Abstract

A structural analysis of Cretaceous alkaline dykes swarm associated with the central segment of the Asunción Rift is reported here. Dykes are generally single near-vertical tabular bodies, less than 5 m wide, although multiple and composite intrusions also occur. Many of these small bodies have been emplaced into Paleozoic sedimentary rocks and exhibit a regional NW-SE orientation pattern. Petrographical and geochemical data allow recognition of two different lineages of potassic dykes: a silica-undersaturated suite ranging from basanite to phonolite (B-P) and a silica-saturated suite ranging from alkali basalt to trachyte (AB-T). The morphological features, the regional en-échelon distribution, and the NW-SE orientation pattern suggest that the dykes were injected along fractures and faults, under a transtensional tectonic regime with σ1 NW/horizontal, σ2/vertical, and σ3 NE/horizontal. Detailed analysis, combining dyke petrography, orientation pattern, and relative chronology reveals a rotation from WNW toward NNW during dyke emplacement. In terms of the paleostress field orientation, the evidence indicates that the dykes were diachronically formed under a similar stress condition. Finally, the pattern of orientation documented for the Cretaceous alkaline dykes of the Asunción Rift is consistent, temporally and spatially, with the phases of regional deformation that occurred during the process of the Atlantic Ocean opening.

1. Introduction

In many volcanic systems, dyke intrusions of variable size and composition are often considered the main channel feeders and represent one of the more important vertical transfer routes of mantle-derived molten material through the lithosphere to the upper crust. In general, dykes occur in diverse geological times and tectonic settings; however, the vast majority of the dyke swarms on continental areas are of the Proterozoic or Late Phanerozoic age [1]. Although their abundance in the crust is less expressive than continental flood basalts and large granitic masses, dykes are excellent tracers for many geological processes. Over the last two decades, numerous geochemical studies have been carried out on these small vertical tabular-like bodies with the purpose of better understanding the cause and nature of partial melting and magmatic differentiation processes. Consequently, many of the predictions made about the thermodynamic behavior and the conditions that govern the lower and upper mantle are based on the isotope and trace element signatures of dykes [24]. Dykes are largely used to examine the relationship involving elastic brittle-fluid-host rocks and magma transport. These studies are mainly targeted at determining the regional paleostress fields and the active mechanism [4, 5]. A theoretical and applied approach about the emplacement mechanism and propagation of individual dykes is well documented in the synoptic paper of Anderson [6].

Eastern Paraguay is an intracratonic region situated in the westernmost border of the Paraná basin. During Early to Late Cretaceous time, this region was affected by an important tectonomagmatic event, the Paraná-Etendeka large igneous province (cf. [7]) related to the opening of the South Atlantic Ocean, which caused a series of alkaline magmatic episodes [8, 9]. Although the distribution of these alkaline bodies is quite widespread around the Paraná basin (Figure 1), in the Eastern Paraguay region, they are confined to distinct well-exposed areas. In the northern (Rio Apa) and northeastern (Amambay) provinces, the alkaline rocks show similar ages of  Ma (cf. [10]), predating the tholeiitic lavas of the Serra Geral Formation (133–130 Ma, cf. [1115]). However, in the Central-Eastern (Central) and southern (Misiones) provinces, they are younger than the basaltic rocks,  Ma and  Ma, respectively (cf. [10]).

One of the most conspicuous occurrences of alkaline rocks in Eastern Paraguay is represented by the Central Province. Geological and geophysical data indicate that the Asunción Rift development, in Cretaceous time, was responsible for multiple diachronous events of potassic alkaline magmatism (Central Province, cf. [10, 16, 17]). The mode of occurrence of the numerous bodies is quite variable. Intrusive formations are mainly represented by stocks of various dimensions; on the other hand, the extrusive units comprise essentially lava flows, domes, and plugs. Nevertheless, the most significant magmatic event of the province is represented by the hypabyssal rocks, occurring largely as individual dykes, which exhibit a wide variation in composition, texture, and size.

The good exposure of the outcrops offers an excellent opportunity to examine in detail the contact relationships, the morphological feature of the walls surface, and the orientation pattern of the dykes. Based on these data, this paper discusses the mechanism of emplacement, as well as the regional distribution of the paleostress field at the time of intrusion.

