Abstract Some 75 years after the visionary work of Wegener and du Toit, Neoproterozoic to Mesozoic geological correlations between South America and Africa are re-examined in the light of plate tectonics and modern geological investigation (structural and metamorphic studies, stratigraphic logging, geochemistry, geochronology and palaeomagnetism). The book presents both reviews and new research relating to the shared Gondwana origins of countries facing each other across the South Atlantic Ocean, especially Brazil, Argentina, Cameroon, Nigeria, Angola, Namibia and South Africa. This is the first comprehensive treatment to be readily available in book form. It covers the common elements of cratonic areas pre-dating Gondwana, and how they came together in late Precambrian and Cambrian times with the formation of the Pan-African/Brasiliano orogenic belts (Dom Feliciano, Brasília, Ribeira, Damara, Gariep, Kaoko, etc.). The subsequent shared Palaeozoic and Mesozoic sedimentary record (Karoo system) prior to Gondwana break-up is also reviewed.

Abstract The first to recognize the complementary shapes of Africa and South America and to suggest that these continents were once joined together was Dutch scientist Ortelius in 1596. He was followed in 1620 by Elizabethan philosopher Sir Francis Bacon, who asserted that the similarity of their shapes could not be accidental. Nearly 200 years later, German naturalist von Humboldt described how the two continents may have fitted together, and in 1860 French geographer Antonio Snyder produced the first map that showed South America and Africa in close contact (e.g., Blankett 1965 ). By 1915 the German meteorologist Alfred Wegener had amassed enough data to publish a comprehensive scientific argument for the past conjunction of these two continents on the basis of similarities in the Palaeozoic–Mesozoic geology on each side of the South Atlantic, and then boldly proposed that ‘horizontal displacements of the continents’ (Horizontal verschiebungen der Kontinente) caused their subsequent separation ( Wegener 1915 ). Wegener's original hypothesis of ‘continental displacement’ ( Krause & Thiede 2005 ) was severely criticized, especially by geophysicists ( Oreskes 1999 ). Nevertheless the concept was successfully transformed into the continental drift hypothesis through the support of, amongst others, two prominent geologists working

Abstract We present two alternative sets of global palaeogeographical reconstructions for the time interval 615–530 Ma using competing high and low-latitude palaeomagnetic data subsets for Laurentia in conjunction with geological data. Both models demonstrate a genetic relationship between the collisional events associated with the assembly of Gondwana and the extensional events related to the opening of the Tornquist Sea, the eastern Iapetus Ocean (600–550 Ma), and the western Iapetus Ocean (after 550 Ma), forming a three-arm rift between Laurentia, Baltica, and Gondwana. The extensional events are probably plume-related, which is indicated in the reconstructions by voluminous mafic magmatism along the margins of palaeo-continents. The low-latitude model requires a single plume event, whereas the high-latitude model needs at least three discrete plumes. Coeval collisions of large continental masses during the assembly of Gondwana, as well as slab pull from subduction zones associated with those collisions, could have caused upper plate extension resulting in the rifted arm that developed into the eastern Iapetus Ocean and Tornquist Sea but retarded development of the western Iapetus Ocean. As a result, the eastern Iapetus Ocean and the Tornquist Sea opened before the western Iapetus Ocean.

