A dike network transecting a basement of intrusive and metamorphic rocks related to the Hercynian orogeny is exposed in the Sila Grande (southern Italy). Dike magmatism, similarly to other regions of the western Mediterranean, such as Sardinia, Corsica, and Catalonia, is of calc-alkaline to alkali-calcic affinity. Zircon U-Pb geochronology indicates that dike magmatism took place between 295 ± 1 to 277 ± 1 Ma, after the main late Hercynian emplacement of granitoids (306 ± 1 Ma). Barometry indicates that the basement underwent exhumation of 8 ± 3 km before dike injection. The dike network has a geometrical arrangement consistent with a transtensional stress regime that resulted in ductile thinning of the lower crust during the late stage of the Hercynian orogeny and concurrent fracturing of the upper crust that made possible magma ascent through dikes. The proposed tectonic evolution is related to dismemberment of the southern Hercynian belt in the central Mediterranean area as a result of dextral transtension of Gondwana in relation to Laurasia during the Pennsylvanian–Early Permian.


The geometry, orientation, and spatial distribution of sheeted magmatic bodies (i.e., dikes and sills) are commonly used to reconstruct pathways of magma ascending through the crust. In principle, dike pattern mainly reflects the interplay between two factors: (1) the orientations of pre-existing (preemplacement) and/or syn-magmatically developed (synemplacement) tectonic discontinuities (e.g., faults and fractures), and (2) magma driving pressure. To evaluate the influence of both these factors, active subvolcanic systems can be successfully analyzed (e.g., Gudmundsson and Brenner, 2004; Gudmundsson, 2006; Tibaldi et al., 2008). The study of areas affected by ancient dike magmatism is equally important in the investigation of the significance and time of activity of tectonic discontinuity arrays. The preferential orientation of dikes is considered as symptomatic of a tectonic strain field controlling the orientation of the synemplacement faults and fractures (e.g., Glazner et al., 1999; Sadeghian et al., 2005; Dini et al., 2008), whose age of formation can be deduced by dating injected magmatic rocks. Therefore, dike swarms can be used as both age and large-scale strain markers.

Widespread dike magmatism affected the Hercynian belt of Europe during the late stages of the orogenic cycle (Wilson et al., 2004, and references therein). Dike swarms postdate most of the plutonic complexes, and are mainly related to calc-alkaline and alkaline magmatic affinity, and more rarely to tholeiitic series (Orejana et al., 2008, and references therein). In the central and western Mediterranean, dike swarms, mainly of Early Permian age, are well developed in different sectors of the southern Hercynian belt of Europe, such as the Catalan Coastal Range (Enrique, 1984, 1990; Ferrés, 1998), Corse (Rossi et al., 1993, 2002; Carmignani et al., 1999; Traversa et al., 2003; Cocherie et al., 2005), and Sardinia (Carmignani et al., 1999; Ronca et al., 1999; Atzori et al., 2000; Traversa et al., 2003) (Fig. 1). In these regions an extensional regime is considered the more appropriate tectonic context for dike magmatism (Enrique, 1984; Rossi et al., 1993; Atzori et al., 2000). Farther east, other remnants affected by dike magmatism are represented by the Sila and Serre Massifs of the Calabria-Peloritani terrane (Fig. 1), even though little is known about serial affinity, age, and tectonic framework of this magmatic event. The purpose of this paper is to close this knowledge gap through new petrological, geochronological, structural, and geochemical data collected on dikes from the central part of the Sila Massif, known as Sila Grande (Fig. 2). The main goal is to define the timing and tectonic context for magma ascent through dikes in the course of the Hercynian cycle. Petrological and geochronological results are used to estimate the amount of exhumation following an approach suggested by Ring et al. (1999). Considering that most of the Sila Grande dikes injected granodiorite (e.g., Liotta et al., 2008), depths of emplacement estimated for dikes and for the granodioritic wall rock can be combined with related intrusion ages. In addition, the analysis of dike orientations could be used to infer the tectonic strain field during exhumation. In this way the tectonic framework of dike magmatism in the course of the orogenic cycle can be successfully defined. This represents a more complete multidisciplinary approach, different from that adopted in most studies focused on dikes of the southern Hercynian belt, that rely in large amount on geochemical and geochronological data to reconstruct the tectonic setting during the late orogenic magmatism (e.g., Rossi et al., 1993; Atzori et al., 2000). Results of this research can be useful in studies devoted to recognition of geological relationships between pieces of the Hercynian basement now dispersed throughout the central Mediterranean (Alvarez and Shimabukuro, 2009).


