In this paper, new paleomagnetic results from the Calabrian Arc are presented, together with a critical review of all paleomagnetic data collected in the last decades in southern Italy. Our study is focused on the upper Miocene to middle Pleistocene deposits of the Crati extensional basin, a sector of the arc where an abrupt change in the sense of paleomagnetic rotations is observed. Paleomagnetic data indicate that the Crati basin underwent a uniform clockwise (CW) rotation of about 15°–20° in its central and southern part, whereas the northern sector is organized in small-scale fault-bounded blocks, which rotated independently. We interpret this pattern of deformation as the evidence of the complex nature of this area, which represents the boundary between two domains characterized by opposite rotations: the southern Apennines, which rotated counterclockwise, and the Calabria and Sicily, which rotated CW. Integrating these new paleomagnetic data with paleomagnetic data from southern Italy, we reconstruct the history of paleomagnetic rotations through time. Paleomagnetic rotations highlight the peculiarity of the formation of the Calabrian Arc curvature and imply that either an oroclinal bending model or a progressive arc model cannot be simply applied to the Calabrian Arc formation. We describe a realistic tectonic-geodynamic model, where the progressive curvature of the Calabrian Arc is framed within the space-time evolution of the Ionian subduction system.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
 During the past few decades, paleomagnetism has been used as a fundamental tool to assess kinematic models of curved orogenic systems around the world because of its great potential in quantifying vertical axis rotations [Carey, 1955; Eldredge et al., 1985; Marshak, 1988; Van der Voo et al., 1997; Weil and Sussman, 2004]. On the basis of the spatial and temporal relationship between deviations in structural trend and the vertical axis rotation that took place within the belt, arcuate belts have been interpreted as primary, secondary or composite features. Primary arcs initiate in their present curved form and their curvature does not increase during subsequent deformation. Conversely, oroclines [e.g., Carey, 1955; Marshak, 1988] are originally linear fold-thrust belts that acquire curvature in a second phase of tectonic deformation, accompanied by opposite large-scale vertical axis rotations shaping the arc. Between these two end-members, progressive arcs represent either orogenic belts that acquire their curvature during progressive deformation or belts that acquire a portion of their curvature during a subsequent deformation phase [Vogt, 1973; Weil and Sussman, 2004]. This intermediate category describes the formation of most of ancient and modern curved orogenic systems. In addition to the kinematics of the arc and the timing of its curvature, a crucial factor for understanding the modality of belts curvature is the knowledge of the geodynamic process governing arc formation. In active convergent margins, slab roll back process is considered the most likely candidate for the formation of arcs [e.g., Schellart and Lister, 2004]. In particular, the nature of both the overriding plate and the subducting slab, their geometry, width and space-time evolution, are elements that must be necessarily taken into account to fully understand the formation and evolution of related arc-shaped belts [Faccenna et al., 2004; Morra et al., 2006; Schellart and Lister, 2004].
 The Mediterranean area shows a large number of narrow arcs, defining an irregular and rather diffuse plate boundary (Figure 1). In particular, the boundary between the Africa and Eurasia plates runs parallel to the North African coast and then turns northward forming the tight arcs of Gibraltar and Calabria, at the western and eastern side of the western Mediterranean, respectively. A common feature of these curved mountain belts is their location on top of very narrow subducting slabs, well-defined by seismicity and seismic tomography [e.g., Wortel and Spakman, 2000], suggesting that the origin of such arcs and related back-arc basins formed as a consequence of trench rollback during Neogene and Quaternary [Faccenna et al., 2004, and references therein]. In southern Italy, the arcuate trend of the Calabrian Arc is defined by the regional variation in the strike of fold axes, striking from NW-SE with a NE vergence in southern Apennines to E-W with a southern vergence across Sicily (Figure 1). From a general point of view, paleomagnetic data show a correlation between the changes in paleomagnetic declinations and the changes in trend of the orogen, indicating that the present-day shape of Calabrian Arc is a secondary feature, achieved throughout circa symmetrical opposite rotations along the two limbs of the belt, with counterclockwise (CCW) rotations in southern Apennines and clockwise (CW) rotations in the Calabria-Sicilian Maghrebides [e.g., Channell et al., 1980, 1990; Gattacceca and Speranza, 2002; Speranza et al., 2003, 1999]. However, despite the agreement on the secondary nature of the curvature of southern Apennine, uncertainty remains about the structural evolution of the Calabrian Arc. In particular, the following aspects need further consideration. Along the Calabrian Arc a peculiar pattern of paleomagnetic rotations is observed. Rather than a progressive change in the sense and magnitude of along-strike paleomagnetic rotations (as expected in a typical orocline), an abrupt change of the direction of paleomagnetic directions occurs along a very narrow region located at the boundary between southern Apennines, where counterclockwise rotations occurred, and Calabria, where clockwise rotations have been measured [e.g., Gattacceca and Speranza, 2002; Scheepers et al., 1994; Speranza et al., 2000]. Furthermore, the core of the Calabrian Arc is constituted by a distinct polymetamorphic orogenic domain (hereafter referred as Calabro-Peloritane Domain, CPD) that recorded the early phases of the Apennine orogenic construction and the subsequent extension-related exhumation phase [e.g., Rossetti et al., 2004]. Since the middle to late Miocene the CPD, originally located in the innermost Apennine chain sector, was affected by postorogenic extension and drifted away from Sardinia to be finally assembled to the southern Apennine-Maghrebide orogenic system [Alvarez et al., 1974; Bonardi et al., 2001; Faccenna et al., 1997; Malinverno and Ryan, 1986; Mattei et al., 2002]. As a consequence, the present shape of the Calabrian Arc is the result of a complex tectonic history juxtaposing a drifted crustal block (CPD) onto the two deforming orogenic belts (southern Apennines and Sicilian Maghrebian chain). These observations indicate that a simple oroclinal model [Eldredge et al., 1985] is not sufficient to describe the complex evolution of the Calabrian Arc curvature.
 In this paper, we present new paleomagnetic results from the upper Miocene to middle Pleistocene sedimentary deposits filling the tectonic depression of the Crati basin in northern Calabria (Figure 2). Paleomagnetic results from this study show that the deformation pattern at the boundary between Calabria and southern Apennines is accommodated by different, small sized fault-bounded blocks, characterized by complex vertical axis rotations. After a critical review of the previous paleomagnetic data available in southern Italy, we propose a tectonic and geodynamic model for the progressive curvature of the Calabrian Arc as a consequence of the geometry and evolution of the Ionian subduction system.
