Laboratories of the Salamanca University (Spain).Whole rock samples of unaltered granite were analyzed for oxygen and hydrogen isotope composition (Table 2). 18O/ 16 O determinations for whole rock samples were carried out by laser fluorination using a conventional vacuum extraction line employing ClF 3 as the reagent (Borthwick and Harmon, 1982;Clayton and Mayeda, 1963).H 2 and H 2 O extraction for D/H isotopic analysis for whole rock samples were carried out by the uranium technique described by Godfrey (1962) with modifications introduced by Jenkin (1988) heating the samples at temperatures up to 1500 ºC.Oxygen and hydrogen isotope ratios were measured in a SIRA-II mass spectrometer and data are reported in the normal denotation relative to V-SMOW.
Sr-Nd isotopic compositions were measured on six representative granites (Table 2) at the Centro de Asistencia a la Investigación (CAI) of Geochronology and Isotope Geochemistry (Complutense University of Madrid, Spain).Whole-rock samples were dissolved in ultra-pure reagents and the isotopes were subsequently isolated by exchange chromatography.Isotope analyses were carried out using a Sector 54 VG Multicollector Thermal Ionization Mass Spectrometer with data acquired in multidynamic mode.Repeated analyses on the NBS-987 standard gave 87 Sr/ 86 Sr = 0.710240 ± 0.00005 (2σ, n = 8) and for the La Jolla standard, values  MS coupled to a New Wave UP213 laser source (213 nm frequency quadrupled Nd-YAG laser).The diameter of the laser beam was 10 µm.A 40 s gas blank was analyzed first to establish the background, followed by 50 s measurements for the remainder of the analysis.Each analysis was normalized to Si using concentrations determined by electron microprobe.Relative element sensitivities were calibrated with a NIST 612 glass standard.
In variation diagrams using TiO 2 , Logrosán samples do not define clear fractional crystallization trends, except for the MgO, which markedly decreases towards the more evolved samples (Fig. 3).Although there is some scatter in the, generally the Al 2 O 3 , FeO t , MgO, CaO and K 2 O contents decrease while the SiO 2 , Na 2 O and P 2 O 5 contents increase.Scatter is mainly found in a slightled albitized sample which behaves discordantly to other samples of the Main Unit (e.g.AQ6, see Table 1 and Figs 3 and 4).However, more evolved granites (Evolved Units) seem to form discrete trends for certain major elements such as CaO and K 2 O. Their depletion would be consistent with plagioclase and alkali feldspar fractionation, but they show different evolution trends for the Main and Evolved Units.
Granite REE patterns show a variable fractionation (La N / Yb N = 8.37-33.6)but are generally highly fractionated, with a negative Eu anomaly for granites of the Main Unit (Eu/ Eu*=0.32-0.42)and slightly more negative towards granites of the Evolved Units (Eu/Eu*=0.37-0.45)(Fig. 4).In trace-element chondrite normalized diagrams, the Logrosán samples show negative anomalies in Ba, Nb, Sr and Ti (Fig. 4).The negative Eu anomaly and the low Sr contents of the granite suggest that plagioclase fractionation has occurred for both Main Unit and Evolved Units.In variation diagrams Ba, Rb, Sr, Eu, Cs, Y, Th, HREE and LREE become depleted with increasing degree of fractionation (Fig. 3).As with major elements, Evolved granite Units show different trends for some elements (e.g., Rb, Fig. 3) compared to the Main Unit set of samples.