A tilted, nearly complete late Hercynian continental crust section is exposed in the Sila Massif (Dubois, 1976; Graessner and Schenk, 2001). It is basically composed of three main portions; from bottom to top, these are: (1) high-grade metamorphic rocks, occupying the Hercynian lower to middle crust; (2) nearly tabular late Hercynian granitoids, located at mid-crustal levels; and (3) low-grade metamorphic rocks belonging to the upper crust (Fig. 2).

High-grade metamorphic rocks are exposed to the west and are mainly represented by migmatitic paragneiss with intervening marble and metabasite becoming progressively more abundant toward the top. In the migmatitic paragneiss, the metamorphic assemblage of biotite + garnet + sillimanite + K-feldspar ± cordierite ± spinel equilibrated at peak pressure-temperature conditions of 400–600 Mpa and 740–770 °C (Graessner et al., 2000; Graessner and Schenk, 2001). The age of peak metamorphism, obtained by the U-Pb method on monazite, spans from ca. 305 to 296 Ma (Graessner et al., 2000).

Late Hercynian high-temperature shear deformation affected middle to lower crust in the zone straddling the contact between foliated granitoids and host migmatitic high-grade metamorphic rocks (Festa et al., 2006; Liotta et al., 2008). The compositions of granitoids range from monzogranite to tonalite (Ayuso et al., 1994), but minor mafic rocks (norite, diorite, quartz diorite) (Caggianelli et al., 1994) and strongly peraluminous leucogranite (Caggianelli et al., 2003) can be also found. Emplacement depths obtained by Al-in-hornblende barometry range from 8 to 18 km, and confirm that the intermediate crust was a preferential site for the emplacement of granitoid magmas (Caggianelli et al., 1997; Liotta et al., 2004). Using the 40Ar/39Ar method on hornblende and muscovite, Ayuso et al. (1994) estimated ages ranging from 293 to 289 Ma for these plutonic rocks. Graessner et al. (2000), using the U-Pb method on monazite, obtained emplacement ages of 304–300 Ma. Comparison of results and techniques indicates that dates by the 40Ar/39Ar method can be considered as cooling ages, by virtue of the distinctly lower closing temperatures of the involved isotopic systems (for a compilation, see Spear, 1993).

In Sila Grande, dikes transect essentially intrusive rocks but can be also found in the metamorphic basement (Fig. 3). In the earliest studies, dikes were generally related to Cenozoic volcanic activity (e.g., Rittmann, 1946; Bertolani, 1957). According to this view, some porphyritic rocks were considered to be crystallization products of lava flows. De Fino and La Volpe (1970) excluded an original subaerial setting, and recognized that emplacement of magma took place exclusively at moderate depths along subparallel fissures transecting the Paleozoic basement. Burton (1970, 1971a, 1971b, 1971c), relates dikes in the Calabria region to Paleozoic magmatic activity. One recent dating by the U-Pb method on zircon indicates an age of emplacement of 284 Ma for a porphyritic dike from Sila Grande (Liotta et al., 2008). This represents the only available age determination for this magmatic activity to date.


