We present data from northern Calabria, in southern Italy, which show that subduction may have initiated beneath a continental margin east of the Corsica-Sardinia-Calabria block during the Eocene. Calabria lacks ophiolites (oceanic rocks) within the upper plate, which are the strongest evidence for intraoceanic subduction initiation. The structurally higher nappes of the Calabrian subduction complex include continental crustal material, a feature that is also inconsistent with intraoceanic subduction initiation. In addition, Calabria lacks an amphibolite-facies metamorphic sole, regarded by some as a consequence of intraoceanic subduction initiation in young oceanic lithosphere. The age of blueschist metamorphism, compared with ages of arc volcanism in Sardinia, indicates preservation of rocks subducted and accreted during or shortly after subduction initiation, precluding significant subduction erosion. Peak subduction-related metamorphism reached blueschist facies, including rocks that apparently accreted early, during or shortly after subduction initiation. The protoliths of these rocks were ≥80 m.y. old at the time of subduction, indicating that subduction began in old and cold lithosphere along a continental margin. Subduction initiation here suggests that the serpentinization of the upper mantle, observed in the Tethyan rocks of Calabria, may have been important in weakening the oceanic lithosphere at the continental margin. Alpine subduction may have initiated in a similar manner along other reaches of the orogen.

Subduction beneath continents is a first-order plate-tectonic process, but the way in which it begins is poorly understood. Subduction initiation beneath a continental margin presents a paradox—although oceanic crust adjacent to passive continental margins is old and dense, it is also thick and modeled as rheologically strong and unlikely to break and allow subduction to initiate (Cloetingh et al., 1989; Mueller and Phillips, 1991; Nikolaeva et al., 2010). This kind of initiation appears likely, however, given the presence of long subduction zones along active continental margins, as well as evidence for subduction initiation along continental or microcontinental margins in the western Pacific (e.g., Hall, 2002; Harris, 2003). The scarcity of field evidence for present or ancient examples of subduction initiation beneath passive margins has hampered studies of this process (Stern, 2004).

In contrast, cases of subduction initiation within an ocean basin—and beneath oceanic lithosphere—are well documented in both modern and ancient examples (Stern, 2004). Upper-plate ophiolites, which overlie paleosubduction complexes, provide the best direct evidence of intraoceanic subduction initiation because they show that the upper plate of the subduction system consisted of oceanic lithosphere (e.g., Moores, 1970; Dewey and Bird, 1971). The presence of upper-plate ophiolites above subduction zones associated with continental-margin arcs has led to the recognition that subduction along some continental margins originated as intraoceanic subduction zones (e.g., Stern and Bloomer, 1992; Guilmette et al., 2009).

Intraoceanic subduction recorded in many orogenic belts may have initiated in young, hot oceanic lithosphere, at or near spreading centers, resulting in a thin high-pressure (HP), high-temperature (HT) sheet, or metamorphic sole, accreted structurally beneath the upper-plate ophiolite (e.g., Williams and Smyth, 1973; Spray, 1984; Nicolas, 1989; Hacker, 1990; Wakabayashi and Dilek, 2000, 2003). As this involves subduction of very young crust, the ages of ophiolite formation in the upper plate and amphibolite metamorphism in the sole are similar (Wakabayashi and Dilek, 2000). Blueschist-facies assemblages overprint the HT assemblages in some soles, likely recording continuing subduction that lowered the geothermal gradient (Wakabayashi, 1990; Dilek and Whitney, 1997). Alternatively, intraoceanic subduction has been proposed to initiate in comparatively old oceanic lithosphere, with development of upper-plate ophiolitic rocks in the early stages of subduction as a result of upper-plate extension associated with slab rollback following subduction initiation (e.g., Stern and Bloomer, 1992) and development of the metamorphic sole shortly afterward as a result of ridge subduction (Shervais, 2001).