The HFSE show no clear trend with granite differentiation.A slight tendency of increasing Nb and Ta can be recognized, while Sn exhibits a wider scatter (Fig. 3).High Sn contents found in this granite with respect to the crustal average (i.e.granites with Sn contents higher than 10 ppm, as defined by interval (for further analytical details see Rubio-Ordóñez et al., 2012).Data reduction was done using the PbMacDat spreadsheet (Isacksen et al., 2007;www.earth-time.org).All isotopic ratios are corrected for mass fractionation, blank and initial common Pb after the model of Stacey and Kramers (1975).Ages and uncertainties were calculated with the decay constants of Jaffey et al. (1971), and are reported at the 2σ level.The concordia age was calculated and the data were plotted with Isoplot 3.0 (Ludwig, 2003).
U-Pb in situ age determinations were carried out on 37 polished zircons using a New Wave Research LUV213 laserablation microprobe, attached to an Agillent 7500 quadrupole ICP-MS at the ARC GEMOC Centre of the Macquarie University, Sydney (Australia).A laser beam of 30 µm diameter with energies of 60-100 mJ/pulse and 5 Hz repetition rate was shot during 100-120 s resulting in pits of about 30 µm deep.Real-time data were processed using the GLITTER ® software package.The correction factors were then checked using the GEMOC-GJ-1 with a TIMS age of 608.5 Ma, the Mud Tank zircon (734 ± 32 Ma, Black and Gulson, 1978) and the 91500 international zircon standard (1064 Ma, Wiedenbeck et al., 1995).Concordia diagrams (2σ error ellipses), concordia ages and upper intercept ages were calculated using the Isoplot/Ex software (Ludwig, 2003).
In situ Hf isotopic measurements were performed on 28 previously dated zircon spots at the ARC GEMOC Centre of the Macquarie University.Analyses were carried out using a New Wave Research LUV213 laser-ablation microprobe, attached to a Nu Plasma multi-collector (MC) inductively coupled plasma mass spectrometer (ICPMS).The laser system delivers a beam of 213 nm UV light from a frequency-quintupled Nd:YAG laser.The laser was fired with energy of 5-7 J/ cm 2 , laser beam diameter was 30 μm and repetition rate was 5 Hz.The laser beam ablated the zircon surface during 100-120 s resulting in pits 30μm deep.The analytical methods are the same as described in detail by Griffin et al. (2002;2004).To evaluate the accuracy and precision of the laser ablation results we have repeatedly analyzed two zircon standards: 91500 and Mud Tank (MT).These reference zircons gave 176 Hf/ 177 Hf = 0.282310 ± 0.000049 (2σ) and 0.282502 ± 0.000044 (2σ), respectively, which are identical to average published values of 0.282306 ± 0.000008 for 91500 and 0.282507 ± 0.000006 for MT (Woodhead and Hergt, 2005).The 2σ uncertainty on a single analysis of 176 Lu/ 177 Hf is ± 0.001-0.002%(about 1 epsilon unit), reflecting both analytical uncertainties and the spatial variation of Lu/Hf across many zircons.The 176 Lu decay constant value of 1.865x10 -11 a -1 was used in all calculations (Scherer et al., 2001).Chondritic 176 Hf/ 177 Hf = 0.282772 and 176 Lu/ 177 Hf = 0.0332 (Bouvier et al., 2008) and the depleted mantle 176 Hf/ 177 Hf = 0.28325 (εHf = +16.4)and 176 Lu/ 177 Hf = 0.0384 were applied to calculate εHf values and model ages used in this work.
The trace element zircon composition was obtained by laser ablation (LA-ICPMS) at the Natural History Museum of London (NHM, London, UK) using an Agilent 7500CS ICP- Flinter, 1971) allow classifying it as stanniferous granite.Ga and Hf data display trends where more fractionated samples show the lowest contents.The overall range for Ga and Hf concentration is quite restricted (Ga: 20-25 ppm and Hf: 1.5-3.0).