2. Geological Setting
2.1. Tectonic Setting of the Calabrian Arc
 The Calabrian Arc defines a mountain belt encircling the Tyrrhenian Sea, from southern Apennine to Sicilian Maghrebides (Figure 1). According to the current geodynamic models, the evolution of the Calabrian Arc during the Neogene and Quaternary was driven by the southeastward retreat of the Ionian slab [Faccenna et al., 2001; Malinverno and Ryan, 1986]. The Calabrian Arc is presently located on top of a narrow (roughly 200 km) and steeply (70°) subducting slab [Anderson and Jackson, 1987; Lucente et al., 1999; Selvaggi and Chiarabba, 1995; Spakman et al., 1993]. No vertical discontinuities have been recognized in such structure by seismic tomography and deep seismicity distribution, suggesting the existence of a continuous subducting lithosphere from the present-day Ionian Sea, for a total length of about 700 km [Piromallo and Morelli, 2003; Wortel and Spakman, 2000]. The presence of seismicity on a well-defined Benioff zone reveals a direct trace of a still active subduction of the Ionian realm, a Mesozoic oceanic lithosphere, intervening between the Apulia and Africa continental margins. Earthquake distribution indicates that the present-day width of the subducting slab is very narrow and that its northeastern boundary corresponds, at the surface, with the boundary between southern Apennines and Calabria (Figure 1).
 Since middle-late Miocene, back-arc extension and rifting processes in southern Tyrrhenian Basin were synchronous with the outward migration of compressional phases in southern Apennines-Sicilian Maghrebides toward the Apulia-Hyblean foreland. Trench retreat of the Ionian slab generated the opening of the southern Tyrrhenian Sea back-arc basin and the overprinting of the early Alpine compressional structures by extensional tectonics [e.g., Rossetti et al., 2004]. The fast retreat of the Ionian slab was responsible of the southeastward drifting of the CPD, which was finally incorporated in the Apennine-Maghrebian chain, to form the present-day Calabrian Arc. As a consequence the Calabrian Arc is characterized by two sectors, the CPD and the southern Apennines and Sicilian Maghrebides thrust and fold belts, which show a different tectonic evolution, structural architecture, and deep lithospheric structure.
 The geological architecture of northern Calabria consists of an Eocene orogenic nappe stack accreted onto the Apulian paleomargin and is made of Hercynian basement rocks and Alpine polymetamorphic sequences derived from the deformation of the southern Tethyan margin. This orogenic pile is unconformably covered by middle Miocene to Recent postorogenic synrift to postrift sediments along the Tyrrhenian side and by a thick Oligocene to Pleistocene forearc sedimentary sequence along the Ionian side [e.g., Bonardi et al., 2001].
2.2. Postorogenic Cycle in Northern Calabria and the Crati Basin
 Northern Calabria is characterized by discontinuous outcrops of Miocene clastic successions ascribed to the basal part of the “postorogenic” cycle [Bonardi et al., 2001, 1976; Ogniben, 1973], which fills extensional basins formed during the collapse of the orogenic wedge. The onset of brittle postorogenic extensional deformation in Calabria is well documented in the Amantea basin [Argentieri et al., 1998; Mattei et al., 2002], located at the southern termination of the Coastal Range (Figure 2), where the postorogenic sedimentary sequence starts with alluvial plain conglomerates that grade upward to fan delta coarse-grained massive marine sandstones and claystones, Serravallian in age.
 The N-S trending Crati basin represents the most prominent structural feature associated with the postorogenic extensional tectonics in north Calabria (Figure 2). The Crati basin is filled by Miocene to Pleistocene sediments, thickening northward. The Miocene clastic deposits show general affinities with the Amantea basin, even if they may differ in the stratigraphic and depositional features and in the age of the basal deposits, which tend to become younger (Tortonian) to the inner sectors of the chain [Di Nocera et al., 1974; Ghisetti and Vezzani, 1982; Lanzafame and Zuffa, 1976; Morrone, 1991; Perrone et al., 1973; Romeo and Tortorici, 1980; Spadafora, 1990]. Furthermore, in the Crati basin Plio-Pleistocene deposits outcrop extensively and are represented by ∼500 m of upper Pliocene–lower Pleistocene marine deposits composed of conglomerates, sands and clays, several hundred meters of lower-middle Pleistocene marine clays and a 100 m of middle Pleistocene beach sands and conglomerates. The last sedimentary sequence observed in the Crati basin is represented by continental fanglomerates and sands, mostly outcropping along the mountain front [Colella, 1988; Lanzafame and Zuffa, 1976; Tortorici et al., 1995; Vezzani, 1967].
 The structural architecture of the Crati basin has been recently described by Cifelli et al. . It is defined by an overall half-graben geometry, bounded to the west by eastward dipping (60°–80°), N-S striking extensional faults, bordering the Coastal Range at its eastern side. In plan view, the border faults show a composite pattern, consisting of subordinate NE-SW and NW-SE striking fault segments (Figure 2). In particular, in the northwestern sector of the Crati basin, the fault array shows a complex pattern of interfering, nearly orthogonal NE-SW and NW-SE striking extensional fault strands (Figure 2). The boundary of the Crati basin is given by the ENE-WSW oriented Pollino fault, which separates the Crati basin from the Pollino structure, a Meso-Cenozoic carbonatic ridge, which forms the southern part of the southern Apennines thrust belt. Geologic, structural and sedimentologic data suggest that this tectonic lineament has been active during the early Pleistocene as a left-lateral strike-slip fault [Colella, 1988] and later on as normal fault [Schiattarella, 1998]. In the central and southern Crati basin, two main normal fault systems have been recognized: a major northwest dipping NE-SW striking extensional fault systems, controlling the Miocene basin stratigraphy and architecture, and a N-S striking high-angle, conjugate extensional fault arrays, which mainly border the Pliocene-Pleistocene tectonic depressions (Figure 2). In Miocene deposits the greater amount of extension in the early stages of basin formation was primarily accommodated by westward dipping NE-SW low-angle extensional faults. Conversely, the N-S striking extensional fault systems overprinted this deformation pattern reconstructed from the Miocene deposits, and it is characterized by high-angle faults [Cifelli et al., 2007].
