Abstract During the Late Carboniferous and Early Permian an extensive magmatic province developed within northern Europe, intimately associated with extensional tectonics, in an area stretching from southern Scandinavia, through the North Sea, into northern Germany. Within this area magmatism was unevenly distributed, concentrated mainly in the Oslo Graben and its offshore continuation in the Skagerrak, Scania in southern Sweden, the island of Bornholm, the North Sea and northern Germany. Available geochemical (major- and trace-element, and Sr–Nd isotope, data) and geophysical data are reviewed to provide a basis for understanding the geodynamic setting of the magmatism in these areas. Peak magmatic activity was concentrated in a narrow time-span from c . 300 to 280 Ma. The magmatic provinces developed within a collage of basement terranes of different ages and lithospheric characteristics (including thicknesses), brought together during the preceding Variscan orogeny. This suggests that the magmatism in this area may represent the local expression of a common tectono-magmatic event with a common causal mechanism. Available geochemical (major and trace element and Sr–Nd isotope data) and geophysical data are reviewed to provide a basis for understanding the geodynamic setting of the magmatism in these areas. The magmatism covers a wide range in rock types both on a regional and a local scale (from highly alkaline to tholeiitic basalts, to trachytes and rhyolites). The most intensive magmatism took place in the Oslo Graben (ca. 120000 km 3 ) and in the NE German Basin (ca. 48 000 km 3 ). In both these areas a large proportion of the magmatic rocks are highly evolved (trachytes-rhyolites). The dominant mantle source componet for the mildly alkali basalts to subalkaline magmatism in the Oslo Graben and Scania (probably also Bornholm and the North Sea) is geochemically similar to the Prevalent Mantle (PREMA) component. Rifting and magmatism in the area is likely to be due to local decompression and thinning of highly asymmetric lithosphere in responses to regional stretching north of the Variscan Front, implying that the PREMA source is located in the lithospheric mantle. However, as PREMA sources are widely accepted to be plume-related, the possibility of a plume located beneath the area cannot be disregarded. Locally, there is also evidence of other sources. The oldest, highly alkaline basaltic lavas in the southernmost part of the Oslo Graben show HIMU trace element affinity, and initial Sr–Nd isotopic compositions different from that of the PREMA-type magmatism. These magmas are interpreted as the results of partial melting of enriched, metasomatised domains within the mantle lithosphere beneath the southern Olso Graben; this source enrichment can be linked to migration of carbonatite magmas in the earliest Paleozoic (ca. 580 Ma). Within northern Germany, mantle lithosphere modified by subduction-related fluids from Variscan subduction systems have provided an important magma source components.

Abstract In the early Carboniferous, final subduction of the Rhenohercynian Ocean, accretion of a magmatic arc and docking of microcontinents caused fault reactivation, extension and fault-controlled basin formation in the foreland of the Variscan Orogen. Lithospheric stretching resulted in generally mildly alkaline basaltic volcanism that peaked in the Visean. In the internal Variscides, rapid uplift and granitoid plutonism shortly followed collision and was probably due to slab detachment(s) or removal of orogenic root material. A regional-scale, E-W-oriented stress field was superimposed on a collapsing orogen and its foreland from the Westphalian onwards. In the Stephanian-Early Permian, a combination of outward-propagating collapse, mantle or slab detachment and the regional stress field resulted in widespread formation of fault-controlled basins and extensive magmatism dated at 290–305 Ma. In the foreland, large amounts of felsic volcanic rocks erupted in northern Germany, accompanied by mafic-felsic volcanics and intrusions in the Oslo Rift, and dolerite sills and dyke swarms in Britain and Sweden. In the internal Variscides, mafic rocks are rare and felsic-intermediate compositions predominate. Their apparent subduction-related signature may have been inherited from metasomatized mantle sources or caused by extensive assimilation of continental crust.