Subduction initiation along a passive margin differs from intraoceanic subduction initiation in that the upper plate should lack oceanic crust. In addition, the oldest crust within ocean basins is adjacent to passive continental margins (Turcotte and Schubert, 1982) so that subduction initiation in such a setting would take place under conditions of low geothermal gradient, in contrast to the high geothermal gradients in young oceanic crust. In such cool conditions, insufficient heat is available to form HT metamorphic soles associated with subduction initiation in paleo-intraoceanic settings (e.g., Hacker, 1990). In contrast to intraoceanic settings, subduction initiation along passive continental margins should result in: (1) no ophiolite in the upper plate of the paleosubduction system, (2) protolith ages of early-subducted rocks that vastly exceed the age of subduction initiation, and (3) blueschist-facies, rather than amphibolite-facies, metamorphism for the earliest subduction-related metamorphic rocks.

The majority of the length of modern subduction zones features nonaccretion or removal of previously accreted or upper-plate material (e.g., von Huene and Scholl, 1991; Clift and Vannucchi, 2004). It is therefore conceivable that subduction erosion may follow initiation of subduction in an ocean basin, resulting in the removal of the metamorphic sole and upper-plate ophiolitic rocks.

In this paper, we present metamorphic petrologic observations, 40Ar/39Ar geochronology, and assessment of the regional geology from the Cenozoic accretionary wedge in northern Calabria, southern Italy. Our analysis suggests that field relationships are not an artifact of subduction erosion and indicates that subduction initiated along a rifted continental margin that was at least 80 m.y. old at the time.

The Calabrian orogeny resulted from complicated microcontinent tectonics associated with Africa-Europe convergence (Amodio-Morelli et al., 1976; Bonardi et al., 2001). Although it geographically links the sedimentary rock orogens of Italy (Southern Apennines) and Sicily (Maghrebides), Calabria primarily consists of igneous and metamorphic rocks and represents a fragment of the Cenozoic Alpine chain, first displaced with the Sardinia-Corsica block during the Oliogene backarc opening of the Ligurian Sea, and then as an independent microcontinent during Miocene opening of the Tyrrhenian Sea (Alvarez et al., 1974; Alvarez, 1976; Dewey et al., 1989).

Calabria can be divided into Northern and Southern Calabria based on the presence or absence, respectively, of ophiolitic rocks and blueschist-facies metamorphism (Bonardi et al., 2001). In this paper, we will focus on Northern Calabria, which preserves a more complete tectonic record (Fig. 1). Here, a subduction complex, composed of blueschist-facies nappes of both oceanic and continental affinity, was emplaced beneath Hercynian continental crust that underwent little or no Alpine metamorphism (Amodio-Morelli et al., 1976). There is along-strike variation in the nappe stack, with coherent units in the south and mélange units in the north.

In the southern part of Northern Calabria, the structurally highest unit is the Sila Unit, which preserves a nearly complete crustal section through the Paleozoic Hercynian orogen (Graessner and Schenk, 2001) (Fig. 2). In the Sila Unit, Alpine tectonic deformation has taken place entirely in the brittle field and most likely represents late, out-of-sequence thrusts (Acquafredda et al., 1994). As it underwent little or no Alpine metamorphism and is separated by a regional mylonitic fault from the subduction complex, it represents the upper plate of the subduction complex. Peridotites, pyroxenites, and serpentinites, interpreted to be ultramafic cumulates or subcontinental upper mantle, are locally associated with the base of the Sila Unit (Schenk, 1984; Piluso and Morten, 2004).

