The Rio Itapicuru Greenstone Belt, Bahia, Brazil: Structure and Stratigraphical Outline

Precambrian Research, 42 {1988) 1-17 1 Elsevier Science Publishers B.V., Amsterdam -- Printed in The Netherlands

THE RIO ITAPICURU GREENSTONE BELT, BAHIA, BRAZIL: STRUCTURE AND STRATIGRAPHICAL OUTLINE

IAN DAVISON 1, JOAO BATISTA GUIMARAES TEIXEIRA 2, MARIA DA GLORIA DA SILVA 1, MANOEL BARRETO DA ROCHA NETO 3 and FERNANDO MARTINS VIEIRA MATOS 2

1Universidade Federal da Bahia, Curso de Pds-gradua¢fio em Geologia, Instituto de Geoci~ncias - Rua Caetano Moura, 123, Federa¢5o, CEP40.120, Salvador, Bahia (Brazil) ~DOCEGEO, Rio Doce Geologia e Minera¢5o S.A. - Rua Guadalajara, 20, Barra, CEP 40.160, Salvador, Bahia (Brazil) :~C.B.P.M., Companhia Baiana de Pesquisa Mineral - 4 °. Avenida s/n, Centro Administrativo da Bahia, CEP 40.000, Salvador, Bahia (Brazil)

(Received and accepted March 15, 1988)

Abstract

Davison, I., Teixeira, J.B.G., Silva, M.G., Neto, M.B.R. and Matos, F.M.V., 1988. The Rio Itapicuru greenstone belt, Bahia, Brazil: structure and stratigraphical outline. Precambrian Res., 42: 1-17.

The Rio Itapicuru greenstone belt is situated within the Silo Francisco craton, and is the largest of a series of N-S trending volcanosedimentarysequences surrounded by gneisses and granitoids. Pb-Pb and Rb-Sr dating of andesitic lavas within the supracrustal sequence yielded ages of 2.1 Ga, with late tectonic granitoids emplaced about the same time. The succession appears to be developed on a sialic basement represented by the Santa Luz gneiss complex. Although the actual contact between basement and supracrustals has not been observed, there is a sharp contrast in strain and metamorphic conditions near to the contact zone. A 2.9 Ga U-Pb age obtained from a gneissic raft within the intrusive Ambrosio dome is interpreted as a basement to the greenstone sequence and was plucked from a lower crustal level by the ascending magma. The Itapicuru River section provides good exposures along an E-W profile and has been used to attempt a reconstruction of the lithostratigraphy in this area, where the supracrustals reach a maximum thickness of 9.5 km. Rare stratigraphic facing and structural data indicate that the base of the sequence is marked by an intrusive granitoid. Above this contact is a 5.0 km thick succession of tholeiitic subaqueous basalts with rare cherts, banded iron formation, pelites, andesitic lavas and tufts. This mafic domain is overlain by a 3.5 km thick sequence of andesites, tufts and volcanogenic sediments, with rare dacites and laminated cherts. A silica gap between 55% and 60% Si02 characterises the lavas as a bimodal sequence. Two small ultramafic bodies with komatiitic chemistry also occur in the andesitic volcanic domain. The uppermost 1.0 km consists of volcanogenic siltstones and shales with rare arkoses, tufts and basalts. Localised bedding-parallel shear zones are the earliest deformation features observed. The shear zones may be up to 200 m thick and were important in controlling subsequent lode gold mineralisation. In the north central sector of the greenstone belt these shear zones were deformed by N-S trending, east verging folds, with ~ 20 km wavelength. The main greenstone sequence in the Itapicuru River section lies on the overturned limb of a syncline, and the Am- brosio dome occupies the adjoining anticline. In the southern part of the greenstone belt, the initial deformation produced a major shear zone striking E-W with a 40 ° S dip and a subhorizontal E-W stretching lineation. North verging folds and thrusts produced the present E-W trend. A late, upright, NE-SW trending fold phase terminated the ductile deformation. Greenschist facies metamorphism affected the major part of the belt during deformation, with amphibolite facies assemblages developed around the margins, and thermal aureoles developed around intrusions. Metamorphic fluids transported gold to higher crustal levels via pre-existing, steeply dipping shear zones and faults and produced the large Fazenda Brasileiro deposit, along with many smaller deposits.

The Rio Itapicuru greenstone belt is the larg- The mafic volcanics crop out poorly, but est of a series of Precambrian volcanosedimen- weather to produce a characteristic red soil, tary remnants preserved along the east margin which was used to map their areal extent. They of the Sgo Francisco craton in NE Brazil (Fig. occupy ~ 60% of the greenstone area and are 1). Since the discovery of the greenstone by principally situated along the marginal zones of Mascarenhas (1973) it has attracted much in- the belt, in contact with granitoids and gneisses. terest because of its metallogenic potential. A Massive, greenish-grey, fine-grained basalts regional map and preliminary lithological de- constitute the main lithotype. Modal estimates scriptions were presented by Kishida (1979) average about 40% actinolite, 25% plagioclase, and Kishida and Riccio (1980), and regional 15% epidote, 10% chlorite, 5% calcite and 5% investigations culminated with the discovery, accessory components. In places, variolitic tex- in 1976, of the Fazenda Brasileiro gold deposit tures occur in irregular zones within the mas- (Teixeira and Kishida, 1980; Teixeira et al., sive flows. Varioles consist of whitish round 1986). spots (up to 1.5 cm in diameter) filled with cal- The tectonic evolution and stratigraphy of the cite, albite and quartz. This texture is inter- belt are described with the aid of detailed maps preted as resulting from the rapid crystallisation produced by local mining companies operating of two immiscible liquids, one felsic and the in the area over the past ten years. The sedi- other basaltic (Silva, 1984). Pillow structures mentary sequence and the tectonic history of were observed in only two localities (A and B the southern E-W trending arm of the belt ap- in Fig. 3) and range in size from 0.15 m to 2.0 pears to be different from the main N-S trend- m. (maximum dimension). Aphanitic chilled ing arm (Fig. 2 ), hence, the two will be described rims and radial cooling cracks are present and separately. triangular tails can be used to determine the stratigraphic way-up (Fig. 5). Flow breccias of angular fragments in an aphanitic basalt matrix also occur close to the pillows (Fig. 6 ). Por- phyritic basalts are rare and occur as irregular Cross-section in the north central sector lenses, where igneous megacrysts can be either amphibole or plagioclase, and may reach up to 1.0 cm in length. Kishida and Riccio (1980) classified the basalts into inferior (high FeO and Structural studies and way-up criteria have TiO2) and superior (low FeO and Ti02) groups; been used to reconstruct the lithostratigraphy both have tholeiitic affinities. However, three of an E-W cross-section exposed along the repetitions of the inferior group along the Itapi- banks of the Itapicuru River (Fig. 3). The stra- curu River profile imply the presence of two tigraphic column indicates outcrop control major fold structures (wavelength 2.0 km), if which was used in the reconstruction (Fig. 4). such an inferior and superior stratigraphy ex- Kishida and Riccio (1980) postulated the same ists. Several marker beds, with great lateral stratigraphic order presented in this paper, but continuity, occur in the basalts. They are not without the supporting structural and strati- repeated and we find no evidence for such fold- graphic facing data. ing. Hence, we believe that the inferior and su- 60 ° W -F- 12o N