Abstract The Carboniferous-Permian evolution of NW Europe has recently been the focus of an EC-funded Training and Mobility of Researchers (TMR) project ‘Permo-Carboniferous-Rifting in Europe’ (PCR). One of the main goals of this project was to produce a new map for this time period showing the distribution of Late Carboniferous-Early Permian (Lower Rotliegend) volcanics, dykes and sills, and the extent of the tectonic structures of the Early-Late Permian (Upper Rotliegend) sedimentary basins (better known as the Southern and Northern Permian Basins). In order to produce this map, an overview of all the available literature was made. The new map was completed based on our own interpretations from seismic and borehole data. Unpublished data were available through industrial partners associated with the PCR project.

Abstract Models of the collapse of orogenic belts imply diachronous tectonism in which crustal uplift and extension may be compensated by peripheral compression. This first-order prediction is tested against published data on Varsican late-orogenic extensional and compressive structures along a 1500 km transect from the Variscan central internides in France to the foreland in the Irish Sea area. The collapse of the orogen is shown to have expanded northward over time, via three main stages: (i) collapse of the central internides (late Visean–mid-Westphalian, c. 335–310 Ma) – crustal thinning took place by NW–SE extension within a relatively narrow (< 500 km) central axis, accompanied to the north by passive infill of basins spanning a broad seaway inherited from extension during closure of the Rheic Ocean; (ii) reorientation and expansion of collapse (mid-Westphalian-late Stephanian, c. 310–300 Ma) – in the mid-Westphalian, a 90° rotation to NE–SW extension in the central internides was accompanied by changes across the northern internides (episodic basin formation and deformation), the externides (onset of thrust propagation) and the foreland (Westphalian C inversion), while from the early Stephanian, basins began to form in the central internides and expanded to the externides, coeval with final nappe emplacement along the orogenic front; and (iii) collapse of the foreland (late Stephanian-Early Permian, c. 300–290 Ma) – km-scale uplift and erosion of the foreland took place, prior to widespread basin formation in the Early Permian (<290 ± 5 Ma). These three stages are argued to support a model of Variscan late-orogenic collapse in response to three successive detachments of negatively buoyant lithospheric material: of a collisionally thickened orogenic root, and of two (Rheic) oceanic slabs subducted, previously, southward (beneath the orogen) and northward (beneath the foreland). Multiple detachments are a predictable consequence of ocean closure and continental collision, so that episodic collapse may be a common process in the rise and fall of orogenic belts and the tectonic evolution of their forelands.

Abstract Compared to other Permo-Carboniferous rift basins of NW Europe, the Oslo Graben has two distinct characteristics. First, it initiated inside cold and stable Precambrian lithosphere, whereas most Permo-Carboniferous basins developed in weaker Phanerozoic lithosphere, and second, it is characterized by large volumes of magmatic rocks despite relatively little extension. Seismic reflection surveys show that the crust thickens from southern Norway to southern Sweden, the most significant Moho deepening occurring from the Oslo Region eastwards. Deep seismic studies also suggest that the base of the lithosphere deepens markedly eastwards from the Oslo Region. Such a long-wavelength lithospheric geometry cannot be explained by the Permian or post-Permian evolution of the area, hence the Oslo Graben appears to have evolved at the transition between two lithospheric domains with contrasting thickness. Numerical thermo-mechanical modelling is applied to test if this transitional position can influence the dynamics of rifting. Different models with varying lithosphere thickness contrast are considered. Model results show that a crust and lithosphere thickness contrast comparable to the Oslo Region can explain rifting and focusing of magmatism in a narrow zone with minor thinning of the crust. Models also account for the major characteristics of the Oslo Graben in terms of location and strain distributions in the crust.

