Shoshonitic magmas emplaced along shear zones are prone to show crustal contamination to different extents since high-strain zones are suitable sites to dilute or even erase the register of least evolved members and highlight crustal assimilation. Also, in some cases, purely crustally-derived rocks that resemble some of the shoshonitic geochemical signatures, as the high Ba-Sr granitic rocks, are likely to be generated in high-strain zones. Such cases are well exemplified in southernmost Brazil by the Neoproterozoic Itapema granite (Martini et al., 2019; Bitencourt and Nardi, 2004; Rivera et al., 2004), Neoproterozoic Estaleiro granodiorite (Bitencourt and Nardi, 1993), Cruzeiro do Sul granodiorite and later dikes (Knijnik, 2018; Knijnik et al., 2012), orthogneisses from the Tinguí Unit (Martins et al., 2016), and in Uruguay (Lara et al., 2017).
The Itapema granite, a Neoproterozoic syntectonic intrusion whose emplacement was controlled by thrusting under high-grade metamorphism, comprises hornblende-biotite monzogranites to granodiorites that host abundant country-rock xenoliths and irregularly-distributed mafic aggregates. According to Bitencourt and Nardi (2004), it was produced by partial melting of amphibolitic orthogneisses under minimum temperatures of 850 ℃ and medium to high pressure. Its lower alkali contents compared to typical shoshonitic associations, the lack of basic or intermediate associated rocks, and intimate temporal association with high-grade gneisses and migmatites indicate that it was crustally derived (Martini et al., 2019).
The Estaleiro granitic complex (Bitencourt and Nardi, 1993), with 612 Ma (Peruchi et al., 2021), comprises a granodiorite intrusion with numerous veins of granitic composition, mafic microgranular enclaves and sheet-like bodies, syntectonic to a Neoproterozoic shear zone related to the post- collisional stage of Brasiliano/Pan-African orogenic cycle in South Brazil. The geochemical signature of these rocks, with high Sr and Ba contents, moderate HFSE and REE contents, whose patterns show no Eu-anomalies, is similar to that of shoshonitic series granitic rocks with increasing amounts of crustal contribution, mainly in the more evolved varieties.
Knijnik et al. (2012) and Knijnik (2018) identified diorites to granites and late granodiorite to monzogranite dikes, both syntectonic to post-collisional transcurrent shear zones from the eastern part of Dom Feliciano belt, southernmost Brazil, with ages of ca. 634 and 605 Ma, respectively. They were recognized as of shoshonitic affinity and interpreted as produced through AFC processes, with mixing of mantle and crustal components. Their slightly lower contents of Sr and alkalies relative to typical shoshonitic associations are ascribed to the assimilation of crustal melts.
The Tinguí Unit, situated in the central Ribeira belt (Martins et al., 2016), comprises orthogneisses and later dikes of shoshonitic affinity, with magmatic protoliths varying from quartz diorite to granodiorite crosscut by tonalitic to granodioritic dikes with crystallization ages of 551±3 Ma, for the orthogneiss, and of 495±4 to 488±4 Ma for the late dikes. Those authors interpreted the magmatism as pre- to post-collisional relative to the Buzios Orogeny (Schmitt et al., 2004) and formed by the interaction of mantle sources previously modified by the Buzios subduction and partial melting of lower crustal sequences.
Lara et al. (2017) described post-collisional high Ba-Sr granitic magmatism of shoshonitic affinity in the southern section of the Dom Feliciano belt and its cratonic foreland in Uruguay, with magmatic crystallization ages varying from 634±7 to 597±4 Ma, all of them intrusive in either high- or low-strain zones of Neoproterozoic shear zones. Hornblende- and titanite- bearing quartz-monzonites, granodiorites and monzogranites are the most abundant rock types. They share geochemical features typical of shoshonitic granitoids, as K2O/Na2O ca. 1, high Ba (> 700 ppm) and Sr (> 500 ppm) contents, moderate fractionation of REE, high LREE contents, negligible Eu anomalies, and moderate contents of HFSE (e.g., Nb, Zr, Hf, Ta, Y). Initial 87Sr/86Sri values of 0.707 7 to 0.709 0, εNd(t) values of -15.8 to -19.3 at 600 Ma, with high Nd TDM (2.2–2.8 Ga) are reported by Lara et al. (2017) and interpreted as indications of recycling of ancient Paleoproterozoic to Late Archaean sources. Younger granitic intrusions, as the Solis de Mataojo granitic complex (584±13 Ma), Las Flores (586±2.7 Ma), and the quartz-monzodiorite intrusion named Sobresaliente Pluton (585±2.5 Ma) described by Oyhantçabal et al. (2012), increase the time span of shoshonitic magmatism in Uruguay (ca. 634–584 Ma). Its stratigraphic correlation with the Lavras do Sul Shoshonitic Association in southernmost Brazil is also supported by their magmatic crystallization ages.
