Late-Variscan multistage hydrothermal processes unveiled by chemical ages coupled with compositiona.

The scheelite skarn from Los Santos and the W-Au veins from El Cabaco district, located in the Spanish Central System Batholith (SCSB), are some of the best-known tungsten ore deposits in Spain. Uraninite is an accessory mineral in both deposits, which underwent several hydrothermal flow events. Chemical and textural characteristics, as well as electron microprobe U-Th-Pb uraninite chemical data from the different stages of the skarn and the vein-type mineralizations, are presented here. Based on these data the uraninite was able to be classified into two groups. Group I uraninite has an octahedral habit and occurs as inclusions in K-feldspar relicts of the leucogranite related to Los Santos skarn formation. It shows high Th (6.95 to 8.51wt.% ThO 2 ) and high Rare Earth Elements (REEs) contents (0.55 to 1.38wt.% ∑REE 2 O 3 ). Group II uraninite occurs i) associated to El Cabaco granite, in a greenish selvage-style greisen and its reddish envelope and in the mineralized rimming quartz veins and ii) in Los Santos high-temperature endoskarn and anorthite skarn, where it is associated with U-rich mica. This uraninite type has lower Th and ∑REE 2 O 3 contents than Group-I uraninite. The mineral chemistry and the assemblage and textural relationships suggest that Group-I uraninite is magmatic and the attained U-Th-Pb chemical age of 300±4Ma is interpreted as the magmatic age of the skarn-forming aplite granites in the western part of the SCSB. Group-II uraninite includes two events: i) hydrothermal uraninite, which yields an age of 295±2Ma, dates a strong alkali mobilization and early tungsten deposition and ii) a later hydrothermal process, around 287±4Ma, that resulted in sulfides and late scheelite precipitation and widespread silicification. Finally, the gold deposition is younger than this silicification according to textural criteria. Therefore, W-Au deposits in the western part of the SCSB were formed by superposition of several processes that took place some 15Ma after the skarn-forming granite crystallized. Comparable W, W-Au and U deposits in the Variscan orogenic belt show a similar timing of hydrothermal events, suggesting that the hydrothermal history was controlled by large-scale Late-Variscan tectonic processes.


INTRODUCTION
Uraninite is widely used for chemical dating as its high contents of Pb and U are easily measured by electron microprobe.Uraninite chemical dating may provide precise ages (e.g.Alexandre and Kyser, 2005;Finger et al., 2017;Förster, 1999;Förster et al., 2012;Kempe, 2003;Pal and Rhede, 2013;Romer et al., 2007;Votyakov et al., 2013), especially if silicate minerals shield uraninite from later oxidation, which would make uranium mobile.The combination of chemical dating with detailed textural and chemical studies of uraninite may provide good constraints on the age and the history of multiple hydrothermal metal redistribution (e.g.Pal and Rhede, 2013).
Magmatic and hydrothermal mineralization and associated altered rocks in the Spanish Central System Batholith (SCSB) include scheelite, wolframite and cassiterite mineralization in sulfide veins and greisens (Martín-Crespo et al., 2004;Vindel et al., 2000), Sn-W skarns (Casquet and Tornos, 1984), episyenites (Caballero et al., 1993), and scheelite sulfiderich veins (Tornos et al., 1993).At many locations, several different types of mineralization occur in sequence.Without isotopic dating, it is a priori unclear whether these different stages of mineralization are different expressions of the same general process and, thus, belong to one single event, or they are related to several different unrelated events.In either case, isotopic dating of particular mineral assemblages provides key information, as it makes possible to link mineralization to heat sources and tectonic events that operate on a local and regional scale.We study two W±Au±U deposits of the SCSB, i.e.Los Santos scheelite skarn and El Cabaco vein and greisen mineralization.Both deposits show uraninite associated with several mineral assemblages (López-Moro et al., 2007;Timón, 2008), making it possible to date different events of hydrothermal activity.
Los Santos scheelite skarn is a major tungsten deposit currently operated by Almonty Industries Inc.It is a mesozonal, calcic and reduced tungsten-bearing skarn deposit (Timón et al., 2007;Timón and Moro, 2007;Timón, 2008;Tornos et al., 2008) that evolved from a magmatic to a hydrothermal system.The hydrothermal system has been assumed to be distinctly younger than Variscan magmatism (Timón et al., 2009).
El Cabaco deposit consists of gold-and wolframitescheelite-bearing sulfide-rich veins that may form quartz veins and stockwork-selvage-style greisen (López-Moro et al., 2007).Geochronological and structural data and the nature of fluids associated to this deposit suggest that the hydrothermal activity is distinctly younger than the emplacement of the granitic rocks (Antona, 1991;Antona et al., 1994;González-Sánchez, 2003).For instance, muscovite from the salbands yielded a K-Ar age of 256±5Ma (Antona, 1991).
In this contribution we present chemical, textural and chemical age data of uraninite hosted by different metasomatic rocks in two contrasted W-Au deposits, namely Los Santos and El Cabaco deposits, which have undergone several events of hydrothermal overprint.We discuss the processes related to the formation of the different uraninite types and discuss their age in the context of other W-Au hydrothermal deposits from the SCSB and the Variscan orogen.