3. Paleomagnetic Analyses
3.1. Paleomagnetic Sampling
 In this study, a paleomagnetic sampling was carried out in the postorogenic sedimentary sequences exposed in northern Calabria (Figure 2). We focused paleomagnetic sampling in the Crati basin, where Miocene to Quaternary deposits crop out extensively. Out of the 51 sampled sites, 31 were sampled in Miocene deposits, 6 in Pliocene deposits and 14 in Pleistocene deposits (Figure 2). Eight to twelve 2.5 cm diameter cores were drilled at each site with a gasoline-powered drill and oriented with a magnetic compass. In Calabria, the geomagnetic declination observed in the last few years is 2°–3° in the average. This can be considered negligible for the tectonic interpretation of paleomagnetic data. When possible, cores were taken from different stratigraphic levels in order to average secular variation of the geomagnetic field and to check the occurrence of polarity reversals. Most of paleomagnetic measurements were made at the paleomagnetic laboratories of the Department of Geological Sciences of the “Roma TRE” University and part of them at ETH Laboratory of Natural Magnetism in Zürich and at the Istituto Nazionale di Geofisica e Volcanologia Paleomagnetic Laboratory in Rome.
3.2. Rock Magnetism
 The magnetic mineralogy of the sampled sediments was investigated using one specimen from each site, with different standard rock magnetic techniques. The stepwise acquisition of a isothermal remanent magnetization (IRM) was carried out using a pulse magnetizer to apply magnetic fields up to 2.1 T. Also, a three-component IRM was thermally demagnetized in one specimen for site, according to the procedure reported by Lowrie . Maximum, intermediate, and minimum fields applied to the x, y, and z sample axes were 1.7 T, 0.6 T, and 0.12 T, respectively, for all the specimens. Low-coercivity ferrimagnetic minerals were identified in most of the claystone sites (Figures 3a and 3b). For some specimens, the maximum unblocking temperature spectra are in the range of 320–360°C, which suggests that the magnetic mineralogy is dominated mainly by iron sulphides (Figure 3d). In other claystone sites, the maximum unblocking temperature is about 580°C, indicating magnetite as the main magnetic carrier (Figure 3e). High-coercivity magnetic carriers were identified in very few sites, which show a blocking temperature of about 680°C, indicating the presence of hematite together with low-coercivity minerals (Figures 3c and 3f).
3.3. Demagnetization of the Remanent Magnetization
 The natural remanent magnetization (NRM) of the specimens was analyzed both by means of progressive stepwise thermal demagnetization and stepwise alternating field demagnetization. For each site, sister samples of the same core were both previously demagnetized, thermally and by alternating magnetic field, in order to find the more efficient demagnetization technique for each site. Similar results for AF and thermal demagnetization were typically obtained on sister samples from the same cores. Demagnetization was stopped when NRM decreased to the limit of the instrument sensitivity or when random changes of the paleomagnetic directions appeared. In some cases samples had a low NRM (lower then 10−5A/m) or were almost totally demagnetized after the first heating (120°–170°C) or AF (20–25 mT) steps. These samples were discarded and not considered for further analyses. Data were plotted using orthogonal demagnetization diagrams and directions of the remanence components were estimated using principal component analyses [Kirschvink, 1980]. Only specimen yielding maximum angular deviation (MAD) < 10° were considered suitable for further analyses. In most of AF demagnetized samples, after removal of small viscous components by 10–20 mT, a well-defined chemical remanent magnetization (ChRM) was isolated and demagnetized at 60–80 mT (Figure 4a). In most of the thermally demagnetized samples, a viscous component was generally recognized, which was removed at about 120–180°C. Afterward, a ChRM was generally isolated, and specimens were completely demagnetized by 520–580°C (Figures 4b and 4c) or by 360–400°C (Figures 4d, 4e and 4f), confirming the presence of magnetite and iron sulphides as main magnetic carriers. In some samples, we were not able to isolate a well-defined ChRM, and two components of magnetization coexisted. In this case, the site-mean paleomagnetic directions were calculated using the great circle method of McFadden and McElhinny .
3.4. Paleomagnetic Data From Crati Basin Rocks
 Site-mean paleomagnetic directions were calculated using Fisher's  statistics on sites where only ChRMs were isolated, whereas the McFadden and McElhinny  method was used on sites where both ChRMs and remagnetization circles were observed. Some of the analyzed sites (CT14, CT15; CT16 and CT25) show a normal polarity, with a mean direction in geographic coordinates close to the present-day GAD magnetic field (Table 1), and, when considered at the basin scale they show a negative fold test according to McFadden and McElhinny . This evidence suggests that these sites may have been remagnetized in recent times and were not further considered for tectonic interpretations.
Table 1. Paleomagnetic Data From Rocks in the Crati Basina
Before Tectonic Correction
After Tectonic Correction
The n1 and n2/N are number of stable directions and number of great circles/total number of studied samples at a site. D, I are site-mean declinations and inclinations calculated before (DBTC, IBTC) and after (DATC, IATC) tectonic correction. The sites mean for each area, not including (for areas 2, 3, and 4) sites considered remagnetized and sites uppermost late Pleistocene–middle Pleistocene in age (see text for explanation); k and α95 are statistical parameters after Fisher . Paleomagnetic sites from Scheepers  are also reported (FIR, Firmo; CSV, Castrovillari; SMA, San Marco Argentano; SCA, Santa Caterina Albanese; ROS, Rossano/Cerisano; REN, Rende; LUZ, Luzzi). S0 is bedding attitude (azimuth of the dip and dip values).
Sites considered remagnetized.
Deduced by AMS tensor.
Sites uppermost late Pleistocene–middle Pleistocene in age.
 Acceptable paleomagnetic directions were obtained from 33 sites and results are reported in Table 1, together with the site-mean calculated paleomagnetic directions. A complex pattern of vertical axis rotations was evidenced along the basin, forming different rotational domains (Figure 2). In the northeastern sector, nine sites were sampled in Tortonian–middle Pleistocene deposits (area 1 in Figure 2). Reliable paleomagnetic data were obtained from four sites sampled in the uppermost part of the lower Pleistocene (CT24 and CT26) and middle Pleistocene deposits (CT37 and CT38). Both normal and reverse polarities have been observed, which show almost antipodal directions. Paleomagnetic declinations can be compared with paleomagnetic data from uppermost part of the lower Pleistocene sites of Scheepers . After tilt correction, sites have both normal and reverse polarity, which are almost antipodal (the reversal test is of type Rc according to McFadden and McElhinny ). The mean paleomagnetic direction, including results from the sites from Scheepers  and transforming all sites to normal polarity is D = 2.5°; I = 53.6°; k = 130.6; α95 = 5.9 (Figure 5a).