The Lavras do Sul Shoshonitic Association (LSSA) (Lima and Nardi, 1998; Nardi and Lima, 1985) is a plutono-volcanic association of minor potassic trachybasalts (absarokites), followed by a thick sequence of shoshonitic lavas and pyroclastic rocks, cut by porphyritic to heterogranular monzonitic rocks and later basic to acidic dikes. Plutonic rocks vary from diorite to biotite syenogranite and are intruded into the basal portion of the volcanic sequence and older granulitic rocks, as described by Barros and Nardi (1994). In the QAPF diagram, the major lithological groups of LSSA plot along the monzonitic trend, as defined by Lameyre and Bowden (1982), also invading the granodiorite field. The emplacement and evolution of LSSA granitic rocks were discussed by Nardi (1984) and Gastal et al. (2015). Hornblende-bearing lamprophyres, spessartites, are mostly associated with the late monzonitic dikes, where several mingling features are recorded (Lima and Nardi, 1998, 1991). The LSSA granodiorites contain mafic microgranular enclaves, which attest to the coexistence of granodioritic liquids of shoshonitic affinity with basic magmas related to the sodic alkaline, silica-saturated series (Nardi and Lima, 2000). Their origin is probably related to A-type granitic magmatism that commonly follows the shoshonitic one. Monzonitic and quartz monzonitic rocks included in the LSSA were described by Liz et al. (2009) and Gastal and Lafon (2006). They form two main intrusive bodies crosscut by several monzonitic dikes with coeval lamprophyres, mostly spessartites (Müller et al., 2012; Liz et al., 2004; Lima and Nardi, 1998). The Cu-Au-sulphide mineralizations were formed by high- and low-temperature hydrothermal events (e.g., Fontana et al., 2017; Mexias et al., 2005; Nardi and Lima, 1988). Geochronological studies (e.g., Gastal et al., 2015; Liz et al., 2009; Gastal and Lafon, 2006; Remus et al., 2000) indicate that the multi-intrusive LSSA system was formed at ca. 600 Ma. Other volcanic and plutonic rocks of shoshonitic affinity have been described in southernmost Brazil and correlated to the LSSA (Lopes et al., 2014; Sommer et al., 2006, 2005; Wildner et al., 1999; Barros and Nardi, 1994).
Most authors assume the origin of LSSA magmatism (Liz et al., 2009; Gastal and Lafon, 2006; Lima and Nardi, 1998) as involving partial melting of a mantle source modified by fluids produced during previous subduction related to the Brasiliano/ Pan-African Orogenic Cycle. The contribution of crustal materials is probably restricted to that resulting from subduction of oceanic crust or delamination. Fractional crystallization dominated by olivine-clinopyroxene-plagioclase+(Fe-Ti oxides+ apatite) has controlled the evolution from basic to intermediate liquids, whilst differentiation to acidic liquids involved fractionation of plagioclase-amphibole-biotite-Ti magnetite (Lima and Nardi, 1998). Syenitic or trachytic rocks have not been described in the LSSA.
The post-collisional Piquiri Syenite Massif (PSM, Rivera, 2019) in southernmost Brazil is composed of four magmatic pulses, resulting in (ⅰ) ultrapotassic syenites and lamprophyres (609 Ma), (ⅱ) shoshonitic syenites (603 Ma), (ⅲ) shoshonitic quartz syenites and granites (584 Ma). Stabel et al. (2001), Plá Cid et al. (2003) and Nardi et al. (2008), amongst other authors, have admitted that the magmatic source of lamprophyric and syenitic magmas from PSM is a phlogopite-bearing, peridotitic veined mantle previously metasomatised by fluids and melts produced during lithosphere subduction. The lamprophyric source was at great depths, corresponding to pressures over 5 GPa (Plá Cid et al., 2003) and mingling occurred with ultrapotassic syenitic melts, generating the first pulse still at mantle depths. The Nd and Sr isotope data indicate a more radiogenic source for the ultrapotassic pulse, whilst the subsequent pulses show similar values for εNd(t) and 87Sr/86Sri indicative of less enriched sources.