Distribution and Orientation

After the surveys carried out by De Fino and La Volpe (1970) and Burton (1970, 1971a, 1971b, 1971c), new intrusive bodies in the Sila Grande area were recognized, in particular within our study. In addition, the geometry of already mapped bodies is herein reinterpreted, in some instances allowing us to recognize both dikes and sills. Dikes are prevalent, and are particularly abundant in an east-west–oriented strip (5 km wide) that extends westward for 22 km from the town of San Giovanni in Fiore (Fig. 3). Sills are only represented by two subhorizontal felsic intrusions of fine-grained two-mica granite located near Silvana Mansio village and Mount Carlomagno (Fig. 3). In map view they show an irregular shape with maximum and minimum dimensions of 1.5 km and 0.8 km, respectively, and are at least ∼35–40 m thick. All sheeted bodies are hosted in the Hercynian crust section, transecting mostly granitoids and more rarely the migmatitic paragneisses (Fig. 3).

A database of orientation and geometrical features has been prepared, including both new and literature field data for 353 dikes. Although some dikes are inclined, the great majority show subvertical and subplanar sharp walls. Map view suggests that most of them have a nearly tabular shape (Fig. 3). A rose diagram indicates a strong preferred orientation of dikes along an east-west strike. Other significant groups are oriented along east-southeast–west-northwest, northwest-southeast, north-northwest–south-southeast, and east-northeast–west-southwest directions (Fig. 4). Outcrop conditions are not suited to define precisely the dike extent. The maximum length estimated on the map is ∼2 km (Fig. 3). However, the possibility that some dikes are part of a longer, single sheeted body cannot be excluded. The thickness of some dikes (up to ∼15 m) argues for a length decidedly >2 km according to the typical thickness/length ratio of dikes in the upper crust of 0.01, given by Lister and Kerr (1991).

Petrography and Chemical Classification

The Sila Grande dikes display a notable compositional variety, as evidenced by the presence of both felsic and mafic types. Felsic dikes are more frequently exposed, and are represented by microgranite and microgranodiorite. In these rocks plagioclase ranges from albite to An-poor oligoclase, and biotite is the only mafic mineral. In microgranite scarce biotite coexists with white mica, whereas in microgranodiorite biotite occurs in larger amounts, and white mica is absent or rare. In the more felsic dikes, quartz and K-feldspar are frequently present in granophyric texture. Accessory magmatic minerals are represented by columnar to acicular apatite and zircon, whereas sphene, chlorite, and opaque minerals appear to be of postmagmatic origin.

Mafic dikes are made up of dark fine-grained rocks consisting of plagioclase, quartz, green amphibole, biotite, and rare augite. An abundance of amphibole, together with the labradorite composition of plagioclase (An55–70), indicates that these rocks are intermediate between microdiorite and microgabbro. For this reason the name “microgabbrodiorite” is adopted here. Accessory minerals include opaque minerals, columnar to acicular apatite, and zircon, whereas epidote, chlorite, and calcite are considered to be of postmagmatic origin.

Textures in both felsic and mafic dikes are commonly porphyritic, but in some cases (especially in microgranodiorite and microgranite) approach an equigranular type (Figs. 5A–5C), as is typical of felsite. The gradual increase of grain size from the margins inward is a common feature in dikes (Figs. 5D and 5E). Clear evidence of multiple magma injections, such as internal contact and irregular grain size distribution, were not observed. On this basis, the dikes appear to be the product of a single magma intrusion event.

Furthermore, a relation of near parallelism between margins, and the elongation of subhedral feldspars, up to 1 mm in length, is observed in some dikes, and particularly in microgranodiorites where flow texture is sometimes observed (Fig. 6A). Quenching textures are represented in felsic dikes by feldspar and biotite spherulites (Fig. 6B), and in mafic dikes by ocelli consisting of hornblende microlites often radiating around resorbed quartz xeno crysts (Fig. 6C). Additional evidence of fast cooling of magma is suggested by the common elongated shapes of micas and of acicular amphibole and apatite. Textures related to fluid exsolution are represented in mafic dikes by amygdules. They are filled by calcite or by alternating chlorite and calcite shells, and have a maximum diameter of 3 mm (Fig. 6D). Amygdules occupy ∼1% rock volume.