Abstract This study focuses on Late Carboniferous-Permian tectonics and related magmatic activity in NW Europe, and specifically in the Skagerrak, Kattegat and North Sea areas. Special attention is paid to the distribution of intrusives and extrusives in relation to rift-wrench geometries. A large database consisting of seismic and well data has been assembled and analysed to constrain these objectives. The continuation of the Oslo Graben into the Skagerrak has been a starting point for this regional study. Rift structures (with characteristic half-graben geometries) and the distribution of magmatic rocks (intrusives and extrusives) were mapped using integrated analyses of seismic and potential field data. For the analysis of the Sorgenfrei-Tornquist Zone and the North Sea, seismic and well data were used. The rift structures in the Skagerrak can be linked with extensional structures in the Sorgenfrei–Tornquist Zone in which similar fault geometries have been observed. Both in the Skagerrak and in the Kattegat, lava sequences were erupted that generally parallel the underlying Lower Palaeozoic strata. This volcanic episode, therefore, pre-dates main fault movements and the development of half-grabens filled with Permian volcaniclastic material. Upper Carboniferous-Lower Permian extrusives and intrusives have also been found in wells in the Kattegat, Jutland and the North Sea (Horn and Central grabens). Especially in the latter area, the dense seismic and well coverage has allowed us to map out similar Upper Palaeozoic geometries, although the presence of salt often conceals the seismic image of the underlying strata and structures. From the results, it is assumed that the pre-Jurassic structures below large parts of the Norwegian-Danish Basin and northwards into the Stord Basin on the Horda Platform belong to the same tectonic system.

Abstract The Permo-Carboniferous evolution of the central North Sea is characterized by three main geological events: (1) the development of the West European Carboniferous Basin; (2) a period of basaltic volcanism during the Lower Rotliegend (latest Carboniferous–early Permian); and (3) the development of the Northern and Southern Permian Basins in late Permian times. The timing of the late Carboniferous–Permian basaltic volcanism in the North Sea is poorly constrained, as is the timing of extensional tectonic activity following the main phase of inversion during the Westphalian, due to the diachronous propagation of the Variscan deformation front. Results of high precision Ar-Ar dating on basalt samples taken from a core from exploration well 39/2–4 (Amerada Hess) in the UK sector of the central North Sea suggests that basaltic volcanism was active in the late Carboniferous, at c. 299 Ma. The presence of volcanics below the dated horizon suggests that the onset of Permo-Carboniferous volcanism in the central North Sea commenced earlier, probably at c. 310 Ma (Westphalian C). This is contemporaneous with other observations of tholeiitic volcanism in other parts of NW Europe, including the Oslo Graben, the NE German Basin, southern Sweden and Scotland. Interpretations of available seismic data show that main extensional faulting occurred after the volcanic activity, but the exact age of the fault activity is difficult to constrain with the data available.

Abstract Extensional tectonics to the north of the Variscan Front during the Early Carboniferous generated fault-controlled basins across the British Isles, with accompanying basaltic magmatism. In Scotland Dinantian magmatism was dominantly mildly alkaline–transitional in composition. Tournaisian activity was followed by widespread Visean eruptions largely concentrated within the Scottish Midland Valley where the lava successions, dominantly of basaltic–hawaiitic composition, attained thicknesses of up to 1000 m. Changing stress fields in the late Visean coincided with a change in the nature of the igneous activity; subsequently, wholly basic magmatism persisted into the Silesian. As sedimentary basin fills increased, sill intrusion tended to dominate over lava extrusion. In the Late Carboniferous (Stephanian) a major melting episode, producing large volumes of tholeiitic magma, gave rise to a major dyke swarm and sills across northern England and Scotland. Alkali basaltic magmatism persisted into the Permian, possibly until as late as 250 Ma in Orkney. Geochemical data suggest that the Carboniferous–Permian magmas were dominantly of asthenospheric origin, derived from variable degrees of partial melting of a heterogeneous mantle source; varying degrees of interaction with the lithosphere are indicated. Peridotite, pyroxenite and granulite-facies basic meta-igneous rocks entrained as xenoliths within the most primitive magmas provide evidence for metasomatism of the lithospheric mantle and high-pressure crystal fractionation.