GEOLOGICAL SETTING
The studied deposits are located in the Central Iberian Zone (CIZ), which is part of the Iberian Massif, an internal domain of the Variscan belt (Fig. 1A).In the CIZ, the Proterozoic Gondwana basement is overlain by mainly siliciclastic sedimentary rocks that form a passive margin sequence (e.g.Fernández-Suárez et al., 2002;Gutiérrez-Alonso et al., 2005).The Precambrian-Cambrian metasedimentary sequence, known as the Schist-Greywacke Complex, consists of two major units: the Lower Unit, a succession of shales and sandstones, and the Upper Unit, which predominantly includes pelitic rocks and comprise two formations: Monterrubio and Aldeatejada formations (Díez Balda, 1986, 1980).The Monterrubio Formation (Fm.) comprises about 2000m of slates with intercalations of sandstones, conglomerates, and volcanoclastic rocks.The conformable overlying Aldeatejada Fm. is 1600m-thick and consists of slates with intercalations of laminated black shales and phosphorite conglomerates.The Aldeatejada Fm. is overlain by the Lower Cambrian Tamames Sandstone, the Tamames Limestone, and the Endrinal Shale.Above a first-rank discordance, these rocks are overlain by Ordovician to Devonian sedimentary rocks, which are exposed in NW-SE trending synclines.
The igneous rocks of the studied area are part of the SCSB (Fig. 1B), one of the most extensive and well-exposed granitoids in the Iberian Variscan belt.Granodiorites and monzogranites constitute more than 90% of the intrusive rocks, whereas the remaining 10% consist of basic and intermediate rocks, all of them defining a typical K-rich calc-alkaline association and outcropping in restricted locations (Moreno-Ventas et al., 1995).The tectono-metamorphic evolution of the northern half of the CIZ includes four main phases of ductile deformation (Díaz-Alvarado et al., 2013 and references therein): D 1 (upright and E-vergent folds coupled with contractional deformation), ~360-337Ma; D 2 (extensional detachments), ~337-316Ma; D 3 (upright folds and a new extensional episode), ~316-300Ma; and D 4 (upright folds and shear bands only described for the eastern part of the Spanish Central System (SCS)).Dating of zircon from Variscan granitoids reveals that most magmatism in the western part of the SCS occurred during the D 3 phase (309 to 303Ma) (Bea et al., 2004;Díaz Alvarado et al., 2013; Gutiérrez-   Carnicero, 1983 andBea et al., 1999).C) Geologic setting of Los Santos scheelite skarn deposit (map from Yenes et al., 1999).D) Geological sketch map of El Cabaco district with the location of the mineralized fractures and the drill-hole projection (Junta de Castilla y León, 1986).

Alonso
In La Alberca-Béjar granitic area (Fig. 1C, D), four texturally and mineralogically contrasting granite types occur; A: porphyritic biotite granite; B: porphyritic biotite±cordierite±muscovite granite; C: inequigranular biotite granite; and D: fine-grained granite and aplites (e.g.Ugidos et al., 1990).Rocks of the porphyritic facies A and B are dominantly biotite monzogranites with a higher content of cordierite and some muscovite in facies B rocks.In contrast to the rocks of the porphyritic facies, the facies C rocks have fewer alkali-feldspar megacrysts and less cordierite.Apatite, zircon, monazite and ilmenite are minor constituents in this granite type.Uraninite is rare and generally related to samples with a superimposed albitization (López-Moro et al., 2007).Facies D was identified at the scale of the mine.The grain size of these rocks is finer than for rocks of the other granite types and alkali-feldspar megacrystals are missing, biotite is less abundant and cordierite, andalusite, muscovite, apatite, zircon, monazite and ilmenite are minor constituents (Yenes, 1996).Spatial relations between different granite facies indicate that porphyritic granites are younger than facies C, as the former enclose enclaves of facies C rocks.The gradual contacts between porphyritic granites (facies A and B) suggest a limited time span between the emplacement of these two facies.Similarly, rocks of granites facies C and D show gradual contacts.Actually, granites of facies C and D may be related via differentiation (Ugidos et al., 1990).
Facies B and D granites occur in Los Santos skarn, but only facies D granite seems to be related to the formation of scheelite skarn, as fresh granite grades into endoskarntype alteration with the development of calcic plagioclase and clinopyroxene (Timón et al., 2009) (Fig. 1C).W-Auquartz veins of El Cabaco district, in contrast, are hosted in facies C and D granites (Fig. 1D).