Abstract Deposition of Palaeo–Mesoproterozoic sedimentary rocks on the São Francisco–Congo craton started during Statherian taphrogenesis (1.8–1.75 Ga), as verified by ages of c . 1.7 Ga determined for volcanic rocks of the lower part of the Espinhaço Supergroup in the states of Minas Gerais and Bahia (Brazil). These basins contain volcanic rocks and conglomerates alternating with sandstones, argillites and dolomites, deposited in continental, transitional and marine environments. The rocks in the westernmost sector of the Congo Craton (Central Africa) compose the Chela Group, comprising sandstones, argillites and dolomites. In the easternmost region of the Congo Craton the Kibaran, Akanyaru, Kagera and Muva supergroups occur: the first three in the Kibaran Belt and the last in the Irumide Belt and on the Bangweulu Block. They consist predominantly of pelites and schists, sandstones and, in lesser proportion, conglomerates, deposited in shallow marine, fluvial and lacustrine environments. Their sedimentation ages are constrained through ages on felsic tuff layers as follows: Chela Group 1790±17 Ma, Kagera Supergroup 1780 ± 9 Ma, and Muva Supergroup 1879±13 Ma. These data show that broadly coeval and sedimentologically similar epi-continental sedimentary basins occurred on the São Francisco and Congo cratons, suggesting the possible existence of a long-lived wide epi-continental sea covering large areas of these cratons during Statherian times.

Abstract The Borborema and Benin–Nigeria provinces of NE Brazil and NW Africa, respectively, are key areas in the amalgamation of West Gondwana by continental collision during the Brasiliano/Pan-African orogenies. Both are underlain by complex basement: Nigeria has c. 3.05 Ga Archaean crust but no known Palaeoproterozoic rocks >2.0 Ga; in NE Brazil, 2.6–3.5 Ga Archaean rocks form small cores within Palaeoproterozoic gneiss terrains affected by plutonism at c. 2.17 Ga. Both regions exhibit Late Palaeoproterozoic ( c. 1.8 Ga) rift-related magmatism and metasedimentary sequences overlying the basement. The Seridó Group of NE Brazil (<0.65 Ga) is similar to the Igarra Sequence in SW Nigeria. The Ceará Group, which may date back to c. 0.85 Ga, is a passive margin deposit on crust thinned during initiation of an oceanic domain. In both provinces, basement and sedimentary cover were involved in tangential tectonics that resulted in crust-thickening by nappe-stacking associated with closure of this ocean. Frontal collision between c. 0.66 and 0.60 Ga later evolved to an oblique collision, generating north–south continental strike-slip shear zones at c. 0.59 Ga. In NE Brazil, the main Pan-African suture is probably buried beneath the Parnaíba Basin. The Transbrasiliano Lineament, interpreted as the prolongation of the Kandi–4°50 Lineament in Hoggar, may represent a cryptic suture.

Abstract The Congo (CC) and the São Francisco (SFC) cratons were joined at about 2.05 Ga; northern parts of Palaeoproterozoic basement subsequently underwent extension at about 1 Ga, forming intracratonic basins. Neoproterozoic metasedimentary rocks in these basins yield detrital zircons as young as 630 Ma. The Brasiliano and Pan-African ( c. 620–580 Ma) assembly of West Gondwana extensively altered this system. The Sergipano domain occurs north of the SFC, and the comparable Yaoundé domain occurs north of the CC. Crust north of the Sergipano domain comprises the Pernambuco–Alagoas (PEAL) domain. The NE–SW-striking Tcholliré–Banyo fault in Cameroon may extend southwestwards between the PEAL and Sergipano domains, defining northern limits of abundant SFC/CC basement. The Adamawa–Yadé domain in Africa does not appear to extend into Brazil. The Transverse domain of Brazil is a collage of Palaeoproterozoic crustal blocks, the 1.0 Ga Cariris Velhos orogen (CVO), late Neoproterozoic basins, and Brasiliano granites. The CVO extends ENE for more than 700 km in Brazil, but eastern continuation into Africa has not been identified. North of the Transverse domain contiguous c. 2.15 Ga gneisses comprise basement of Rio Grande do Norte and Ceará domains, which continue eastwards into western Nigeria and western Sahara.

