The Montalet granite crops out along the northwestern edge of the Montagne Noire gneiss dome (Fig. 1). The granite is composed of sills or laccoliths ranging in thickness from a few tens of metres to several hundreds metres (Demange et al., 1995). The deformation of the granite increases toward the major tectonic contact that limits the gneiss dome to the north and northwest. This contact is referred to as the Lacaune fault (Demange, 1996) or the Espinouse detachment (Brun and Van Den Driessche, 1994), respectively. According to Demange (1996), a magmatic foliation developed first and was then crosscut by a second foliation when approaching the Lacaune fault, as both planar structures are related to early thrusting events. For Brun and Van Den Driessche (1996), these two planar structures correspond to S-C structures (Berthé et al., 1979) (Fig. 2a) that developed during a single extensional shearing deformation along the Espinouse detachment. But in both cases, these authors consider that the Montalet granite is syn-kinematic in nature. We chose to sample this granite at the Col de Picotalen (or Piquotalen) (Fig. 2a), the location that these two conflicting interpretations are based upon, as it shows a well-developed deformation, (43° 41′ 16. 53″ N, 2° 39′ 33. 38″E).

The Montalet granite is a garnet-bearing, two-mica leucogranite. The selected sample is composed of K-feldspar, plagioclase, quartz, biotite, muscovite, garnet and accessory minerals. The main foliation S is marked by the shape preferred orientation of feldspar porphyroclasts (up to 10 mm long), micas (up to 3 mm) and quartz aggregates (Fig. 2b). Subhedral K-feldspar generally lacks strong internal deformation and displays Carlsbad twin planes, generally parallel to the foliation S. The feldspars include quartz droplets and biotite and muscovite crystals inherited from the early magmatic stage. Very small (0.01–0.03 mm) grains of quartz and micas are contained within regular narrow bands, spaced by 5 to 10 mm in general, interpreted as shear bands C (Figs. 2a and b). In the vicinity of these shear bands, biotite and garnet are locally replaced by chlorite. Quartz ribbons affected by these shear bands display progressive reorientation and grain size reduction towards the shear band. The foliation strikes east-west and dips ca. 20° to the north. The stretching lineation L underlined by the long axis of quartz, micas and feldspar clasts plunges about 15° NE. The C-planes dip ca. 50° to the north, some 20 to 30° steeper than the foliation, indicating a deformation associated with a normal movement to the northeast. They bear fine striae plunging about 45° to the northeast.

The preferred orientation of the subhedral feldspar crystals may be attributed to deformation affecting a partly crystallised magma. Pervasive S-C type structures suggest intense, solid state deformation. Myrmekites that developed at the K-feldspar high-pressure sides suggest diffusion and consequently high-temperature deformation (Simpson and Wintsch, 1989). Quartz constitutes polycrystalline ribbons parallel to the foliation. Recrystallised grains are slightly elongated, forming an internal shape fabric oblique to S. Lobate boundaries suggest some recrystallisation by grain boundary migration, common at high temperatures (e.g. Gapais and Barbarin, 1986). Biotite crystals have typical fish-like shapes and tend to be replaced by chlorite, which is interpreted as a consequence of cooling. The grain size evolution from the foliation to the shear bands also suggests progressive deformation at decreasing temperature.

The S-C structures indicate a top to the northeast shear as supported by internal quartz schistosity and mica fishes. This means that the Montalet granite was deformed syntectonically during its emplacement into the footwall of a normal, ductile shear zone.

3.1 Analytical techniques

3.2 Results

Monazite grains were generally euhedral and yellowish. Thirty grains were analysed and the data are reported in Table 1. In a ²⁰⁶Pb/²³⁸U versus ²⁰⁸Pb/²³²Th diagram, they all plot in a concordant to sub-concordant position (Fig. 4a) and define a Concordia age (Ludwig, 1998) of 293.66 ± 0.96 Ma (MSWD = 18), which is within the error with the mean average ²⁰⁸Pb/²³²Th date of 291.9 ± 3.3 Ma (MSWD = 0.89).