EL CABACO AND LOS SANTOS ORE DEPOSITS W-Au-quartz veins from El Cabaco ore deposit
El Cabaco ore deposit is located in a NW-SE trending antiformal structure composed of Variscan granites at the core (facies C and D) and pelitic hornfels and metasedimentary rocks of the Monterrubio Fm. in the external parts.The entire area is discordantly overlain by Neogene sediments.El Cabaco deposit includes four steeply NE-dipping, SE to SSE striking swarms of quartz-veins (Fig. 1D).The thickness of the quartz-veins varies from 2mm up to 40cm, with most veins falling in the range of 2 to 10cm-width.These mineralized quartz-veins are hosted in the Variscan granitoids (facies C and D) and in rocks of the Monterrubio Fm.Quartz-veins in granitoids commonly show abundant muscovite that forms selvage-style greisen, which typically are enveloped by a reddish K-feldspar-enriched rim (due to hematite microinclusions) with traces of new albite and chloride, which are typical features of episyenites.Voids that formed during episyenitization were later filled by secondary silicification.Rare alkali-feldspar veins commonly do not show alteration haloes in the host rock.Quartz-veins hosted in the metapelites of the Monterrubio Fm. typically show a selvage of tourmaline with sulfides (mainly pyrrhotite).Quartz veins show evidence of a complex geological history with repeated openings and closings, and healing by quartz and carbonates.Basically, four types of ore assemblages can be distinguished in these vein swarms (Fig. 2 et al., 2007).
Uraninite is restricted to greenish selvage-style greisen and its reddish envelope but is absent in the core of the quartz-veins.In these quartz-vein envelopes, uraninite is hosted by plagioclase, muscovite or quartz, whereby uraninite is strikingly more abundant in the reddish and greenish envelopes than in the surrounding host granite.

Los Santos scheelite skarn deposit
Los Santos scheelite skarn is located along the contact between the granite and the Upper Vendian to Lower Cambrian metasedimentary rocks on the northern limb of the Tamames syncline (Fig. 1C).The skarn is made up of a series of meter-thick lenticular bodies of mineralized pyroxene skarn, which occur in a zone less than 1km-long.They display a very complex geometry that is controlled by northeast-southwest-trending faults and dikes.The scheelite skarn of Los Santos occurs in the massive limestone and dolostone beds of 50 and 200m-thick, respectively, Tamames Limestones Fm. and in the transition zone to the Tamames Sandstones Fm., which consists of dark fine-grained greywacke.The mineralized skarn includes prograde skarn and its retrograde alteration.The prograde exoskarn is made up of a thick, almost monomineralic mass of hedenbergite, accompanied by grossular and Mo-rich scheelite (Sch I).It is associated with granite facies C and D, which developed endoskarn with calcic plagioclase and clinopyroxene.The retrograde skarn has a complex paragenesis that has been divided into two substages (Timón et al., 2009).The first substage is made of anorthite, zoisite, apatite, titanite, quartz, sulfides (mainly pyrrhotite linked to strong silicification and epidotization) and subcalcic garnet associated with the leaching of early scheelite and the precipitation of Mo-poor scheelite (Sch II).The second substage is characterized by the formation of clinozoisite, prehnite, quartz, calcite, chlorite, white mica, zeolite and mineral phases belonging to the Bi-Te-Ag-S-Au system (Fig. 2).
In Los Santos skarn zone, uraninite occurs in the aplitic fragments, as well as in Ca-metasomatic rocks.Uraninite is hosted by K-feldspar crystals in the prograde endoskarn and is included in calcic plagioclase and associated with mineral phases typical of the retrograde stage of the skarn development.

METHODS
Samples were collected from Los Santos and El Cabaco ore deposits.A Geiger counter was used to select samples for alpha-autoradiography using cellulose nitrate film (Fig. 3A) (Basham and Easternbrook, 1977;Basham, 1981).Each sample was initially studied in polished thin sections (Fig. 3B).The petrographic observations to identify uraninite and determine its mode of occurrence were carried out using a Leica DM RD polarized microscope.High-resolution backscattered electron images were generated to examine within-grain compositional heterogeneity.
Major, minor and trace element compositions of uraninite were determined using a JEOL Hyperprobe JXA-8500F with a thermal field-emission cathode at Deutsches GeoForschungs Zentrum (GFZ) in Potsdam, Germany.The following analytical conditions were used: 20kV acceleration voltage, 20nA beam current, 1-2μm beam diameter, and 20s peak counting times for Pb, U and Fe, 30s for Y and Si, 40s for Ca and Th, and 50s for the REEs.The CITZAD routine in the JEOL software, which is based on the Φ(ρZ) method (Armstrong, 1995), was used for data processing.
There are several methods for chemical age calculation on the basis of the chemical composition of uranium-rich minerals (Bowles, 1990).All these methods assume that all lead contained in the mineral, uraninite in this case, is radiogenic and that there was no mobility of lead, uranium and thorium.Single point ages were obtained using the iteration of the formula: where t is the time in years, λ 232 =4.9475x10 -11 y -1 , λ 235 =9.8485x10 -10 y -1 and λ 238 =1.55125x10 -10 y -1 are the decay constants of 232 Th, 235 U and 238 U, respectively (Steiger and Jäger, 1977).The constants k 0 =0.848485, k 1 =0.816367 and k 2 =0.0059475 account for the molecular weight ratios and isotopic abundances.Reproducibility was tested by repeated measurements of the same spot (e.g.Finger et al., 2017;Förster et al., 2012).
For chemical dating of uraninite, initial incorporation of Pb is irrelevant in comparison with in situ Pb growth.For that reason, this approach makes it possible to date single uraninite crystals and to combine the chemical composition with the age of different parts of individual grains to characterize and possibly date the hydrothermal alteration (Pal and Rhede, 2013).Uraninite ages were calculated with ISOPLOT 3.0 software (Ludwig, 2003).The accuracy of the dating procedure was monitored using uraninite of known ID-TIMS age (e.g.Finger et al., 2017).For age calculation we assigned an uncertainty of ±10Ma to each spot.A c t a , 1 7 . 1 , 1 -1 9 ( 2 0 1 9