Abstract Twenty-one new 40 Ar/ 39 Ar step-heating experiments on mineral separates from intrusive and extrusive Carboniferous and Permian igneous rocks in the Midland Valley of Scotland yielded 17 concordant experiments with a relative age precision better than 1% (2σ). These ages resolve inconsistencies between existing K-Ar dates on the same samples and their stratigraphical constraints correlated to recently published timescales. The precise 40 Ar/ 39 Ar dates are stratigraphically constrained to stage level and can contribute to Carboniferous timescale tie points at the Tournaisian-Visean boundary, within the Visean and at the Carboniferous-Permian boundary. Situated in the extending Variscan foreland, two distinct phases of extension-related transitional-alkaline volcanism have been resolved in the Dinantian: the Garleton Hills Volcanic Formation in the eastern Midland Valley near the Tournaisian-Visean boundary, 342.1 ± 1.3 and 342.4 ± 1.1 Ma; and the Clyde Plateau Volcanic Formation in the western Midland Valley during the mid-Visean, 335 ± 2329.2 ± 1.4 Ma. Alkaline basic sills near Edinburgh, previously thought to be Namurian, appear to be coeval with the Clyde Plateau Volcanic Formation at 331.8 ± 1.3–329.3 ± 1.5 Ma. The new ages allow correlation between these short-lived Dinantian magmatic pulses and extensional and magmatic phases in the Northumberland-Solway and Tweed basins to the south. After late Westphalian, end-Variscan, compression and a regionally important tholeiitic intrusive phase at c. 301–295 Ma, alkaline magmatism related to post-Variscan extension occurred in the central and western Midland Valley during the latest Carboniferous or Permian from 298.3 ± 1.3 to 292.1 ± 1.1 Ma. This correlates well with post-Varsican extension and magmatism observed across the NW European foreland from 300 to 280 Ma.

Abstract Noble gas studies of well-characterized spinel-peridotite-facies lithospheric mantle xenoliths and garnet megacrysts from Scottish Permo-Carboniferous dykes, sills and vents demonstrate that the mantle beneath Scotland during the late Palaeozoic was more radiogenic than the source of mid-ocean ridge basalts (MORB). The samples were collected from the Northern Highland Terrane and the Midland Valley Terrane, which vary from Archaean-Proterozoic to Proterozoic-Palaeozoic in age. Helium isotope ratios of between 3 R a and 6 R a ( R a = atmospheric ratio) indicate that there has been time-integrated U-Th enrichment of the subcontinental mantle. This enriched mantle was preferentially melted following the transition from early Palaeozoic compression to late Palaeozoic extensional tectonics. Helium isotope ratios provide no clear evidence for the presence of undegassed plume-type mantle beneath this part of Scotland during the Permo-Carboniferous. The measured helium ratios do not discount the presence of a low-helium plume similar to those of the European Cenozoic volcanic province. A passive origin, however, is preferred for the Permo-Carboniferous magmatism due to the protracted activity, relatively small-extruded volumes of mafic magma and the low-helium isotope ratios measured.

Abstract Permo-Carboniferous rifting in Europe was accompanied by the widespread emplacement of mantle-derived magmas forming regional dyke swarms and sills in northern England, Scotland, Norway and southern Sweden during the late Stephanian and early Autunian. The trends of the dyke swarms intersect at a focal point in the Kattegat south of the Oslo Graben, and are probably all related to a single magmatic centre that could be plume-related. The WNW- to NW-trending dyke swarm at the SW margin of the Fennoscandian Shield in southern Sweden is composed mainly of tholeiitic dolerites, with lesser amounts of alkaline mafic rocks (camptonites, alkali basalts and spessartites) and trachytes. The alkaline mafic rocks are enriched in Ba, Sr, Nb, P and CO 2 , implying a metasomatic enrichment of their upper-mantle source prior to melting. After generation of alkaline melts by relatively small degrees of partial melting, increased extension was accompanied by the formation of subalkaline tholeiitic magmas. Whole-rock compositions (Mg-numbers between 55 and 30) and mineral chemistry (olivine Fo 60 -Fo 40 ; clinopyroxene approximately Wo 30 En 45 Fs 15 ; plagioclase An 70 -An 50 ) indicate relatively evolved melts that have undergone crystal fractionation of olivine and clinopyroxene. Two groups of dolerites can be distinguished on the basis of bivariate element plots, e.g. Zr-TiO 2 and La-Sm. Although both groups show enrichment in the whole range of incompatible trace elements, slight differences in mantle-normalized trace-element patterns and different Nb/La ratios suggest that they were generated from two different sources. Group I dolerites were formed from a (re-)enriched, but isotopically mildly depleted, sublithospheric garnet-bearing mantle source (Nb/La mostly > 0.9, εNd i = +4 to +3, where εNd i is the initial Nd isotope ratio), whereas group II dolerites seem to indicate mixing of the asthenospheric-derived magmas with lithospheric mantle melts (Nb/La mostly <0.9, εNd i = 0 to −1). Increasing Th/Ta ratios together with decreasing U/Nb ratios from group I towards group II dolerites further reflect progressive crustal contamination.