2. Geological Setting

The central area of Eastern Paraguay is characterized by an important tectonomagmatic-sedimentary event of Early Cretaceous age [1820]. The major faulting zone is represented by the Asunción Rift [21], a tectonic feature of roughly 200 km length and 25–45 km wide, with NW-SE general orientation (Figure 2). According to Velázquez et al. [17] and Riccomini et al. [22], the rift consists of three different segments. The western segment, with a well-defined NW-SE-trend, extends more than 90 km between the localities of Benjamin Aceval and Paraguarí, and it comprises an extensive Cenozoic sedimentary deposit and several Tertiary intrusive bodies of sodic ultra-alkaline rocks (Asunción Province, cf. [17, 20, 23]). The central E-W-trending segment measures approximately 70 km, extends from the town of Paraguarí to the locality of Villarrica, and represents the region of major potassic alkaline magmatism. Largely, those rocks cut siliciclastic deposits of Silurian and Permian ages and locally the Early Cretaceous aeolian deposits of the Misiones Formation. Finally, the less defined NW-SE-trending eastern segment, with about 40 km of extension, is developed from the locality of Villarrica until the Ybytyruzú mountains. In the last area, the potassic alkaline rocks intrude both the Cretaceous aeolian deposits of the Misiones Formation and the overlying tholeiitic flows of the Paraná basin volcanism.

From a geodynamic viewpoint, at least two tectonic episodes of importance led to the current shape of the relief in the area. The first event, of Early Cretaceous age, which was induced by an NE-SW extensional tectonic regime, provoked major graben faulting and expressive potassic alkaline magmatism. The second one was also of extensional regime, with the major period of faulting taking place in the Paleocene. However, it continues until today causing small seismic movements of low amplitude in the region [24]. Possibly, a generalized lithospheric thinning and the emplacement of a hot mantle closer to the upper crust were responsible for the significant changes of the geothermic gradient occurred in this epoch. A detailed study of fission tracks in Silurian deposits and Cretaceous alkaline rocks (cf. [25]) indicates two different periods of thermal activity for the whole area, between 90–60 Ma and 60–10 Ma. These cooling ages are significantly younger than the tectonomagmatic event of the Central Province, but, in part, are consistent with the intrusion of the nephelinitic alkaline rocks of the Asunción Province, which mainly formed at  Ma (cf. [10]). According to Riccomini et al. [20], the last period of time corresponds to an important tectonic phase, with generation of deep faulting that served as conduit for the mantle material to migrate to the surface of the crust.

3. Morphology, Occurrence and Field Relations

The potassic alkaline dyke intrusions are well exposed throughout a large section of the central segment of the Asunción Rift. More than 200 bodies of mafic, intermediate, and felsic dykes are randomly distributed in that region. The largest continuous exposures of individual dykes are found in the Sapucai district [16, 26] (Figure 3). Other important outcrops are confined to near the villages of Potero Ybaté and Gral. Bernardino Caballero is not so far from the town of Sapucai. Some dykes were also recorded in the Ybytymí region.

Most of the dykes show a continuous tabular aspect or are composed of several separate segments, resulting in an apparent sinuous intrusion. In both cases, they occur essentially as single intrusions, but multiple and composite injections are also present. Here, the multiple term is used to characterize a repeated injection of dyke with the same or similar compositions (alkaline basalt, trachybasalt, and trachyandesite), and composite when the repeated injections are of different compositions (alkaline basalt, tephrite, and phonolite).

The width of the dykes can vary from 0.15 m to 10 m, but in most of the cases is between 0.30–3 m; in general, the length cannot be traced for more than a few kilometers. Regionally, the dykes cross-cut Paleozoic deposits. In the Sapucai and Potrero Ybaté regions, they are intruded into lavas and plugs of alkaline rocks. The dyke-wall rock contact is usually vertical to subvertical, with chilled margins well preserved, and with prevalence for normal dilatation. Oblique relative displacement between the walls, in a dextral or sinistral sense, was also observed.

In a regional distribution, the dykes show a parallel to subparallel orientation and, in some cases, an en-échelon arrangement (Figure 3). However, there are localities where the orientation is less regular, indicating a significant change of direction and trajectory. In this case, the intrusion relationship is complex. Dykes of distinct composition and generation display oblique and/or perpendicular intersections.

4. Nature and Composition

The main petrographical and geochemical features of the dykes have been extensively discussed by many authors [16, 18, 2629], and the relevant features are summarised here.