The chemical compositions of the sampled felsic and mafic magmatic rocks are given in Table 1. Chemical classification according to the TAS (total-alkali-silica) diagram is proposed in Figure 7A. Dikes display a wide compositional range from basaltic andesite to rhyolite, and are mainly represented by dacite and rhyolite. In comparison to host calc-alkaline granitoids, dike magmatism drifts toward trends of alkali-calcic serial affinity (Frost et al., 2001) (Fig. 7B), characterized by a moderately high Na2O content (Fig. 7C), consistent with late orogenic to postorogenic timing.

Rare earth element patterns, from the results of the analyses shown in Table 2, are provided for three dikes and host granodiorite. All patterns are fractionated, with CeN/YbN increasing from the mafic (CeN/YbN = 6.69) to the felsic types (8.55 and 12.66) and to granodiorite (14.26). A negative Eu anomaly increases in magnitude from the mafic (Eu/Eu* = 0.747) to felsic compositions (0.508 and 0.345), with granodiorite having an intermediate value (0.611) (Fig. 7D). These variations are compatible with a variable extent of plagioclase and amphibole fractionation. The spider diagram in Figure 7E shows a significant depletion in K, Ta, Nb, Sr, and, Ti for all rock samples. A major role of plagioclase in magma-fractionation processes is confirmed by the more pronounced Sr negative spike observed in felsic dikes.


Most dikes transect granitoids, and this suggests that magma ascent reached at least the intermediate level of the continental crustal section. However, relevant textural features indicate an even shallower level of emplacement, and a significant time lag from the intrusion of the main granitoids. Magma quenching textures, such as alkali-feldspar spherulites (Fig. 6B), acicular hornblende around quartz ocelli (Fig. 6C), and vesiculation textures, such as amygdules (Fig. 6D), point to fast magma cooling at a shallow crustal level. In addition, the presence of amygdules, restricted to mafic dikes, is in agreement with the lower solubility of volatiles in basic melts with respect to acidic melts.

In a tentative attempt to define quantitatively the emplacement level, a petrological approach was adopted. A depth estimate was obtained by combining Al-in-hornblende barometry (Anderson and Smith, 1995) with hornblende-plagioclase thermometry (Holland and Blundy, 1994). This approach was applied both for a mafic dike and for the granodiorite wall rock (Table 3). In the mafic dike, amphibole composition in the groundmass ranges from edenite to hornblende, whereas plagioclase is labradorite (An59–69). Simultaneous solution of the barometric and thermometric (edenite-richterite exchange) equations allowed us to estimate a level of emplacement for the mafic dike of 5 ± 2 km (pressure, P = 130 ± 60 MPa). In granodiorite, amphibole rims are edenitic hornblende, and plagioclase rims are andesine (An39–45). By the same approach, we calculated a level of emplacement for granodiorite of 13 ± 2 km (350 ± 60 MPa). Results are considered reliable, because the pressure-temperature equilibrium values (130 ± 60 MPa at 840 ± 40 °C and 350 ± 60 MPa at 760 ± 40 °C) plot beyond wet solidus conditions for andesite to basalt (Wyllie, 1978; Green, 1982) and for a granodiorite (Schmidt and Thompson, 1996), respectively. Therefore, the significant difference in emplacement depths indicates that intrusion of the mafic dike took place after considerable exhumation of the host granodiorite, that, according to the more conservative choice allowed by error ranges, is not <4 km. In addition, as indicated by the presence of amygdules, basic magma reached saturation in volatiles somewhere during ascent. Saturations at the pressure-temperature conditions of the emplacement level both in H2O and CO2 can be set by VolatileCalc software (Newman and Lowenstern, 2002), and are, respectively, 3.3 wt% and 600 ppm for a basaltic magma. These are minimum estimates for original volatile content of the magma. By lowering magma density, volatiles may have been the determining factor in favoring magma ascent up to shallow crustal levels. If water was the dominant volatile, as suggested by the elevated content of hydrous minerals, then it contributed also to depression of the liquidus temperature, allowing a considerable cooling of magma before full crystallization (Sisson et al., 1996).