Oxygen isotopes
Whole rock and quartz δ 18 O results on Logrosán samples gave a narrow range between 14.1 and 15.0‰ (vs.SMOW) showing a positive correlation with SiO 2 of the rock (Table 2).δ 18 O values higher than 10 ‰ have been observed in many Variscan granites from western Europe (Tartèse and Boulvais, 2010;Hoefs and Emmermann, 1983) and are typical for high-SiO 2 peraluminous granites (Taylor, 1978).Similar δ 18 O values have been found in other granites from the southern CIZ (e.g., Castelo Branco Batholith: 12.2-13.7‰,Antunes et al., 2008).Nevertheless, these values are significantly higher than those found in the peraluminous S-type granites of the Spanish Central System, which mostly range from 8.3 to 10.2‰ (Villaseca and Herreros, 2000;Recio et al., 1992).

Sr-Nd isotopes
The measured Sr and Nd isotope ratios were recalculated back to 300 Ma based on the intrusion age determined in the section 5.5 (U-Pb geochronology).The Logrosán granite is characterized by a large variation in initial 87 Sr/ 86 Sr ratios (from 0.7125 to 0.7286) whereas εNd is much less variable (between -4.3 and -4.0) (Table 2).The Sm-Nd model ages were calculated using the equation of Liew and Hofmann (1988) (Table 2) (Fig. 5).The obtained model ages (1.31-1.67Ga) are similar to those given for other peraluminous monzogranites and leucogranites from the Central Extremadura Batholith (e.g., Antunes et al., 2008;Castro et al., 1999).
The BSE images show that most zircon grains have homogeneous or sector-zoned inner cores with dark thin rims (Fig. 6, e.g.L24-01).Stubby zircons sometimes present texturally discordant dark cores and/or fine euhedral oscillatory zoning (Fig. 6, L30-22).Logrosán zircons show negative Eu anomalies (Eu/ Eu*<0.3) and relatively high REE (647-2435 ppm), Nb (1.17-5.24ppm), Ta (0.31-3.92 ppm), Sc (157-939 ppm) and Hf (7631-11800 ppm) contents, characteristic of zircon from granitoid rocks (Hoskin and Schaltegger, 2003;Belousova et al., 2002) (Table 3).Zircon REE abundances normalized to chondrite values (McDonough and Sun, 1995) show steeplyrising slopes from the LREE to the HREE.Euhedral elongated bipyramidal primatic crystals (type-1) are generally Variscan-age zircons and usually richer in Hf than inherited zircons, which are mainly stubby (type-2) (Fig. 6).This is in accordance with the fact that the abundance of Hf in igneous zircons is considered as a marker of the degree of magma differentiation (Hoskin and Schaltegger, 2003).Hf correlates positively with P, Y, Th, U, Nb, Ta, HREE and LREE, whilst it shows a negative correlation with Zr/Hf (Fig. 7).The positive correlation of P, Y and HREE with Hf reveals that "xenotime" substitution mechanism is the dominant substitution in the Logrosán zircon (Speer, 1982).Experiments on Zr/Hf fractionation in zircon-crystallizing melts unravel a decrease of the Zr/Hf ratio of the residual melt related to an increase of the abundance of HfO 2 in zircon for fractional crystallization of peraluminous granitic melts (Linnen and Keppler, 2002).Then, the decrease in Zr/Hf with increasing Hf observed in the Logrosán zircon can be explained by crystal fractionation.Likewise, the behavior of incompatible elements, such as Th, U, Nb and Ta, which increase correlatively with Hf in zircon, denotes an enrichment of these elements in the melt as the magma evolves.

U-Pb geochronology
Seven zircon fractions and one monazite fraction were analyzed by ID-TIMS.All zircon fractions were pre-treated to remove Pb loss, before final dissolution, with the chemical abrasion (CA) method of Mattison (2005).Only one zircon fraction (Z7), a single grain, plots in a discordant position which could be attributed to partial secondary Pb-loss (see Krogh, 1982).The remaining data are either concordant or scattered due to the presence of older xenocrystic zircon.
Five fractions (zircon fractions Z3, Z5, Z8, Z6 and monazite fraction M1) are concordant and plot between 307 and 310 Ma.The remaining zircon fractions (Z1, Z2 and Z4) are discordant due to the presence of inherited zircons.These are multigrain (made up of 25 crystals) to some single crystal fractions (Table 4).The concordant fractions Z3 and Z5 and the discordant fractions Z2 and Z4 define a mixing line (Line 1; The age of the intrusion of the Logrosán granite is constrained by the cluster of analyses on the Concordia curve between 307 and 310 Ma (fractions Z3, Z5, Z8, Z6 and M1; Table 4).Fraction Z6, a single zircon, is slightly older suggesting the possible presence of an inherited component.Zircon fractions Z3, Z8 and Z5 and the monazite fraction M1 overlap at 307-308 Ma (Fig 8a).These four fractions provide a combined "Concordia" age of 307.88±0.86Ma with an MSWD of 1.8 (decay constant errors included).This age     ** Temperatures estimated using the Ti-in-zircon geothermometer recalibrated by Ferry and Watson (2007).Temperatures uncertainty for each data is ±4.5% Fig. 5.-Initial Sr-Nd composition of the Logrosán granite and representative samples of other granites from the Central Extremadura Batholith (Castro et al., 1999;Antunes et al., 2008;González-Menéndez and Bea, 2004), the Spanish Central System (Villaseca et al., 1998) and crustal protoliths of the Central Iberian Zone (Beetsma, 1995;López-Guijarro et al., 2008;Ugidos et al., 1997;Villaseca et al., 2014).All data have been recalculated to a reference age of 300 Ma.
is considered the age of zircon and monazite crystallization, and therefore our most accurate estimate for the age of the granite intrusion.
The LA-ICPMS data set is listed in Table 4 and plotted in a concordia diagram (Fig. 8b).Ages younger than 1,000 Ma are 204-corrected 206 Pb/ 238 U, whereas older ages are 204-corrected 207 Pb/ 206 Pb.Thirty seven analyses were performed, 32 of which yielded concordant ages ranging from 294 to 1975 Ma.Five analyses were rejected due to high common-Pb or degree of discordance.Nineteen analyses yielded Variscan ages and 13 analyses yielded older-than-Variscan ages.A total set of 19 Variscan zircons provide a weighted average 206 Pb/ 238 U age of 303.0 ± 2.3 Ma (Fig. 8b).This weighted mean age is similar to that calculated using ID-TIMS data (307.88± 0.86 Ma) within uncertainty.We consider the ID-TIMS age of 307.88 ± 0.86 Ma as the most accurate estimate for the age of the intrusion.