0 500 lO00 km I I I 54 ° W -Jf- 8ON

48ow -tL 4ON +++

_,_ IbNA~+ _,_ _,_ + V + ) .~ 0 o + + + :...':..:

• :.....-.: .:... ' A:M:A:Z-'.O'N:~ iiS::BASl "": • .. :'..:....:::!:..::...~ .:: :.. -- : ..'-::...:.:..: - - .:.....::-,:,'. . 56ow : . *." "'... .':: .. + + + + + + + :i : ...... • • . ,.-...• .... , "... :.. -Jr- 4Os + + + + + !::.(+ ~...~+ + + + + + ,,, ,~ + + + + + ]:~ ; +~-:..-~ TRA L -IB~A~ I L~ + + + + I::::) -+ ++~+SHIELD + + • .': :.:. "-:':.'...'..': + + T + + + +/-/ -+++++++~~+ +:: + • :."" :~ +++ ++ "~: )+++ ~J ++ ++++ /]'SA'VA00R ++ ~-++ ++) •.:.:~_ ~ + + + + +2+~u,4:.: LEGEND •.. :...... ~.. ~ATLAN 1 [I C+SH IE LD(..5." •..-.:.:: : :.~ + + +, ~J+ + +j::) (o \ x, ~ 2 P.ARANA-::/+ + ~,+~ + [;:~ • + ÷ / V ~ 3 ::. '...%::.7++ +'-.e+/ x~ I:.: :?-C-~+/"- + ~o~ "7 ": "'.':~ +f 4a ° w

Fig. 1. Major geological units of the east-central part of South America. Simplified after Almeida (1977 ), Gibbs and Barron (1983), Schobbenhaus (1984) and other maps. Legend: (1) Phanerozoic cover, (2) greenstone belts, the Itapicuru and Guiana Shield occurrences are Proterozoic and the others are Archaean, (3) Precambrian terrains, (4) outline of the Silo Francisco craton, {5) study area (Rio Itapicuru greenstone belt). perior terms do not have any stratigraphic schists (in places graphitic and associated with significance. small massive sulphide bodies) and rare car- Intercalations of chemical sediments and bonates are the main lithotypes. One marker fine-grained clastics occur throughout the horizon ofpelitic sediments, ~ 100 m thick, can MVD. Fine, parallel-laminated, white and black be traced along strike for ~ 100 km, without any cherts, banded iron formation (BIF), pelitic change in facies, indicating that the large + + + LEGEND

+ + + 3

+ NO R BEST I N+A X..~ + +DOME'+ + + [i ~9

+ + +J[ 12 4- + 13 N 14

÷,~. 0 SCALE 10 i\+ I I km \+2 \"~ / /

,--¢~ \\ \ • +

~"-(~ \~ ~ l~'~AMBROSIO ~-- ~ ~.~'~_IJ ~ "~OME+

\ + + \ ' ., + + \ ~_f. + +

\k\+

~,, ~ ~

;ALVADOf .... : ::F +

2;# + + 1 ~ + + 4- ,oo 4O° B9020' ~ ~ + ~r + 4- 39000 ' + Fig. 2. Geological map of the Rio Itapicuru greenstone belt. The two thick rectangles correspond to the areas shown in Figs. 3 and 11. Legend: ( 1 ) Mesozoic sediments (Tucano basin ), (2) late tectonic granitoids (mainly granodiorites ), (3) meta- diorites, (4) metagabbro sills, (5) metadolomites, (6) metasiltites and metapelites from the sedimentary domain, (7) banded iron formation and cherts, (8) metaconglomerates and meta-arkoses, (9) calc-alkaline metavolcanics (lavas and pyroclastics ), ( 10 ) metapelites and chemical sediments in the mafic volcanic domain, ( 11 ) tholeiitic metabasalts, ( 12 ) deformed granitoids and gneisses, ( 13 ) gold mines, ( 14 ) gold deposits and occurrences, ( 15 ) locality where basement is inferred (see text). END

I-/18

• 12

-I S 15 ~LE 2I n

PEDRA ~ "ALTA+ + -I .DOME+ + + ,.+ + + +

+ + -I

Fig. 3. Geological map of the Itapicuru River section. Legend: (1) late to syntectonic granitoids ( mainly granodiorites), (2) metagabbro, (3)metadiorite, (4)felsic metatuffite, (5)andesitic meta-agglomerate, (6)meta-andesite and metadacite, (7) metapelite and metasiltite, (8) chemical metasediments (BIF and chert ), (9) tholeiitic metabasalt, ( 10 ) deformed granitoid and gneiss, (11) younging direction (way-up criteria), (12) attitude of foliation, (13) pillow structures, (14) locality of photographs (see text), ( 15 ) geological section (see Fig. 9 ). Labels for thin units: sp -- serpentinite; go -- gondite (spessartine + quartz ); js -- jaspilite. developing basin had a constant sedimentary (SD). One localised volcanic centre is marked environment in a N-S direction (Fig. 2 ). These by andesitic agglomerates which approach 400 lithologies suggest that the MVD developed in m in thickness; this centre lies astride the Itapi- an immersed basin with very little clastic sedi- curu River (Fig. 3). mentary influx, but with major subaqueous ba- The felsic volcanic edifice has a N-S extent salt effusions in tranquil water conditions. of 30 km. Agglomerates are surrounded by finer Water depth during deposition is difficult to de- tufts and andesitic lavas on all sides. The rela- termine, but is believed to be deeper than wave tive proportions of pyroclastics, tufts and lavas base, owing to the absence of any current-in- are still difficult to determine accurately, but duced sedimentary structure in the interca- lavas and tufts appear most important, occur- lated sediments. Both oxidising and reducing ring in approximately equal proportions. conditions are witnessed by the intercalations The grey-coloured andesites can be massive of oxide facies BIF and graphitic shales with or schistose, depending on the amount of strain. minor massive sulphides. They are fine-grained to porphyritic with horn- Felsic volcanic domain (FVD) blende, plagioclase and, rarely, quartz phenocrysts; the matrix consists of finely saussuritised This domain is irregular and grades laterally plagioclase, chlorite and quartz. Alteration into the MVD or into the sedimentary domain products are mainly epidote, sericite and albite. PRESENT DAY THICKNESS IOkm PRESENT EROSION LEVEL BLACK STRIPS ~ -L OUTCROP CONTROL ),FINE GRAINED SILTS AND SHALES, PARALLEL LAMINATED SEDIMENTARY ),WITH GRADED BEDDING (PROBABLY DISTAL TURBIDITES)/DOMAIN (SD) -~-~ -x-x- GONDITE HORIZON , GABBROIC SILLS