The nappes below the Sila Unit form an apparently inverted metamorphic stack, with peak metamorphic grade varying from amphibolite facies in the structurally high units to blueschist facies in the lowest units (Piccarreta, 1981). The uppermost unit that has undergone Alpine subduction—the Castagna Unit—is composed of three discontinuous members: an upper augen gneiss, a middle biotite gneiss, and a lower mica schist–dominated Zangarona Schist unit, all of which are Hercynian and pre-Hercynian metamorphic rocks deformed during the Alpine phase (Colonna and Piccarreta, 1976). Only the Zangarona Schist underwent epidote-blueschist-facies metamorphism during the Alpine orogeny (Colonna and Piccarreta, 1975). The other two members underwent a low-grade cataclastic-mylonitic event during the Alpine event (Colonna and Piccarreta, 1975; Paglionico and Piccarreta, 1976).

The phyllitic Bagni Unit underlies the Castagna Unit. It is not clear whether it is predominantly a Hercynian unit (Amodio-Morelli et al., 1976), or if it represents the cover of the underlying Calabrian ophiolitic units (Filice et al., 2011). Below, the Jurassic–Early Cretaceous Calabrian ophiolitic units consist of discontinuous exposures of mid-ocean-ridge basalt (MORB)–type basalt (Liberi et al., 2006), all containing some evidence of blueschist-facies metamorphism (Liberi et al., 2006; De Roever, 1972), locally associated with serpentinite and a Calpionella-bearing limestone and Jurassic radiolarite cover. These exposures represent off-scraped fragments of Tethyan Ocean crust incorporated into the Calabrian accretionary wedge (Lanzafame et al., 1979; Liberi et al., 2006; Tortorici et al., 2009). The use of the term “ophiolite” is somewhat misleading, since they do not represent extensive sheets of ultramafic and mafic rocks commonly associated with the term (e.g., Dilek, 2003).

Lower in the nappe stack, there are, in descending structural position, the Cretaceous quartzite and phyllite Frido Unit (Bonardi et al., 1988), and the Triassic–Miocene carbonate-platform Verbicaro Unit (Iannace et al., 2007), both of which preserve Tethyan marine sediments (Amodio-Morelli et al., 1976). It should be noted that the sequence described here generally agrees with the classic picture of Amodio-Morelli et al. (1976), which fit our field observations better than does the reinterpretation of Rossetti et al. (2001, 2004).

In the northernmost part of Calabria, the Frido Unit transitions along strike into several distinct mélange zones. These units are shale-matrix mélange with blocks of lower continental crust, metabasalt, and serpentinite (Spadea, 1982). In the structurally high western zone, the included blocks have undergone a blueschist overprint; in the structurally lower eastern zone, they are subgreenschist facies (Monaco and Tortorici, 1995). As higher levels of the tectonometamorphic stack are not present, there is disagreement as to what is the structurally highest unit in this region. Some consider fragments of Jurassic mafic rocks near Terranova di Pollino to represent remnants of an ophiolite that comprised the upper plate above the subduction complex (Knott, 1987, 1994), as they consist of <1 km metabasalt underlain by a serpentinite-matrix mélange bearing <10 m blocks of amphibolite and lower continental crust (Bonardi et al., 1988; Knott, 1987, 1994). Others consider these metabasalts to be blocks in mélange rather than klippen as interpreted earlier (Monaco et al., 1995; Monaco and Tortorici, 1995). We test this idea with a new 40Ar/39Ar age and find that these mafic fragments are not remnants of an upper-plate ophiolite; instead, they are blocks in mélange within the subduction complex, rather than the upper plate.

The timing of the start of Alpine subduction in Calabria is controversial. Many workers consider the Calabria-Corsica subduction system to have started in the Cretaceous (Amodio-Morelli et al., 1976; Principi and Treves, 1984; Knott, 1987; Faccenna et al., 2001). However, the first evidence of magmatic activity in Sardinia, associated with the Calabrian subduction zone, occurred in the Eocene (Lustrino et al., 2009), and there is increasing evidence that Cretaceous ages of HP rocks are erroneous (Brunet et al., 2000; Molli, 2008). Newer dates suggest that the Corsica-Calabria system may have started as late as the Eocene (Brunet et al., 2000; Lustrino et al., 2009).