URANINITE CHARACTERIZATION
Based on chemical, textural and morphological characteristics, which are directly related to mineral assemblages, two uraninite groups are differentiated: Group-I and Group-II in Los Santos skarn and Group-II in El Cabaco quartz veins.Characteristics of the two types of uraninite are described below.Their chemical composition is shown in Table 1 and in Figures 4 and 5.

Group-II uraninite
Group-II uraninite differs from Group-I uraninite both in chemical composition and texture.The total oxide content of the Group-II uraninite varies between 88.72 and 95.15 (Table 1).Group-II uraninite have higher contents of UO 2 (84.29-89.54wt.%)(Fig. 4A, B) than Group-I Dating of Late-Variscan hydrothermal activity in NW Spain Dating of Late-Variscan hydrothermal activity in NW Spain 8 uraninite and lower contents of ThO 2 (0.41-3.61wt.%and ∑REE 2 O 3 (0.94-0.21wt.%)(Fig. 4A, B).Ca and Si contents are near or below detection limit (Table 1) and the FeO content is low (average=0.31wt.%);its relationship with ∑REE 2 O 3 and UO 2 is shown in Figure 5A, B.
In Los Santos district, Group-II uraninite forms small (10-15μm) round grains and occurs in the high temperature endoskarn, next to relicts of aplite granite (Fig. 6A), which was affected by K-feldspar alteration.Group-II uraninite grains from the endoskarn are included K-feldspar crystals and the uraninite radiation-damage aureole often contains U-rich micas and Fe-oxide-hydroxides (Fig. 6B, C).Group-II uraninite also occurs in the anorthite skarn, the retrograde alteration of Los Santos endoskarn (Fig. 6D).In the anorthite skarn, this uraninite is included in anorthite crystals, which are altered to epidote group minerals (Fig. 6E, F).In the metamict aureoles of the uraninite from the anorthite skarn, K-feldspar is missing.In Los Santos deposit, a large uraninite grain included in K-feldspar crystals, without sulfides in the mineral association, displays two compositional zones (Fig. 7).The compositional variation observed within single uraninite grains is indistinguishable from the groups defined on the deposit scale.
In El Cabaco ore deposit, Group-II uraninite occurs predominantly in reddish and greenish halos surrounding quartz veins (Fig. 8A-C) and more rarely in altered granite.Uraninite of this district can be hosted by plagioclase, white mica or quartz, and individual uraninite crystals may reach up to 120μm (Fig. 8B).
In both districts, Group-II uraninite also occurs in silicified zones typically associated with sulfides and scheelite (Figs. 6, 8) and the composition of these uraninite grains is similar to that of the rest of Group-II, except for the lower PbO (3.33-3.52wt.%)contents, which reflect the younger age of this uraninite.

CHEMICAL DATING OF URANINITE
Electron probe dating of uraninite shows age variations from 282 to 309Ma (Table 1).Group-I uraninite yields chemical ages in the range of 297-309Ma.The average age of Group-I uraninite is 300±4Ma.The oldest age of 309Ma is obtained from a Group-Ia uraninite included in a K-feldspar crystal in the leucogranite.Group-II uraninite yields generally younger chemical ages that range from 282 to 302Ma (Table 1) with some overlap with Group-I uraninite.Differentiating Group-II uraninite according to mineral paragenesis and textural relations indicated that there may be two younger groups of uraninite with average ages of 296±2Ma, specifically uraninite in association with anorthite in Los Santos Skarn and secondary albite in El Cabaco, and 287±4Ma for uraninite associated with the silicification event with sulfides and tungsten in both deposits (Fig. 5C).
Chemical ages also vary within individual uraninite grains (Fig. 7).Zone A, with Group-I composition, yields a chemical age of 298Ma, and Zone B, with Group-II composition, yields a chemical age of 286Ma.The older age is obtained for a compositionally homogeneous zone that shows octahedral crystal habit, whereas the younger age is recorded in a zone where uraninite crystals are anhedral to subhedral.

