serie NOVA TERRA nº 49

132 during weathering, transport, diagenesis or metamorphism ( McLennan et al., 1983; Nesbitt et al., 1980; Taylor and McLennan, 1985; Wronkiewicz and Condie, 1987 ), thus offering a robust approach for prov- enance studies and geodynamic setting discrimination ( Bhatia and Crook, 1986; McLennan et al., 1990; Taylor and McLennan, 1985; Thorogood, 1990 ). This way a combination of structural, metamorphic, stratigraphic, and geochemical data seems adequate to overcome the uncertainties derived from paleogeographic studies dealing with high-P terranes (e.g., Albert et al., 2015b; Fuenlabrada et al., 2012; Mahlen et al., 2005 ). Following the closure of various oceanic basins, the collision between Gondwana and Laurussia formed the Variscan orogen during the Devonian and Carboniferous ( Franke, 1989; Matte, 1991 ). This orogen occupied the core of Pangea ( Bambach et al., 1980 ), which formed by a stepwise process of collision affecting terranes dispersed along the Gondwana margin (e.g., Stamp fl i et al., 2013; von Raumer et al., 2015 ). A complex internal zone separates the two main land- masses, and contains terranes with continental and oceanic af fi nity ( Arenas et al., 2007a; Ballèvre et al., 2009; Díaz García et al., 1999; Faure et al., 2009; Kroner and Romer, 2013; Quesada et al., 1994; Rossi et al., 2009; Schulmann et al., 2009 ), many of which are suspected to de fi ne a rootless suture zone that may be extended across the orogen ( Arenas et al., 2016a; Díez Fernández and Arenas, 2015 ). The southern part of the Variscan orogen, the so-called Iberian Massif, has been divided into allochthonous and autochthonous terranes according to structural and tectonometamorphic criteria ( Fig. 1 a; Arenas et al., 1986; Díez Fernández and Arenas, 2015; Martínez Catalán et al., 2007; Ribeiro et al., 1990 ). Separated by lithosphere-scale thrusts, a nappe stack of peri-Gondwanan continental and oceanic (ophiolitic) terranes was progressively juxtaposed to mainland Gondwana during the Variscan collision (e.g., Díez Fernández et al., 2016 ). Located at the base of the allochthonous tectonic pile, the so-called basal allochthonous units (commonly referred to as basal units) have been proposed as a co- herent continental terrane that experienced high-P metamorphism. This basal allochthonous terrane spreads under different ophiolitic units across the Iberian Massif and over sections of the Variscan orogen characterized by medium- to low-P metamorphism ( Fig. 1 b). The structural position of this terrane across Iberia was presented by Díez Fernández and Arenas (2015) (see also extended discussion by Díez Fernández and Arenas, 2016 as well as previous contributions by Arenas et al., 1986; Azor et al., 1994; Castro, 1987; Fonseca et al., 1999; Martínez Catalán et al., 1996; Pereira et al., 2009; Ribeiro et al., 1990; Ries and Shackleton, 1971 ). Díez Fernández and Arenas (2015) followed the trace of the tectonic pair made by this particular set of high-P units together with the allochtho- nous ophiolitic units through NW and SW Iberia, describing what they believe would represent an intra-Gondwanan suture zone related to the closure of a short-lived oceanic basin. According to Díez Fernández and Arenas (2015) , the basal allochthonous units of SW Iberia crop out in two domains ( Fig. 1 ): (i) a northern domain, referred to as the Central Unit ( Azor et al., 1994 ) and included in the Coimbra-Cordoba shear zone ( Burg et al., 1981; Pereira et al., 2010 ), and (ii) a southern domain including the Évora Massif ( Pereira et al., 2009 ). A fi rst quantitative approach to the internal coherence of the basal allochthonous terrane can be obtained by means of the equivalent timing of the high-P metamorphism that characterizes its initial Variscan orogenic record in NW and SW Iberia (e.g., Abalos et al., 1991; Abati et al., 2010; Booth-Rea et al., 2006; Moita et al., 2005; Ordóñez Casado, 1998; Rodríguez et al., 2003; Rosas et al., 2008; Rubio Pascual et al., 2013 ). However, this aspect does not provide information about whether the basal allochthonous terrane is a composite set of different continental units experiencing a similar Variscan evolution, or a large, yet single piece of continental crust involved in a complex subduction-exhumation system and then incorporated to the base of an allochthonous tectonic stack. In this contribution, new geochemical data (whole-rock and Sm – Nd) from the basal allochthonous units of SW Iberia (Évora Massif) are faced against data from equivalent tectonometamorphic units located in the allochthonous complexes of NW Iberia. Their comparison should prove if these particular allochthonous units de fi ne a single or a composite terrane, the answer having considerable implications for the amalgamation of Pangea. 