The highest structural level of the Calabrian subduction complex, which may represent its earliest subducted and accreted nappe and the rock record of subduction initiation, includes rocks metamorphosed in the amphibolite facies. If these amphibolite-facies assemblages formed during the ca. 300 Ma Hercynian orogeny, they would have no relationship to Alpine subduction, whereas an Alpine age of metamorphism might indicate the development of a metamorphic sole during initiation of subduction. In the south, the Zangarona Schist member of the Castagna Unit is the highest blueschist-facies unit. The structurally higher augen gneiss member of the Castagna Unit is locally present; however, it has not undergone blueschist-facies metamorphism and may not have been fully subducted. In the north, the inferred structurally highest levels of the subduction zone are no longer present. However, amphibolites present as blocks in mélange may represent fragments of the upper plate or the structurally highest part of the subduction complex.

In the western Sila Piccola region, the uppermost nappe unit that experienced Alpine blueschist-facies metamorphism is the Zangarona Schist member of the Castagna Unit, a garnet mica schist with rare blocks of amphibolite metabasalt (Colonna and Piccarreta, 1975, 1976). Amphibolite is well exposed in a road cut (39°2′55.07″N, 16°16′6.50″E, WGS84) near the town of San Fili (Fig. 3) and as a tectonically dismembered block at Cozzo Volante (39°2′55.97″N, 16°10′28.73″E), near the town of San Mango d’Aquino.

At San Fili, petrographic and electron microprobe analyses show that the garnet amphibolite is composed of blue-green amphiboles (Mg-hornblende to ferrotschermakite; GSA Data Repository Table DR11) with almandine-grossular garnets (Table DR2 [see footnote 1]). Late glaucophane rims the early amphiboles (Fig. 4). Ti-bearing phases preserve rutile cores, with an intermediate ilmenite zone, surrounded by titanite rims. These amphibolites have been interpreted as Hercynian in age (Colonna and Piccarreta, 1975, 1976), but they have not previously been dated using radiogenic methods.

The Cozzo Volante amphibolite is similar to the San Fili exposure, with blue-green amphiboles (Mg-hornblende) rimmed by glaucophane (Table DR1 [see footnote 1]); however, it does not contain garnet. Zoning in Ti phases is likewise similar to the San Fili amphibolite. It is not clear if this exposure is part of the coherent Zangarona Schist nappe, or if it is a block within a sedimentary mélange. It was previously described as an actinolite schist and considered to be a part of the Tethyan Calabrian ophiolite unit (Piccarreta and Zirpoli, 1970).

About 100 km north of these localities, the tectonic stack is dominated by several structurally high levels of mélange, some of which preserve amphibolite blocks. Near Terranova da Pollino, an outcrop at Timpa Pietrasasso (39°59′59.34″N, 16°15′56.94″E) exposes a serpentine-matrix mélange with blocks of amphibolite and gneiss beneath a stratigraphically intact gabbro-basalt-radiolarite unit (Spadea, 1982; Bonardi et al., 1988). The high-Ti content of brown amphibole from the amphibolites (Table DR1 [see footnote 1]) indicates equilibration temperatures of nearly 1000 °C, using a semiquantitative thermobarometer (Ernst and Liu, 1998); this is consistent with the presence of melt textures and leucosome segregations. These blocks have been interpreted to be Hercynian lower crust, based on analogies with the diorite-kinzigite unit present throughout Calabria, but have not previously been radiometrically dated (Spadea, 1982; Monaco et al., 1995). As these are the highest-grade rocks in the area (Lanzafame et al., 1979; Spadea, 1982), we sampled them to test whether they are remnants of a metamorphic sole developed beneath oceanic crust.