AN DESITIC AGGLOMERATES x x x MARBLE HORIZON, MASSIVE , HIGHLY FRACTURED x x x ~-QUARTZ DIORITE INTRUSION REMOVED FROM SECTION ~- x x \Mg-RICH ULTRAMAFICS (SERPENTINISED) FELSIC x x x ANDESITES AND ANDESITIC TUFFS AND PELITES - VOLCANIC I x x x RARE DACITES AND CHERT DOMAIN x...,.~ CHERT (FVD) X X X X X X X X X j{ X • x_ FAULTED CONTACT FLOW BRECCIAS + PILLOWS 5 "Q! ANDESITES AND ANDESITIC TUFFS I THOLEIITIC BASALTS 4 I GRAPHITIC SHALES WITH MASSIVE PYRITE I BLACK CHERTS AND TUFFACEOUS SEDIMENTS MAFIC ,- -,- ANDESITES -VOLCANIC ~.o. BIF DOMAIN I (MVD) ), PILLOW BASALTS ( FAZENDA REBO LO ) I,._ ""'' ANDESITES + DACITES I "'" BIF THOLEIITIC i BASALTS

0 ++ + + + INTRUSIVE GRANODIORITE ),, WAY- UP CRITERIA I- ++++

Fig. 4. Reconstructed lithostratigraphic column of the Itapicuru River section, composed along the geological profile indicated in Fig. 3. Present day thicknesses are shown along the left-hand side, and black strips show the outcrop control used in the reconstruction.

Modal estimates indicate that the andesites affinity, and there is a bimodal silica distribu- contain 40% plagioclase (An3o_3~), 35% chlo- tion between the tholeiitic basalts (MVD) and rite, 10% epidote, 10% quartz and 5% accessory felsic lavas (FVD), with a gap between 55% and phases (for chemical analyses see Kishida and 60% Si02 (Kishida and Riccio, 1980). Riccio, 1980). Lenses of spherulitic andesite, Felsic tuffs are schistose, with strongly ori- which have a rugose surface in outcrop, occur ented chlorite and sericite flakes in the matrix. in a few places. The spherulites contain aggre- The clasts are fragments of andesitic to dacitic gates of plagioclase and quartz and measure up lavas, broken plagioclase, quartz crystals and to 3.0 mm in diameter. These are believed to be volcanic glass shards. The matrix is very fine- immiscible globules which formed prior to ex- grained and probably composed of ash fall ma- trusion (Silva, 1984). Minor porphyritic grey terial with an andesitic composition. Volcanic dacites also occur in this domain but their areal bombs of andesitic composition which reach up abundance is unknown owing to poor exposure. to 2.0 m in length are common in the agglom- The andesites and dacites show a calc-alkaline erates. Some helicoidal shapes have been iden- i

Fig. 5. Pillowed metabasalts near the Fazenda Rebolo, in Fig. 7. Rhythmic graded bedding in the metaturbidite from the mafic volcanic domain, with way-up criteria well pre- the sedimentary domain. The stratigraphic top is indicated served due to the low strain state. Rocks young from right by the arrow. Consistent younging directions with struc- to left. Exposure locality labelled A in Fig. 3. The hammer tural facing confirm the sequence was only affected by one is 30 cm long. phase of folding. Locality D, Fig. 3. tiffed, indicating that subaerial eruptions the volcanic centres, making this an ideal en- occurred, although it is still debatable whether vironment for massive sulphide deposition, but the bombs finally settled in water or not. as yet only barren massive pyrite has been Epiclastic sediments occur throughout the discovered. domain, with textures and mineralogy similar The only ultramafic rocks yet found in the to the fine tufts, but with sedimentary struc- greenstone belt occur in the FVD near the tures indicating reworking. Chert layers indi- Itapicuru River (Fig. 3). Two ellipse-shaped cate that low sedimentary deposition rates outcrops 1.0 km long and 200 m wide were iden- occurred between volcanic eruptions. Thus, al- tified. These rocks are serpentinites with eu- though the felsic volcanic agglomerates and hedral, equidimensional olivine pseudomorphs, tufts were the results of subaerial eruptions, the 1.0-2.0 mm long, totally transformed to serpen- intercalated cherts suggest water surrounded tine, talc and carbonate, with interstitial augite and minor chromite. Chemical analyses indicate these rocks have komatiitic affiliation, although spinifex textures have not been observed (Silva, 1984 ).

Sedimentary domain (SD)

The main rock types in this domain are fine- grained psammites and pelites, which are rhythmically banded with planar parallel strat- ification (Fig. 7 ). Individual siltstone beds may reach up to 15.0 cm thick and contain pelite rip- up clasts. Graded bedding and rare small-scale Fig. 6. Pillow breccia from the mafic volcanic domain. Lo- cross-bedding provide way-up criteria, which cality B in Fig. 3. The hammer is 30 cm long. were used in the structural interpretation (Fig. 7). White laminated chert, jaspilite, BIF and ~go 181 39°09 '_ , manganese-rich sediments (gondites with spessartine and quartz) also occur throughout ~D this domain, marking periods of low clastic sediment influx. This sequence is interpreted as volcanically derived distal turbidites with in- tervening chemical sediments. Rare arkoses and ]2 conglomerates (with clasts up to 10.0 cm) occur near the base of the sequence, indicating more proximal conditions. Minor basalts are intercalated with the sed- 74 iments along with gabbroic sills, representing ]5 the magmatic activity which may have been in- volved in the production of charged hydrothermal fluids and resulting chemically precipitated sediments.