Owing to outcrop conditions, no geological method can be applied with confidence to infer the emplacement sequence of the various dikes. Thus geochronology is the only option to shed light on this subject. To this end, we selected three different dike samples representative of the main compositional and textural types. In addition, we dated a granodiorite wall rock to define the time lag of dike magmatism with respect to the main late Hercynian magmatic event.


Zircons were concentrated using standard mineral separation techniques from fresh representative samples of ∼3–6 kg. Zircons free from cracks were preferentially selected, mounted in epoxy resin, and polished to obtain the exposure of their maximum surface along the c-axis. Grains were examined with cathodoluminescence (CL) imaging techniques using the JEOL scanning electron microscope at the CNR-IGG (Consiglio Nazionale della Ricerche–Istituto di Geoscienze e Georisorse), Unità di Pavia (Italy). CL images of zircons were used in order to select the most suitable location of the analytical spots for U-Pb geochronology (Fig. 8).

Age determinations of zircons were performed (CNR-IGG, Unità di Pavia) using a 193 nm ArF excimer laser ablation microprobe (GeoLas200QMicrolas) coupled to a magnetic sector inductively coupled plasma–mass spectrometer (Element from ThermoFinnigan). The analytical details are given in Tiepolo (2003). Analyses were carried out in single spot mode and with a spot size of ∼25 μm. The laser was operated with a frequency of 5 Hz and with a fluence of 12 J cm−2. Mass bias and laser induced fractionation were corrected by adopting zircon 91500 (1062.4 ± 0.4 Ma; Wiedenbeck et al., 1995) as the external standard. The same spot size and integration intervals were considered on both standard and studied samples. Data reduction was carried out through the GLITTER software package (van Achterbergh et al., 2001). Time-resolved signals were carefully inspected to detect perturbation of the signal related to inclusions, cracks, or mixed age domains. Within the same analytical run, the error associated with the reproducibility of the external standard was propagated to each analysis (see Horstwood et al., 2003), and after this procedure each age determination was accurate within the quoted error. The concordia test was performed for each analytical spot from 206Pb/238U and 207Pb/235U ratios using the function in the software package Isoplot/Ex3.00 (Ludwig, 2003). Concordia ages and associated 2σ errors are reported in Table 4. The discordant data were not taken into consideration because of doubtful interpretation. The Isoplot/Ex3.00 software was also used to draw probability density plots and concordia diagrams (Fig. 9).

Analytical Results and Age Interpretations

Ages were interpreted in light of internal features (Fig. 8). Euhedral, concentric, oscillatory zones have been considered as evidence of magmatic growth (Paterson et al., 1989; Vavra, 1990; Miller et al., 1992; Hanchar and Miller, 1993). Inner zones with truncations and/or with abrupt changes of the zoning pattern with respect to the surrounding areas have been interpreted as inherited domains.

Sample FLV2 (Granodiorite)

FLV2 is characterized by very large zircons, often having dimensions of ∼200 × 80 μm (3.2 aspect ratio). Crystals preferentially developed the steep pyramid forms in longitudinal sections. Usually, the zircons contain inclusions and the larger crystals are fractured.

CL images of ∼110 crystals revealed that the zircons are very luminescent. They show a well-developed growth zoning with pronounced fine oscillatory zoned bands. In some cases these bands enclose one large uniform central zone. Generally, the outer part of the crystals is less luminescent with respect to the inner part. Zircons could contain small rounded inclusions of apatite and/or inherited cores (Zr110; Fig. 8A). In few cases, the oscillatory zoning is interrupted by discontinuities (Fig. 8A).