We selected amphibolite samples from the San Fili, Cozzo Volante, and Timpa Pietrasasso outcrops for 40Ar/39Ar analyses. The San Fili amphibolite resulted in a nonplateau integrated age of 214.7 ± 0.9 Ma, possibly as a result of partial reheating during Tethyan rifting (Fig. DR1 [see footnote 1]), whereas the Cozzo Volante amphibolite yielded a forced plateau age of 319 ± 4 Ma (Fig. 5), and an isochron date of 304 ± 4 Ma. This date is broadly consistent with a late Hercynian event (Graessner and Schenk, 2001). Previous 40Ar/39Ar work on muscovites from Castagna Unit rocks from the Sila Piccola returned mixed ages rising from 23 to 47 Ma at lower-temperature steps to 177–205 Ma at higher-temperature steps, attributed to partial reheating during a low-grade Alpine event (Rossetti et al., 2004). Hornblendes from the amphibolite block-in-mélange from Timpa Pietrasasso yielded a plateau age of 193 ± 2 Ma (Fig. 5). This post-Hercynian age is consistent with an episode of extension that preceded the Jurassic rifting in this area (Piccardo, 2009; Liberi et al., 2011).

The structurally lower Tethyan ophiolitic units have a Jurassic to Early Cretaceous age of ophiolite formation. The Fuscaldo (De Roever, 1972) and Diamante-Terranova (Amodio-Morelli et al., 1976; Lanzafame et al., 1979; Cello et al., 1996) units are at least Tithonian-Neocomian (older than 130 Ma) in age, based on the presence of Calpionella-bearing limestone overlying the Fuscaldo Unit metabasites. These ages are further supported by a 136 Ma zircon fission-track age in the Fuscaldo Unit (Thomson, 1994). In the northern part of Calabria, at the Timpa Pietrasasso locality, Jurassic radiolarites directly overlie oceanic basalt (Bonardi et al., 1988).

All of the nappe units were subducted to Alpine blueschist-facies conditions. We obtained a new 40Ar/39Ar integrated age of 46.7 ± 0.4 Ma on HP phengites from the Diamante-Terranova unit at the La Guardiola locality, 1 km south of the town of Diamante (Figs. 6A and 6B). These occur in lawsonite-blueschist-facies metabasalts that did not undergo a greenschist-facies overprint (Shimabukuro et al., 2009); hence, we consider it to be a HP crystallization age. Because this is the oldest Alpine blueschist date obtained from Calabria, we use this as an estimate of the age of subduction for Calabrian units. This is consistent with the estimated 49–42 Ma age for the beginning of subduction beneath Sardinia, based on the presence of volcanic activity in Sardinia at 38 Ma, and estimated subduction velocities of 1–3 cm/yr, at a slab dip of 45°, with an 80 km depth of melting (Lustrino et al., 2009). It is also consistent with geochronology further north in Corsica, which suggests that subduction beneath the Corsica-Sardinia margin began as late as the Eocene (Brunet et al., 2000; Lustrino et al., 2009).

An examination of the regional geology shows that subduction in Calabria began beneath a continental margin. In the southern part of our study area, the upper plate of the Calabrian subduction zone is the continental Sila Unit, which was not metamorphosed during the Alpine orogeny (Acquafredda et al., 1994). Before the Miocene opening of the Tyrrhenian Sea, this unit was continuous with the Hercynian rocks of Sardinia (Kastens et al., 1988; Alvarez and Shimabukuro, 2009), and represented the forearc of the Eocene–Miocene continental arc in Sardinia (Lustrino et al., 2009).

In the northern part of our study area, it has been suggested that fragments of Jurassic oceanic crust represent the upper plate of the paleosubduction zone (Knott, 1987). Our new Jurassic age for the Timpa Pietrassaso amphibolite indicates that the high-temperature event is too old to be an Alpine metamorphic sole (or any other sort of Alpine metamorphic product) directly beneath an upper-plate ophiolite remnant. In addition, the discontinuous nature of the mafic fragments (Monaco et al., 1995; Monaco and Tortorici, 1995) and presence of Alpine HP metamorphism (Lanzafame et al., 1978; Spadea, 1982) indicate that these mafic fragments represent oceanic crust scraped off the subducting plate. The widespread presence of blocks of lower continental crust (Spadea, 1982; Monaco et al., 1995), similar to the Sila Unit, suggests that the upper plate—now absent—may have been part of the Hercynian Sila Unit. We interpret the Timpa Pietrasasso amphibolite as a remnant of the continental upper plate in the north part of our study area.