Hf isotope zircon composition
The zircon Lu-Hf isotopic data collected during this study are summarized in Table 6 and plotted as a function of their crystallization ages in figure 9.This figure also includes published data for zircon from the Spanish Central System Batholith (Villaseca et al., 2012) and the Schist Greywacke Complex (Teixeira et al., 2011).Depleted-mantle model ages (T DM ) are useful to estimate the crustal residence age for the granite protolith (Andersen et al., 2002).T DM were calculated using the measured 176 Lu/ 177 Hf ratios, referred to a model depleted mantle with a present-day 176 Hf/ 177 Hf = 0.28325 and 176 Lu/ 177 Hf = 0.0384 (Griffin et al., 2000;2002).These T DM ages represent only a minimum age for the source of the host magma.Thus, a more realistic two-stage model (T DM2 ) has been used to estimate model ages of the source of the Logrosán granite.T DM2 were calculated assuming a 176 Lu/ 177 Hf ratio of 0.015 for the average continental crust (Griffin et al., 2000).
The Variscan zircon population yields initial 176 Hf/ 177 Hf ratios of 0.282299-0.282758which correspond to εHf (t) varying from +5.7 to -10.5, a range well outside of analytical uncertainties.The T DM2 range for Variscan zircon is accordingly wide and ranges from 928 to 1957 Ma but mostly between 1179 and 1594 Ma with a mean value of 1368 Ma which encompasses the values given by the whole rock Nd depleted model age (1.31-1.67Ga).Neoproterozoic inherited zircons show initial 176 Hf/ 177 Hf ratios of 0.281549 to 0.282735 which corresponds to εHf(t) of +14.7 to -29.7, with more frequent εHf(t) between +6.0 and -3.2.Meso to Paleoproterozoic zircons have 176 Hf/ 177 Hf initial ratios of 0.281326-0.282007and εHf(t) between -3.4 and -7.6.

Zircon saturation and Ti-in-zircon thermometries
Zircon saturation thermometry has been calculated based on the relationship between zircon solubility, temperature, and major element composition of the melt (Watson and Harrison, 1983).Zircon saturation thermometry of the Logrosán granite ranges from 699ºC to 777ºC and yields an average temperature of 742ºC (Table 1).Apparent temperatures for zircon crystallization have been also estimated using the Tiin-zircon thermometer (Ferry and Watson, 2007).The Logrosán granite petrography suggests simultaneous crystallization of quartz, zircon and ilmenite, and thus aSiO 2 = 1 and aTiO 2 < 1.An a(TiO 2 ) value of 0.6 based on the presence of ilmenite has been assumed.This would lead to underestimation of zircon crystallization temperatures by ≤50 °C (Watson and Harrison, 2005).Ti-in-zircon thermometry yields values from 744ºC to 1024ºC with an average temperature of 836ºC (Table 3), these values are markedly higher than the ones obtained by zircon saturation thermometry.