Abstract In the area of the Cenozoic Rhine rift system, crustal and lithospheric thicknesses range between 24 and 35 km, and 60 and 120 km, respectively. This rift system transects the deeply truncated Variscan Orogen and superimposed Permo-Carboniferous wrench-induced troughs, and Late Permian and Mesozoic thermal sag basins. At the time of its Westphalian consolidation, the Variscan Orogen was probably characterized by 45–60 km deep-crustal roots that were associated with its Rheno-Hercynian-Saxo-Thuringian, Saxo-Thuringian Bohemian and Bohemian-Moldanubian sutures, all of which are transected by the Cenozoic Rhine rift system. During the Stephanian-Early Permian wrench-induced disruption of the Variscan Orogen, subducted lithospheric slabs were detached causing upwelling of hot mantle material. During the resulting thermal surge, partial delamination and/or thermal thinning of the continental mantle-lithosphere induced regional uplift. At the same time the Variscan orogenic roots were destroyed and crustal thicknesses reduced to 28–35 km in response to the combined effects of mantle-derived melts interacting with the lower crust, regional erosional unroofing of the crust and, on a more local scale, by its mechanical stretching. Towards the end of the Early Permian, the potential temperature of the asthenosphere returned to ambient levels. With this, regional, long-term thermal subsidence of the lithosphere commenced, controlling the development of a new system of Late Permian and Mesozoic thermal sag basins. However, the evolution of these basins was repeatedly overprinted by minor short-term subsidence accelerations that reflect the build-up of far-field stresses related to rifting in the Tethyan and Atlantic domains. Comparison of observed and modelled subsidence curves suggests that in the area of the Rhine rift system the lithosphere had equilibrated with the asthenosphere at the end of the Cretaceous at depths of 100–120 km, before it became thermally destabilized again by Cenozoic rifting and plume-related magmatism. Modelled subsidence curves indicate that by the end of Early Permian times the thermal thickness of the remnant mantle-lithosphere ranged between 10 and 50 km in areas that were later incorporated into Mesozoic thermal sag basins; this corresponds to mid-Permian thermal lithosphere thicknesses of 40–80 km.