Abstract Geological and geochronological data for the northwestern part of the Brasiliano Borborema Province are described and compared with their counterparts in the Pan-African Dahomey (Pharusian) belt that flanks the southeastern margin of the West African Craton, where outcrops are sufficiently continuous to discern the nature of the collision during West Gondwana assembly. In the Médio Coreaú domain, NW Borborema Province, U–Pb and Sm–Nd data have revealed unusual basement rocks representing 2.35–2.30 Ga juvenile crust, along with large tracts of 2.15–2.10 Ga juvenile gneisses in the Ceará Central domain. These basement blocks were affected by two pulses of intracratonic extension at 1785 and 775 Ma. Prior to West Gondwana collision, a continental arc (the Santa Quitéria batholith) developed between 665 Ma and 620 Ma. The presence of this arc strengthens the hypothesis that convergence between the Borborema Province and the São Luis craton involved closure of an oceanic basin. New geochronological data are presented showing that Palaeoproterozoic orthogneisses (U–Pb upper intercept 2288±2 Ma) were affected by a major late Neoproterozoic event (554±4 Ma U–Pb lower intercept, 558±3 Ma Sm–Nd whole-rock and mineral isochron). Exhumation and cooling of granulite rocks between 568 and 550 Ma in the Médio Coreaú domain and between c . 587 and 576 in the West African Dahomey Belt indicate that the final tectonic phase was not simultaneous along this front of the orogen.

Abstract Structural, geochronological, geochemical and mineralization patterns in the Nigeria–Borborema province of western Africa and NE Brazil reflect a complex Proterozoic evolution culminating in the Neoproterozoic Pan-African/Brasiliano orogenesis ( c . 600 Ma). Reworking of the Archaean–early Proterozoic crust produced heterogeneous deformation exemplified by prevalent shears, migmatization, granitization and intrusion of large volumes of granitoids typical of a Himalayan-type thickened crust resulting from continent–continent collision. Dominant north–south to east–west structures, with prominent penetrative fabric and mylonitised wrench faults, refolded, transpressed, or even obliterated older structural trends, which are preserved in nappes of the central Sahara region (NW Africa to Nigeria) and in NE Brazil. Anatexis and recrystallisation were coeval with emplacement of Pan-African granitoids throughout this mobile belt. Bulk chemical modification, especially affecting magmatophile elements and REE patterns, attest to chemical exchange between Archaean basement and Pan-African/Brasiliano rocks. Older crust is present in both regions, including early (3.6–3.5 Ga), mid (3.1 Ga) and late (2.7–2.5 Ga) Archaean, as well as large areas of Palaeoproterozoic rocks reworked by the c . 600 Ma tectono-thermal events. The extent and interpretation of Eburnian/Transamazonian (2.1–2.0 Ga) events have not yet been resolved due to inadequate structural and isotopic data. Litho-structural control of Au, Sn, Nb and Ta mineralization relates to main or late-stage Pan-African deformation.

Abstract The São Luís Craton and the Palaeoproterozoic basement rocks of the Neoproterozoic Gurupi Belt in northern Brazil are part of an orogen having an early accretionary phase at 2240–2150 Ma and a late collisional phase at 2080±20 Ma. Geological, geochronological and isotopic evidence, along with palaeogeographic reconstructions, strongly suggest that these Brazilian terrains were contiguous with the West African Craton in Palaeoproterozoic times, and that this landmass apparently survived subsequent continental break-up until its incorporation in Rodinia. The Gurupi Belt is an orogen developed in the southern margin of the West African–São Luís Craton at c . 750–550 Ma, after the break up of Rodinia. Factors such as present-day and possible past geographical positions, the timing of a few well-characterized events, the structural polarity and internal structure of the belt, in addition to other indirect evidence, all favour correlation between the Gurupi Belt and other Brasiliano/Pan-African belts, especially the Médio Coreaú domain of the Borborema Province and the Trans-Saharan Belt of Africa, despite the lack of proven physical links between them. These Neoproterozoic belts are part of the branched system of orogens associated with amalgamation of the Amazonian, West Africa–São Luís, São Francisco and other cratons and minor continental blocks into the West Gondwana supercontinent.