Petrographical parameters (mineralogical assemblage and texture) and field relations (relative chronology of events and spatial distribution) allow recognition of three compositionally types of dykes showing distinct relationship: (a) tephritic (basanites-tephrites-tephriphonolites), (b) phonolitic (phonolites-peralkaline phonolites), and (c) basaltic (alkaline basalts-trachybasalts-trachyandesites-trachytes/trachyphonolites). The first two types are geochemically grouped together forming the B-P (basanite-phonolite) suite, whereas the third one into the AB-T (alkali basalt-trachyte) suite. Mineralogical and petrographical data reported in Comin-Chiaramonti et al. [18, 26] and Cundari and Comin-Chiaramonti [30] show that these rocks are usually porphyritic in texture, the groundmass being fine-grained or even aphyric and variable from holocrystalline to hyaline. Phenocrysts (mega and micro) are mainly represented by olivine, clinopyroxene, feldspars, (plagioclase and alkali feldspar) and feldspathoids; occasionally amphibole, biotite, garnet, and sphene can also be found. Some other general mineralogical features are: clinopyroxene composition (diopside to ferrosalite) consistently yielding Al enrichment trends; Fo81–69 content of the olivine in tephrites and alkali basalts decreasing up to 65% in phonolites; zoned megacrysts of hastingsitic hornblende (core) to kaersutite (rim), associated with accessory groundmass pargasite in tephrites and phonotephrites; K-rich hastingsite and K-rich ferropargasite in phonolites; accessory groundmass mica (annite-phlogopite series), consistently yielding insufficient (Si+Al) to satisfy the T site position; phenocryst, that is, xenocrystal plagioclase An70–20 and An74–42 in tephrites and phonolites, respectively; coexisting plagioclase (An14–22) microlites associated with soda-sanidine and sanidine; feldspathoids include analcimized leucite and nepheline; Ti-magnetite, rarely ilmenite or hematite, as Fe-Ti oxides; and Ti-andradite and sphene as main accessory minerals.

The mineralogical assemblage of the tephritic-type includes phenocrysts of clinopyroxene (Wo40–50Fs10–19), olivine (Fo60–85) and leucite pseudomorphs (sanidine + nepheline) all set in a glassy groundmass consisting of microlites of clinopyroxene ± olivine, Ti-magnetite ± ilmenite, Ti-phlogopite-biotite, alkali feldspar (Or15–88), and nepheline-analcime (Ne44–59Ks17–26). Plagioclase (up to An74) is also present as phenocrysts in some samples. Accessory phases are amphibole (pargasite-kaersutite), apatite, and zircon. The phonolitic-type is characterized by phenocrysts of alkali feldspar (Or47–75), leucite pseudomorphs, clinopyroxene (Wo48–50), ferropargasite, nepheline ± biotite ± sphene ± melanite (Ti-andradite up to 68%) ± magnetite, or hematite. Glassy groundmass has alkali feldspar, nepheline, and clinopyroxene ± melanite ± opaques microlites. The composition of the basaltic-type for the alkali basalts, trachybasalts, and trachyandesites shows pheno and/or microphenocrysts of clinopyroxene (Wo44–49Fs7–15), olivine (Fo65–83), plagioclase (An28–76), magnetite, and biotite. The glassy groundmass contains microlites of clinopyroxene (Wo46–49Fs13–18), magnetite, ilmenite, biotite, plagioclase (An20–45), alkali feldspar (Or52–65), nepheline-analcime (Ne37–73Ks22–38), amphibole, and apatite ± sphene ± zircon. On the other hand, the mineralogy for the trachyphonolites and trachytes includes phenocrysts of alkali feldspar (Or60–65), clinopyroxene (Wo46–49Fs14–20), plagioclase (An14–16), leucite pseudomorphs, amphibole, and biotite in hypocrystalline to glassy groundmass having microlites of alkali feldspar, biotite, and clinopyroxene ± biotite ± amphibole ± magnetite ± Ti-andradite ± hematite.

Geochemical analyses and Sr-Nd isotopic signatures indicate that the dykes are quite variable in composition. Gomes et al. [31] and Comin-Chiaramonti et al. [26, 28], using chemical data of minerals and bulk rocks and petrochemical classification, grouped all the potassic dykes into two main lines of evolution: (a) a silica-undersaturated lineage ranging from basanite to phonolite (B-P) and (b) a silica-saturated lineage ranging from alkaline basalt to trachyte (AB-T) (Figure 4), probably related to distinct parental magmas. In general, both suites show similar elemental enrichment patterns for large ion lithophile elements (LILE), that is, Rb and Ba, as well as for other incompatible elements, such as La, Ce, Sm, and Tb. Nevertheless, the major differences between the two groups are the higher concentrations of K2O, TiO2, Zr, Nb, Y, and REE of the B-P rocks in comparison to those of the AB-T suite. Mg-values (Mg*) and Ni and Cr contents indicate that all the magma-types representative of the less evolved dykes in the investigated area can be considered to some extent to be derivatives [26].

