Physiographic, Stratigraphic and Structural Development of the Quadrilatero Ferrifero Minas Gerais, Brazil

GEOLOGICAL SURVEY PROFESSIONAL PAPER 641-A

Prepared in cooperation with the Departamento Nacional da Producao Mineral of Brazil under the auspices of the Agency for International Development of the United States Department of State Physiographic, Stratigraphic and Structural Development of the Quadrilatero Ferrifero Minas Gerais, Brazil By JOHN VAN N. DORR 2d

REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL

GEOLOGICAL SURVEY PROFESSIONAL PAPER 641-A

Prepared in cooperation with the Departamento Nacional da Producao Mineral of Brazil under the auspices of the Agency for International Development of the United States Department of State

UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1969 UNITED STATES DEPARTMENT OF THE INTERIOR WALTER J. HIGKEL, Secretary

GEOLOGICAL SURVEY William T. Pecora, Director

For sale by the Superintendent of Documents, U.S. Government Printing Office Washington, D.C. 20402 CONTENTS

Page Stratigraphy Continued Abstract.. ______Al Minas Series Continued Page Introduction-______3 Caraga Group Continued Location and importance of area______3 Environment of deposition. A38 Development of project______.______4 Metamorphism. ______39 Acknowledgments. ______6 Itabira Group..______39 Transportation______7 Caue1 Itabirite.______40 Climate._____---_-_---_-__-___.______7 Gandarela Formation _____ 44 Topography and physiography.______7 Environment of deposition. 47 Drainage systems.______7 Metamorphism. ______48 Influence of rock types on topography______Piracicaba Group.______48 Granitic rocks______Cercadinho Formation. _ _ _ _ 49 Argillaceous rocks.______9 Fe~cho do Funil Formation. 52 Quartzose rocks______9 Tabooes Quartzite. ______54 Dolomite ______10 Barreiro Formation. ______54 Iron-formation ______10 Sabara Formation. ______55 Canga_ _-_-_-_--__----______.______10 Environment of deposition. 58 Hematite ______10 Metamorphism______59 Physiographic development-______10 Itacolomi Series.______59 Stratigraphy, ______14 Rio das Velhas Series______15 Type Itacolomi. ______59 Nova Lima Group.______16 Santo Ant6nio f acies. ______60 Contacts and thickness.__--___-__-______16 Relation between the type Itacolomi and the Lithology______16 Santo Ant&nio f acies______-__----______62 Schists and phyllites______17 Environment of deposition. ______62 Iron-formation, ______18 History and problems of regional stratigraphic Quartz-dolomite and quartz-ankerite correlation. ______63 rock______19 Unmetamorphosed sedimentary rocks______68 Quartzite ______19 Fonseca Formation. ______68 Conglomerate. ______20 Other Tertiary rocks. ______68 Graphitic and carbonaceous phyllite. 20 Mudstone______69 Environment of deposition and nature of Weathering products. ______71 original sediments.______20 Saprolite______71 Metamorphism. ______21 Splash rock______72 Maquine' Group ______21 Laterite______--____---_---_ 72 Palmital Formation.______22 73 Casa Forte Formation______23 Alluvium. ______75 Maquine" Group undivided...______24 Talus and colluvium. 76 Environment of deposition______24 Landslides. ______76 Metamorphism. ______25 Igneous and granitic rocks. 76 Interpretation of the Rio das Velhas Series_____ 25 Granitic rocks______76 Minas Series.______27 Ultramafic rocks___._. 78 Contact with the Rio das Velhas Series. ______27 Distribution, weathering, and metamorphism_ 28 Distribution. ______79 Tamandu& Group..______t______28 Age______79 Cambotas Quartzite..______30 Gabbroic rocks and diabase______79 Unnamed formation______31 Porphyry dikes.______80 Rocks tentatively correlated with the Pegmatite dikes______80 Tamandud, Group______32 Quartz veins.______81 Environment of deposition______33 Quartz associated with steatitic bodies___ 81 Caraga Group.______33 Cleavage and fibrous quartz______81 Moeda Formation.______34 Quartz as metamorphic segregations___ 81. Batatal Formation.,______37 Gold-bearing hydrothermal quartz veins- 82 HI rv CONTENTS Page Regional structure: Continued Regional structure-______A82 Post Itacolomi orogeny Continued Pre-Minas structure______83 Major folds Continued! Page Pre-Itacolomi faulting.______84 Rio das Velhas uplift.______A92 Post-Minas-pre-Itacolomi deformation______84 Vargem do Lima syncline. 93 Post-Itacolomi orogeny______84 Conceig ao-Caraga uplift- _____ 93 Major folds______85 Other uplifts.______94 Santa Rita syncline______-______85 Faults______-______95 Dom Bosco syncline______86 Thrust faults__-_-_------___. 95 Moeda syncline______87 Tear faults___-____--____. 97 Serra do Curral______88 Cross faults.__-_-__-_-____-- 98 Gandarela syncline-______90 Geologic evolution of the Quadrilatero Ferrifero. 99 Ouro Fino syncline_-_--__--_--_____---_- 91 Structures in the Itabira and Monlevade Ref erences. ______104 districts. ______---____--______-___ 91 Index.______--_-_---____--__-____. 107