Abstract The Araçuaí–West Congo orogen encompasses orogenic domains located to the SE of the São Francisco Craton in Brazil, and to the SW of the Congo Craton in Africa. From the opening of the precursor basin to the last orogenic processes, the evolution of the orogen lasted from the very beginning of the Neoproterozoic up to the Cambrian–Ordovician boundary. After the spreading of the South Atlantic Ocean in Cretaceous time, the Araçuaí–West Congo orogen was split into two quite different but complementary counterparts. The Brazilian side (Araçuaí orogen) inherited two thirds of the whole orogenic edifice, including all the Neoproterozoic ophiolite slivers, the entire magmatic arc and syn-collisional to post-collisional magmatism, and the suture zone. The African counterpart (West Congo Belt), a fold–thrust belt free of Neoproterozoic ophiolite and Pan-African orogenic magmatism, inherited the thick pile of bimodal volcanic rocks of the Early Tonian rift stage, implying that the precursor basin was an asymmetrical rift with the thermal–magmatic axis located in the West Congo Belt. Both counterparts of the Araçuaí–West Congo orogen include Neoproterozoic glaciogenic deposits, allowing tentative lithostratigraphic correlations, but identification of the ice ages remains uncertain because the lack of sufficient well-constrained geochronological data.

Abstract Provenance studies on metasedimentary rocks of the Baixo Araguaia Supergroup of the Brasiliano Araguaia Belt, central Brazil, yield 207 Pb/ 206 Pb zircon evaporation ages for detrital zircons from quartzites concentrated around 1000–1200 Ma and 2800–2900 Ma; Sm–Nd T DM model ages of schists and phyllites scatter around 1600–1700 Ma. Facies analysis of low-grade metasedimentary rocks from drill cores suggests a sedimentary environment of basin floor and lower- to upper-slope turbidites. Nearby sources are indicated by the textural and mineralogical immaturity; together with structural geological data indicating tectonic transport of the supracrustal pile towards the NW, this suggests probable provenance from the southeastern portion of the Araguaia Belt and not from the Amazonian Craton as usually believed. The Goiás Massif, Goiás Magmatic Arc, São Francisco Craton and Paranapanema block are considered to be the best candidates. They may have formed a larger continental mass during West Gondwana amalgamation, prior to their collision with the Amazonian Craton to form the Araguaia Belt. Final timing of this collision is constrained by c . 550 Ma syntectonic granites. Similar ages for high-grade gneisses in the Rokelide Belt suggest coeval collision and coetaneous metamorphism of the Araguaia and Rokelide belts, but more geological and geophysical data are required for a decisive correlation between these belts.

Abstract The Brasília Belt comprises terranes and thrust-sheets that were tectonically transported towards the western passive margin of the São Francisco–Congo palaeocontinent during an orogenic episode resulting from collision of the Paranapanema and Goiás blocks and the Goiás magmatic arc against São Francisco–Congo at 0.64–0.61 Ga. The tectonic zones of the belt are, from east to west: a foreland zone with Archaean–Palaeoproterozoic granite–greenstone basement covered by Neoproterozoic anchimetamorphic sedimentary rocks (Bambuí Group); a low metamorphic grade thrust-fold belt of proximal shelf successions, mostly siliciclastic, containing rare basement slivers; metamorphic nappes in upper greenschist to granulite facies of distal shelf and slope metasediments and subordinate tholeiitic metabasalts; the Goiás massif, possibly a microcontinent; and the Goiás magmatic arc. The accretion of these terranes against the western margin of the São Francisco–Congo palaeocontinent took place during an early phase of Gondwana supercontinent amalgamation, when terranes accreted around São Francisco–Congo to create a proto-West Gondwana landmass, around which subsequent collisional and accretionary events followed, such as those in the Borborema–Trans-Saharan province ( c. 0.62–0.60 Ga); in the Ribeira–Araçuaí belt ( c. 0.58 Ga); along the Araguaia and Paraguay belts (collision of Amazonia, c. 0.54–0.52 Ga); and the accretion of Cabo Frio terrane in the Ribeira Belt ( c. 0.53–0.50 Ga).