Abstract During Late Carboniferous-Early Permian times dextral transtensional movements along the NW-trending Franconian Fault System and parallel faults caused complex block faulting in the Thuringian Forest region, Germany, accompanied by intense magmatism. This is well constrained by geochronological data ( 207 Pb/ 206 Pb single zircon, SHRIMP, 40 Ar/ 39 Armica, zircon fission-track ages), field relations, and the sedimentary record from the Ruhla Crystalline Complex (RCC) and surroundings. The combined dataset indicates that the Ruhla Crystalline Complex was faulted into three nearly N–S-trending segments, which underwent different exhumation histories during Late Carboniferous–Permian times. The central segment of the RCC was exhumed by several kilometres as a horst block, while the eastern and western segments subsided simultaneously, forming the basement to the Oberhof and Eisenach molasse basins, respectively. Late Carboniferous–Permian uplift of the central segment is constrained by 40 Ar/ 39 Ar cooling ages of 311 ± 3 (muscovite) and 294−288 ± 3 Ma (biotite), a weighted zircon fission-track age of 272 ± 7 Ma and overlying Zechstein sediments. In contrast, the eastern segment shows much older 40 Ar/ 39 Albiotite cooling ages between 336 ± 4 and 323 ± 3 Ma, was exposed at c. 300 Ma, and subsequently covered by molasse sediments and volcanic rocks between 300 and c. 275 Ma. A similar Late Carboniferous evolution is inferred for the western segment, as it is also overlain by Lower Permian volcanic rocks and has a 297 ± 29 Ma single zircon fission-track age. Simultaneous horst and basin formation is additionally constrained by granite pebbles in conglomerates of the Oberhof and Eisenach basins. These pebbles can partly be derived from granites in the central segment of the RCC. Age data and the orientation of granitoid bodies and dykes in the Ruhla Crystalline Complex and its surroundings provide evidence for the opening of NE-trending structures between 300 and 294 Ma, and formation or reactivation of W- to NW-trending structures between 290 and 275 Ma. Magmatic activity in the Thuringian Forest region may have been caused by widespread mantle upwelling in central Europe during the Late Carboniferous-Early Permian.

Abstract 40 Ar/ 39 Ar step-heating dating of mineral separates from a series of lamprophyre dykes in the Saxothuringian Zone of the Variscan Orogen yielded Viséan-Namurian (334–323 Ma) and Stephanian–early Permian (297–295 Ma) crystallization ages indicating magma generation over a period of 30 Ma. In many cases, dyke emplacement was controlled by faults. Many are composite or show evidence for mingling of primitive and evolved magmas, and, to a certain degree, contamination with crustal melts. The high MgO (6–7 wt%), Ni (75–270 ppm) and Cr (140–1250 ppm) contents and mafic phenocryst assemblage are evidence for derivation from a mantle source. Kersantites and minettes have similar incompatible trace-element and rare earth element (REE) patterns (light REE (LREE)- and medium REE (MREE)-enriched and heavy REE (HREE)-depleted) and high, but varying Th, Zr and Hf contents. Positive Ni v. Mg# (FeO=FeO tot ) correlations suggest early fractionation of olivine, and the general absence of negative Eu anomalies makes feldspar fractionation improbable. For the lamprophyres of the Spessart, the variations of Ba, Rb and TiO 2 indicate phlogopite fractionation. Negative Ta, Nb and Ti anomalies are common, and may be an artefact of the high large ion lithophile element (LILE) and REE contents, but are more likely to reflect derivation from a mantle source that was metasomatized during a previous (Devonian?) subduction event. The generation of the parent melts was possibly triggered by partial melting of metasomatized mantle due to lithosphere detachment, removal and replacement of metasomatized lithospheric mantle by upwelling hot asthenospheric mantle. Compared to the spessartites, the minettes and kersantites appear to have originated by partial melting of deeper-mantle sources. Lithospheric mantle detachment may have caused post-collisional Namurian uplift and cooling of the crust, and facilitated emplacement of lamprophyre dykes along fault zones at high crustal levels.

Abstract The Saar–Nahe Basin is a late Variscan intermontane basin that developed on the site of an earlier island arc, the Mid-German Crystalline Rise. Within the c. 6500m-thick continental sedimentary fill of the basin, a large variety of igneous rocks was emplaced over a period of c. 4 Ma from 296 to 293 Ma as high-level intrusions and lava flows, extrusive domes, diatremes and pyroclastic deposits, ranging in composition from basalt and basaltic andesite to rhyodacite, rhyolite and trachyte. Composite intrusive–extrusive complexes consist of andesite, rhyodacite and alkali feldspar trachyte with up to 10 wt% K 2 O. The geochemical characteristics of the most primitive mafic magmas indicate a slightly enriched upper mantle source modified by subduction-related fluids. Nd–Sr–O isotope data indicate that crustal contamination was important in the petrogenesis of the more differentiated magmas.