Compositional variations in uraninite
Uraninite, UO 2 , may contain considerable amounts of ThO 2 (up to 11wt.%),REEs (up to 4wt.% ∑REE), Y (up to 1.3wt.%),and depending on age-radiogenic lead (Finch and Murakami, 1999;Förster, 1999;Frimmel et al., 2014;Janeczek and Ewing, 1992).Uraninite is a chemically active mineral and may change its composition as it recrystallizes during interaction with fluids (Alexandre and Kyser, 2005;Finch and Ewing, 1992;Kempe, 2003;Pal and Rhede, 2013).The two uraninite-types differentiated in this work could represent two different events of uranium mineralization or successive modifications of pristine uraninite by superimposed hydrothermal processes.It is commonly accepted that thorium and REEs are not mobile in fluids (Keppler and Wiley, 1990) and high contents in Th and REEs have been documented from magmatic uraninite [ThO 2 ranging from 5.6 to 11wt.% in biotite granites (see Förster, 1999); REEs between 0.05 and 4wt.% (see Frimmel et al., 2014)].Therefore, Group-I uraninite from Los Santos, with contents in thorium between 7 and 8.5wt.% and REEs ranging between 0.6 and 1.8wt.%,exhibits values of magmatic uraninite and may represent primary uraninite related to the magmatic system rather than hydrothermal uraninite.Other chemical features, such as the absence of CaO and SiO 2 , which points to the lack of coffininite or various uranyl silicates typical of alteration (Frimmel et al., 2014) and a U/Th ratio ca. 10 (magmatic uraninites have values <200, see Frimmel et al., 2014), also seem to be typical for uraninite that formed from granitic residual fluids (Hazen et al., 2009;Plant et al., 1999).Additionally, Kempe (2003) interpreted the association of uraninite with zircon, similar to that of Group-I uraninite and zircon in Figure 3D, as a magmatic feature, which also supports the magmatic character of the Group-I uraninite.Moreover, Group-Ib uraninite has slightly lower Th, Y and REEs contents than Group-Ia.Group-Ib uraninite could represent a late uraninite in relation to K-feldspar alteration, a modification of the pristine uraninite by fluids related to feldspatization, or a magmatic uraninite that is depleted in Th, Y and REEs because of the strong fractionation of these elements by Group-Ia uraninite or monazite.A magmatic formation of uraninite seems probable for two reasons: i) REEs and ThO 2 contents are in the range of magmatic uraninites; and ii) recrystallization of magmatic uraninite would release Th, Y and REEs, which are little mobile in fluids and, thus, should be sequestered in Th-Y-REEs-rich minerals close to the depleted uraninite.In contrast, Group-Ib uraninite is rimmed by Group-II uraninite, i.e. anhedral to subhedral Th-Y-REEs-poor uraninite (Fig. 7), which is not compatible with fluid interaction.Group-II uraninite from Los Santos skarn exhibits significantly lower Th and REEs contents than pristine Group-I uraninite.The low mobility of these elements in fluids suggests that Group-II uraninite is related to skarn formation.The low analytical totals of pristine uraninite probably reflect the presence of structural H 2 O or hydroxyl groups (Alexandre and Kyser, 2005) or the higher oxidation state of elements present (Zachariáš et al., 2008).
In El Cabaco deposit, no Th-REEs-enriched uraninite was found, and all crystals exhibit chemical features typical of Group-II uraninite of Los Santos skarn.Uraninite from El Cabaco deposit and Los Santos skarn shows the same age groups, even though ages are related to different locations and very different deposit types.This cluster of ages indicates that element redistribution occurred during distinct hydrothermal events related to large-scale tectonic processes, possibly operating at the scale of the Variscan orogen.
The alteration type in uraninite is a function, among other factors, of the chemical and physical characteristics of the available fluids, which in turn are related to hostrock lithology and structure.Of primary importance is the availability of elements in the fluid.Group-II uraninite shows Ca concentrations below detection limit.The low Si and Ca contents of uraninite in skarn may reflect the crystallization of calcium silicate minerals, such as epidote and anorthite, which acted as Si and Ca sinks, impoverishing the mineralizing solution and, thus, the uraninite in these elements.
Another factor to take into account is the redox state of the fluids.Uranium silicates tend to form under reducing conditions, whereas uranyl minerals form under oxidizing conditions (Finch and Murakami, 1999).Although later fluids in Los Santos scheelite skarn are oxidizing, the gradual interaction of magmatic fluids with the graphitic rocks of the Schist-Greywacke Complex resulted in a reduction in ore fluids (Timón et al., 2009) and enhanced formation of uranium silicates such as the crystals that appear in the radiation damage of Group-II uraninite (see Fig. 6B).Thus, the chemical composition of uraninite is partly controlled by the composition of the host-rocks that may affect the redox state of the fluids.