We collected 52 U-Pb data from 43 different crystals; 36 U-Pb analyses on oscillatory zoning domains yield concordant ages ranging from 326 ± 6 Ma to 272 ± 5 Ma (Table 4; Fig. 9), whereas a single spot on the core yields an U-Pb concordia age of 361 ± 6 Ma (Fig. 8A), suggesting an inherited origin for this core. U-Pb concordia ages show a major cluster with weighted average age of 306 ± 1 Ma (mean square of weighted deviates, MSWD = 0.81, probability = 0.70; n = 20), and two minor clusters ca. 325 and 280 Ma (Fig. 9). The two ages of ca. 325 Ma refer to bright oscillatory cores surrounded by dark oscillatory rims, and thus they have been interpreted as inherited ages. The younger ages of ca. 280 Ma may be the result of partial U-Pb resetting due to a subsequent tectonometamorphic event.

Sample FLV 19 (Mafic Dike)

Zircons in this sample show well-developed prismatic faces with steep pyramidal terminations. The typical dimensions range from 100 to 200 μm in length, and from 40 to 80 μm in width with aspect ratios between 2 and 4. Generally, the zircons are very luminescent. Rarely, low luminescent crystals or areas are visible. CL images revealed well-developed oscillatory growth zoning without significant luminescence variations from core toward rims. In few cases, zircons contain small rounded inclusions and/or inherited cores (Fig. 8B).

We collected 50 U-Pb data from 36 different crystals; 45 U-Pb analyses yield concordant ages ranging from 313 ± 7 Ma to 273 ± 8 Ma (Fig. 9), and three ages were 556 ± 13 Ma, 428 ± 24 Ma, and 353 ± 8 Ma (Table 4). These three older ages clearly correspond to inherited cores. A large population of U-Pb concordia ages defines a prominent peak in the density diagrams ca. 296 Ma with a weighted average age of 295 ± 1 Ma (MSWD = 0.93, probability = 0.56; n = 24) (Table 4; Fig. 9).

Sample FLV 16I (Felsic Dike)

Combining optical observations with CL images, zircons in this sample are characterized by longitudinal sections with well-developed prismatic faces and flat pyramidal terminations. The dominant dimensions are ∼140 × 50 μm (aspect ratio = 3) with rare cases of 250 × 50 μm. CL images revealed that they are very luminescent with variable intensity from outer toward inner portions. The oscillatory growth zoning from core toward rims is a typical internal feature of the crystals, and is locally interrupted by discontinuities related to dissolution. Inherited cores showing internal features discordant with respect to the regular growth zoning of the surrounding rims rarely occur. Alternatively, they can be spongy or may show irregular zoning (Fig. 8C).

We collected 50 U-Pb data from 37 different crystals; 20 U-Pb analyses yield concordant ages ranging from 309 ± 6 Ma to 266 ± 6 Ma, with a single value at 393 ± 9 Ma. The larger population of U-Pb data defines a major and minor cluster ca. 278 Ma and 304 Ma, respectively. The ages of the major cluster refer mainly to rims with well-developed oscillatory zoning or to cores of zircon grains showing continuous oscillatory zoning, and have a weighted average age of 278 ± 2 Ma (MSWD = 0.69, probability = 0.70; n = 9) (Table 4; Fig. 9). The ages of the minor cluster refer mainly to core with different luminescence properties with respect to the rims, except for two data obtained from near-rim domains enclosed within the dissolution surface. According to zircon internal features and to U-Pb data, the older ages refer to the zircon core and/or domain of inherited origin. The inherited ages are coherent with the dominant ages of the granodiorite.

Sample FLV 13 (Felsic Dike)

In accordance with the lower grain size in the entire matrix, zircons are smaller than those of the other samples, with dimensions ranging from 150 μm × 100 μm to 65 μm × 50 μm. Optical and cathodoluminescent investigations revealed that the zircon grains have longitudinal sections with well-developed prismatic faces, and poorly developed flat pyramidal terminations. CL images indicate that the zircons are weakly luminescent, and do not show a pronounced oscillatory growth zoning. In some cases, outer domains are more luminescent with respect to the inner domains, and, rarely, inverse luminescence properties have been observed (Fig. 8D).