Granite gneiss domes 8 The Ambrosio dome is the largest body oc- ]9 curring within the greenstone belt sequence and is the most carefully studied to date (Fig. 8). ]1o Lithologies range from tonalite to granite and notable amounts of tourmaline-rich pegmatite 11 occur near the contacts with the greenstone.

Jardim de S~i ( 1982 ) suggested that the gneisses 12 in the dome represent basement, based on observations that these rocks have a more com- 13 plex deformation history than the greenstone sequence. Detailed mapping of the Ambrosio 14 structure indicates that most of the body was 15 intrusive, although large basement rafts were plucked from lower crustal levels (Matos and 16 Davison, 1987a,b ). The most striking feature of the Ambrosio dome is its highly deformed mar- Fig. 8. Map of the geology around the Ambrosio dome. Leg- gins, which are composed of deformed grani- end: (1) pegmatite, (2) granite and granodiorite, (3) me- toids, pegmatites and gneiss rafts. These appear tadiorite, (4) metagabbro, (5) granitic gneiss and migmatite, (6) banded iron formation, (7) metapelite, (8) to have deformed at temperatures high enough felsic metalava and pyroclastics, (9) metabasalt and am- to produce ductile deformation of K-feldspar phibolite, (10) banded gneiss, (11 ) Maria Preta gold mine, phenocrysts and thus deformation is thought to ( 12 ) gold deposit, ( 13 ) attitude of foliation, ( 14 ) plunge of be synchronous with intrusion. The finite strain mineral lineation, ( 15 ) anticlinal axial trace, ( 16 ) locality of dated gneiss sample (see text). fabric at the dome margins, in the Itapicuru River section, is oblate, which suggests the present outcrop level is the flattened roof zone the dome. Rare exposures and magnetic sur- of an intrusion (Matos and Davison, 1987a,b). veys suggest the BIF almost completely sur- An 80.0 m thick BIF horizon is in contact with rounds the dome and it is envisaged that the iron formation acted as a horizontal competent formation in the gneisses may be older (Fig. 2 ). barrier which arrested the magma ascent and Further indirect evidence for a pre-existing produced extensive horizontal flow. The Pedra sialic basement include the megaxenoliths of Alta and Nordestina domes (Fig. 2) consist of banded gneiss within the highly deformed bor- similar lithologies and have similar character- ders of the Ambrosio dome which have been istics to the Ambrosio intrusion. dated using U-Pb on zircons at 2930 + 32 Ma (Ga~l et al., 1987). Later granitoids Deformation in the northern sector Smaller granitoid bodies which are clearly intrusive into the greenstone supracrustals are The deformation history in the N-S trending common (Fig. 2). They are mainly grey grano- arm of the greenstone belt is best displayed in diorites without deformation textures and are the rocks along the E-W Itapicuru River sec- presumed to be late tectonic. As yet the struction (Fig. 3), although the exposure is far from tures and lithologies of the granitoids have not continuous. In the MVD the cleavage dips 50- been mapped or studied geochemically. 70 ° W with an average of 65 ° (Fig. 9). Bedding structures are rarely seen in the basalts. Pillow Gabbroic sills lavas (Fig. 5) at Fazenda Rebolo (Fig. 3) indicate a low strain state (X:Z = 2.1:1), with the Several narrow sills with gabbroic composi- Z axis of the pillows perpendicular to a weak tion have been mapped in the northern sector. foliation which is only observable in the inter- These are preferential sites for gold mineralis- pillow matrix. This suggests that the orienta- ation when associated with shearing and quartz tion of the foliation is close to that of the bed- veining. Most of the gabbros were intruded near ding planes, assuming the pillows had an the contact between the FVD and SD. Relict original bedding-preferred orientation. Gra- ophitic textures with plagioclase and intersti- phitic schists and cherts within the basalt se- tial hornblende are commonly preserved. quence indicate the strike of the bedding is parallel to the schistosity. Minor folds in these Basement sediments have been observed in drill cores, where an axial planar cleavage dips parallel to Unequivocal evidence of the basement to the the cleavage in the surrounding basalts. The greenstone sequence does not exist and there felsic lavas do not usually exhibit minor defor- are no data which support the hypothesis of re- mation structures. In the tufts and volcano- working of basement fragments into the sedi- genic sediments of the FVD and SD, minor folds ments of the greenstone. However, the Santa with up to 3.0 m wavelength and interlimb an- Luz gneiss complex, which lies to the southwest gles of 10-60 ° have been observed. These folds of the Rio Itapicuru greenstone belt is inter- have an axial planar cleavage which can be: ( 1 ) preted as a sialic basement to the supracrustal slaty, with quartz, chlorite or biotite alignment; sequence (Fig. 2). Here, the contact between (2) a crenulation cleavage with or without new the banded gneisses and the greenstone has not muscovite and biotite along the axial planes; and been observed directly, but exposures of pelites, (3) a centimetre-scale, spaced, solution cleav- which lie 50.0 m above the gneisses (locality A, age with dark solution seams. The axial planes Fig. 2), do not show contact metamorphism. of minor folds dip 60 ° W and axes vary with a Furthermore, there is a strong structural dis- plunge of 0-40 ° towards 000-020 ° or to 180- cordance between the gneissic banding and the 200 °. Where younging directions have been ob- cleavage in the pelites, which suggests the de- served in the sediments there were no strati- 10

SEDIMENTARY AMBROSIO PEDRA ALTA MAFIC VOLCANIC DOMAIN )~ELSIC VOLCANIC DOMAIN X ~DOMAJN X DOME ) DOME -- 'x +/+ t..:.:

! :. w + +/+ + +/+ ×~× × /+ + + /+ + +

×,~, x/+ + + + + + +F"