Subsequent researchers have been seeking to establish the mechanism of generation and evolution of the alkaline magmatism in Central-Eastern Paraguay. Based on textural, mineralogical and petrochemical evidence and utilizing quantitative mass balance calculations on major oxides, the latter authors [8, 18] concluded that fractional crystallization was a potentially important process in the formation of these rocks. Furthermore, Sr and Nd isotope data suggested that the parental magmas responsible for the alkaline potassic series (B-P and AB-T) and sodic suites occurrences, found in some areas of Eastern Paraguay (northern, central, and southern: Alto Paraguay, Asunción, and Misiones provinces, respectively (cf. [10]), would have derived from a heterogeneous subcontinental mantle lithospheric source, submitted to different melting degrees and variously enriched in incompatible elements during Proterozoic times. Significant H2O, CO2, and F are expected in the mantle source (s) considering the occurrence of coeval carbonatites [7, 18, 32]. Comin-Chiaramonti et al. [26] also suggested that the complex compositional variation registered in the dykes of the Central Province would have resulted from multiple re-equilibrium of the crystallized phases at shallow levels, in a volcanic pressure regime.

5. Emplacement Ages

The dykes occur associated spatially and temporally with the alkaline lavas, stocks, and plugs. In order to better separate these events, a critical review of the geochronological and paleomagnetic data available in the literature for the dykes will be done following.

The K/Ar whole rock and mineral radiometric ages on a few dykes from the Sapucai area are reported in Velázquez et al. [33] and Gomes et al. [16]. Similar to the data listed in Palmieri [34] and Bitschene [23], these ages display a peak interval between 130–125 Ma. On the other hand, Rb/Sr internal isochron for some intrusive bodies reveal ages of  Ma and  Ma [23, 33]. 40Ar/39Ar ages determined on phlogopite separates from Sapucai-Villarrica lamprophyre dykes indicate an average value of 127.5 Ma [7].

More recently, a detailed geochronological work using the 40Ar/39Ar method was performed by Gomes et al. [35] and Comin-Chiaramonti et al. [10]. The complete set of radiometric data obtained on biotite and plagioclase separates, as well as whole rocks samples, indicate that the dykes and some early alkaline intrusions (stocks and plugs) occur within a time interval of 128–126 Ma, with the latter authors proposing an average value of  Ma for the alkaline potassic magmatism of the Central Province.

An extensive paleomagnetic study on the alkaline rocks and associated dykes of the Sapucai region was driven by Ernesto et al. [36]. According to those authors, the dykes acquired two opposite polarities of primary magnetization, normal and reversed, during the emplacement. The cross-cutting intrusion relationships between the dykes indicate that the normal primary magnetizations are younger than reversed.

Although the geochronological data available in the literature for the alkaline potassic magmatism of the Central Province point to wide interval (130–125 Ma), high-precision 40Ar/39Ar ages indicate a short period of time (126-127 Ma, cf. [7, 10]). The field evidence suggests that the event would have occurred in a successive way, through a series of magmatic pulses. The lavas and other major intrusions (ring complexes, plugs, and stocks) predate the dyke intrusions. This fact increases the possibility that all the dyke swarm was formed in a short period of time (less than 1 Ma) and, mainly, during two opposite polarities of primary magnetization, as suggested by the available paleomagnetic data (Figure 5).

6. Orientation of the Dykes and Their Relation with the Regional Structures

Figure 6 shows the orientation of all the dykes measured in the central segment of the Asunción Rift. The measurements refer to dykes with intrusive contacts, showing clear chilled margins in 90% of cases. Two preferential orientations are recognized, NW-SE and NE-SW, the first being more frequent. Interposed between the two main orientations, there are two other less common directions, N-S and E-W. On a regional scale, joint and fault systems of similar orientation also occur. The relative chronology between faults and dykes is not always of easy distinction. In some cases, because the interval between the events was relatively short and, in others, because there was an overlap of events, obliterating the former structure. However, in more favorable outcrops, the structural relationships indicate that the dyke intrusion was preceded by synthetic normal and some strike-slip faulting. Similarly, the determination of the temporal relation of dyke intrusions was based mainly on their cross-cutting relationships. This correlation indicates that basaltic-type predate the tephritic-type and both are cut by phonolitic-type.