Granite fractional crystallization
The whole-rock geochemistry described above provides no evidence for a unique fractional crystallization trend to link the Main granite suite with the Evolved leucogranite Units.Hence, the Main and the Evolved Units are not re-lated by simple crystal fractionation.Sequential restite fractionation can be dismissed as mafic xenoliths are absent in the Logrosán granite and most elements do not vary linearly with SiO 2 contents (Chappell et al., 1987).The low Ba and Sr contents together with high Rb concentrations found in the granites suggest that they have undergone some fractional crystallization (Breaks and Moore, 1992), which is consistent with a cupola stockwork scenario.Source-rock heterogeneities could explain the compositional evolution of the Logrosán granite since a significant number of major and trace elements do not show any linear correlations (Fig. 3).Moreover, the great variability in the whole-rock initial Sr isotopic ratios ( 87 Sr/ 86 Sr: 0.7125-0.7286)and the Hf-isotope composition of the Variscan-age zircons from the Logrosán granite (from +5.7 to -10.5) indicates a heterogeneous magmatic system.The variation of more than 15 εHf units (Table 6) is the maximum range obtained in zircons from the Iberian Variscan granites (Villaseca et al., 2012;Teixeira et al., 2011).Besides, the lack of a single evolutionary path in terms of Sr, Ba, Rb, Eu and CaO between the Main and Evolved Units indicates that fractional crystallization of plagioclase and K-feldspar did not occur in a simple closed magmatic system.Magmatic recharge from a deeper and chemically heterogeneous magma reservoir could produce the more residual and evolved granites of the Logrosán cupola.Thus a complex multi-pulse granite system inefficiently mixed, formed by isotopically heterogeneous fractionated magmas that episodically replenish the Logrosan cupola is more likely than a single magma batch following closed in situ fractional crystallization processes.

Zircon saturation and Ti-in-zircon thermometries
The average temperatures obtained for the Logrosán Variscan-age zircon crystallization is 836 ºC, similar to other temperature estimates on S-type granites of the Central Iberian Zone (788 to 844ºC after Orejana et al., 2012).Lower temperature estimates (699-777 ºC, Table 1) have been obtained based on whole-rock zircon saturation.The average wholerock zircon saturation temperature (742ºC) is below the average Ti-in zircon temperature (836ºC) for the Logrosán granite.Granitoids rich in zircon inheritances are probably undersaturated with respect to zircon at the source and consequently their calculated zircon saturation temperatures are most likely underestimations of the actual temperature of crystallization (Miller et al., 2003).The zircon inheritances of the Logrosán granite may yield unrealistic zircon saturation temperatures.Hence, the average Ti-in-zircon temperature (836ºC) provides a better estimation of the temperature of the magma from which the zircon crystallized.