Abstract Four main classes of tectonic entities may be considered for the Ribeira Belt and southwest African counterparts: (1) cratonic fragments older than 1.8 Ga and their passive margin successions, (2) reworked basement terranes with Mesoproterozoic and/or Neoproterozoic deformed cover, (3) magmatic arc associations, (4) terranes with Palaeoproterozoic basement and deformed Neoproterozoic back-arc successions. Based on comparative investigation, a tectonic model of polyphase amalgamation is proposed with c . 790 and 630–610 Ma major episodes of intra-oceanic and cordilleran arc magmatism along both sides of the Adamastor Ocean. Subsequent diachronous collision of the arc terranes and small plates followed at c . 630, 600, 580 and 530 Ma. The tectonic complexity reflects an accretionary evolution from Cryogenian to Cambrian times. The São Francisco–Congo and Angola palaeo-continents did probably not behave as one consolidated block, but rather may have accommodated considerable convergence during the Brasiliano/Pan-African episodes. The final docking of Cabo Frio and Kalahari in the Cambrian was coeval with the arrival of Amazonia on the opposite side, resulting in lateral reactivation and displacement between the previously amalgamated pieces. The transition between the Cambrian and the Ordovician is marked by the extensional collapse of the metamorphic core zones of the orogens.

Abstract Neoproterozoic–Cambrian amalgamation of West Gondwana involved the collision of several terranes of older crust that are now in eastern South America and western Africa. U–Pb (SHRIMP) detrital zircon ages from representative metasedimentary units of the Ribeira and Dom Feliciano belts (South America) and Gariep and Damara belts (Africa) provide constraints on the possible sediment source areas across probable suture zones. Ribeira detrital zircons are Palaeoproterozoic and Archaean. For the Dom Feliciano Belt, a contribution of Meso- and Neoproterozoic zircons is present, which definitely indicate Neoproterozoic sedimentation. It is proposed that the inflow of material to the Ribeira basin was essentially derived from the Paranapanema and Rio de la Plata cratons, whereas for the Damara and Gariep–Rocha belts source areas were from the Namaqua Belt. The Dom Feliciano Belt received sediments from the South American side and to a lesser degree from African sources. These results highlight the differences in the detrital zircon signatures across a proposed West Gondwanan suture, with those in the west being derived from distinctive South American basement sources and those in the east from distinctive African sources.

Abstract The Pan-African Damara orogenic system records Gondwana amalgamation involving serial suturing of the Congo–São Francisco and Río de la Plata cratons (North Gondwana) from 580 to 550 Ma, before amalgamation with the Kalahari–Antarctic cratons (South Gondwana) as part of the 530 Ma Kuunga–Damara orogeny. Closure of the Adamastor Ocean was diachronous from the Araçuaí Belt southwards, with peak sinistral transpressional deformation followed by craton overthrusting and foreland basin development at 580–550 Ma in the Kaoko Belt and at 545–530 Ma in the Gariep Belt. Peak deformation/metamorphism in the Damara Belt was at 530–500 Ma, with thrusting onto the Kalahari Craton from 495 Ma through to 480 Ma. Coupling of the Congo and Río de la Plata cratons occurred before final closure of the Mozambique and Khomas (Damara Belt) oceans with the consequence that the Kuunga suture extends into Africa as the Damara Belt, and the Lufilian Arc and Zambezi Belt of Zambia. Palaeomagnetic data indicate that the Gondwana cratonic components were in close proximity by c. 550 Ma, so the last stages of the Damara–Kuunga orogeny were intracratonic, and led to eventual out-stepping of deformation/metamorphism to the Ross–Delamerian orogen ( c. 520–500 Ma) along the leading edge of the Gondwana supercontinental margin.