Monzonitic, syenitic, dioritic and granitic rocks of shoshonitic affinity, and ultrapotassic lamprophyres form the Arroio do Silva Pluton (ASP) (Padilha et al., 2019) with ca. 578 Ma age (LA-MC-ICP-MS, U-Pb zircon). The ASP shoshonitic rocks are compositionally similar to pulses (ⅱ) and (ⅲ) of the Piquiri Syenite Massif. The main differentiation process from dioritic to syenitic rocks was mixing and mingling of dioritic, syenitic and lamprophyric magmas, as proposed by Padilha et al. (2019). Magmatic sources are similar to those proposed for the Piquiri Syenite Massif by Plá Cid et al. (2003).
Based upon a selection of 115 representative samples from southern Brazil shoshonitic associations, including basic to acidic rocks and shoshonitic-like granitic rocks, some geochemical parameters were calculated (Table 1 and Fig. 2).
Geochemical parameters Syntectonic granitoids Southernmost Brazil (A) ASP dioritic- monzonitic rocks (B) Monzonitic rocks, LSSA (C) Qz syenites, MSP (D) LSSA late dikes (E) LSSA lamprophyres (F) SiO2 (wt.%) 66.76 55.64 58.33 67.44 64.62 56.06 K2O/Na2O 1.04 1.36 1.00 1.66 0.96 0.37 K2O+Na2O (wt.%) 6.8 7.74 7.76 10.11 8.56 7.06 K2O (wt.%) 3.45 4.41 3.83 6.31 4.19 1.89 Sr (ppm) 495 1 312 889 1 356 936 528 Ba (ppm) 932 2 700 1 508 2 516 1 555 1 209 Nb (ppm) 12.4 20 18 35 9.10 6.7 La/Nb 3 3.96 3.25 2.87 4.62 3.25 LaN/YbN 20 31 32 59 42 32 LaN 157 330 246 418 175 246 (A) Syntectonic Estaleiro Granodiorite (Bitencourt and Nardi, 1993) and Cruzeiro do Sul Granodiorite (Knijnik et al., 2012); (B) Arroio do Silva Pluton dioritic and monzonitic rocks (Padilha et al., 2019); (C) Lavras do Sul Shoshonitic Association (LSSA) monzonitic rocks (Liz et al., 2009); (D) Piquiri Syenite Massif syenites and quartz-syenites (Rivera, 2019); (E) LSSA late latitic dikes and (F) LSSA lamprophyres (Müller et al., 2012).
Table 1. Representative values of diagnostic geochemical parameters of shoshonitic rocks from southernmost Brazil
Figure 2. Illustration of some diagnostic geochemical parameters of shoshonitic rocks from southernmost Brazil.
It is of particular relevance the average values of K2O/Na2O ca. 1 for all samples, K2O from 1.89 wt.% in lamprophyres, 3.5 wt.% to 4.5 wt.% in granitic and monzonitic rocks, and 6.3 wt.% in syenitic ones, Sr > 500 ppm, Ba > 900 ppm, LaN > 150, Nb < 35 ppm. Samples from associations that are unrelated to shear zones are metaluminous and plot in the TAS field of silica-saturated alkaline series (Figs. 1 and 3). In the R1-R2 diagram (Fig. 4), they plot along the monzonitic or shoshonitic trend. Their REE patterns, in general, show LaN varying from 100 to 500, YbN from 2 to 10, and no Eu anomalies (Fig. 5). Trace element patterns are closer to those of OIB than to other types of sources, with enrichment of LILE and LREE and prominent negative Nb, Ti, and P anomalies. Spidergrams displayed in Fig. 6 show very coherent patterns for all samples, suggesting that they represent rocks formed through similar processes from sources of similar composition.
Figure 3. Distribution of shoshonitic rocks from southernmost Brazil in Shand's diagram (after Maniar and Piccoli, 1989).
Figure 4. Shoshonitic associations from southernmost Brazil plotted on R1-R2 diagram. 1. Monzo-gabbro; 2. monzonite; 3. monzo-diorite; 4. quartz monzonite.