Inheritances and source constraints
Mineralogical and geochemical features of the Logrosán granite (e.g., high peraluminosity and low CaO contents) and the absence of mafic microgranular enclaves suggest a crus- tal-derived melt origin with a major contribution of aluminous metasedimentary sources (Dias et al., 2002;Chappell et al., 1991).Likewise, high whole rock δ 18 O values point to an 18 O-enriched sedimentary or metasedimentary protolith.Perphosphorous character may be inherited from a P-rich source (Villaseca et al., 2008;Rodríguez-Alonso et al., 2004) as indicated in other P-rich granites of the area (Antunes et al., 2008;Villaseca et al., 2008;Ramírez and Menéndez, 1999).This P-rich character is also consistent with the presence of phosphate mineralizations surrounding the granite and with dispersed phosphate mineralizations in the regional Schist Greywacke Complex (Vindel et al., 2014).Zircons from the Logrosán granite have Zr/Hf ratios ranging 41-55 with an average value of 48.According to Pupin (2000), the estimated signature of zircon from the continental crust is 36-45.However, Pérez-Soba et al. (2007) estimated the Zr/Hf ratios of zircons from migmatites and augen orthogneisses of the Spanish Central System and suggested a slightly higher range (Zr/Hf = 36-56) for pure crustal signatures.Hence Zr/Hf ratios of the Logrosán zircon fit well with crustal signatures.It is expected that the Zr/Hf ratio in a single magma series should be approximately constant (Wang et al., 2010).Therefore, the wide range of Zr/Hf values for such a small granitic body may indicate not only a different degree of differentiation but also a participation of several magma inputs.
The high initial 87 Sr/ 86 Sr ratios of the Logrosán samples are attributed to a significant participation of the Sr derived from crustal material as established for similar granitoids of the CIZ (Ruiz et al., 2008;Castro et al., 1999;Villaseca et al., 1998).Variability in 87 Sr/ 86 Sr data for the Logrosán granite (Fig. 5) may well indicate an origin by partial melting of a compositionally heterogeneous continental crust.The 87 Sr/ 86 Sr ratios of the Logrosán granite plot in an intermediate space in the Sr-Nd isotopic field drawn by other monzogranites and leu-cogranites from the Central Extremadura Batholith.The large variability in 87 Sr/ 86 Sr data at almost constant εNd values can hardly indicate mixing of different proportions of mantlederived magmas with crustal sources (Castro et al., 1999).The low negative initial Nd isotope ratios of the Logrosán granite (around -4.2), typical of other Central Extremadura Batholith granites (e.g., Castelo Branco, Antunes et al., 2008;Alburquerque, González Menéndez and Bea, 2004;Montes de Toledo, Villaseca et al., 2008), contrasts with other S-type granites from northern areas of the CIZ (Fig. 5).The metasediments of the SGC also show low Nd isotope signatures (at Variscan times) and might be the most appropriate protolith of the Logrosán granite (Fig. 5).
Inherited zircons in the Logrosán granite define an age range of ca.447-1975 Ma displaying predominant Neoproterozoic populations (Table 5).Despite the limited dataset (13 inherited zircons), it is still possible to distinguish two main groups: (1) Cambro-Ordovician and Neoproterozoic ages and (2) Meso-and Paleoproterozoic ages (Fig. 10).The first group of inheritance (Early Paleozoic and Neoproterozoic) is the most abundant and fits well with materials derived from the Cadomian orogeny.The minor amount of Mesoproterozoic zircon inheritances and the distribution of Nd and Hf model ages around 1.45 ± 0.2 Ga, either in Variscan granites and Neoproterozoic metasediments, have been attributed to the Saharan and Arabian-Nubian shields as possible supplying provinces for the Central Iberian Zone (e.g., Pereira et al., 2012;Villaseca et al., 2011;Bea et al., 2010;Henry et al., 2009).The only Ordovician age (ca.447 Ma) could be related to the Ordovician magmatism described along the Central Iberian Zone (e.g., Rubio-Ordóñez et al., 2012;Neiva et al., 2009;Solá et al., 2008;Bea et al., 2007).In general, the probability density curve of the Logrosán granite inheritances shows a broad overlap with the zircon U-Pb age distribution previously reported for the Schist Greywacke Complex ages  (Talavera et al., 2012;Teixeira et al., 2011;Gutiérrez-Alonso et al., 2003) (Fig. 10).This comparison again points to the Schist Greywacke Complex as the most probable protolith of the Logrosán granite.Magmatic zircons from the Logrosán granite have a significant wide range of εHf (t) values showing variations up to 15 εHf units, similar to those observed in samples of Variscan S-type granites from northern Portugal (Teixeira et al., 2011).The large range of negative εHf (t) values is in agreement with the S-type nature of the granite but there is also a significant proportion of positive εHf (t) values (Fig. 9).The large range of negative εHf (t) values denotes an origin by partial melting of heterogeneous crustal sources or by crustal contamination of a mantle-derived parental magma as proposed for other Variscan S-type granites from the Central Iberian Zone (Teixeira et al., 2011;Neiva et al., 2013).Although positive εHf (t) values tend to be interpreted as mantle-derived, this cannot be easily applied to the Logrosán granite.The absence of coeval mafic magmatism and the clearly peraluminous and perphosphorous character of the Logrosán cupola is more in agreement with the involvement of recycled juvenile material in the genesis of this granite.From this prospective, εHf values of Neoproterozoic zircons from the SGC metasediments (sample from Sabugal area, spots 21 and 70 of Teixeira et al., 2011), recalculated at the age of the Logrosán granite emplacement, yield positive values (up to +4.5) that overlap with those found in Variscan-age zircons from the Logrosán granite.Moreover, some inherited Neoproterozoic zircons within the Logrosán granite show strong positive εHf (t) values (up to +14.7, e.g.L24-33, Table 6) similar to zircon populations found in metasediments from the SGC (up to +11.7, spot Malcl-21 of Teixeira et al., 2011).Such correlation reinforces the genetic relationship with this probable protolith (Fig. 9) and is in agreement with the presence of a recycled mantle-derived component in the granite crustal source.
The comparison of the ages of the granites from the southern CIZ (those of the Central Extremadura Batholith) and the Logrosán granite shows a good overlap: 310 Ma for Castelo Branco (Antunes et al., 2008), 308 Ma for Nisa-Alburquerque (Solá et al., 2009), 314-298 Ma for the Montes de Toledo (Orejana et al., 2011), and 309 Ma for the Cabeza de Araya and Trujillo plutons (Gutiérrez-Alonso et al., 2011).They are also similar in age to the emplacement of the Los Pedroches batholith (314-304 Ma, Carracedo et al., 2009)   granite batholiths and plutons from the Spanish Central System and the southern CIZ are practically coeval.

