Nova Terra 52

foliation (S V ). The older set must be pre-Variscan (Cadomian) and its trend, according to the type-2 fold interference, should be orthogonal to superimposed Variscan folds. The largest identifiable fold of the Cadomian set is the open antiform (as inferred from the dome structure) that separated Cadomian nappes in the W from Cadomian nappes in the E, i.e. rocks affected by the Trujillanos detachment from rocks affected by the Carija detachment. Other minor folds of this set can be deduced from the ellipsoidal pattern of poles to Cadomian foliations (S C ) observed in stereo- graphic projection (Figure 8(b)), which results from the interference between Variscan and Cadomian folds with orthogonal axial planes. The lack of foliation parallel to Cadomian axial planes suggests this type of folds resulted from bending (crustal upwarping) during exten- sion, a process that fits well into the bivergent nature of late Cadomian extensional structures that affected the study area (note Cadomian folds trend normal to the extensional flow inferred for the detachments; Figure 9). Former interpretations of the lithostratigraphy of the Ediacaran Serie Negra Group did not consider any of its mafic-ultramafic ensembles as individual tectonic slices, but as part of a complex sequence deposited and then intruded in an arc setting (e.g. Sánchez-Lorda et al . 2016 and references therein). Recent works on some of these mafic-ultramafic ensembles have revealed that, besides their suprasubduction geochemical affinity (e.g. Arenas et al . 2018), their contacts with the rest of the Serie Negra Group are mechanical, they show different metamorphic evolution compared to the rest of the sequence around, and are discordantly covered by Cambrian or younger strata (Díez Fernández et al . 2019). These observations apply to our study case and preclude the classical inter- pretation of the Serie Negra Group as a coherent lithos- tratigraphic series. Previous models proposed that the mafic rocks of our Mafic-ultramafic Unit represent the foundations of an immature arc. Such model is based on the calc-alkaline geochemistry of the metaigneous rocks of the region plus isotopic data, and on the assumption that the entire set of rocks represents a single tectonic block (Bandrés et al . 2004). However, there are not rocks similar to those of the Mafic-ultramafic Unit elsewhere in the study area that may be seen as intrusive bodies within either the overlying or underlying units. There is no evidence of contact metamorphism in those same units either, and there are no xenoliths from the overlying or underlying units within the Mafic-ultramafic Unit. The negative anomalies in Nb, Ti, and Ta relative to REE, plus 87 Sr/ 86 Sr i ratios higher and ڙ Nd values lower than present-day N-MORB, suggest a subduction-related setting for the generation of magmas in the Mafic-Ultramafic Unit, with variable proportions of mafic, felsic and sedimen- tary rocks in their genesis (Bandrés et al . 2004). The current regional boundaries between all of the pre- Ordovician lithostratigraphic units are mechanical and crustal-bearing (formerly assumed as intrusive). Some of these contacts could reflect the functioning of rather different faults through time, which may explain part of their kinematic complexity observed in a preliminary analysis. In this sense, the Mafic-ultramafic Unit of the Mérida Massif separates two continental crustal blocks, each of them made by a different set of rocks, according to their lithostratigraphy. The internal structure of the mafic-ultramafic ensemble is defined by tectonic slices, including several layers of ultramafics at different struc- tural positions. The Lower Gneiss Unit lacks of mafic rocks, and yet it rests below hundreds of metres of mafic and ultramafic rocks. If the mafic-ultramafic ensemble is to represent the lower sections of an arc- related batholith (Bandrés et al . 2004), the lack of mafic rocks in the units underneath is likely due to their sub- sequent tectonic juxtaposition via underthrusting beneath the mafic ensemble. This idea, together with the qualitative strain distribution within the Mafic- ultramafic Unit, could also explain the internal repetition of the metaperidotites as the result of thrust imbricates (Figures 3 and 4). Overall, there is no reason to consider the entire set of pre-Ordovician rocks as a single tectonic block prior to Ordovician times, and the mafic-ultramafic ensemble emerges as a good candidate to represent a suture zone. This model does not conflict with the arc- related nature of any of the rocks on the region, but implies the development of an accretionary thrust (now observed as a major shear zone) within an arc-setting to close (suture) a marginal basin. The recognition of a full, non-modified section of an ophiolite is rare in the geological record. Most of the ophiolitic ensembles lack of some or many parts from a typical sequence (e.g. Munhá et al . 1986; Díaz García et al . 1999; Arenas et al . 2007, 2018; Sánchez Martínez et al . 2007, 2012), or these parts could even occupy anomalous structural position due to superimposed deformation. The Mafic-ultramafic Unit of the Mérida Massif includes most of the lithologies expected for the lower part of an ophiolite. However, a layer containing metaperidotites is several times repeated, thus occupy- ing different structural positions within the mafic- ultramafic ensemble relative to sections dominated by metagabbros and amphibolites. Such layout is compati- ble with the internal structure of the mafic-ultramafic ensemble as derived from tectonic imbrication of lower sections of an ophiolitic ensemble (Mérida Ophiolite). Based on geochronological and geochemical grounds, previous works suggested a Cadomian active setting for INTERNATIONAL GEOLOGY REVIEW 13 Tectonostratigraphy of the Mérida Massif