 In the northwest sector, two different rotational patterns are observed. In the northwesternmost edge of the basin (area 2 in Figure 2), out of the 12 sampled sites, six sites from Miocene deposits (CT10-11, CT18, CT20-21, CT31) and two sites from upper Pliocene–early Pleistocene deposits (CT22 and CT32) give reliable results. Both normal and reverse polarities were observed (reversal test is indeterminate, according to McFadden and McElhinny ). The site-mean paleomagnetic directions are better grouped after (D = 339.6, I = 43.3, k = 17.2, α95 = 13.7) than before tectonic correction (D = 349.2, I = 44.8, k = 10.6, α95 = 17.9) and the fold test is positive at 99% of confidence, according to McFadden and McElhinny . The mean paleomagnetic direction, when all the sites are reported to the normal polarity, is D = 339.6° and I = 43.3° (Figure 5b). South of this area, an opposite rotational pattern is recorded. Five sites (CT28-30, CT34 and CT39) from the six sites sampled in this area give reliable results. Sites from Tortonian to early Pleistocene clays give a consistent CW paleomagnetic rotation, whereas one site from middle Pleistocene clays (CT39) shows no significant rotation and is considered separately (Table 1 and Figure 5c). After tilt correction, the four rotated sites show either normal or reverse polarity (reversal test is indeterminate, according to McFadden and McElhinny ). After tectonic correction and including the two sites sampled from Scheepers , the mean ChRM from the Tortonian to early Pleistocene sites gives a magnetic declination D = 29.3° and a magnetic inclination I = 52.5°, with k = 49.7 and α95 = 9.6° (Figure 5c). The fold test is positive at 99% of confidence according to McFadden and McElhinny .
 In the central and southern sector of the Crati basin, 19 sites were sampled in Tortonian to middle Pleistocene claystones. In order to define the amount of paleomagnetic rotations we excluded site CT41 taken from middle Pleistocene deposits, where no paleomagnetic rotations were observed. Both normal and reverse polarities have been observed, which show antipodal directions. The reversal test is of type Rc (γo = 14.5°; γc = 19.5°) and the fold test is positive (with a maximum K at 84% of unfolding) at 95% of confidence. Paleomagnetic results indicate that upper Miocene deposits (except CT08 and CT04 that show CCW and no rotation, respectively) underwent a CW rotation. Including results from Scheepers , the mean sites direction before tectonic correction is D = 14.0, I = 54.0, k = 16.8, α95 = 9.3, whereas after tectonic correction is D = 17.2, I = 48.5, k = 20.9, α95 = 8.3 (Figure 5d).
4. Postorogenic Tectonic Architecture of Northern Calabria
 Paleomagnetic data collected in southern Italy (Figure 6) show that the Crati basin is located between two geologic sectors, southern Apennines to the north and Calabria and Sicily to the south, affected by opposite tectonic rotations (CCW and CW, respectively) [Channell et al., 1980, 1990; Gattacceca and Speranza, 2002; Speranza et al., 2003, 1999]. In particular, along this narrow area, an abrupt change of the directions and values of paleomagnetic rotations occurs, suggesting a quite complex tectonic evolution. New paleomagnetic data obtained in this work show that the Crati basin can be subdivided into various rotational domains, each characterized by a near-uniform amount and sense of paleomagnetic rotations. The dispersions in paleomagnetic directions observed in area 2 and area 4 are due to small-scale rotations related to local tectonic structures.
 In the northwest corner of the basin, both CCW rotations (area 2 in Figure 2) and CW rotations (area 3 in Figure 2) were observed. This different rotational pattern can be related to the tectonic architecture of the northwest sector of the Crati basin, characterized by the activity of different fault systems (striking NW-SE and NE-SW, respectively). These faults subdivide this part of the basin into smaller sized blocks, filled by marine and continental sedimentary sequences, including tectonically controlled fan delta systems [Colella, 1988]. The occurrence of fault-bounded block rotations in north Calabria, and the consequent segmented architecture of the postorogenic extensional faulting, implies that the postorogenic evolution of this sector is quite complex from a tectonic point of view. Rather, the extensional domain is subdivided into different structural compartments, individuated by the adjustment in the faults strike imposed by the regional rotation pattern [Cifelli et al., 2007].
 In the southern and central parts of the Crati basin (area 4 in Figure 2), only CW rotations have been observed. In this area, structural data indicate that N-S striking faults cut the preexisting NE-SW ones in response to the regional CW rotations. Such structural evolution is a common feature in Calabria and has been also recognized in the Amantea basin [Mattei et al., 2002], located toward the west, along the Tyrrhenian coast. In this area, the entire Serravallian to late Tortonian basin evolution was controlled by NE-SW oriented normal faults, which were later dissected by N-S oriented normal faults, which represent the most prominent geologic and morphologic feature of such basin. Paleomagnetic data suggest that these complex relationships among fault bounded blocks, recognized both in the southern or central Crati and in Amantea basin, could be the result of CW rotation of the Calabria terrane in this area, which caused the NE-SW oriented normal fault systems to become unfavorably oriented with respect to the active stress field. As a consequence, this fault system became locked. At the same time N-S oriented normal faults started to form and to control the Pliocene and Pleistocene basin evolution in the Crati area. The superposition of differently oriented fault systems is a typical feature in active regions characterized by rotating fault-bounded crustal blocks, such as, for example, central Greece [Jackson, 1994; Mattei et al., 2004a; McKenzie and Jackson, 1986]. There, seismologic, structural, geomorphologic and paleomagnetic data show that the complex deformation associated with postorogenic extension on top of an active subducting slab is accommodated by rotations about vertical axis of independent fault bounded blocks, and that vertical axis rotations are able to control the duration of normal faults segments, tens of kilometers long, and their structural architecture. These data indicate that the upper part of the continental crust undergoing postorogenic extension was formed by distinct, fault-bounded rigid blocks, which differently rotated during continuous extensional shearing. Paleomagnetic data from this study confirm the nature of transitional region of the Crati basin evidenced by different senses of rotation moving from the southern Apennines to Sicily. The northwest sector (area 2 in Figure 2) shows strong similarities with the counterclockwise rotational pattern detected for the southern Apennines block [Gattacceca and Speranza, 2002; Mattei et al., 2004b; Sagnotti, 1992; Scheepers and Langereis, 1994; Scheepers et al., 1993]. On the other hand, the southern and central sector of the study area (areas 3 and 4 in Figure 2) behaves coherently, with the clockwise rotational pattern defined for both Calabria and Sicily [Aifa et al., 1988; Besse et al., 1984; Channell et al., 1990; Duermejier and Langereis, 1998; Scheepers and Langereis, 1993; Speranza et al., 1999].