We collected 34 U-Pb data from 21 different crystals; 15 U-Pb analyses yield concordant ages ranging from 284 ± 5 Ma to 260 ± 4 Ma with a major concordant population with weighted average age of 277 ± 1 Ma (MSWD = 0.58, probability = 0.86; n = 13) (Table 4; Fig. 9).

On the basis of these data, the Sila Grande can be interpreted to be the result of a sequence of late Hercynian magmatic events, starting with emplacement of granodiorite at 306 ± 1 Ma, followed by the injection of the mafic dikes at 295 ± 1 Ma, and ending with the felsic dikes at 277 ± 1 Ma.


Tectonic Exhumation of the Sila Grande Basement

Results obtained in this study can be used to constrain amount of exhumation of the Sila basement from Pennsylvanian to Early Permian time. Toward this end, depths of emplacement estimated for the mafic dike and for the granodioritic wall rock are combined with related emplacement ages. Assuming that intrusion of the mafic dike took place with a time lag of 11 ± 1 m.y. with respect to granodiorite, and that the difference in emplacement depth amounts to 8 ± 3 km, an average exhumation rate of 0.7 ± 0.3 mm/yr is estimated. This value can be compared with the duration of the late Hercynian decompression documented for Hercynian lower continental crust exposed in Calabria. According to Schenk (1989) and Graessner and Schenk (2001), the lower crust underwent decompression of 200–140 MPa in a time span of 10 m.y. (from 300 to 290 Ma; Schenk, 1989). These data indicate an exhumation rate of 0.5–0.7 mm/yr. Therefore, the late Hercynian exhumation documented here for the upper crust is consistent with the decompression rate of the lower crust.

Given that dike magmatism and exhumation were simultaneous, we examine here the question of whether tectonic deformation is the main cause for both processes. As suggested for other locations, the preferred orientation of dikes is strong evidence in favor of a tectonic control on the pattern of brittle discontinuities that favored magma ascent through dikes (e.g., Glazner et al., 1999; Sadeghian et al., 2005; Dini et al., 2008). Following this view, it can be shown that dikes in the Sila Grande are oriented along a geometrically coherent system of brittle discontinuities. The angular relations among the main east-west strike and the four secondary orientations (east-southeast–west-northwest, northwest-southeast, north-northwest–south-southeast, and east-northeast–west-southwest) are shown in Figure 4. It can be seen that east-west strike can be interpreted as a principal displacement zone (Sylvester, 1988), while the secondary orientations can be interpreted as the associated brittle structures: extensional discontinuities, and synthetic or antithetic shears (see Fig. 10) (Petit, 1987; Sylvester, 1988). Thus the Sila Grande network of discontinuities may be related to a tectonic event with a significant horizontal component of dextral simple shear. In addition, to produce tectonic exhumation, a vertical orientation of the principal stress is required for the late Hercynian Calabria continental crust (Caggianelli et al., 2007). Consequently, we propose a dextral transtensional setting as the tectonic framework for Pennsylvanian–Early Permian exhumation and infiltration of magmas in the upper crust. In agreement with this interpretation, the synoptic sketch of Figure 10 depicts the evolution of the continental crust exposed in the Sila Grande from 306 Ma (age of emplacement of the granodiorite) to 277 Ma (age of the latest dike). The block diagram in Figure 10A represents the starting point of the evolution. The plutonics of the Sila Grande occupied the intermediate part of the continental crust. A surface located at a paleodepth of ∼13 km is traced on the basis of igneous barometry. The block diagram in Figure 10B is a snapshot of the situation ca. 295 Ma, the age of emplacement of the mafic dikes. The paleosurface, originally located at 13 km, has now reached the shallower level of ∼5 km as constrained by barometric estimates on the mafic dike. Exhumation was activated by transtensional tectonics with a dextral component that produced a coherent network of brittle discontinuities, representing a preferential path for magma ascent into the upper crust. The block diagram in Figure 10C represents a more advanced stage of the evolution to 277 Ma, when the network of discontinuities acquired the current extent, and injected magmas were dominantly acidic. Figure 10D reproduces the present situation, after Mesozoic to Cenozoic tectonic episodes responsible for the northeastward tilting of the Sila crustal section.