Zircons were translucent to pinkish, euhedral and generally elongated. Cathodoluminescence imaging of the grains revealed a bright core with magmatic zoning surrounded by darker rims (Fig. 3). Twenty-two analyses were performed on twenty grains (Table 1). During the course of the analyses, several zircons showed the presence of common Pb, but no correction was applied. Plotted in a Terra Wasserburg diagram (Fig. 4b), the data do not define a simple trend with apparent dates ranging from ca. 500 Ma down to 295 Ma. A first group with two analyses (24130410b and 25130410b) defines a date around 496 ± 15 Ma. A second group with three analyses (04130410b, 05130410b and 50130410b) defines a date of 323 ± 5 Ma. The last group defined by six concordant analyses allows one to calculate a Concordia age of 294.4 ± 2.6 Ma (MSWD = 0.51; Fig. 4b, insert). The remaining points plot in a sub-concordant to discordant position. We believe that their position can be linked to the combined effect of the presence of “common” Pb incorporated in some of the grains and a slight Pb loss. The presence of initial “common” Pb in some of the grains is attested by the fact that several zircon grains present a slight (few tens of counts) positive ²⁰⁴Pb value. This common Pb might have been present in some small inclusions not detected during imaging or related to fractures or areas of radiation damage. In the Terra Wasserburg diagram (Fig. 4b), these data, together with the previous six concordant allow a discordia to be drawn with a lower intercept of 293.7 ± 2.3 (MSWD = 0.88). In light of the present data, which are all consistent within error regardless of whether common Pb was detected or not, we thus conclude that zircon crystallisation took place 294 ± 3 Ma ago. In this scenario, the remaining older ages must be attributed to inheritance. This age of 294 ± 3 Ma is also identical to the age of 294 ± 1 Ma found for the monazites. Consequently, we conclude that the Montalet granite was emplaced during the Early Permian, ca. 294 Ma ago.

Petrological and structural data show that the Montalet granite is a syntectonic intrusion that cooled progressively during a continuous deformation from magmatic to solid states. The 294 ± 1 Ma age of the Montalet granite demonstrates that it was emplaced during the Late Paleozoic extension in the ductile lower crust, contemporaneously with the: (1) ductile shearing event related to the Espinouse detachment and; (2) sedimentation within the Stephanian-Permian Graissessac-Lodève basin.

Indeed, similar ⁴⁰Ar/³⁹Ar ages of 297 ± 3 Ma have been obtained by Maluski et al. (1991) on biotite and muscovite from sheared orthogneisses, in the footwall of the Espinouse detachment, and were also interpreted as S-C mylonites (Beaud, 1985; Burg et al., 1994; Echtler and Malavieille, 1990; Van Den Driessche and Brun, 1992). Bruguier et al. (2003) published an age of 295 ± 5 Ma for zircons extracted from a volcanic ash layer interbedded in the Late Carboniferous sedimentary fill of the Graissessac-Lodève basin, developed in the hangingwall of the Espinouse detachment (Fig. 1). Similar ages, ranging from 295 to 300 Ma have also been reported by the same authors in other Late Carboniferous extensional basins of the southern French MC, demonstrating an intense magmatic activity during this period (Bruguier et al., 2003).

On the other hand, Bé Mézème (2005) and Faure et al. (2010) have analysed monazites and zircons from the same pluton, but which was sampled in three sites located several kilometres farther to the south-west (Fig. 1).

Zircons were dated at 324 ± 3 Ma by SIMS and monazites at 327 ± 7 Ma by EPMA. At this stage, it has to be noted that four concordant to sub-concordant zircons analysed by Faure et al. (2010) plot in a “younger” position than the nine grains used to calculate the age of 324 ± 3 Ma (Figure 12C of Faure et al., 2010). A mixing discordia between common Pb and radiogenic Pb for these four “younger” points yields a date of 305 ± 29 Ma (MSWD = 0.78). The significance of these four points has not been discussed in the paper. However, it is interesting to note that this date of ca. 305 Ma is within the error of the age of 294 ± 3 Ma found for the zircons in this study. Furthermore, we also demonstrate the existence of inherited cores within our zircon population that are dated at ca. 323 Ma.

As for the monazites, Faure et al. (2010) analysed three different samples of the Montalet granite. Two of them yielded EPMA ages of 333 ± 4 Ma and 327 ± 7 Ma, respectively. For this second sample, it is interesting to note the discrepancy between the U-Pb age (289 +43/–49 Ma) and the Th-Pb age (349 +28/–25 Ma) and the fact that the theoretical isochron is barely within the error envelope. The last sample yielded an EPMA age of 499 ± 6 Ma. This indubitably shows the presence of inherited monazites within at least one of their samples (which they acknowledge on page 665). Because of the intrinsic principles of the EPMA technique, one cannot therefore exclude the presence of inherited Pb (²⁰⁸Pb?) within some of the monazites from the remaining two samples, which would have resulted in a meaningless older age.

Another explanation to account for the difference between the ages published by Faure et al. (2010) and the ages found in this study could be that we dated two different intrusions corresponding to two magmatic pulses that are separated in time, yet mapped as a single intrusion known as the Montalet granite.

Nevertheless, irrespective of the origin of the discrepancy between these ages, the age of 294 ± 1 Ma found for the syntectonic Montalet granite confirms the existence of a major magmatic event that accompanies crustal extension during the Late Palaeozoic in the southern French MC (Van Den Driessche and Brun, 1989).

The cause and the modes of the extension that is supposed to be responsible for the decay of the Variscan belt are still debated, but all studies agree that the period between 330 Ma and 290 Ma is a critical one: it corresponds to the vanishing of plate convergence and the onset of crustal extension (e.g. Burg et al., 1994). The present study emphasizes the absolute need for precise dating to refine this evolution and decipher the possible causal relation between these two major tectonic processes.