5. A Critical Review of Paleomagnetic Data From Southern Italy
 In order to describe a complete rotational history of the Calabrian Arc, we integrated new paleomagnetic data from this study with published results from southern Apennines, Calabria and Sicily. A total of 132 paleomagnetic directions have been computed from about 500 sites, collected either for paleomagnetic or magnetostratigraphic investigations from Middle Jurassic to Pleistocene strata (Table 2). In this study we use the African apparent polar wander path as a proxy for Adriatic geodynamics seems reasonable, as paleomagnetism in the Tethyan domain has robustly proven a post-Permian Africa-Adria paleomagnetic coupling [e.g., Channell and Tarling, 1975; Muttoni et al., 2001].
Table 2. Paleomagnetic Data From Rocks in Southern Italya
 In Mesozoic sites, paleomagnetic rotations were computed in relation to the Adriatic and Hyblean forelands, using the coeval African paleopoles from Besse and Courtillot . Rotation values and associated 95% confidence limits were calculated according to the method of Demarest . Conversely, the declination values observed for the Tertiary and Quaternary strata have been compared with the local north to obtain rotation values, as the African plate underwent no significant rotation during the same time span [Besse and Courtillot, 2002]. The overall data set is reported in Table 2 and in Figure 6, where paleomagnetic rotations have been differentiated according to the age of sampled rocks. Finally, paleomagnetic rotations versus age diagrams have been produced separately for southern Apennines, Sicily and Calabria (Figures 7 and 8) .
 The paleomagnetic data (Figure 6) show that rotations are confined to the orogenic wedge, and do not extend to the Apulia and Hyblean forelands, which did not undergo tectonic rotations (see also Table 2). Accordingly, in the following we focus our analyses on paleomagnetic data collected in Sicily and the southern Apennines orogenic nappes, and then we analyze results from the CPD.
5.1. Southern Apennines and Sicily
 Paleomagnetic data from Mesozoic and lower Tertiary sedimentary sequences in southern Apennines and in Sicily come from different paleogeographic units of the southern Tethys passive margins, which presently form the fold and thrust belt of the southern Apennines and Maghrebian chain. Paleomagnetic results show that such units record large-magnitude, opposite sense, rotations either in Sicily or in southern Apennines, CW and CCW, respectively (green arrows in Figure 6).
 In Sicily, paleomagnetic data have been collected in the “Ammonitico Rosso” red nodular limestones (Middle Jurassic), in “Scaglia Rossa” pelagic limestones (Upper Cretaceous), and in “Scaglia” pelagic limestones (middle Eocene–Oligocene) [Channell et al., 1980, 1990; Nairn et al., 1985; Schult, 1976; Speranza et al., 2003]. Results show a decrease in paleomagnetic rotations from the more internal Panormide-Imerese and Sicanian units (up to 145° CW), to the intermediate paleogeographic units of the Trapanese and Internal Saccense (50° of CW), whereas the external, undeformed units (External Saccense) are substantially unrotated (Figure 7a). In the southern Apennines, paleomagnetic data come from Cretaceous to middle Miocene carbonate platform sediments (Apenninic or Internal Platform) and from Jurassic to Cretaceous pelagic calcareous-marly sediments related to shelf-to-basin transition [Catalano et al., 1976; Gattacceca and Speranza, 2002; Jackson, 1990]. These paleogeographic realms now outcrop along the internal part of the Apenninic chain and have been folded and thrusted during the Apenninic orogeny. Results show a constant amount of CCW paleomagnetic rotations (70°–80° in average) which have been measured either in Jurassic to Oligocene units or, locally, in middle Miocene sediments.
 Paleomagnetic results from upper Miocene and Pliocene deposits (blue arrows in Figure 6) have been mostly sampled in foredeep and piggyback basins lying in the external part of the Maghrebian and southern Apennines thrust belts (Table 2 and Figures 6 and 7). Samples from Sicily come from upper Miocene strata, which consist in Tortonian siliciclastic units and Messinian carbonates outcropping in the Caltanisetta thrust top basin [Butler et al., 1992; Speranza et al., 2003]. Results from Tortonian sites show either CW or CCW rotations (the latter considered as related to local tectonic phenomena), whereas Messinian sites show almost uniform CW rotations (66° in average). Pliocene data come from the lower Pliocene Trubi Formation, extensively sampled in the piggyback basins developed on top of the Gela Nappe during its southward thrusting [Duermejier and Langereis, 1998; Scheepers and Langereis, 1993; Speranza et al., 2003, 1999]. Most of the data have been collected for magnetostratigraphic analyses, and show a significant magnitude of CW rotation, ranging from 20° up to 60°, giving an average value of about 40° (Figure 7a). In southern Apennines, only one result is available from upper Miocene foredeep siliciclastic sediments outcropping along the Tyrrhenian side of the chain, which show about 40° of CCW rotations, whereas most of the data have been obtained from Pliocene strata collected in claystones from the Potenza, Calvello, and Sant'Arcangelo foredeep and piggyback basins [Sagnotti, 1992; Scheepers et al., 1993]. In these basins, results show a variable amount of CCW rotation, ranging from 14° to 37° (Figures 6 and 7b and Table 2).
 In the Calabro-Peloritane Domain (CPD), paleomagnetic data have been mainly collected in upper Miocene to Pleistocene sedimentary sequences cropping out both in the postorogenic basins along the Tyrrhenian coast and in the forearc basins located along the Ionian side of the region (Figure 6). Notwithstanding the large amount of paleomagnetic results, no data are available from Mesozoic and lower Tertiary units, hindering the possibility to reconstruct the older rotational history of the CPD crustal block. With the exception of the northwestern sector of the Crati basin (area 2), where CCW rotations were measured, upper Miocene to Pleistocene paleomagnetic sediments show a general CW rotational pattern along the entire CPD. Serravallian to upper Tortonian sediments sampled from the Tyrrhenian and Ionian side of Calabria show similar 20° CW rotations. This same value has been also obtained from Pliocene to lower Pleistocene strata in Calabria and Peloritani area [Aifa et al., 1988; Cifelli et al., 2004; Mattei et al., 2002; Scheepers et al., 1994; Speranza et al., 2000; Tauxe et al., 1983] (Table 2 and Figures 6 and 8). Similar to southern Apennines and Sicily, in the CPD the uppermost lower Pleistocene–middle Pleistocene sediments generally do not show appreciable rotations. This is well documented in the Crati basin, where paleomagnetic data from the uppermost early Pleistocene–middle Pleistocene sedimentary strata are not affected by tectonic rotations (Figure 9), whereas a small amount of clockwise rotation has been registered in central Calabria (Figure 6). These data indicate that no significant vertical axis rotations affected sediments of that age, suggesting that the major episodes of the Calabrian Arc bending were almost completed by that time.