Detailed analysis, combining dyke petrography, orientation pattern, and relative chronology, reveals that the basaltic-type is orientated preferentially N50-70W, the tephritic-type N30-60W, and the phonolitic-type N20-45W (Figure 7). Such disposition shows clearly a rotation from WNW toward NNW during the emplacement of the dykes, accompanied by a progressive compositional change of the magma. Field observations are consistent with continuous extensional deformation and right-lateral rotation, which led to a dynamic interrelation between faulting and dyke injection. The absence of internal solid-state deformation in the dykes and the nature of the kinematic indicators present in the walls suggest that the fault generation and the dyke injection would have occurred simultaneously, under an extensional tectonic regime, followed by a short period of progressive shearing, affecting the walls of the partially solidified dykes (Figure 8).

7. Discussion and Conclusion

The evolution of the regional stress field is frequently determined by structural analyses of faults, fractures, cleavages, and other microtectonic evidence. Unfortunately, these methods do not permit the precise dating of the specific period in which the stress field has been acting. Igneous rocks, however, can be isotopically dated and frequently display distribution and internal structural patterns related to the crustal stress field at the time of magmatism. Therefore, it is possible to establish an absolute chronology of the variation in the trajectory of the regional paleostress-fields [25]. This manner, volcanic features are frequently utilized as paleostress indicators, including alignments of volcanoes or ash cones and axes of elongation of volcanic edifices [37, 38].

Dykes are also very important kinematic indicators, considering that magma may invade coeval fractures under the action of a regional paleostress-field [39, 40]. Because magma-driven fracturing is preferentially normal to the principal compressive stress σ1 (Anderson’s prediction), vertical dykes can only be generated when σ3 is horizontal, that is, when the regional paleostress of the crust is dominantly extensional [6, 4145].

The Cretaceous alkaline dyke swarms of Central-Eastern Paraguay show a large compositional variation, petrographic facies, and texture, offering important constraints on magma generation and dyke intrusions. Field observations indicate that the dykes represent majorly single vertical intrusions, but composite and multiple intrusions are also recognized.

The wide mineralogical variation and heterogeneous geochemical composition of the dykes require a dynamic generation mechanism during their emplacement, with constant and progressive crystal-liquid fractionation from the magmatic chamber. Such rapid rates of magma generation, in short periods of time, argue that partial melting took place during an extensional tectonic regime and by adiabatic decompression. Many of the dykes contain a significant proportion of primary hydrous minerals (biotite and/or amphibole) suggesting that hydrodynamic force (water-vapor pressure) played an important role in the magma generation. As a result of a constant modification of the thermodynamic conditions, the absolute majority of the dykes presents a well-marked change in texture, ranging from holocrystalline to hypocrystalline. In contrast to Komar’s prediction [46], no gradual concentration in phenocrysts from the walls to the center or internal structural zoning was observed in those dykes. This fact seems to confirm that flow-differentiation processes were not an important mechanism responsible for their generation. Thus, it is reasonable to admit that the dykes were emplaced in liquid state and consolidated under very fast cooling rates. In this case, an unstable condition of the flow regime due to variables such as the velocity gradient across the fractures, width of the fractures, magma viscosity, cooling rates, and crystal supply caused different effects on the crystals growth during emplacement and broadly influenced textural variations.

Because thin dyke channels may only last for hours to days (cf. [47]), liquids generated in the mantle and subject to continuous removal toward the surface would be under the constant influence of the regional stress fields [4852]. As previously stated, vertical dykes occur as a product of magma-filled fractures when the minimum compressive stress, σ3, is perpendicular to the direction of the propagation plane. Such conditions seem to be shown by the field relations of the studied dykes. This assumption is based on the careful examination of the chilled margins of the dykes and their geochemical data, which indicate no obvious significant interaction with the country rocks during magma injection. The walls of the dykes usually display normal dilatation and planar morphology. Only a few bodies show oblique opening, in either a dextral or sinistral sense, with overall extension preferentially orthogonal to the main propagation trajectory of the dykes. Additional evidence, such as the limited length (less than 3 km) and narrow width (average of 1–3 m) of the sheet-like bodies, the lack of a flux orientation of phenocrysts aligned parallel at the dyke margins, and the en-échelon regional distribution, indicates that vertical injection was more common than lateral magma injection.