2. Geological setting 2.1. The Évora Massif 2.1.1. Regional structure The Évora Massif (after Carvalhosa, 1983 ) is de fi ned by a two-layer crustal structure consisting of an underlying high-grade domain man- tled by a low- tomedium-grade domain ( Fig. 2 ). Both layers are exposed in an open, dome-like mega-structure that is transected by strike-slip shear zones and reworked by upright folds trending NW-SE ( Pereira et al., 2003, 2007, 2009 ). The low- to medium-grade domain occupies the cores of upright synforms and includes gneisses, schists, and am- phibolites formed during the Variscan evolution ( Chichorro, 2006; Pereira et al., 2007, 2008 ). Coring upright antiforms, the exposures of the high-grade domain are made of gneisses and migmatites that are closely associated with Variscan syn-kinematic granitic rocks and gabbro-diorites dated at Carboniferous ( Lima et al., 2013; Moita et al., 2009, 2015; Pereira and Silva, 2002; Pereira et al., 2009; Pin et al., 2008 ), some of which may also reach the structural levels of the low- to medium-grade domain. Those two domains are separated by exten- sional shear zones, which are responsible for regional telescoping of Variscan metamorphic isograds and the development of syn-orogenic sedimentary basins ( Pereira et al., 2009, 2012b ). In the central western part of the Évora Massif ( Fig. 2 a), the upright geometry of the Cabrela-Carvalhal synform reveals the structural stack of all the distinctive tectonostratigraphic units of this region ( Fig. 3 a and b; Chichorro, 2006 ). At the bottom of the tectonic pile, the high- grade domain does not contain evidence of a high-P metamorphic event, rather it shows a not well-preserved Variscan Barrovian evolution (garnet/staurolite isograd) followed by pervasive shearing at low-P/high-T conditions ( Chichorro et al., 2003; Pereira and Silva, 2002 ). On top of it, the lower part of the low- to medium-grade domain is occupied by a dominantly terrigenous sequence associated with orthogneisses and minor amphibolites and marbles ( Carvalhosa, 1965; Chichorro, 2006 ). This sequence contains lenses of metabasites trans- formed into eclogites in the Late Devonian (ca. 371 Ma, ca. 1.8 GPa; Leal, 2001; Moita et al., 2005; Pedro, 1996 ). Although the nature of the current tectonic contact between this sequence and the underlying high-grade domain is extensional ( Chichorro et al., 2003; Pereira et al., 2009 ), the juxtaposition of such a high-P domain on top of another domain that has only experienced medium- to low-P conditions (high- grade domain) requires a preexisting major thrust ( Díez Fernández and Arenas, 2015 ). The upper contact of the high-P sequence is also an extensional fault that separates it from an overlyingma fi c series ( Pereira et al., 2007 ). The whole sequence resting on top of this fault lacks of high-P metamor- phism (medium- to low-P Variscan evolution; Chichorro, 2006; Pereira et al., 2007 ), what points again to the existence of another major thrust, which would separate the overlying sequence from the underlying high-P sequence in the fi rst place. This preexisting fault would probably represent an accretionary thrust formed during the continental subduction process that affected the eclogite-bearing sequence. 2.1.2. Tectonostratigraphy of the allochthonous units The basal allochthonous unit with high-P rocks exposed in the Cabrela-Carvalhal synform comprises albite-bearing schists and paragneisses, banded metagreywackes, mica schists, and interbedded layers of black quartzites, amphibolites, orthogneisses and minor mar- bles ( Fig. 3 c). The age of the sedimentary protoliths of this sequence, particularly those occupying the lower parts, has been estimated as 286 R.D. Fernández et al. / Lithos 268 – 271 (2017) 285 – 301