Figure 5. REE normalised by chondritic values (Boynton 1984), for (a) Arroio do Silva Pluton and Piquiri Syenite Massif samples, (b) Lavras do Sul Shoshonitic Association rocks, and (c) Granitoids emplaced along active shear zones.
Figure 6. Spidergrams for trace elements normalised by OIB values (Sun and McDonough, 1989) for (a) Arroio do Silva Pluton and Piquiri Syenite Massif samples, (b) Lavras do Sul Shoshonitic Association rocks, and (c) granitoids emplaced along active shear zones.
Syntectonic shoshonitic associations within shear zones show lower contents of alkalies. They are subalkaline according to the TAS diagram (Fig. 1), metaluminous to peraluminous (Fig. 3), and most of them plot in the R1-R2 fields of granodioritic rocks (Fig. 4). In addition, they show lower LREE, MREE, and higher LILE contents, but the HREE keeps the same values as displayed by samples from associations unrelated to shear zones (Figs. 5 and 6). These features can reveal the participation of crustal melts and, when they are predominant, the connection with the shoshonitic parental magma can be lost; in such cases, the granitoids should be classified as high Ba-Sr rocks.
The good correlation shown by pairs of elements such as K, Nb, La, Zr in the entire set of samples suggests that all magmas emplaced either within or outside shear zones are derived from similar sources. On the other hand, crustal assimilation is more significant in syntectonic shoshonitic-like granites, as indicated by increased peraluminousity, decreased alkalies, HFSE, LREE, MREE, and some LILE, such as Sr and K.
The Sr-Nd isotopic signatures of the shoshonitic and shoshonitic-like rocks emplaced outside and within shear zones (Fig. 7) show variable 87Sr/86Sri values from 0.704 up to 0.719, with mean concentration of 0.705 to 0.710. The LSSA rocks have the lowest 87Sr/86Sri values around 0.704, distinct from all the other associations. The εNd(t) values are all negative, with mean concentrations around -8 to -12. Again, LSSA is an exception with the least negative εNd(t) values (-9 to 0). The Estaleiro granitic complex also shows distinct εNd(t) values (-3 to -5) compared to other shoshonitic-like syntectonic granites.
2.1. Magmatism Syntectonic to Collisional or to Post- Collisional Shear Zones
2.2. Magmatism Unrelated to Deformational Sites
2.3. Elemental and Sr-Nd Isotope Geochemistry
The Sr-Nd isotope data from 141 representative samples from shoshonitic, ultrapotassic, shoshonitic-like and high Ba-Sr rocks, including basic to acidic, of varied ages were used to discuss the isotopic signature and the origin of granitoids and syenitic rocks. Most 87Sr/86Sri values vary from 0.704 to 0.710, whilst εNd(t) values show a large range from 0 to -24. No correlation of 87Sr/86Sri and εNd(t) isotope values is observed for different post-collisional magmatic associations displayed in Table 2, as it would be expected if crustal assimilation were the only cause for the variation of both parameters. When crustal melting is recognized as significant in the generation of intermediate to acidic shoshonitic-like and/or high Ba-Sr rocks, frequently when they are syntectonic associations, there is a tendency to increase 87Sr/86Sri and decrease εNd(t).