ILLUSTRATIONS

[Plates are in separate volume] PLATE 1. Geologic map and sections of the Quadrilatero Ferrffero, Minas Gerais, Brazil. 2-10. Geologic maps: 2. Santa Luzia quadrangle, by B. P. Alves. 3. Serra da Piedade quadrangle, by B. P. Alves. 4. Caete" quadrangle, by B. P. Alves. 5. Gandarela quadrangle, by J. E. O'Rourke. 6. Rio de Pedras quadrangle, by J. E. O'Rourke. 7. Sao Bartolomeu quadrangle, by A. L. M. Barbosa. 8. Ant6nio Pereira quadrangle, by A. L. M. Barbosa. 9. Ouro Pr£to and Santa Rita do Ouro PrSto quadrangles, by A. L. M. Barbosa. 10. Mariana and Rio das Bandeiras quadrangles, by A. L. M. Barbosa. 11. Geologic sections of the Sao Bartolomeu, Ant6nio Pereira, Ouro Pre'to, Santa Rita do Ouro Pr£to, Mariana, and Rio das Bandeiras quadrangles. 12. Diagram showing development of stratigraphic column of the Quadrilatero Ferrffero, Brazil, 1915-65. Page FIGUBE 1. Index map of the Quadrilatero Ferrffero-______-____-__------_---_----_----_----_--- A3 2. Photograph showing post-Gondwana surface preserved on canga sheet, Serra da Rola Moga -____--____ 12 3. Photograph showing ancient surfaces on the Moeda Plateau____------13 4. Photograph showing panoramic view, Macacos quadrangle.______.__.___-______--____-_-_--_---_-__-- 14 5. Aerial photograph of Minas-Rio das Velhas Series unconformity.____-_-__--_- ______29 6. Photograph of typical bare outcrops and cliffs of the Cambotas Quartzite in the Serra do Caraga______30 7. Photograph of typical cliffy outcrops of the Cambotas Quartzite in the Serra das Cambotas______30 8. Map showing variations in thickness of the Moeda Formation-______35 9. Photograph of cobbles in basal conglomerate of the Moeda Formation.______-_____-___-_-_-___--___- 36 10. Photograph of puddingstone conglomerate in the Moeda Formation.______37 11. Photograph of the typical topographic expression of the Moeda Formation in the Serra das Gaivotas. Ibirite1 quadrangle..-______-__-____-___---_------_------_-- 37 12. Map showing regional variations in thickness of the Cau§ Itabirite______.______--____ 41 13-25. Photographs: 13. Sill in itabirite-____--_. __.______-______-.__-____----_----_----.---- 43 14. Cau6 Itabirite between Alegria and Conta Historia_-______-____--_-----__-- 44 15. Cercadinho Formation in the Gandarela syncline.______._ 50 16. Basal conglomerate of the Sabard, Formation-______-___---_--__----- 56 17. Boulder of granite in tilloid of the Sabar£ Formation._._-______--___---_-----_--_ 57 18. Festoon crossbedding in the Santo Antonio facies of the Itacolomi Series.____-____-___-_-_-_--_ 61 19. Basal conglomerate in "mudstone". ______.___--_-___- 69 20. Outcrop of "mudstone".______--__-__-----_-_ .______--____--__---_-----_-----_--- 70 21. Laterization profile, Mutuca bauxite mine.. ______-____--_----___-- 73 22. Rich normal canga fragments on a typical canga surface.-______75 23. Canga rica at the Aguas Claras deposit.-______-____--__-___-__-__--__-_------75 24. Paragneiss forming from schist-______-____-______-____-______-_-_------77 25. Granite intrusive into Piracicaba Group schist-- _-______--_------_--- 78 CONTENTS V

ILLUSTRATIONS Continued FIGURE 26. Sketch map showing alternative explanation of structure at junction of the Moeda syncline with the Serra Page do Curral ___-__--_---.---_------__------_---_---_----_---_---_---_---_---_--_--_ A89 27. Photograph showing thrust fault, Cau6 Itabirite over Moeda Formation. ______.____^______._ 96 28. Photograph showing tear fault in the Mutuca II iron-ore deposit.______98