6.1. Timing of Paleomagnetic Rotations in the Calabrian Arc
 The paleomagnetic rotations estimated for Sicily, southern Apennines and CPD and their correlation with their age (Figures 7 and 8) allow us to unravel the history of paleomagnetic rotations through time and to depict the evolution of the Calabrian Arc curvature.
 In Sicily and in southern Apennines, paleomagnetic rotation versus age diagrams are characterized by two different trends: (1) in the early segment (highlighted by dashed ellipses in Figure 7), which spans Jurassic to Oligocene–middle Miocene time, values of paleomagnetic rotations are very high but remain almost constant during the entire time interval; and (2) in the second segment, which corresponds to the middle-late Miocene to Pleistocene time interval, a progressive decrease in paleomagnetic rotations with time occurs. These different trends define a change in the geodynamic setting, and mark the progressive incorporation of different paleogeographic domains in the Apennines and Maghrebian orogenic wedge. The early segment of the diagrams show that the internal Apenninic and Sicilian units did not rotate respect to Africa during this time span, when they were still part of the passive margin of the African-Apulia plate or represented the foreland of the former Apenninic and Maghrebian chain, before their paleogeographic realms were incorporated in the Apennine and Maghrebian fold and thrust belt. The magnitude of paleomagnetic rotations in these rocks appears to be a function of their position within the thrust belt. The more internal tectonic units are systematically more rotated with respect to the external ones [Oldow et al., 1990; Speranza et al., 2003]. The transition between the two segments of the diagrams, even if not yet well defined in detail, allows defining the beginning of rotational processes in Sicily and in southern Apennines. Speranza et al.  showed that in eastern Sicily, two Tortonian sites are rotated about 30° CW, while nearby Eocene-Oligocene sites are rotated 100° CW, implying a major phase of about 70° of CW rotations between Oligocene and Tortonian (Figure 7a). In southern Apennines, Oligocene–middle Miocene sites are rotated more than 60° CCW [Gattacceca and Speranza, 2002], whereas upper Miocene strata show about 40° of CCW rotations, implying that about 20–30° of CCW rotations occurred in the internal units of the chain between early middle Miocene and late Miocene (Figure 7b). Both in southern Apennines or in Sicily, late Miocene sediments have been deposited in foredeep or piggyback basins, and they are indeed coeval to the main phases of thrust emplacement in the Apenninic-Maghrebian chain. This suggests that the observed Miocene paleomagnetic rotations initiated during the middle to late Miocene, when internal Apenninic and Maghrebian strata started to be incorporated in the orogenic wedge.
 The recent history of paleomagnetic rotations in Sicily and southern Apennines indicates that there was a progressive decrease in the magnitude of rotations from late Miocene to early Pleistocene. During this time interval, thrust sequences forming the southern Apennines and Maghrebian chains were progressively emplaced and translated toward the Apulia and the Hyblean foreland, respectively (enlargements in Figure 7). This suggests that vertical axis rotations systematically accompanied deformation and migration of upper plate rocks during the progressive bending of the Calabrian Arc. It is also worth noting that the significant magnitude of opposite sense rotations measured in lower Pleistocene sediments in Sicily and southern Apennines indicate that vertical axis rotations played a very important role during the recent history of the Calabrian Arc [e.g., Scheepers and Langereis, 1993; Scheepers et al., 1993; Speranza et al., 1999].
 In the CPD, the progression of paleomagnetic rotations with time shows some notable differences with respect to southern Apennines and Sicily thrust belts, which reflect the peculiar tectonic evolution of the CPD (Figure 8). In particular, in the CPD, measured paleomagnetic rotations are almost uniform along the entire region, and in particular there is no significant difference between paleomagnetic rotations measured in the Ionian fore-arc basins and in the Tyrrhenian postorogenic extensional basins. Also, there is no evidence of a decrease of paleomagnetic rotations with time, because an almost constant value of about 20° of CW rotations has been obtained from Serravallian to lower Pleistocene strata. These data suggest that during the early Pleistocene the CPD rotated as an almost homogeneous block, which underwent significant 15–20° CW rotations, and that no significant rotations occurred during late Miocene–Pliocene time. The history of paleomagnetic rotations may be considered the result of the different tectonic evolution of the CPD, which has been progressively drifted to the southeast as an almost rigid block during the Neogene and Quaternary, without evidence of internal shortening accommodated by thrusting and folding, as observed in Sicily and southern Apennines. Finally, a common feature of Calabria, Sicily and southern Apennines is the fact that in all these regions, uppermost lower Pleistocene and middle Pleistocene sediments have not been deformed, marking the end of the main rotational phases in the Calabrian Arc at about 1 Ma [Cifelli et al., 2004; Mattei et al., 2004b; Scheepers and Langereis, 1993].