The significance of the chemical uraninite ages from Los Santos and El Cabaco in the context of the Spanish Central System Batholith
Taking into account the compositional and textural features shown above, it may be inferred that the ages between 309 and 297Ma (average=300±4Ma) calculated for Group-I uraninite display the age of magmatic uraninite crystallization.This age overlaps with zircon ages (309 to 303Ma) from granites surrounding Los Santos skarn (Bea et al., 2004;Díaz Alvarado et al., 2013;Gutiérrez-Alonso et al., 2011;Valle Aguado et al., 2005;Zeck et al., 2007).
The ages of hydrothermal Group-II uraninite (296±2 for K-alteration and 287±4Ma for assemblages with associated sulfides and tungsten) at Los Santos and El Cabaco may be placed within a broader scenario.Geochronological constraints of hydrothermal episodes in the western sector of the SCSB are scarce and not accurate, but several hydrothermal episodes related to the waning Variscan orogeny have been recognized in the eastern sector of the SCSB (Caballero et al., 1992;Galindo et al., 1994;Martín-Crespo et al., 2004;Tornos et al., 2000).These include W-Sn sulfides veins and greisens, Sn-W skarns, episyenites, scheelite sulfides-rich veins and greisens and barren quartz veins.
The oldest ages of the hydrothermal uraninite (ca.296Ma could be related to the development of the prograde skarn at Los Santos and the oxide stage at El Cabaco (Fig. 9).In both districts, uraninite yielding this age is associated with new alkali and early tungsten deposition.The earliest metasomatic alteration coincides in time with the first episode of hydrothermal activity recognized in the eastern SCSB during the Upper Carboniferous, ca.295Ma (Tornos et al., 2000).The older hydrothermal rocks are restricted to minor skarns, as is the case of Sn-W Carro del Diablo skarn, formed by the replacement of marbles near the contact with a ca.302Ma biotite granite stock (Casquet and Tornos, 1984).Another group of hydrothermal rocks is represented by showing two distinct zones.The brighter area (zone A) has higher ThO 2 (7.16wt.%)and ∑REE 2 O 3 (0.55wt.%) contents than the darker area (zone B) 3.61wt.%ThO 2 and 0.21wt.%∑REE 2 O 3 .Zone B appears surrounding the octahedral crystals of zone A. The composition of zone A uraninite resemble Group-Ib uraninite, whereas zone B uraninite is similar to Group-II uraninite.Numbers correspond to analysis numbers and to the calculated ages on individual spots given in Table 1.t a , 1 7 . 1 , 1 -1 9 ( 2 0 1 9 (Vindel et al., 1995) that have provided K-Ar ages between 302±6 and 291±8Ma (Caballero et al., 1992).Hydrothermal muscovites coeval with sulfides, using the 40 Ar/ 39 Ar method at As-(Ag) sulfide veins at Mónica Mine (Bustarviejo) provided 286±4Ma for this mineralizing event (Martín-Crespo et al., 2004).These rocks are commonly found in relation to apical zones of leucogranites (Tornos et al., 1993) and they formed during strike-slip faulting after the Variscan orogeny (González Casado et al., 1996).
Uraninites yielding a younger age (close to 287Ma) are associated with the greenish muscovite-enriched selvage of quartz veins of El Cabaco and the mineral association typical of the retrograde skarn at Los Santos (Fig. 9).They are associated with the precipitation of late scheelite linked to the main sulfides deposition and silicification process.A widespread precipitation of scheelite, arsenopyrite and metal base sulfides developed during the reactivation of W-quartz vein which occurred during the second hydrothermal event identified in the eastern part of the SCSB.A Permian age of this event (267±7Ma) was determined from K-Ar on muscovite (Tornos et al., 1993).Probably synchronously, another group of hydrothermal metasomatic rocks represented by barren episyenites developed ca.277Ma in the SCSB (Rb-Sr internal isochron ages; Caballero et al., 1996Caballero et al., , 1993)), an age that is like that reported in gold-and uraninitebearing episyenites close to the SCSB (270±12Ma, From the contact with the vein to the fresh granite, there is a selvage-style Greisen (Gr Zone), a Chlorite-rich zone (Chl Zone) and the granite transformed into episyenite (Reddish Zone).B) Backscattered electron image of a Group-II Uraninite (Urn) associated with Pyrite (Py).C) Backscattered electron image of Group-II Urn hosted by Albite (Ab).López-Moro et al., 2013).A similar age has also been reported for a subsolidus overprint recorded in allanite of vaugnerite rocks from the Tormes dome (275Ma, U-Pb ID-TIMS; López-Moro et al., 2017). 40Ar/ 39 Ar dating of muscovites of the granite-and gneiss-hosted barren quartz veins at the Guadarrama sector in the SCSB has provided ages of 274±5Ma (Martín-Crespo et al., 2002).Thus, the younger ages yielded by uraninite associated with abundant sulfide and silica deposition in Los Santos and El Cabaco districts could correspond to a second hydrothermal event recognized on a regional scale.
The last stage of mineralization, i.e. the gold stage, does not have associated uraninite (Fig. 2; González-Sánchez, 2003;Timón et al., 2009).Therefore, its age is not known.A maximum age is given by the ca.287Ma old sulfides stage.Gold mineralization, however, may be distinctly younger, as indicated by K-Ar muscovite ages, which range in the eastern SCS between 265±4 and 216±4Ma (Caballero et al., 1992).This younger reactivation may also account for the resetting of K-Ar muscovite ages from salbands of El Cabaco mineralized veins (255-256±5Ma;Antona, 1991), 30Ma younger than the U-Pb chemical ages of uraninite.(Fig. 9).