TABLES Page TABLE 1. Erosion levels, Quadril&tero Ferrffero_____-_-______-_--___--__--_-_-_--_-----_----____-_-_-- All 2. Analyses of "mudstone"______--_-__-_-______-__-_--_-_.-----__-__--__-_-_---_--__-___-___--__-- 70

REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL

PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL

By JOHN VAN N. DORR

ABSTRACT eugeosynclinal suite now dominantly composed of chlorite, The Quadrilatero Ferrifero is an area of some 7,000 square sericite, quartz-chlorite, and quartz-sericite phyllite and schist. kilometers in central Minas Gerais, Brazil, centered about lat. Metavolcanic rocks are common. Other rock types include ftray- 20°15' S., long. 43°30' W. For 250 years the region has poured waeke, subgraywacke, quartzite, carbonate-facies iron-forma­ forth a variety of mineral riches, now totaling more than tion, metachert, carbonaceous and graphitic phyllite, and $2 billion, and future production will undoubtedly be even tilloid. The Nova Lima is the host for widespread gold mineral­ greater. The main products are iron ore, manganese ore, and ization and, to the south of this area, for important manganese gold. To assist this development, the Brazilian and American deposits in the Lafaiete district. The base of the Nova Lima Governments in 1946 jointly undertook the first detailed geologic Group has never been found, but the minimum stratigraphic study of the region; this report is a synthesis of the results of thickness must be on the order of 4,000 meters. work and mapping by 17 Brazilian and American geologists Overlying the Nova Lima Group with locally gradational, under this program; other reports discussing the economic locally unconformable contact lies the Maquine Group, a geology, the metamorphic geology, and the igneous geology of molasse suite subdivided in the type locality in the Rio das the region as a whole are being prepared. Velhas valley into two formations, the Palmital and the Casa Forte. The Palmital Formation, composed dominantly of The region is one of the highest in Brazil and forms the quartzose chloride and sericitic phyllite and schist, wedges watershed between two major drainage systems. Major streams out to the southeast; and the overlying Casa Forte Formation, are deeply incised, and total relief is about 1,400 meters. Inter­ composed dominantly of protoquartzite, grit, and conglome-ate, mittent erosion has produced a number of prominent erosion sur­ there overlies the Nova Lima Group with apparently conform­ faces, the oldest persistent surface possibly dating back to the able contact. In much of the eastern part of the region the middle Mesozoic. These erosion levels are of economic impor­ Maquine Group cannot be subdivided, but is composed of rocks tance, for they control the development of certain supergene ore similar to those in the type locality, with less conglomerate. deposits. The maximum thickness of the Maquine Group is perhaps 1,600 All rocks of the region are strongly weathered, most to depths meters. of 50 meters or more. With the exception of an unjointed, Overlying the Maquine Group with angular unconformity is unfoliated granite, the more complex the mineral composition, the Tamandua Group, here included with the Minas Series. The the more intense the weathering, other factors being equal. group crops out in a relatively small part of the region ard is Products of lateritic weathering and saprolite are widespread. known only in the east-central part. The lower formation, the The region is underlain by granitic gneisses, granites, and Cambotas, is an orthoquartzite perhaps 1,000 meters in nxaxi- similar coarsely crystalline rocks; mafic and ultramafie intru­ mum thickness, and it is overlain by an unnamed forme tion sive rocks of various ages and the alteration products thereof; composed of sericite and quartzose phyllite and dolomitic iron- and three series of metasedimentary rocks of Precambrian age formation with a maximum thickness of perhaps 250 meters. The with a minimum aggregate thickness of about 14,000 meters. group is believed to represent prismatic paralic sedimentation. A few small areas of Tertiary continental sedimentary rocks Conformably overlying the Tamandua Group and, in most are known. of the region, unconformably overlying the Rio das Velhas Series The metasedimentary rocks of the region are divided into is the Caraqa Group of the Minas Series, divided into two for­ the Rio das Velhas Series, first metamorphosed some 2,800 mations, the Moeda and the Batatal. The Moeda Formation million years ago, uiiconformably overlain by the Minas Series, varies greatly but systematically in thickness and in lithology. which in turn is unconformably overlain by the Itacolomi In the west-central part of the region the formation consists of Series. The Minas Series was metamorphosed by intrusive rocks as much as 1,000 meters of grit, protoquartzite and ortho- dated by radioisotope methods as being formed some 1,300 mil­ quartzite, and conglomerate, with lenticular phyllite bodies. In lion years ago. The age of the Itacolomi Series is indeterminate, the east-central and much of the northern, far western, eastern, but it was metamorphosed to the same degree as the Minas and southeastern parts of the region it is less than 100 meters Series and is thus believed to be also older than 1,300 million thick and consists of quartzite or sericitic quartzite. It is a years. transgressive blanket formation. The conformably overlying The Rio das Yelhas Series is composed of two groups, the Batatal Formation, commonly less than 100 meters thick, is Nova Lima and the Maquine. The Nova Lima is a flyschoid composed of phyllite; it also is a blanket formation. Al A2 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL

The Itabira Group, conformably overlying the Caracal Group, eastern part of the region, where the highest metamorphic contains the major economic mineral resources of the region, grade is found. which consist of various types of iron ore, manganese ore, Unmetamorpliosed sedimentary rocks, some dated as late bauxite, building stone, and water; the water is largely Tertiary by fossils, include fossiliferous lignite and claystone unexploited. Two iutergradational formations form the Itabira beds found in valleys and ancient sinkholes. A rock called "mud- Group: the Caue Itabirite, composed of metamorphosed oxide- stone," found principally on the higher erosion surfaces, is facies iron-formation, and the Gandarela Formation, composed thought by the writer to have been formed by alteration of vol­ of dolomitic and calcitic marble, carbonate-rich phyllite, and canic ash, possibly of Mesozoic age. thin units of itabirite. The Caue is intergradational with the The first recorded orogeny was that after Ric das Velhas Batatal Formation and averages perhaps 250 meters in thick­ time. The Rio das Yelhas strata were folded and tilted, but it is ness ; the Gandarela averages perhaps 300 meters in thickness. not now possible to evaluate closely the degree of structural Both are blanket formations thought to have been deposited in disturbance 011 a regional basis because of intense later defor­ a stable-shelf environment. mation. The unconformity 'between the Rio das Velhas and the The Piracicaba Group of the Minas Series overlies the Itabira Minas Series is locally as sharp as 90°. Granitic rocks were Group with erosional but not angular unconformity to the west, formed in part of the Rio das Velhas uplift. It is believed but and with a gradational contact in the east. The Piracicaba not yet proved that granitic rocks to the south and west of the Group has been divided into five formations. The lowermost, region here discussed were the source area of much of the the Cercadinho Formation, found throughout the region, con­ coarser sedimentary rocks of Minas Series and Itacolomi age. sists of ferruginous quartzite, grit, and phyllite, and has a The Precambrian nietasedimentary rocks of the region were maximum thickness of about 400 meters. The Fficho do Funil For­ folded into a peculiar pattern of bordering syncliues around a mation conformably and gradationally overlies the Cercadinho central uplifted block. The central block is known as the Rio and is composed of siliceous dolomite, dolomitic phyllite, and das Velhas uplift, a north-south structure some 100 kilometers quartzose phyllite. Its maximum thickness is about 410 meters, long and about 26 kilometers wide at its narrowest point. It and it wedges out to the northeast and east. Overlying the veers to the northeast at the north end. This is probably the F&cho do Funil Formation with gradational contact is the Ta- first structure to have formed in the post-Minas orogeny and booes Quartzite, an orthoquartzite with a maximum thickness probably started to form during the pre-Minas orogeny. Around of about 120 meters. The Tabooes is found only in the western it were formed the Santa Rita syncline to the east, north-south and northwest central part of the region and wedges out to the in elongation; the Dom Bosco syncline to the sorth, east-west south and east. It is conformably and gradationally overlain in elongation; and the Moeda syncline to the west, north-south by the Barreiro Formation, composed of phyllite and carbon­ in elongation. The north boundary of the region is the over­ aceous and graphitic phyllite, with a maximum thickness of turned structure trending N. 70° E. called the Serra do Ourral; about 120 meters. The Barreiro Formation either wedges out its conjugate limb to the north has been destroyei by younger to the west or has been eroded away in that area; it is in granite except at its east end. In the central ane1 east-central gradational contact with the overlying unit in the north-central parts of the region lie the Gandarela syncline and the Con- part of the region and is found through the central part and ceiQao-Caraga uplifts. Peripheral to the synclin?s lie major much of the eastern part of the region. All formations of the domal uplifts now occupied by granitic rock of several ages. Piracicaba Group through the Barreiro are stable-shelf blanket In the extreme northeastern part of the region are synclinal deposits. structures that preserve relatively small areas of metasedimen­ At the close of the Barreiro sedimentation an abrupt change tary rocks surrounded by granitic rocks. in the sedimentary environment occurred, for the uppermost Cross folds are found on all major folds, cro^wcutting the formation of the Piracicaba Group is a typical eugeosynclinal major structuers at angles from 30° to 45°. All major folds are assemblage composed of graywackes, chloritic and sericitic overturned toward the northwest and west in at least part of schists and phyllites, tilloid, and tuffaceous rocks. This assem­ their trace. Thus, the deforming stresses came from the south­ blage is named the Sabara Formation and has a maximum east and east. Directions of movement were strongly influenced known thickness of perhaps 3,500 meters. The formation is by crystalline buttresses to the west, to the south, and in the structurally conformable with the underlying Piracicaba Group southernmost part of the Rio das Velhas uplift. formations, and it is locally intergradational with the Barreiro Continuation of stresses from the southeast ttereafter pro­ Formation in much of the region; however, in the west and duced major thrust faults in the eastern and southern parts of south it seems to be on an erosion surface. The formation the region. In the southern part of the Quadrilatero Ferrifero, probably once extended over the whole region, but it has been movement was channelized by the central and southern but­ removed by erosion in many places. tresses, and the Dom Bosco syncline was fragmented into a Overlying the Minas Series with small angular unconformity series of imbricate thrust blocks along its entire length. In the but deep erosional unconformity lies the Itacolomi Series, com­ east and east-central parts of the region, the Fundao fault, a posed of grit, protoquartzite, quartzose sericitic phyllites, and master structure which has been mapped for a length of 88 conglomerate. It was very probably a nearshore sediment The kilometers, formed; and to the east of this fault imbrication total thickness may be about 2,000 meters; its top is always is common. Only one major thrust fault is known to the west limited by the present erosion surface. and north of the Fundao fault, although some minor thrusting The Precambrian sequence was metamorphosed to the green- is found. The Santa Rita syncline, on the east edge of the schist facies or higher. It is moot whether there was more than region, is severely sliced by many north-trending thrust faults one episode of regional metamorphism; the youngest and oldest with displacement toward the west. rocks do not differ appreciably in metamorphic grade. The The root area of the orogeii cannot lie far to the east of the youngest stratigraphic units are not certainly known either in Quadrilatero Ferrifero, as also attested by the systematic east­ the contact aureoles of intrusive granitic rocks or in the far ward rise in grade of metamorphism in the mapped area. PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A3