Rock associations εNd(t) 87Sr/86Sri Age (Ma) 1 0.5– -6 0.705–0.712 567 2 -5– -7 0.708–0.709 67–80 3 0.5 0.703 1–0.704 7 1 857–1 770 4 -9– -14 0.719–0.735 18–25 5* -2.9– -6.5 0.707–0.712 ca. 320 6 -1.6– -6.8 0.708–0.719 ca. 330 7a 0– -5 0.704 7–0.706 4 Tertiary 7b 0– -9 0.705 9–0.711 5 Tertiary 8 -4– -13 0.706–0.714 0.5–23 9 -5 0.708 ca. 10 10 -1.5– -6.8 0.708–0.710 ca. 330 11* -10– -15 0.706–0.710 605–638 12* -15.8– -19.3 0.708–0.709 598–604 13 -0.2– -9 0.704–0.705 ca. 592 14 -10– -14 0.701–0.714 ca. 592 15 -2.4– -2.8 0.703–0.704 ca. 2 028 16 -8– -10 0.707–0.708 ca. 610 17* -4– -13 0.705–0.707 ca. 488–551 18* -14– -16 0.710–0.715 573 19 - 8– -12 0.709–0.711 609 20 -8– -12 0.706–0.708 589–603 21 -2– -3 0.707 ca. 611 22* -3– - 5 0.709–0.711 ca. 611 1. Tismana Pluton (Duchesne et al., 1998); 2. Baranadag (Ilbeyli et al., 2004); 3. Svecofennian post-collisional shoshonitic plutonism in the Fennoscandian Shield (Eklund et al., 1998); 4. potassic and ultrapotassic magmatism in Southwest Tibet (Miller et al., 1999); 5. Tormes Dome (López-Moro and López-Plaza, 2004); 6. Black Forest, Germany, post-collisional, Variscan potassic magmatism (Hegner et al., 1998); 7. Serbia, shoshonitic (7a) and ultrapotassic (7b) Tertiary rocks (Cvetković et al., 2004); 8. latitic rocks, Kunlun, Tibet (Williams et al., 2004); 9. northwestern Iran post-collisional minettes and spessartites; 10. post-collisional lamprophyres and rhyodacites, Germany (Hegner et al., 1998); 11. Cruzeiro do Sul granodiorite and granite dikes (Knijnik, 2018); 12. high Ba-Sr granodiorites (Lara et al., 2017); 13. Lavras do Sul shoshonitic association rocks of ca. 592 Ma (Gastal et al., 2005; Remus et al., 1997); 14. syenites and monzonites, NE Brazil (Guimarães and Silva Filho, 1998); 15. ultrapotassic syenites, Morro do Afonso, NE Brazil (Rios et al., 2007); 16. potassic and ultrapotassic syenites, São Paulo, South Brazil (Carvalho et al., 2014); 17. shoshonitic orthogneisses, Paraná, South Brazil (Martins et al., 2016); 18. crustally-derived, shoshonitic-like monzonitic-granitic rocks, Pernambuco, NE Brazil (Ferreira et al., 2015); 19. Piquiri Syenite Massif ultrapotassic syenites (Rivera, 2019); 20. Piquiri Syenite Massif shoshonitic quartz syenites (Rivera, 2019); 21. dioritic dykes, Estaleiro granitic complex, Santa Catarina, South Brazil (Chemale et al., 2012); 22. granodiorite and leucogranite dykes from Estaleiro granitic complex, Santa Catarina, South Brazil (Peruchi et al., 2021; Chemale et al., 2012); *. high Ba-Sr granitoids or shoshonitic rocks highly contaminated with crustal material.
Table 2. Sr-Nd isotopic parameters of collisional to post-collisional shoshonitic and ultrapotassic magmatism products and high-Ba, Sr granitoids of distinct ages (approximate values)
The origin of shoshonitic (or potassic) and ultrapotassic plutonic intermediate to acidic rocks related to the shoshonitic and ultrapotassic magmatic series is addressed by two main genetic models, both proposing that their primary magmas represent melting products of mantle regions metasomatised by fluids and melts derived from subducted lithosphere (e.g., Miller et al., 1999; Guimarães and Silva Filho, 1998; Ferreira and Sial, 1993; Peccerillo, 1992). Syenitic or monzonitic shoshonitic and ultrapotassic magmas can be generated by direct partial melting of mantle sources. Liquids of such compositions, including granitoids, could also be derived from lamprophyric or intermediate magmas through fractional crystallization, sometimes with concurrent assimilation of crustal materials. Fractional crystallization of lamprophyric magmas leading to the generation of syenitic rocks has been admitted by many authors (e.g., Rios et al., 2007; Melzer and Foley, 2000; Janasi et al., 1993; Leat et al., 1988), whilst others assume that syenitic and monzonitic magmas would be directly extracted from metasomatised mantle regions (e.g., Carvalho et al., 2014; Weaver et al., 2013; Plá Cid et al., 2012, 2003, 2000; Nardi et al., 2008, 2007; Conceição et al., 2000; Peccerillo, 1992; Venturelli et al., 1984). The association of both ultrapotassic and shoshonitic magmatism has been recognized by several authors (e.g., Prelević et al., 2012; Guo et al., 2006; Miller et al., 1999; Silva Filho et al., 1993; Rogers, 1992; Thompson and Fowler, 1986; Venturelli et al., 1984; Civetta et al., 1981), sometimes including lamproitic rocks (Conticelli et al., 2009) when the influence of subduction is less evident (Avanzinelli et al., 2009). Ultrapotassic, shoshonitic and lamproitic magmas, therefore, may represent melts derived from subduction- affected mantle regions with different contents of phlogopite and K-amphibole (e.g., Weaver et al., 2013; Prelević et al., 2012; Maria and Luhr, 2008; Nardi et al., 2008; Plá Cid et al., 2003; Conceição et al., 2000). The experimental support for such models comes from Conceição and Green (2000), Williams et al. (2004) and Conceição and Green (2004).