6.2. A Paleotectonic Reconstruction of the Calabrian Arc
 In this section, we propose a schematic model of the tectonic evolution of the Calabrian Arc from early middle Miocene to Recent time (Figure 10). The change in orientation of the main orogenic fronts is derived from paleomagnetic data reported in Figures 6, 7, and 8. These data show that the foreland domains remained almost unrotated with respect to the Africa plate while large magnitude tectonic rotations occurred in the orogenic wedge, where systematic CCW and CW rotations have been measured in the southern Apennines, and in Calabria and Sicily, respectively. This opposite pattern of rotations corresponds with the different trend of structural domains, which range from a NW-SE orientation in southern Apennines to an E-W orientation in Sicily. We consider that vertical axis rotations accompanied the entire deformation history of the orogenic wedge and the propagation of the thrust sheets in southern Apennines and Sicily, whereas the CPD was considered as a rigid block and rotated mainly during Pleistocene. Similarly to other tectonic models [Argnani et al., 2006; Bonardi et al., 2001; Faccenna et al., 2004; Guegen et al., 1998; Gvirtzman and Nur, 1999; Lonergan and White, 1997; Malinverno and Ryan, 1986; Patacca and Scandone, 1989], we consider the presence of an oceanic (deep sea) Mesozoic seaway, intervening between the Adriatic and Africa continental margin, as the major geologic element in controlling Neogene to Recent tectonic evolution of the Calabrian Arc. The oceanic or continental character in the subducting lithosphere results in a different efficiency in the subduction processes, which in the central Mediterranean was responsible for the differential southeastward roll-back of the Ionian oceanic lithosphere with respect to the Apulia and African continental lithospheres, located to the northeast and southwest, respectively [Faccenna et al., 2004; Malinverno and Ryan, 1986; Patacca and Scandone, 1989]. Finally, the timing and geometry of back-arc extension is mainly derived from Oceanic Drilling Project results and magnetic anomaly analysis in the southern Tyrrhenian Sea and from structural and stratigraphic data in Calabria and southern Apennines [Ascione and Romano, 1999; Kastens et al., 1988; Mattei et al., 2002; Nicolosi et al., 2006, and references therein]. This paleotectonic reconstruction is consistent with the present-day structure of the Ionian subducting slab, which shows a well-defined narrow tongue, underneath the Calabrian Arc, interpreted as a subducted plate of oceanic lithosphere (Figure 1).
 The progressive curvature of the Calabrian Arc is described starting from Burdigalian (about 16 Ma), at the end of CCW rotation of the Corsica-Sardinia block [Speranza et al., 2002] (Figure 10a). In the southern Tyrrhenian Sea, extensional tectonics initiated along the eastern margin of Sardinia and western Calabria during Serravallian-Tortonian time (Figure 10b). Later on, extensional tectonics migrated toward the southeast in the southern Tyrrhenian Sea and was accompanied during early Pliocene and early Pleistocene by important episodes of back-arc extension, which also caused the Vavilov and Marsili oceanic seamounts formation (Figures 10c and 10d). In the southern Apennines, extension started during late Pliocene time along the western Tyrrhenian margin and progressively migrated toward the east to be active today along the Apenninic watershed [Valensise and Pantosti, 2001] (Figures 10d and 10e). The evolution of extensional tectonics in Calabria is more complicated, reflecting the peculiar tectonic history of the CPD. In fact, along the Tyrrhenian side of Calabria, late Pliocene to Pleistocene normal faults are superposed as an older, Serravallian-Tortonian extensional tectonic phase, which is related to the earliest episodes of back-arc formation in the southern Tyrrhenian Sea. During the incipient phases of southern Tyrrhenian extension, late Miocene normal faults, and associated extensional basins acted as the conjugate extensional margin of Sardinia. Later on, during the major spreading phases, the extensional basins passively drifted toward southeast, together with the whole CPD and assembled to the southern Apennines and Maghrebian chain. Extensional tectonics was active at that time, along an almost continuous belt, extending from central Apennines to CPD, which is the most seismically active region in southern Italy [Galli and Bosi, 2003; Monaco and Tortorici, 2000] (Figures 10e and 10f).
 The timing of the curvature of the Calabrian Arc may be assessed using the paleomagnetic presented in this paper. The shape of the arc at the end of Corsica-Sardinia CCW rotation is considered to have an almost N-S orientation, represented by the Apennines and Sicilian Maghrebian chain (Figure 10a). This trend has been obtained considering the large amount of opposite CW and CCW rotations, which occurred later in the internal units of Sicily and southern Apennines, respectively. Because of the lack of paleomagnetic data from Tunisia, we hypothesize an E-W orientation of the tectonic structures that form the Tell chain only on the base of the rectilinear shape of orogenic structures all along the northern African margin [Frizon de Lamotte et al., 2000]. No information is available on the hypothetical continuity of the Calabrian Arc between western Sicily and Tunisia during this time. Later on, the Apennine-Maghrebide orogenic wedge started to bend by means of opposite sense vertical axis rotations along the two arms of the arc, which involved the more internal units of the Apenninic and Sicilian chain. The beginning of such rotations is constrained by paleomagnetic data, which are middle Miocene in age (Figure 7), coeval with the early phases of extensional processes in the southern Tyrrhenian Basin [Mattei et al., 2002]. Conversely, the frontal part of the arc, represented by the CPD, migrated toward the SE as a consequence of the faster roll-back of the subducting Ionian lithosphere, without any significant rotation (Figure 10b). The drifting of the CPD was more active during Pliocene and early Pleistocene, as a consequence of back-arc extension, which caused the formation of new oceanic crust in the Vavilov and Marsili basins [Nicolosi et al., 2006]. At the same time along the edges of the arc, vertical axis rotations involved the more external units of the chain, causing a progressive tightening of the arc and increasing the obliquity between the orientation of the trench and the main shortening direction, which was responsible of the increasing of strike-slip component of motion along the two edges of the arc (Figures 10c and 10d). During the Pleistocene, vertical axis rotations affected Sicily and southern Apennines, accompanying the last compressional phases along the outer front of the chain. The CPD also underwent an almost homogeneous CW vertical axis rotation, reaching its present-day configuration (Figures 10d and 10e). Rotations ceased at about 1.0 Ma together with the ending of compressional tectonics in southern Apennines and Sicily (Figures 10e and 10f).
6.3. Calabrian Arc and the Models for Arc Formation
 One of the main results from the analysis of paleomagnetic data discussed in this study is that the tectonic evolution of the Calabrian Arc cannot be explained using classical concepts of primary, secondary or progressive arcs, which do not account for the complex rotational history of this area. The presence of opposite paleomagnetic rotations in Sicily and southern Apennines demonstrates that the tectonic evolution of the Calabrian Arc does not fit well with a primary arc model. In fact, the Calabrian Arc did not form with its present-day shape and the belt curvature increased during deformation, with large magnitude vertical axis rotations. The orocline model is not consistent with the tectonic evolution of the Calabrian Arc curvature as well, because large magnitude vertical axis rotations occurred contemporaneously with fold-thrust belt formation and not during a second phase, when the orogen were already developed. Consequently, as a first approximation the Calabrian Arc could be better described as a progressive arc [Weil and Sussman, 2004], which acquired its curvature during the development of the belt. However, the progressive arc model does not completely explain the evolution of the Calabrian Arc curvature. In fact, the present-day shape of the Calabrian Arc results from a complex tectonic and rotational history, juxtaposing a drifted crustal block (CPD) between two deforming orogenic belts (southern Apennines and Sicilian Maghrebian chain), which were originally located in different paleogeographic domains. The curvature of the arc increased progressively only along its edges, whereas the frontal part, represented by the CPD, did not undergo significant magnitude paleomagnetic rotations during late Miocene and Pliocene, but was mainly affected by SE drifting. Consequently, an abrupt change in the sense of rotations between southern Apennines (CCW) and CPD and Sicily (CW) is observed. This different sense of rotation defines a tectonic boundary located in the northern part of the Crati basin, where a complex pattern of block rotations has been observed, and where seismic and tomographic data locate the northeastward boundary of the subducting slab of the Ionian oceanic lithosphere (Figure 1).