Late-Variscan hydrothermal mineralization
Paleozoic tin-tungsten mineralization in Central and Western Europe and Atlantic North America define a discontinuous belt of several thousand kilometers in length.Mineral deposits within this belt are related to the Variscan-Appalachian orogenic belt, which was formed by the closure of the Rheic Ocean during the collision of Gondwana with Laurussia and the final assemblage of Pangea, and define three age groups, i.e. 400-360Ma, 335-310Ma, and 300-275Ma (Romer and Kroner, 2016 and reference therein).Mineralization of the 300-275Ma age group is related to a reorganization of the plate movement between Gondwana and Laurussia, The ages of granodiorites are from Orejana et al. (2012) andDíaz Alvarado et al. (2013).All Rb-Sr ages were recalculated using the decay constant recommended by IUGS (Villa et al., 2015).
which resulted in widespread extensional tectonics in Central Europe (Chen et al., 1993;Cuney et al., 2002;Marignac and Cuney, 1999;Romer et al., 2010).Crustal extension was associated with localized heat input from the mantle that accounts for Sn and W mineralization in the Variscan foreland, i.e.Cornwall and Slovakia, and locally also in the internal parts of the Variscan orogeny, i.e. the Saxo-Thuringian Zone, the French Massif Central and NW Iberia (Fig. 10).However, in internal parts of the Variscan orogen, there is also an older (325-318Ma), significantly more important, generation of Sn and W mineralization that is related to granitic magmatism after the continental collision (Förster et al., 1999;Romer et al., 2007).
For high-temperature Sn and W mineralization of Central and Western Europe, isotopic dating shows close correspondence between the age of granite emplacement and mineralization (Chesley et al., 1993;Chicharro et al., 2016;Darbyshire and Sheperd, 1985;Romer et al., 2007).There is, however, increasing evidence of important hydrothermal mineralization that shows spatial closeness to igneous rocks, but which is much younger than the igneous activity.The spatial distribution of these hydrothermal mineralizations may indicate that the igneous rocks were involved as metal sources for these younger mineralizations or affected the distribution of the structures reactivated during younger tectonic processes.These younger hydrothermal mineralizations are structurally controlled.As these regionally widely separated mineralizations show a similar age distribution that closely corresponds to changes in the regional stress pattern affecting Central and Western Europe, they may have been controlled by the same overall process.For instance, the ca.300-275Ma old mineralization, formed when Central and Western Europe experienced crustal extension as result of the changing movement between Gondwana and Laurussia to close the remmants of the Rheic Ocean between North America and Africa (e.g.Kroner et al., 2016).Even younger reactivations of the same structural elements, in some regions with related mineralization, occurred during the opening of the Tethys and the North Atlantic (Cathelineau et al., 2012;Romer et al., 2010).
Examples for late-Variscan mineralization in the Saxo-Thuringian Zone include 300-290Ma old Sn-W greisen at Gottesberg, which shows a time gap of at least 20Ma with respect to the Eibenstock granite intrusion (Kempe et al., 2004) and in particular Permian fault-bound U mineralization that became redistributed during later tectonic events (e.g.Förster and Haak, 1995;Romer et al., 2010), as well as Triassic and younger Ba-F veins.Late-Variscan mineralization in the French Massif Central include tungsten mineralization from the Vaulry, Leucamp and Enguialès deposits (298-274Ma) (Harlaux et al., 2018) and polymetallic mineralizations such as W-Au-As and As-Sb-(Au) veins of Brioude-Massiac, Pb-Ag and As-Sn veins of Pontgibaud, and As-Pb-Sb-Au veins of Labessette (Bril et al., 1991;Marignac and Cuney, 1999).In Cornwall, high-temperature Sn-W mineralization related to granite yields ages of 293.3±1.2Ma,whereas hydrothermal mineralization is younger (Late Jurassic/Early Cretaceous; Chen et al., 1996).In the Iberian Massif, W-Mo-Sn quartz veins of the Carris deposit, north Portugal, yield a Re-Os molybdenite age of 279.4±1.2Ma,indicating that circulation of mineralizing fluids took place up to 1-4Ma after the emplacement of the Carris granite (Moura et al., 2014).Thus, although late-Variscan mineralization in part is related to intrusions, in part to hydrothermal veins, it is controlled by the same extensional processes that acted on the large scale.
The late-Variscan ages obtained for the hydrothermal activity at Los Santos and El Cabaco deposits suggest that the W-Au mineralization is younger than the granites of the SCSB and is related to hydrothermal processes along shear zones that became reactivated as the stress field changed during late-Variscan plate reorganization.Actually, available age data indicates that there may also be mineralizing events related to Jurassic and Cretaceous tectonic reactivation of these shear zones (e.g.Caballero et al., 1992;Cathelineau et al., 2012;Romer et al., 2010).As these later tectonic events may have resulted in the redistribution of metals within existing deposits, it is quite possible that different stages of mineralization were not related to the same mineralizing event.If these repeated reactivations of old structures occur under contrasting stress regimes, subsequent metal mobilization within a particular deposit may be controlled by different structural elements.