Abstract Permian volcanic rocks from the Gorzów Wielkopolski region (NW Poland), although pervasively altered by low-grade metamorphism, still preserve the geochemical characteristics of continental mafic volcanic rocks formed by partial melting of an enriched mantle source. The metamorphic assemblage comprises corrensite, pumpellyite, laumontite, quartz and chalcedony, albite, calcite and solid bitumen (major components). Petrological studies combined with microthermometric determinations indicate a low-pressure zeolite-greenschist-facies transitional zone metamorphic grade with a clockwise (pressure-temperature) P-T path: earliest event 140–210 °C and 630–760 bar; metamorphic peak 220–300 °C and 950 bar; youngest episode: T ⩾ 130 °C and 630–760 bar. The metamorphism of the Rotliegend volcanic rocks is generally ascribed to penetration of upwelling fluids released from clastic rocks underlying the extrusive Permian unit. However, the ubiquitous occurrence of anhydrite in the altered volcanic rocks suggests an influence of pore water from the overlying Zechstein evaporite sequence. The source of metamorphic heat can be tentatively assigned to abnormal heat flow and/or exothermic reactions during magmatic mineral alteration processes. Dating of metamorphism in neighbouring areas suggests an Upper Jurassic thermal event related to the upwelling of a mantle diapir during the initiation and early evolution of the North Atlantic rift.

Abstract Carboniferous-Permian magmatism in the Spanish Central System, Iberian Ranges, Cantabrian Chain, Pyrenees (Maladeta plutonic complex) and the French Massif Central includes a range of mafic calc-alkaline and shoshonitic rock types, as well as amphibole-bearing lamprophyres (spessartites) and minor alkaline lamprophyres (camptonites). Subalkaline basalts with intermediate characteristics between enriched mid-ocean ridge basalts (E-MORB) and the mafic calc-alkaline rocks also occur in the Pyrenees (Panticosa, Cinco Villas and La Rhune). The incompatible trace-element characteristics of the least differentiated subalkaline rocks and lamprophyres indicate that variably enriched mantle sources were involved in their genesis. High large ion lithophile element/high-field-strength element (LILE/HFSE), light rare earth element (LREE) HFSE and low Ce/Pb ratios in the calc-alkaline and shoshonitic rocks require either assimilation of crustal rocks plus fractional crystallization (AFC) of the parental mafic magmas or melting of a previously subduction-modified mantle source. In the Cantabrian Chain and the Massif Central, melting of a subduction-modified mantle source seems more likely. In the Central System, Iberian Ranges and Maladeta area the lack of any evidence for a contemporaneous subduction system suggests that AFC processes were more likely to be responsible for the crustal signature of the magmas. The alkaline camptonites from the Central System were generated from an enriched mantle source, which had lower LREE/HFSE and LILE/HFSE ratios than the source of the older calc-alkaline magmas from the same area. The incompatible trace-element patterns and ratios (e.g., Y/Nb, Zr/Nb) of the subalkaline basalts from Panticosa, Cinco Villas and La Rhune suggest that they were generated from similar parent magmas, formed by mixing of partial melts of an asthenospheric source and a crustal component.

Abstract Permian magmatism in the Pyrenees is characterized by two compositionally different and temporally consecutive magmatic episodes: a calc-alkaline–transitional phase (andesites) and a mildly alkaline phase (basalts and dolerites). These two magmatic episodes were related to the attenuation of late Variscan transtensional tectonics and the onset of extension related to regional rifting. The strike-slip fault systems that affected the Pyrenees in late Variscan times initially controlled the development and morphology of the sedimentary basins. These were periodically affected by phases of extension, which controlled the subsidence of the basins, and, in addition, the emplacement of magmas. The whole-rock trace-element and isotopic signature of the andesites suggests that they were derived from the upper mantle and variably hybridized with late orogenic crustal melts, whereas the alkali basalts could have been derived from a lithospheric mantle source, enriched as a consequence of Variscan subduction processes with the contribution, in some areas, of an enriched (asthenospheric) component.