Tectonic processes may be responsible for Pennsylvanian–Early Permian exhumation of the entire Hercynian crust, by dominant ductile thinning in the hot lower levels, and by fracturing in the cool upper levels (Ziegler and Cloetingh, 2003, and references therein).

Regional Significance of Dike Magmatism

The Sila Grande dikes were fed by magmas similar in composition to those found in other pieces of the Hercynian basement now dispersed throughout the central Mediterranean. In Figure 11 data points of the Sila Grande dikes are plotted together with those of dikes from Sardinia, Corse, and the Catalan Coastal Range (Enrique, 1984; Atzori et al., 2000; Traversa et al., 2003). In the TAS diagram (Fig. 11A), data points related to Sila Grande, Sardinia, and Catalan Coastal Range are spread over a wide compositional range, from basaltic andesite to rhyolite. Dikes from Corse display an even larger compositional interval, including basalt dikes with silica content as low as 46 wt%. In terms of the modified alkali-lime index (MALI of Frost et al., 2001), dikes from the selected locations mostly show calc-alkaline to alkali-calcic characteristics (Fig. 11B).

This magmatism took place in a time span from the Pennsylvanian to Early Permian. A close match of the ages in the various regions cannot be established owing to the different dating methods adopted (Ferrés, 1998; Carmignani et al., 1999; Ronca et al., 1999; Atzori et al., 2000; Rossi et al., 2002; Cocherie et al., 2005). Notwithstanding this problem, it can be deduced that most of this magmatic activity affected the southern Hercynian domain between 295 and 270 Ma (Fig. 1). Compositional and timing similarities suggest a common tectonic episode of regional importance was responsible for the widespread Pennsylvanian to Early Permian dike magmatism in the southern Hercynian range.

The dextral transtensional tectonics recognized for the Sila Grande appear to be compatible with Pennsylvanian to Early Permian geodynamical evolution characterized by a major dextral translation between Gondwana and Laurasia (Arthaud and Matte, 1977; Ziegler, 1993), as sketched in Figure 12. This process affected a large region, and was responsible for the clockwise rotation of the European Hercynian belt from ca. 300 to 270 Ma (Stampfli and Borel, 2002; Deroin and Bonin, 2003; Vai, 2003). In this tectonic context, magmas transitional from calc-alkaline to alkaline series were emplaced in the basement now exposed in the Mediterranean area both as volcanic and plutonic rocks, marking the end of the Hercynian orogenic cycle (e.g., Rossi et al., 1993; Bonin, 1998; Rottura et al., 1998; Orejana et al., 2008).

Dike magmatism in the Sila Grande fits well in this geological scenario and provides further evidence of the Pennsylvanian to Early Permian transtensional regime between Gondwana and Laurasia.

This study was supported by MIUR-PRIN (Ministero Istruzione Università RicercaProgrammi di Ricerca di Rilevante Interesse Nazionale) 2007 research funds (to R. Spiess, Università degli Studi di Padova, Italy), and Fondi Ateneo (to A. Caggianelli, Università degli Studi di Bari, Italy). We are grateful to P. Enrique and P. Mazzoleni for comments that helped us to improve the manuscript, and to D. Harry and A. Dini for their editorial work. We thank S. Rocchi for his considerable efforts in organizing the LASI III (laccolith and sill conference), and D. Liotta for encouraging discussions on the topics covered in this study.