CONCLUSIONS
The integration of mineral assemblages with textural, chemical and age data of uraninite from W-Au mineralization demonstrates their close genetic link with spatially associated granites and multiple metal redistribution during recurrent hydrothermal activity.Using chemical and textural information, uraninite is distinguished into several distinct groups.Uraninite with high contents of Th and REEs, typical of magmatic uraninite, yields a chemical age (300±4Ma) that overlaps with zircon ages of late-to post-Variscan granites from the western part of the SCSB (309 to 303Ma).Hydrothermal uraninite is typically depleted in Th and REEs.It occurs in several different assemblages.Uraninite related to alkali mobilization and early tungsten precipitation yields the oldest ages.Subsequent to this mineralizing phase, the hydrothermal rocks underwent a second event of hydrothermal fluid flux that resulted in uraninite associated with the main stage of tungsten and sulfide deposition and the silicification process in both districts.Uraninites related to this hydrothermal process yield the youngest ages (287±4Ma), which are similar to the ages of the hydrothermal processes recognized at a regional scale.Although the ages of the two hydrothermal uraninite types are not clearly separated, mineral and textural constraints are consistent with the age relation among different uraninite types, i.e. old uraninite ages for alkali deposition and young uraninite ages for sulfide, tungsten and silica deposition.Based on mineral assemblages, deposition of gold occurred during this event, but it cannot be dated isotopically.Systematic dating in the Variscan belt of Europe has shown that the late Variscan hydrothermal activity was widespread.Some of these hydrothermal events were not related to magmatic sources, but to high geothermal gradients in an extensional tectonic setting during the Permian reorganization of plate movement.The large time span between hydrothermal activity and magmatism does not reflect a prolonged cooling history of the associated granites but reflects separate thermal events.D O I : 1 0 . 1 3 4 4 / G e o l o g i c a A c t a 2 0 1 9 . 1 7 .1 Dating of Late-Variscan hydrothermal activity in NW Spain ) D O I : 1 0 . 1 3 4 4 / G e o l o g i c a A c t a 2 0 1 9 . 1 7 .1 Dating of Late-Variscan hydrothermal activity in NW Spain 6 FIGURE 3. A) Autoradiography using Cellulose Nitrate (CN) film showing alpha track clusters of uraninite grains.B) Photomicrograph illustrating the metamict aureole of a Group-I uraninite.C) The endoskarn of Los Santos with relicts of aplite granite (sample S47-3).The dashed line delimits the aplite granite and the darker metasomatized zone (endoskarn).D) Backscattered electron image of Group-Ia Uraninite (Urn) hosted in a K-feldspar (Kfs) crystal.Zircon (Zrn), Anorthite (An) and aggregates of white Mica (Ms) can be observed.

FIGURE 5 .FIGURE 6 .
FIGURE 5. A-B)FeO vs. REE 2 O 3 and UO 2 plots allow to distinguish Group-Ia uraninite from the other groups.C) Group-Ia, Group-II display different age ranges in age vs. REE 2 O 3 plot.Group-Ib shows similar values as Group-II uraninite.Grey zone corresponds to uraninite from samples enriched in silica.

FIGURE 7 .
FIGURE 7. Backscattered electron image of a large uraninite grain showing two distinct zones.The brighter area (zone A) has higher ThO 2 (7.16wt.%)and ∑REE 2 O 3 (0.55wt.%) contents than the darker area (zone B) 3.61wt.%ThO 2 and 0.21wt.%∑REE 2 O 3 .Zone B appears surrounding the octahedral crystals of zone A. The composition of zone A uraninite resemble Group-Ib uraninite, whereas zone B uraninite is similar to Group-II uraninite.Numbers correspond to analysis numbers and to the calculated ages on individual spots given in Table1.
) D O I : 1 0 . 1 3 4 4 / G e o l o g i c a A c t a 2 0 1 9 . 1 7 .1 Dating of Late-Variscan hydrothermal activity in NW Spain 12 the W-(Sn)-bearing quartz veins and fracture-controlled greisens FIGURE 8. A) Hand specimen taken from the El Cabaco SA-8-1 drill-hole illustrating a mineralized quartz-vein, healed by Quartz (Qz), Arsenopyrite (Apy), Pyrite (Py) and Carbonates (Cb), hosted by inequigranular biotite granite.The vein mineralization includes As-Fe sulfides, scheelite and uraninite.From the contact with the vein to the fresh granite, there is a selvage-style Greisen (Gr Zone), a Chlorite-rich zone (Chl Zone) and the granite transformed into episyenite (Reddish Zone).B) Backscattered electron image of a Group-II Uraninite (Urn) associated with Pyrite (Py).C) Backscattered electron image of Group-II Urn hosted by Albite (Ab).