Liz et al. (2009) used a sliding normalization technique for comparing trace element signatures of post-collisional shoshonitic and ultrapotassic associations, which has led to the identification of two different trends: (ⅰ) Zr-Nb-Ti-Y-REE enrichment and (ⅱ) increasing of Rb-K2O with concomitant Nb decrease. The former was taken as indicative of an asthenospheric component, whilst the latter indicates crustal contribution through lithospheric subduction. A strong positive correlation of K and Nb indicates that the increase of both elements is related to the variation of melt fractions or with the magnitude of enrichment in the mantle source.
The high incidence of 87Sr/86Sri ratios at ca. 0.708, particularly when a mantle origin is assumed, suggests that shoshonitic magmatism in different regions and ages comes from the same type of source and does not involve significant crustal assimilation. Some authors (e.g., Rivera, 2019; Miller et al., 1999) observed that the ultrapotassic magmatism frequently shows higher 87Sr/86Sri ratios in relation to the associated shoshonitic and high-K subalkaline ones. The ratios of 87Sr/86Sri in ultrapotassic lamprophyres and syenitic rocks may reach values greater than 0.720, and this has been explained in some cases as due to the influence of a strongly radiogenic mantle source (ITEM, Italian Enriched Mantle) with values of 87Sr/86Sri ca. 0.724 and εNd(t) ca. -13 (Stoppa et al., 2014; Bell et al., 2013). The more common values of 87Sr/86Sri ratios, ca. 0.708 and negative εNd(t) frequently reported for shoshonitic magmatism, as well as trace element patterns with high LREE/HFSE and LILE/HFSE ratios probably reflect mantle sources compositionally close to EM1-type, as observed by several authors (e.g., Nardi et al., 2008; Sommer et al., 2005). The trace element ratios proposed by Weaver (1991) for characterising mantle sources, on the other hand, indicated that the sources of the south Brazilian associations—LSSA, MSP and ASP—were more LILE-, LREE- and Th-enriched than EM1 sources approaching crustal values. Nb/Th and Zr/Nb ratios of less differentiated rocks (SiO2 < 56 wt.%) indicated their derivation from hydrated mantle, a source type characteristic of arc-related settings (Condie, 2015). The ratios of Th/Yb, Nb/Yb and TiO2/Yb proposed by Pearce (2008) for identifying magma sources showed that the associations from southernmost Brazil are derived from OIB-like sources with a strong influence of subduction, which causes an increase of Th/Yb and TiO2/Yb ratios. Shoshonitic lamprophyres plot near the EMORB field. Intermediate and acidic rocks plot approximately in the same field (Figs. 8 and 9).
Figure 8. Trace element ratio diagram (TiO2/Y vs. Th/Yb) for determination of sources, according to Pearce (2008): less differentiated rocks from Arroio do Silva Pluton, Piquiri Syenite Massif, and Lavras do Sul Shoshonitic Association lamprophyres.
Figure 9. Trace element ratio diagram (Nb/Y vs. Th/Yb) for determining sources, according to Pearce (2008): less differentiated rocks from Arroio do Silva Pluton, Piquiri Syenite Massif, and LSSA lamprophyres.
According to data discussed in this review, the plutonic rocks, either syenitic, monzonitic or granitic related to the shoshonitic series, were produced dominantly by fractional crystallization of basic to intermediate, sometimes lamprophyric magmas, whose source is the mantle modified by previous lithosphere subduction, probably a veined peridotite with abundant pyroxene+phlogopite+K-richterite (e.g., Weaver et al., 2013; Conceição et al., 2004, 2000; Plá Cid et al., 2003; Peccerillo, 1992; Venturelli et al., 1984).