Granite formed during the major orogeny some syntectonie, 7,000 square kilometers, takes its name from the abun­ some posttectonic. The age of the latter was determined by dant occurrence of iron-formation and iron ore. rubidium-strontium methods on feldspar as 1,230 million years. Cross faulting occurred after the orogeny ended. Most of it The geology of this region was mapped by a team of was minor, with displacements measured in tens of meters, but geologists during the years 1946-62. The physiographic, some of its was major with displacements nieasxired in hun­ stratigraphic, and structural geology and the rock rela­ dreds 'Of meters, at a few places displacement was more than tions of the region as a whole are the subjects of this a thousand meters. Such faults are at angles of 30°-45° or about report, which is in part a compilation of the data re­ 90° with the trend of the major fold structures which they transect, regardless of fold orientation. Neither cross faults nor ported by these and other geologists, in part a presenta­ earlier thrust faults are significantly folded, although both may tion of data collected by the writer, and in part an be refracted in passing between rocks of different physical interpretation by the author of all these data. properties. Minor faults tend to swing into the planes of bedding or foliation of more plastic rocks. The Quadrilatero Ferrifero played an essential role in the development of both the Portuguese Empire of INTRODUCTION the 18th and early 19th centuries and in the evolution of the Brazilian nation. During the Colonial period, LOCATION AND IMPORTANCE OF AREA several hundred tons of gold produced from the rich Most of the Quadrilatero Ferrifero is included in the placer deposits of the region formed a large part of the rectangular area between lat 19°45' S. and 20°30' S. and financial foundation for rapid expansion of the Portu­ long 43°22'30" W. and 44°7'30" W. in central Minas guese Empire and the influence of that nation in world Gerais, Brazil; but outliers and continuations of the affairs. Major engineering works were constructed by rocks, structures, and ores continue to the east and west large gangs of slaves to bring water 10 km (kilometers) some tens of kilometers beyond those limits, as shown and more to the higher workings; streams and rivers in figure 1 and on plate 1. The region, an area of about were diverted to expose the rich gravel deposits. John

70° 62" 46° 38'