Within the accretionary wedge, the two structurally highest, and thus most proximal to the continental margin, nappe units are Hercynian crustal slivers (Fig. 2). These likely represent fragments of continental crust stranded near the Calabrian margin during Tethyan rifting (e.g., Schmid et al., 1996). Thus, the lack of ophiolitic rocks in the upper plate anywhere in Calabria, and continental material accreted during the early phases of subduction, implies that the Calabrian subduction zone initiated beneath a continental margin.

In intraoceanic settings where subduction initiates along or near a spreading ridge, newly formed oceanic crust may be subducted and metamorphosed to amphibolite-facies conditions, leading to the formation of a metamorphic sole beneath upper-plate ophiolitic rocks (Hacker; 1990; Wakabayashi and Dilek, 2003). The subduction of young crust beneath equally young crust is necessary to raise geothermal gradients high enough to produce amphibolite-grade metamorphism (Hacker, 1990). In such cases, the ages of the ophiolitic igneous protolith and its metamorphic sole are similar, with less than 10 m.y. separating the age of two units (Fig. 7A) (Wakabayashi and Dilek, 2000).

This is not the case in Calabria, where there is a large difference between the protolith age and the subduction metamorphic age (Fig. 4). The 40Ar/39Ar ages from the Zangarona Schist and Timpa Pietrasasso amphibolite blocks show that they record a metamorphic event far older than, and not associated with, Alpine subduction, so they do not represent a metamorphic sole formed at the inception of Alpine subduction. Instead, the amphibolites record the age of the crust, which, at the time of subduction initiation, was at least 260 m.y. old (304 Ma–47 Ma) and 150 m.y. old (193 Ma–47 Ma), respectively. Similarly, the Fuscaldo ophiolite—representative of the oceanic crust subducted beneath Calabria—was at least 80 m.y. old (130 Ma–47 Ma) at the time of subduction. At all nappe levels, cold crust, at least 80 m.y. old was subducted.

As a result of the subduction of old and cold crust, there does not appear to have been the necessary heat to cause an early amphibolite-grade metamorphic event. Our new dates indicate that structurally high amphibolites are associated with Hercynian or Tethyan events, and not with the early stages of Alpine subduction. There is a report of bluish tschermakite rims on amphiboles from the Castagna Unit that were interpreted as Alpine in age (Langone et al., 2006); however, we believe these to be Hercynian in age, as they are nearly identical to the blue-green tschermakites of the San Fili amphibolites. To the south, in the Aspromonte region of Calabria, amphibolite-facies metamorphism exists; however, this metamorphism is not restricted to a limited structurally high horizon as expected for a metamorphic sole, and it is most likely related to reheating following the continental collision that has taken place between the Calabrian nappes and the Sicilian continental crust (Bonardi et al., 2008; Heymes et al., 2010).

Alpine amphibolite-grade metamorphism, whether from collisional reheating, or by processes shortly before or during subduction initiation, is precluded in the remainder of Calabria by the complete lack of a high-temperature event recorded in the isotopic ages obtained from Calabrian metamorphic rocks (40Ar/39Ar white mica and hornblende) as well as the zircon fission-track date that shows that the Fuscaldo Unit has been below ∼250 °C since 136 Ma (Thomson, 1994). Instead, the first metamorphic evidence of Alpine subduction is the widespread blueschist-facies metamorphism present in Calabria (De Roever, 1972).