6.4. Origin of the Calabrian Arc Curvature
 The abundant paleomagnetic data obtained in the Calabrian Arc region give a unique opportunity to investigate the mechanism responsible for the formation and evolution with time of arcuate orogens located on top of narrow subducting slabs. In general terms, the arcuate shape of arc-back arc systems reflects many factors, such as the presence of buoyant ridges and plateaus in the subducting plate, which will resist rollback [e.g., Nur and Benavraham, 1982; Schellart and Lister, 2004], a laterally heterogeneous lithosphere [e.g., Morra et al., 2006], lateral mantle flow which can influence the shape of the trench [Faccenna et al., 2005], and along strike transition from subduction of oceanic lithosphere to continental collision [Wallace et al., 2005].
 In the central Mediterranean, the main factor controlling the bending of the Calabrian Arc is the presence of a small oceanic lithosphere plate (the Ionian Sea), intervening between the Apulia and Africa continental margins [Malinverno and Ryan, 1986]. This configuration of the subducting plate caused a differential southeast rollback of the trench during the Neogene and Quaternary, and the progressive drifting of a continental region (CPD), now impinged within the core of the Calabrian Arc. Paleomagnetic data clearly show a different behavior between the CPD, located on the top of the Ionian slab, and the southern Apennines and Sicilian Maghrebides units, which represent the part of the Calabrian Arc where continental lithosphere was partially subducted. The boundary between the CPD, which underwent a general uniform CW rotation, and the southern Apennines that rotated CCW, is a very narrow deformation belt (northern Crati and Pollino), which corresponds to the northeast boundary of the Ionian actively subducting slab, which in turn corresponds to the boundary between the Ionian oceanic realm and the Apulian continental margin (Figure 1).
 In the Calabrian Arc, the tightening of the arc was also accompanied by the decreasing of the width of the active trench through time. In fact, structural and stratigraphic data show that the southern Apennine foredeep basins and the thrust activity progressively shifted toward the southeast, parallel to the front of the chain, during late Pliocene to Pleistocene, and are currently located offshore in the Ionian Sea [Tropeano et al., 2002] (Figures 10d–10f). Consequently, the outer front of the southern Apennines was progressively sealed, marking the cessation of active subduction, as testified by seismicity and seismic tomography (Figure 1). The progressive increase of areas affected by continental collision caused a lowering of the efficiency of the subduction process, which became inactive along almost the entire length of the arc, and is now confined to the frontal part of the Calabrian Arc, corresponding to the present-day Ionian oceanic domain (Figures 1 and 10). The progressive decrease in the width of the trench increased the role of lateral shear along the two edges of the arc and could be ultimately responsible of the present-day tight curvature of the Calabrian Arc.
 In conclusion, we speculate that slab retreat could be not sufficient alone to form tight arcs like those that characterize the Mediterranean region, and we suggest that a progressive decrease in the width of the active trench should also be considered as an active mechanism to explain the timing and distribution of crustal rotations in the Calabrian Arc and its huge curvature.
 The tectonic evolution of northern Calabria was elucidated by means of a detailed paleomagnetic investigation carried out on upper Miocene to middle Pleistocene deposits of the Crati extensional basin. Integration of this new data set with the available paleomagnetic data allow us to describe the Neogene tectonics of this part of the Mediterranean region as a consequence of the evolution of the Ionian subduction system. In particular, paleomagnetic data indicate that (1) rotations are confined to the orogenic wedge, and do not extend to the Apulia and Hyblean foreland, which are almost unrotated; (2) large magnitude and opposite sense paleomagnetic rotations were measured in the southern Apennine and Sicily orogenic wedge; (3) in Sicily, and probably in southern Apennines, the magnitude of paleomagnetic rotations decrease moving from the inner to the external chain sectors, suggesting a fundamental structural control of the rotational processes during the formation of the Apennine and Maghrebide orogenic wedge; (4) the Calabro-Peloritane Domain (CPD) rotated as an almost rigid block during its southeastward drifting, suggesting that the present shape of the Calabrian Arc is the result of a complex tectonic history juxtaposing a drifted crustal block (CPD) between two deforming orogenic belts (southern Apennines and Sicilian Maghrebian chain); (5) paleomagnetic rotations measured in upper Miocene to middle Pleistocene sedimentary deposits of the Crati basin (northern Calabria) indicate that at the boundary between Calabria and southern Apennines the deformation pattern is accommodated by different, small sized fault bounded blocks, characterized by complex vertical axis rotations; these data confirm the complex character of this area, located between two sectors, southern Apennine to the north and Calabria-Sicily to the south, respectively, characterized by distinct tectonic features and corresponding to the northeastern termination of the Ionian subducting slab; and (6) paleomagnetic rotations in the Calabrian Arc ceased at about 1.0 Ma, together with the ending of compressional tectonics in southern Apennines and Sicily, suggesting a decrease in the efficiency of the tectonic processes related to the long-lived subduction of the Ionian slab.
 The proposed paleogeographic reconstruction of the Calabrian Arc tectonic evolution shows that either the oroclinal bending model or progressive arc model cannot be simply applied to the Calabrian Arc formation and that the curvature of mountain belts in active subduction margins is mostly governed by the presence of lateral lithosphere heterogeneities in the subducting slab.
 The authors would like to thank Associated Editor and the two reviewers John Geismann and Fabio Speranza for the accurate and constructive reviews of the manuscript. A.M. Hirt and L Sagnotti are warmly thanked for hosting the Authors in ETH (Zurich) and INGV (Rome) paleomagnetic laboratories. We would also like to thank Giorgio Ranalli for reviewing the English style of the paper. Financial support for this work was provided by 01-LECEMA22F [WESTMED]; by the European Science Foundation under the EUROCORES Programme EUROMARGINS, through contract ERAS-CT-2003-980409 of the European Commission, DG Research, FP6; and by COFIN- 2004 MIUR research program.