QUADRANGLE NAMES 1 Itabira 22 Gandarela 2 Santa Luzia 23 Concelcao do Rio Acima 3 Serra da Piedade 24 Catas Alias 4 Antonio dos Santos 25 Piedade do Paraopeba 5 Cocais 26 Lagoa Grande 6 SaoGonealo 27 Itabirito 7 Monlevade 28 RiodePedras 8 Belo Horizonte 29 Capanema 9 Nova Lima 30 Santa Rita Durlo lOCaete 31 Martnho da Serra 11 Congo Soco 32 Bacao 12 Santa Barbara 23 Cachoeira do Campo 13 Floralia 34 Sao Bartolomeu 14 Rio Piracicaba 25 Antonio Pereira 15 Itauna 36 Casa de Pedra 16 Serra Azul 37 SaoJuliao Itatiaiucu 38 DomBoeco 17 Igarape Sousas 39 OuroPreta 18 Fecho do Funil Santa Rita do Ouro Preto Brumadinbo 40 Mariana 19 Ibirite Rio dasBandeiras 20 Macacos 41 Jeceaba 21 RioAtima 42 Congonhal

10 10 20 KILOMETERS

FIGURE 1. Index map of the Quadrilatero Ferrifero, showing location and names of the quadrangles mapped. A4 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL

Mawe (1816) described and illustrated the methods of subsistence agriculture. Most have great charm but few extraction, which were highly 'advanced for the time. facilities for the passerby. Early in the 19th century the placer deposits were ex­ Within the region certain cities stand out for their hausted and lode mining commenced, largely with Euro­ specialized activities. Mariana (pi. 1), with many beauti­ pean capital. By midcentury, most of the lode mines had ful churches, is an ecclesiastical center and l^as a small reached the limits of economic exploitation, and a period textile factory. Ouro Preto, the first capital cf the State of decay set in which lasted nearly a century. In the and the focus of former gold mining activity, is a tourist middle of the 20th century iron mining began to ex­ center and the site of the National School of Mines. pand; at the present time both quantity and value of Nearby is Saramenha, the site of a large aluminum re­ production are rapidly increasing and will soon far sur­ duction plant and ferroalloy plant. Congonh? s is a tour­ pass the value of the gold mined in earlier years. On this ist center because of a famous church with exceptional solid base, the industrialization of the region and of the baroque statuary and is also a busy mining center. nation is now progressing rapidly, aided by significant Itabirito is a textile and shoe factory city, specializing in production of aluminum from local bauxite deposits and football shoes; the nearby Esperanga plant makes very hydroelectric power from the rivers of the region and high quality cast iron. Rio Acima manufactures ceramic nearby regions. products, as does Caete, which also has a nearby iron During the early 20th century the Catalan forges of plant specializing in centrifically cast iron pipe. Nova the 19th century were displaced by small (15-50 tons Lima is the site of the famous Morro Velho gold mine, per day) charcoal blast furnaces, which were in turn which has produced for 150 years and in 1931 was the outstripped by larger and more modern integrated steel deepest mine in the world. Barao de Cocais is the site of plants in, or dependent on, the region. A small but sig­ a small but active steel plant. Monlevade is the site of nificant production of manganese ore has come from the largest privately owned steel plant in Brazil and the region, for export during the first half of the 20th the largest steel plant based on charcoal in the world, century 'and largely for domestic use thereafter. with a production of about 500,000 tons per year; it is a model modern city. To the north lies Itabira, which in The long history of mining and the isolation first 15 years was changed from a sleepy 19th-century town imposed by law by the Portuguese Empire and in later to a bustling modern city by the successful large-scale years by the rugged topography have set their marks mines of the Cia. Vale do Rio Doce, which produced on the population of the region; the people are inde­ about 8 million tons of iron ore in 1965. pendent, freedom loving, and individualistic, although The metropolitan center of the region is Belo Hori- conscious and proud of the traditions of the past. This zonte, which grew from about 200,000 to nearly 1,000,000 area cradled the first stirrings of revolt against the inhabitants in the period from 1945 to 19^5. All the Crown and toward the creation of a free Brazil. amenities of life are there available, and the city is the Within the Quadrilatero Ferrifero proper the soils political, commercial, and industrial center of the State. are poor, and farming is generally on a subsistence Industrial products range from steel, pipe, structural basis. Deforestation for the charcoal blast furnaces 'and forms, refractory brick, cement, textiles, tractors, and the railroads, which used wood for fuel until the decade radio tubes to prefabricated kitchens and other of 1950, has destroyed the fine virgin forests which are commodities. so vividly described in the accounts of the early explor­ ers and travelers; reforestation by eucalyptus has been DEVELOPMENT OF PROJECT accomplished on a large scale in the last decade, how­ In 1944 Brazilian and American officials began dis­ ever. Much of the region is sparsely populated, and the cussions directed toward a joint study of the geology standard of living outside the industrialized towns is and evaluation of the mineral resources of this region. low by Brazilian standards. In 1945 an agreement was reached, and in 1946 the first Picturesque towns relict from the days of gold min­ American geologists and topographic engineers arrived ing, such as Catas Altas, Agua Q,uente, Santa Rita in Brazil under the auspices of the U.S. Department of Durao, and Antonio Pereira, attest former prosperity State to begin the project with their Brazilian colleagues by their large and beautiful baroque churches, some of of the Departamento Nacional da Produc/w Mineral. which are now being restored and preserved, others of Lack of adequate base maps in the region seriously ham­ which are in ruins. Some of these old towns are taking pered operations; the areas around Congonhas do on new life with the advent of new mining activity; Campo and Itabira were chosen to start the work because others are off the path of development, and their eco­ base maps of parts of these areas were available and nomy depends precariously on charcoal gathering and because both were sites of active mining enterprises for PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A5 which a geological background was needed. Thanks to another by the present writer on the economic geology the continuing efforts of officials of the Departamento of the region. Regional reports are necessary because the Nacional da Produgao Mineral, aerial photographs of reports on the individual areas within the region were part of the region at a scale of 1: 10,000 were made avail­ prepared at widely different times by geologists thor­ able, and later aerial photographs of the entire region oughly familiar only with parts of the area. As the at a scale of 1:25,000 were furnished by the Commissao knowledge of the region grew and evolved, interpreta­ Vale do Sao Francisco and the Departamento Nacional tions made early in the work, although seemingly well de Producao Mineral. These were used by Services Aero- based at the time, became obsolescent, just as many in­ fotogrametricos Cruzeiro do Sul, SA, a Brazilian map- terpretations made in this report will also become obso­ making company, to map the topography of the region lescent after more detailed work is done in the future. at a compilation scale of 1: 10,000 for publication at Furthermore, much three-dimensional exploratory work 1:25,000, with a 10-meter contour interval. The cost of has been done on ore deposits in the past few years, and these maps was defrayed in part by the Brazilian Gov­ many area reports were written without the benefit of ernment, in part by the American Government. Topo­ information gained thereby. The detailed area reports graphic maps of small but economically important areas were written from many points of view, and diversity of were made at a scale of 1:2,000 by Geofoto, SA, from interpretation was encouraged within broad limits in aerial photographs. order to have as wide a range of ideas considered as With adequate base maps in hand, work could be possible. In the regional reports prepared by this writer, accelerated, and in 1951 the mapping team was enlarged. there is no attempt to give the details of much of the The region was divided into areas with some geological geology and ore