LEGEND I-~-11 I~2~2 I~3 I-~]4 F-~-Is ~6 I-~17-~ a ~-9+io+11

Fig. 9. E-W structural cross-section along the Itapicuru River. Location shown in Fig. 3. Legend: (1) undeformed granite and granodiorite, (2) metadiorite, (3) volcanogenic fine-grained metasediment, (4) calc-alkaline meta-andesite and pyroclastic, (5) chemical metasediment (chert and B IF), (6) tholeiitic metabasalt, (7) banded gneiss and granitic gneiss, (8) pillow structures, (9) stratigraphic younging, (10) syncline axial plane, (11 ) anticline axial plane. The black strips at the top of the section indicate outcrop control. graphic inversions before this folding phase. tently 50-70 ° W. The cleavage observed in the However, localised bedding-parallel shears oc- basalts and andesites is perhaps a composite cur up to several hundred metres wide. During fabric of compaction and tectonic flattening. shear an intense deformation fabric developed The folds are very consistent in axial strike, along with a mineral stretching lineation which and rectilinear continuity of marker horizons is affected by the first phase of folding. These parallel to the axial trace of these folds, is illus- shears appear to be important in controlling the trated in Figs. 2 and 3. Examination of the stra- localisation of lode gold-quartz veins. tigraphic succession indicates that there are no There is a pre-existing mica foliation (mus- obvious large scale repetitions along the Itapi- covite, sericite and biotite) parallel to bedding curu River profile. However, smaller-scale rep- in the sediments and which is folded by the mi- etitions cannot be discounted as the exposure nor folds described above. This foliation has is not continuous. been interpreted as having been produced by an Younging directions indicate that the major earlier phase of recumbent folding (Teixeira, part of this cross-section is on the overturned 1985 ). However, several field observations sug- limb of a syncline (Fig. 9). The centre of the gest this may be a compaction cleavage since no syncline is mapped in the SD, and to the east minor folds have been observed associated with an anticline is occupied by the Ambrosio dome, the foliation, which is consistently parallel to which truncates the fold limb (Fig. 9). Minor bedding. If the limbs of the postulated recum- folds (up to 10 m wavelength) of the same style bent folds are perfectly parallel to the cleavage, and orientation affect the gneissic fabric along this implies very high strain states. However, the margins of the dome. A penetrative biotite finite strain estimates indicate that the de/br- fabric is observed, parallel to the axial planes of mation is moderate to weak (Fig. 10). Struc- these minor folds, and a mineral lineation of tural inversions have not been recorded and biotite and quartz is developed parallel to their there is no observed repetition of mapped axes. This suggests that the Ambrosio dome was marker horizons. It is assumed, therefore, that intruded before, or early in the deformation a bedding-parallel compaction cleavage was de- phase which produced these folds. Thin sec- veloped which was locally sheared, metamor- tions of pelites near to the dome contact, indi- phosed and then refolded by first-phase folds, cate that metamorphic porphyroblasts of garnet which have N-S axial planes and dip consis- and andalusite are syntectonic, with the axial 11

.>, r I1 ÷+ +4 MAFIC 56~11 FELSIC 12 AMBROSIC VOLCANIC Tq~21 VOLCANIC ,SEDIMENTARYDOMAIN DO+ME DOMAIN ','X , DOMAIN N I I/

[] [ FAZENDA 4.,: CELULOSE / lac~w#o FAZENDA REBOLO I L SCALE I GRANDE I 0 km 2 I+ ll°O0 '

Fig. 10. Estimates of maximum flattening across the cleavage in the Itapicuru River section. Numbers are X:Z ratios and the line through the circle is the horizontal projection of the X axis. X~- Y at all localities. X is the maximum, Y the intermediate and Z the minimum axis of the strain ellipsoid. planar cleavage of the major folding phase Measurements were made using pyroclastic wrapping around them and with internally fragments, pillow lavas (Figs. 5 and 6), detrital folded inclusion trails. Minor fold asymmetry quartz grains and conglomerate pebbles. Very indicates the Ambrosio dome is an asymmetric little long-axis angular variation was observed periclinal fold structure rather than an original in the deformed objects, presumably because a dome shape due to intrusion (Figs. 8 and 9). tectonic fabric was superimposed on an initial Earlier it was stated that the Ambrosio dome bedding plus compactional fabric. The har- lies in contact with a BIF horizon which is in- monic mean method of Lisle (1977) was used terpreted to be localised near the top of the to estimate the axial ratios of the 'sediment- greenstone pile; if this interpretation is correct, ary fabric + compaction + deformation' ellipse. the magma must have penetrated up through Measurements which were made in basalts most of the greenstone succession before (X:Z=2.1:I) are considered to be the closest spreading laterally below the BIF horizon. to a true tectonic flattening strain across the A second phase of deformation was observed cleavage, but these are still maximum esti- in the tufts and sediments, but not in the lavas. mates. The pyroclastics and intercalated sedi- This is a crenulation phase, with maximum fold ments show much higher X:Z ratios (up to amplitude up to 30.0 cm and wavelengths of a 12: 1) because of original sedimentary pre- similar scale. There is no mineral growth asso- ferred orientations and higher compaction, as ciated with this deformation phase. Crenula- well as larger tectonic strains (Fig. 10). tion axes generally dip 60-280 ° with variably The possibility of any minor folds producing oriented axial planes. Detailed mapping at a tectonic thickening where there is no exposure scale of 1 : 10 000 does not reveal any larger-scale could result in a reduction of the estimate of the structures associated with these crenulations. original thickness of the greenstone succession. It is significant that these structures are only However, exposed minor folds are rare in the developed in less competent units, and crustal basalts and andesites and therefore fold repe- shortening is thought to be minimal. tition is not thought to be important, except in Strain measurements indicate that the finite the sediments and tufts. This relatively simple fabric is oblate, with X being approximately deformation scheme facilitated the reconstruc- equal to Y in the majority of cases. The flatten- tion of the greenstone succession using the ing across the cleavage (X:Z in Fig. 10) is con- overturned limb of the megasyncline. The total sidered to be a maximum value for the tectonic thickness of the succession is 9.5 km (Fig. 4). deformation, as bedding is subparallel to cleav- This is considered an approximate value since age in all of the samples studied. Thus, any sed- an intrusive granitoid marks the base of the imentary fabric is also included in the greenstone sequence and greenschist facies measurement (see Ramsay and Huber, 1983). sediments occur at the top, suggesting that a 12