Conclusions
The Logrosán cupola is a felsic, perphosphoric and strongly peraluminous (ASI=1.2-2)tin-granite (Sn=11-67ppm).Two distinct units can be distinguished using field, petrographic and geochemical evidences: (1) a coarse-grained monzogranite (Main Unit) and (2) highly fractionated leucogranite bodies mostly concentrated on the top of the pluton (Evolved Units).U-Pb zircon analyses yield a concordia age of 307.88 ± 0.86 Ma which is considered the age of the emplacement of the Logrosán granite, coeval with other post-Variscan granites of the Central Iberian Zone.The Ti-in-zircon thermometry provides an average estimated temperature of 836ºC for the parental magma.The whole-rock geochemistry does not indicate that simple fractional crystallization may produce both the Main and the Evolved granite suites.Complex Sr, Ba, Rb, Eu and CaO trends rather suggest inputs of different felsic magma batches from a deep magma reservoir.The integration of whole-rock geochemistry with O, Sr, Nd and Hf isotopic signatures suggests that the Logrosán granite is the result of partial melting of heterogeneous metasedimentary materials.And finally, zircon inheritance (mostly Neoproterozoic zircons) combined with the metasedimentary nature of the proposed protolith, allow to suggest that the Schist Greywacke Complex is the most appropriated source of the Logrosán granitic magma.(Gutiérrez-Alonso et al., 2003), sample of the Sabugal area (Teixeira et al., 2011), samples PNC1 and PNC2 (Pereira et al., 2012), and sample Ctc67 (Talavera et al., 2012).

Fig. 1 .
Fig. 1.-(a) Schematic geological map of the Logrosán granite and its location within the Iberian Massif.(b) Regional geological sketch displaying the studied area (modified from Rodríguez et al., 2008).(c) Detail of the Logrosán granite.Sampling points are indicated by open circles

Fig. 2 .
Fig. 2.-Field features of the Logrosán granite.(a) Two-mica granite with coarse to medium grain of the Main Unit.(b) Porphyritic sectors in the Main Unit granite.(c) Aplite of the Evolved Units and barren quartz veins cutting through it.(d) Greisen vein selvage.Note the pervasive muscovitization of the original granitic rock.
Fig 8a) which has an upper intercept of ca.1.1 Ga suggesting a Mesoproterozoic inheritance.The concordant fractions Z3 and Z5 and the discordant fraction Z1 also define a mixing line (Line 2; Fig 8a) which has an upper intercept of ca.550-560 Ma pointing to an additional Late Neoproterozoic inherited component.These are very long projections, so the validity of the age of the inherited component has been constrained by in-situ U-Pb ages (see LA-ICPMS data below).

Fig. 7 .
Fig. 7.-Selected trace elements plotted versus Hf content of Variscan-age and inherited zircons from the Logrosán granite

Fig. 8 .
Fig. 8.-(a) Concordia diagram of the ID-TIMS data.Zircon (Z) and monazite (M) fractions, respectively; within brackets, number of crystals in each fraction.(b) U-Pb LA-ICPMS data showing weighted average

Table 2
.-Sr-Nd and O-D isotopic data of the Logrosán granite

Table 4
.-U-Pb ID-TIMS data of zircon and monazite fractions from the Logrosán granite

Table 6 .
-Lu-Hf LA-MC-ICPMS data of zircons from the Logrosán granite