The regional N20-70W-trending of the Cretaceous alkaline dykes suggests that the opening fractures and the magma injection were not perturbed significantly by pre-existing lines of weakness in the country rocks. Therefore, the well-expressed structural fingerprints of the dykes indicate an overall remote stress regime characterized by a NW-trending maximum horizontal shortening axis, an NE-oriented maximum horizontal extension axis, and vertical intermediate stress. Slickenside lineation analysis on NNW-sinestral and WNW-dextral conjugate strike-slip faults plane that outline the central segment borders of Asunción Rift (Figure 2) is also compatible with this paleostress orientation [19, 20, 22]. Basic criteria, such as the geometry, the orientation, and distribution pattern and the relative chronology of the dykes and regional faults, are widely consistent with an E-W dextral transtensional tectonic regime (Figure 9). In the context of a simple Riedel shear model [53], the basaltic-type dykes were preferentially injected along R fractures during the initial stage of the deformation. Some tephitic-type dykes also filled in part these fractures. With strain increase, R′ followed by P fractures were generated, which served as conduits for the tephitic-type dykes. Isolated occurrences of phonolitic-type dykes display consistent orientation with this phase of deformation. At an advanced stage of strain, finally, T fractures, perpendicular to the maximum extension axis, and Y fractures, parallel to the principal displacement zone direction, controlled the intrusion of the phonolitic-type dykes. Pattern orientation corresponding to X fractures is relatively uncommon in the study area. Assuming that the dykes were injected during an event of progressive deformation, the regional distribution in right- and left-stepping en-échelon corresponds to shear fractures (R, R′ and P) and tensional fractures (T). This framework structural has significant implications with regard to the right-lateral global rotation during dyke intrusions. The local parallel orientation of composite and multiple dykes confirm that the trajectory of the regional paleostress fields was relatively constant at the time of the dyke intrusion.

The relationship between alkaline magmatism associated with the Asunción Rift and basaltic melt generation in the Paraná-Etendeka large igneous province is well recognized [79]. Several geological evidence suggest that this episode of instability was occasioned by the lithospheric extension, continental breakup, and mantle plume, provoking a wide range of structural discontinuities, the installation of new sedimentary basins, and the intrusion of numerous alkaline bodies [8, 5456]. High-precision 40Ar/39Ar age determinations indicate that the alkaline magmatism in Eastern Paraguay can be grouped in two main events. The early stage (~145 and 138.9 Ma, cf. [7, 10]), located in the north region of the country, was before the first eruption of the flood basalts at 133 Ma (cf. [14, 15]) and a late phase (~127 and 126.4 Ma), concentrated in the Asunción Rift, which coincides with the youngest flood basalts that occur near the Brazilian and Uruguayan shoreline [7, 10]. In terms of petrogenetic evolution, the Paraná-Etendeka large igneous province and associated alkaline intrusive rocks are commonly interpreted as derived from a similar mantle source region by upwelling Tristan plume [5759]. Gibson et al. [7] postulated that the late event (~127 Ma) of Paraguayan alkaline magmatism would be associated with the final phase of the opening of the South Atlantic Ocean. According to Riccomini et al. [60] and Comin-Chiaramonti et al. [10], the regional spatial distribution of the Cretaceous and Tertiary alkaline magmatism along the coast lines of the South Atlantic continental margins and West Africa reveal that major dextral and sinestral shear zones exerted significant control on the alkaline bodies alignment. This evidence suggests an analogous intraplate deformation in both South American and African plates, as previously proposed by Vink [61] and Fairhead [39]. In this context, a lateral perturbation of the regional stress intensity induced by the differential motion of the plates at the time of continental breakup would favor the propagation of brittle structures toward the interior of the continents (Figure 10). Finally, the presence of a dextral strike-slip motion along the South American second-order plate boundary, proposed by Unternehr et al. [62] and N. Eyles and C. H. Eyles [63], the south-eastwards migration of Paraná-Etendeka large igneous province and associated alkaline rocks, suggested by Turner et al. [14] and Gibson et al. [7], and the Cabo Frio magmatic lineament [60] are strongly indicative of a global motion towards WNW of the South American plate, consistent with the pattern of orientation documented for the Cretaceous alkaline dykes of the Asunción Rift.

Acknowledgments

The authors thank Fapesp for financial support (Proc.: 97/01210-4 and Proc. 07/57461-9). They also thank the two anonymous reviewers for their constructive and perceptive commentaries of the manuscript.