I'l'l'l' ~' '1'1'1'1'1' '1 I'l'l'l I'r + ~ + + + ~o ÷ + area also makes it difficult to reconstruct the stratigraphy. Tholeiitic basalts and calc-alkaline felsic volcanics and sediments are repre- I li If111 I • -- sented, although there is a predominance of II I I I I I I I I I I I I ':''~,-~ ~ <::~< -- coarser-grained clastics at the top of the sedi-

llll i I .... mentary sequence. Polymictic matrix-supported conglomerates, with volcanic-derived + + IIII i III I + + ~ detritus, are well developed south of Araci (Fig. tl / + + 11). Five deformation phases have been recog- / l + /3 ~ 9 Nil;li~/It'71~ll/ I / BARROCAS ~ / b~:~°A4 10 nised in the Fazenda Brasileiro mine area II!l'.lh)~.l:~i~ q/ + + + + + ~ I+ ' '1 I',l',lh~J,l'l~'x ~/ DOME )~ / ~ililllll5 ~ 11 (Teixeira, 1985; Gadl and Teixeira, 1988). The ~ii~il// x j + + + + + /J+ + ~ "~l°~'- I+~. + f~/7-+ + + l',,J~ first tectonic phase was an intense shearing ac- l+7 + \ companied by isoclinal folding. This shear zone extends for at least 8.0 km E-W along the We- ber belt (Fig. 11). A strong ductile fabric pro- Fig. 11. Geological map of the southern sector of the Rio Itapicuru greenstone belt. After Teixeira (1985). Legend: duced in the sediments and igneous rocks ( 1 ) Phanerozoic cover (Tucano basin ), ( 2 ) intrusive gran - destroyed most of the primary textures. An E- odiorite and tonalite, (3) differentiated mafic sill (Weber W, subhorizontal stretching lineation was pro- belt), (4) sedimentary domain; polymictic metaconglom- duced on the $1 shear planes and is interpreted erate, (5) sedimentary domain; metapelite, meta-arkose and metagreywacke, (6) felsic volcanic domain: calc-alkaline as the main movement direction. An F1 anti- metadacite and metarhyodacite associated with metatuf- form which folds the shear fabric is inferred faceous and metasedimentary layers, (7) mafic volcanic from repetition of the differentiated units of a domain: tholeiitic metabasalt interbedded with metasedi- layered mafic sill in the Fazenda Brasileiro mine mentary layers, (8) banded gneiss, (9) attitude of $2 fol- area (Fig. 12; Teixeira, 1985). Further evidence iation, ( 10 ) fault/shear zone, ( 11 ) Fazenda Brasileiro gold mine. of localised F1 folding lies in stratigraphic inversions identified in the sediments. The iso- considerable amount of erosion has occurred to clinal folding (F~) and shearing events are expose these metamorphic rocks at the surface. classified as the D1 deformation in this paper. There is also an offsetting effect of flattening The D1 event is not penetrative outside the We- and, hence, a thinning of the volcanosedimen- ber belt, and the D2 event is the first recognised tary pile caused by compaction and tectonic regional phase in the southern sector. deformation. The D2 event produced E-W trending, tight to isoclinal folds, with axial planes dipping 40 ° S and subhorizontal undulating axes. The fold- Lithologies and deformation in the ing produced an axial planar slaty cleavage out- southern sector side the areas affected by D1, and a finely spaced crenulation cleavage in the mine area. Chlorite The area which was studied in detail in the and white mica are the main fabric-forming southern sector of the Rio Itapicuru greenstone minerals. The F2 folds have >/500 m amplitude belt is outlined in Fig. 2 and shown in more de- in the mine area (Fig. 12). The D:~ event rep- tail in Fig. 11. Regional mapping is not as ad- resents a continuation of the same south to vanced in the southern domain as it is in the north movement pattern indicated by the D2 northern domain, except in the Faixa Weber structures. The D:~ phase produced subvertical, gold district and along principal road and rail E-W trending folds which are open, asymmet- sections. The complexity of the structure in this ric monoclines with a spaced mica crenulation 13

"P2

NOT TO SCALE Ap 3 X s Fig. 12. Geologic section (a) and stratigraphic column (b) of the Fazenda Brasileiro gold mine. After Teixeira (1985). Legend: ( 1 ) axial plane trace of F~ folds, (2) axial plane trace of Fe folds, (3) axial plane trace of F:~ folds, (4) fault, (5) F antiform, (6) F, antiform, (7) F:~ antiform, (8) F~ synform, (9) orebodies (gold-quartz vein network), (10) Fe-rich metagabbro, (11 ) metagabbro, (12) graphitic metapelite, (13) metaturbidite, (14) tholeiitic metabasalt. The gold-quartz veins are hosted by the Fe-rich gabbroic horizon which is part of the differentiated mafic sill (see text). cleavage. These folds refold the axial planar D2 E-W in the north to N-S in the south is poorly cleavage. Northward-directed thrusting which understood. The southern domain appears to is also correlated with the D2-D:3 events oc- be an anomalous area, as all the greenstone curred in the Fazenda Brasileiro area (Fig. 13). remnants further north show a well-defined The maximum shortening direction changed N-S orientation. during the D 4 event, when open, upright NE- SW trending folds developed, with an axial Metamorphism and structural planar crenulation cleavage. F4 fold axes gen- relationships erally plunge 55 ° towards 225 °. As yet the amplitude and wavelength of these late folds have not been determined. Detailed metamorphic studies undertaken in Finally, NE-SW and NW-SE trending brit- the northern sector and field observations in the tle faulting occurred which produced kilometre southern sector indicate that the greenstone is scale horizontal offsets of lithological contacts, generally in greenschist facies, except near to but net slip of these faults is not known. granite-gneiss domes where amphibolite facies Structural correlations between the northern metamorphism developed. and southern domains are still not clear and a Contact metamorphism is observed around change in principal shortening direction from all the granitoid intrusions, with development 14

reoriented by mechanical rotation and pressure solution along with new mica growth (sericite, biotite and chlorite). The F1 folds were thus produced during greenschist facies metamorphism, although mica growth had already com- menced before the D1 phase. Along the Ambrosio dome margins the higher temperatures produced hornblende-hornfels facies assemblages (Matos and Davison, 1987b). The pelites show a rapid increase in metamorphic grade near the eastern border of I//vv~ ~ N the dome where, 500 m from the dome contact, biotite-grade sediments occur. These same sed- / / / / iments undergo partial melting to produce a coarse-grained muscovite + quartz assemblage . t,' ,' ,' A I along the dome margin. This zonation suggests I /,' :( 'i v>V. '/''''/j that within 500 m of the dome contact the thermal gradient during intrusion exceeded 200 °C km-1. The porphyroblastic textures of garnet and andalusite indicate that peak metamorphism was syntectonic during intrusion of the dome. The porphyroblasts have tectonic inclu- Fig. 13. (a) Northward directed thrusting producing a du- sion fabrics which may or may not be deformed, plex structure in the Fazenda Brasileiro gold mine. (b) Line drawing of (a). Broken lines represent $2 foliation traces, and also exhibit pressure shadows and 'wrap solid lines are thrust contacts and the V-pattern is a quart- around' cleavages. A syntectonic metamorphic zofeldspathic marker horizon. peak is supported by evidence of pegmatite in- jection into the high-grade zone during deformation, with the pegmatites deformed to of hornblende hornfels in the basalts and an- varying degrees depending on the time of injec- dalusite and garnet porphyroblasts in the pel- tion (Matos and Davison, 1987a,b). ites. Where the granitoids are not deformed and Because the major part of the greenstone belt appear to be late-tectonic, supporting mineral appears to be greenschist facies it may be ar- textures can be found. Porphyroblasts from the gued that a simple megasyncline in the Itapi- thermal aureole overprint earlier minerals de- curu River section is unlikely, since fining the regional schistosity produced during metamorphic grade should progressively in- the regional metamorphism. crease towards the base of the sequence. The The textures in the sediments allow the tim- mapped vertical distance (normal to bedding) ing of metamorphism to be related to defor- between the uppermost and lowermost expo- mation. All the sediments exhibit a bedding- sures of greenschist assemblages is 7.0-8.0 km. parallel mica fabric of sericite and chlorite. As However, deformation occurred during meta- outlined in the structural section, this fabric is morphism and, perhaps, uplift of the synclinal believed to be caused by deep burial metamor- limbs occurred at the same time as subsidence phism during basin subsidence and is a 'com- in the hinge. This would allow rocks of similar pactional' cleavage. This cleavage was folded by metamorphic grade to crop out over apparently F1 folds which produced an $1 axial planar large vertical profiles, if metamorphic reaction cleavage consisting of mica and quartz grains rates could keep pace with subsidence rates. 15