Are the field relationships an artifact of subduction erosion that has removed the metamorphic sole and/or ophiolite? Removal of oceanic materials from the upper plate would have required subduction erosion of the entire forearc region along the entire length of the orogen (Fig. 7A, lower frame). This does not appear likely based on the location of volcanic arc rocks in Sardinia relative to the restored position of the subduction complex as well as the age of the HP metamorphism compared to the age of arc volcanism. As noted previously herein, the 47 Ma age of the Diamante-Terranova unit appears to closely correspond to the time of subduction initiation based on the age of arc volcanic rocks in the region, and indicates preservation of units accreted early in Calabrian subduction. Such units would be removed by comparatively minor subduction erosion, let alone much more extensive subduction erosion that could remove the entire oceanic forearc region. The lack of Cenozoic (Alpine) amphibolite-grade metamorphism, coupled with the preservation of old, early-accreted units, also precludes a scenario of hot subduction initiation along a rifting continental margin, followed by removal of high-grade units by subduction erosion (Fig. 7B).

We conclude that subduction recorded in northern Calabria initiated in crust >80 m.y. old at the time, and, because of the cold conditions of subduction initiation, the peak metamorphism related to this event reached only blueschist grade. The subduction accretion of nappes of continental affinity as the structurally highest horizons of the subduction complex indicates that subduction initiated along a rifted continental margin that included slices of thinned continental crust (Figs. 7B and 7C).

If subduction took place within old and cold crust adjacent to a continental margin, what allowed this presumably strong and thick lithosphere to break? It has been proposed that an initial episode of extension adjacent to a continental margin may provide the thermal weakening to allow subduction to start (Kemp and Stevenson, 1996; Gurnis et al., 2004). Subduction beginning in newly extended crust may also result in high-temperature metamorphism at the initiation of subduction owing to subduction initiating in thin lithosphere with a high geothermal gradient (Fig. 7C). The metamorphic rock formed in such a process might be expected to resemble a metamorphic sole, except that it would be found structurally beneath continental crustal material instead of oceanic crust. In addition, the newly rifted continental crust should also undergo a widespread regional HT metamorphic event as a result of an increase in geothermal gradients due to asthenospheric upwelling, as well as likely synrifting magmatism (Wickham and Oxburgh, 1985). However, the lack of a metamorphic sole, or any Alpine HT metamorphic event, suggests that thermal weakening associated with active continental rifting did not occur prior to subduction initiation in Calabria.

Instead, recent studies of existing rifted continental margins show the presence of significant volumes of serpentinized mantle (Manatschal et al., 2007; Reston, 2009). Such serpentinized lithosphere may be weak enough to allow subduction initiation (Stern, 2004; Regenauer-Lieb et al., 2001; Hilairet et al., 2007). Furthermore, lithosphere in the Alpine Tethys appears to have formed in a slow-spreading ridge environment, with serpentinite commonly exposed at the ocean floor (Lagabrielle and Lemoine, 1997). In the Calabrian ophiolite units, serpentinite and ophicalcite bodies are common (Alvarez, 2005), suggesting that oceanic lithosphere adjacent to the paleo–continental margin in Calabria was serpentinite rich and hence weak enough for subduction to initiate. Analogous units in the Platta Nappe of the Alps (e.g., Deutsch, 1983; Manatschal et al., 2007) may suggest a similar mechanism for subduction initiation in the Alps.

We would like to thank Francesca Liberi and Eugenio Piluso for hospitality in the field and for wide-ranging discussions on Calabrian geology. We also would like to thank Rob Hall and an anonymous reviewer for helpful comments on the manuscript.

1GSA Data Repository Item 2012338, Figure DR1, additional 40Ar/39Ar analysis, and Tables DR1 and DR2, electron microprobe analyses of amphiboles and garnets, is available at www.geosociety.org/pubs/ft2012.htm, or on request from editing@geosociety.org, Documents Secretary, GSA, P.O. Box 9140, Boulder, CO 80301-9140, USA.