Geochronology deposition is associated with hydrothermal alteration processes which overprint greenschist Brito Neves et al. (1980) dated the felsic vol- facies regional metamorphism and caused car- canics in the southern part of the belt at bonation, albitisation and sulphidation, along 2080 +_ 90 Ma, with a five point Rb-Sr whole with quartz veining (Marimon et al., 1986). The rock isochron, with an initial ratio of main host rock to gold mineralisation is an Fe- 0.7017_+ 0.0014. This age is corroborated by a rich differentiated gabbroic horizon which is three point Pb-Pb mineral isochron on ande- part of the sill (Fig. 12). The main orebodies sites taken from the Itapicuru River section are localised within a shear zone (Teixeira, which gave a tentative age of 2178+_12 Ma 1985). However, mineralisation occurred after (Ga~l et al., in preparation). These ages are in- shearing, when the shear zone had been rotated terpreted as the original magma crystallisation to a 45 ° S dip by later folding. event in the FVD. The presence of mafic rocks is not a prereq- Brito Neves et al. (1980) also published a 13 uisite for gold deposition in the belt, since most point Rb-Sr whole rock isochron on intrusive of the other smaller deposits occur in felsic tufts, granites and migmatites from the western and andesites, pelites and intrusive diorites, fre- southern parts of the belt which gave an age of quently associated with steeply dipping shear 2090_+27 Ma, with an initial ratio of zones which are cross-cut by quartz veins pro- 0.704 + 0.0008. Recent U-Pb zircon ages of syn- duced by hydraulic fracturing. The major quartz tectonic granites in the western part of the belt veins occur along lithological contacts which dip (5.0 km north of Santa Luz; Fig. 2) also gave a parallel to the shear planes. Pervasive small- similar age of 2107+23 Ma (Gadl et al., 1987; scale fracturing in the more competent lithol- Gadl et al., in preparation). ogles was responsible for providing the neces- Gadl et al. (1987) also presented an age of sary surface area for effective fluid-wall rock 2930 +_ 32 Ma from a highly deformed gneiss in reactions to produce economic gold concentra- the marginal zone of the Ambrosio dome (lo- tions. These deposits appear to be very similar cality A; Fig. 8) which we interpret to be a xeno- to Australian Archaean gold deposits described lith plucked from a lower crustal level. The by Groves et al. (1984). geochronological data point to a Proterozoic age for deposition, deformation and intrusion Conclusions within the greenstone succession, with a probable Archaean basement already in existence The Rio Itapicuru greenstone belt is consid- before greenstone formation. The isochron and ered to be the remnant of an Early Proterozoic concordia plots will be presented in more detail lava and volcanoclastic filled basin, which was elsewhere (Ga~l et al., in preparation; Silva et subsequently metamorphosed and deformed al., in preparation). within ~ 200 Ma of formation. Unequivocal sialic basement has not been Gold mineralisation observed, but indirect evidence suggests that granodioritic gneisses did exist before green- The principal gold districts presently known stone deposition. Exposures along the Itapi- in the Rio Itapicuru greenstone belt are shown curu River section permit a crude restoration of in Fig. 2. The dominant geological feature in the stratigraphic pile in the thickest part of the the Fazenda Brasileiro (the largest deposit) is greenstone, but rapid lateral variation in thick- an E-W trending, 8.0 km long differentiated ness and sedimentary facies do not allow ex- mafic sill which is bounded by basalts and tur- trapolation of this stratigraphy to other parts bidites (Figs. 11 and 12; Teixeira, 1985). Gold of the belt. Subaqueous tholeiitic basalts with 16 intercalated chemical sediments occupy the were controlled by steeply dipping shear zones, lowermost 5.0 km of the pile, followed by 3.0 km but mineralisation occurred in hydrothermally of calc-alkaline andesitic lavas, pyroclastics and altered zones after shearing. Host rocks to the epiclastic sediments. A localised felsic volcanic gold-quartz veining are variable, with gabbro, centre occurs in the Itapicuru River section, dacite, andesite, tuff and pelite recorded. which is marked by a 400 m thick agglomerate The large-scale tectonic significance of the unit and a subvolcanic intrusion of dioritic relatively narrow greenstone sandwiched be- composition. Finally, 1.0 km thick distal tur- tween extensive granulite and granite terranes bidites terminate the sequence. However, these of the S~o Francisco craton (Fig. 1 ) is still not sediments may have been much thicker as ero- fully understood. We intend to discuss the gen- sion has exhumed greenschist facies sediments. eral crustal evolution of this region with the aid The sediments are much coarser in the south- of more laboratory data (Ga~l et al., in prepa- ern portion of the belt where polymictic con- ration; Silva et al., in preparation). glomerates occur, but the more complex The Rio Itapicuru greenstone succession has structure and paucity of exposures does not many features in common with Archaean permit a stratigraphic reconstruction in this greenstones, and constitutes another example area. of the growing number of Proterozoic occur- The basin was compressed in an E-W direc- rences identified in South America. tion along most of its present length which produced localised bedding-parallel shearing and Acknowledgements eastward-verging, large-scale folding with a 000-020 ° trend. Farther south, in the Weber I.D. acknowledges financial assistance from gold district, the predominant structures are E- the Secretariat of Mines and Energy of Bahia W striking folds which have a northerly verg- State and a research grant from the Conselho ence. This change in movement pattern is still Nacional de Desenvolvimento Cientlfico e Tec- poorly understood and structural correlations nolSgico (CNPq). J.B.G.T. and F.M.V.M. were between the N-S domain and E-W domain are financially supported by DOCEGEO and not clear, the two being separated by a late N- M.B.R.N. by the Companhia Baiana de Pes- S fault, the displacement on which is not known. quisa Mineral (CBPM). C.R. Anhaeusser, A. E-W shortening in the N-S trending domain KrSner and an anonymous reviewer provided was accompanied by intrusion of massive gran- useful criticism of this paper. odiorites. These syntectonic intrusions assumed elongated domal shapes due to E-W References compression. Most of the greenstone belt was metamor- Almeida, F.F.M., 1977. 0 craton do S~o Francisco. Rev. phosed to greenschist facies, except at the mar- Bras. Geoc., 7: 349-364. ginal contacts with the granite-gneiss domes Brito Neves, B.B., Cordani, U.G. and Torquato, J.R.F., where amphibolite facies metamorphism and 1980. Evoluq~o geocronolSgica do Estado da Bahia. In: partial melting occurred. H.A.V. Inda and F.B. Duarte (Editors), Geologia e Re- cursos Minerais do Estado da Bahia, Textos Bdsicos. Gold deposits are important in the green- Sec. Minas e Energia, Salvador, 3: 1-10. stone, with > 50 t of proven Au reserves and Ga~l, G. and Teixeira, J.B.G., 1988. Brazilian gold: metal- > 100 t of probable Au reserves. Annual pro- logeny and structures. Krystalinikum, 19 (in press). duction from opencast mines currently stands Ga~l, G., Teixeira, J.B.G., D'el Rey Silva, L.J.H. and Silva, at 2.0 t, with underground production sched- M.G., 1987. New U-Pb data from granitoids, reflecting Early Proterozoic crustal evolution in northeast Brazil. uled to start in 1988 which will boost produc- International Symposium on Granites and Associated tion to 6.0 t of Au per annum. The main deposits Mineralisations, Salvador, Bahia, Brazil (abstract). 17

Gibbs, A.K. and Barron, C., 1983. The Guiana Shield Re- ites and Associated Mineralisations, Salvador, Bahia, viewed. Episodes, 2: 7-14. Brazil, pp. 133-135. (extended abstract). Groves, D.I., Phillips, G.N., Ho, S.E., Henderson, C.A., Matos, F.M.V. and Davison, I., 1987b. Basement or intru- Clark, M.E. and Woad, G.M., 1984. Controls on distri- sion? The Ambrosio Dome, Rio Itapicuru greenstone bution of Archaean hydrothermal gold deposits in belt, Bahia, Brazil. Rev. Bras. Geoc, 17: 442-449. Western Australia. In: R.P. Foster (Editor), Gold '82: Ramsay, J.G. and Huber, M., 1983. The Techniques of Geology, Geochemistry and Genesis of Gold Deposits. Modern Structural Geology, Volume 1: Strain Analysis. A.A. Balkema, Rotterdam, pp. 689-712. Academic Press, London, 307 pp. Jardim de S~, E.F., 1982. Nota sobre o estilo estrutural e Schobbenhaus, C. (General Coordinator), 1984. Geologia relaqSes gnaisses vs. supracrustais no greenstone belt de do Brasil. Departamento Nacional de Produ~o Min- Serrinha (Ba). Ci~nc. Terra, 2: 8-13. eral, Brasilia, Brazil, 501 pp. Kishida, A., 1979. Caracterizaq~o geoldgica e geoqufmica Silva, M.G., 1984. A seqfi~ncia vulcanossedimentar do m~- das seqfi~ncias vulcanossedimentares do m~dio Rio dio Rio Itapicuru, Bahia: caracterizaq~o petrogrdfica, Itapicuru, Bahia. M.Sc. Thesis, Univ. Federal da Bahia, considera95es petrogen~ticas preliminares e zoneogra- Salvador, 98 pp. (unpublished). fia metamdrfica. In: P.V.S.V. S~ and F.B. Duarte (Ed- Kishida, A. and Riccio, L., 1980. Chemostratigraphy of lava itors), Geologia e Recursos Minerais do Estado da Bahia, sequences from the Rio Itapicuru greenstone belt, Bahia, Textos Bdsicos. Sec. Minas e Energia, Salvador, 5: 6- Brazil. Precambrian Res., 11: 161-178. 42. Lisle, R., 1977. Estimation of tectonic strain ratio from the Teixeira, J.B.G., 1985. Geologia e controles da mineraliza- mean shape of deformed elliptical markers. Geol. ~o aurffera em Fazenda Brasileiro, Serrinha (Ba). In: Mijnbouw, 56: 140-144. P.V.S.V. Sd and F.B. Duarte (Editors), Geologia e Re- Marimon, M.P.C., Kishida, A. and Teixeira, J.B.G., 1986. cursos Minerais do Estado da Bahia, Textos B~sicos. Estudo da alteraq~o hidrotermal relacionada h miner- Sec. Minas e Energia, Salvador, 6: 9-49. aliza~o aurffera na mina Fazenda Brasileiro (Ba). Teixeira, J.B.G. and Kishida, A., 1980. Geologia das mi- Congr. Bras. Geol. 34th, Goi~nia, Go, Ann. Soc. Bras. neraliza~Ses aurfferas estratiformes da Faixa Weber, Geol., 4: 1556-1570. Araci, Ba. Congr. Bras. Geol. 31st, Camboritl, SC, Ann. Mascarenhas, J.F., 1973. A geologia do centro-leste do Es- Soc. Bras. Geol., 3: 1802-1811. tado da Bahia. Congr. Bras. Geol. 27th, Aracajfi, Se, Ann. Teixeira, J.B.G., Kishida, A. and Marimon, M.P.C., 1986. Soc. Bras. Geol., 2: 35-66. The Fazenda Brasileiro Gold Deposits, Bahia State, Matos, F.M.V. and Davison, I., 1987a. Basement or intru- Brazil. In: A.M. Chater (Editor), Gold '86: Interna- sion? The Ambrosio Dome, Rio Itapicuru greenstone tional Symposium on the Geology of Gold Deposits. belt, Bahia, Brazil. International Symposium on Gran- Toronto, Canada, pp. 151-154 {abstract).