deposits; the individual area reports homogeneity, composed of two to four quadrangles each. must be consulted for them. Those areas were mapped by the individual geologists The quadrangle geologic maps accompany area re­ as specified on plate 1. Each geologist was reponsible ports, the first of which was U.S. Geological Survey for mapping the assigned area, for evaluating the min­ Professional Paper 290, by P. W. Guild (1957). The eral resources therein, and for preparing an area other area reports have appeared or will appear as report. The areas mapped before 1951 had proved to be chapters in U.S. Geological Survey Professional Pgper among the more complex in the region, and it was only 341. It had been planned that the chapters covering all during this second phase that the details of the Pre- the individual areas would be in final form before the cambrian stratigraphy were worked out. general reports were written, but this proved impossi­ The igneous and metamorphic rocks of the region were ble, and this paper was prepared before final, or, for of a complexity that demanded special attention, and two reports, first-draft manuscripts had been preps.red therefore a petrologist was assigned to the party in 1957 on certain specific areas. Quadrangle maps accompany­ to study those problems. Although not given specific ing manuscripts which may not be published in the fore­ mapping responsibility, he reconnoitered 1% quad­ seeable future are included in the pocket of this report as rangles largely composed of granitic rocks for the re­ plates 2-10, in order to have them available to the public gional map, in addition to his petrographic responsibili­ in a reasonable time; geologic sections indicated on ties. plates 7-10 are given on plate 11. By the end of 1961, essentially all field mapping was References to unpublished manuscripts are cited as finished, and laboratory work and the preparation of "written communications"; the use of different dates reports has occupied the staff since then. More than 100 of authors'1 contributions may refer to different dn.fts man-years of geologists' time were spent on this project of the given manuscripts or to separate unpublished between 1946 and 1965. manuscripts. A few final reports omitted data or de­ scriptions that appeared in earlier drafts material con­ Reports discussing individual areas are being pub­ sidered by this writer to be important. In such situations lished by the U.S. Geological Survey as professional the draft is referred to instead of the published report. papers. In addition to these larger publications, about In this report, specific localities are referred to a co­ 45 other papers appeared in professional and technical ordinate system based on measurements in meters from journals. the southwest corner of individual quadrangles. Thus This report is an attempt to describe the physio­ "N. 1,000, E. 6,500, Gandarela" refers to a point 1,000 graphic, stratigraphic, and structural geology of the m (meters) north and 6,500 m east of the southwest cor­ region. Companion regional reports were prepared by ner of the Gandarela quadrangle. This point can be Norman Herz on the metamorphic geology of the region found on plate 1, which is at a scale of 1:150,000, by and on the igneous and granitic rocks of the region, and using the index map showing the location of the Gan- A6 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL darela quadrangle and the bar scale below. More ac­ group and with other geologists working in the area curate locations will be obtained by using the individual was stimulating and led to better understanding of all quadrangle maps published at a scale of 1:25,000. facets of the many problems. Wallace, Pomerene, Gair, Herz, Simmons, and Maxwell assisted in giving field ACKNOWLEDGMENTS experience to geological students of several Brazilian Admiration is here expressed for those officials of the universities. Barbosa and Alves worked only part time Department of State and of the Administration for on the project, devoting much of their effort to teaching International Development and its predecessor organi- in nearby universities during their work for this proj­ zations who, although not scientists, had the understand­ ect. The writer was in charge of the project through­ ing and appreciation of the need and potential results out (1946-64). of scientific and economic inquiry to finance an under­ In addition to the above-mentioned geologists, other taking as large as the present one. Vision was required professionals have contributed greatly to the project. to continue with an infrastructure project that could M. F. Denault, Edward Earner, and John Collins, give tangible results only after some years. Special ap­ topographic engineers, were indispensable in mapping preciation is due Mr. William Warne, Dr. Howard topography by planetable during the difficult initial Cottam, Mr. Eobert Groves, Mr. Leonard Saccio, all the years of the project, and Denault laid out fight lines, AID predecessors, and to Ambassador Lincoln Gordon. obtained the aerial photographs, and initially planned Others too numerous to name were also instrumental the topographic-mapping program. in vitalizing this project. Most geologic and topographic maps prnted with A succession of Directors-General of the DNPM de­ these reports were prepared in Brazil through the stage voted much time and effort to making this program a of color proof. Francis X. Lopez, cartographer of the success, particularly Dr. Mario da Silva Pinto, Dr. U.S. Geological Survey, introduced into Prazil new Avelino Ignacio de Oliviera, and the late Dr. Antonio techniques of preparing complex multicolor maps for Jose Alves de Souza. The work was administered within publication, assisted by Celio Macedo, now Had of the the DNPM by the Divisao do Fomento de Producao Seccao de Cartografia of the DNPM. Geologists of this Mineral; Directors who contributed most to the program project are deeply grateful to Messrs. Lopez and Macedo were Drs. Alberto Ildefonso Ericksen and Irnack for their assistance. The material was prepared by Amaral de Carvalho, later Director-General of the Brazilian companies under contract with the AID, un­ DNPM. Many other officials of the DNPM also contrib­ der the quality control of Lopez and Macedo. The late uted greatly to the work. Eobert Kojima was also of notable assistance in pro­ Among the Brazilian geologists who contributed graming and preparing illustrations for printing. freely of their knowledge were Glycon de Paiva, Othon Without the full cooperation of the many mining Henry Leonardos, Octavio Barbosa, Francisco Jose companies active in the Quadrilatero Fernfero, the Pinto de Souza, Iphygenio Scares Coelho, James Biichi, work would have been much more difficult. In many of Luciano Jacques de Moraes, Elisario Tavora, Claudio the areas studied, no living facilities exist ercept in the Dut^ra, Manoel Teixeira da Costa, and Djalma guesthouses of the active companies; hospitality was Guimaraes. This list is far from complete. freely extended. In remote areas, mules were provided In charge of the Divisao do Centre of the DFPM for transportation, and sometimes woodsmer who knew during the fieldwork was Dr. Jose Alves, who was of the trails and problems of the country were assigned to unfailing assistance in innumerable ways to all the work with project geologists. The Companl ia Vale do geologists taking part in this project. Eio Doce had a number of collapsed adits opened and The geologists who took part in the project are: B. P. pits dug to expose hidden geology; the St. John del Eey Alves (1947-64), Burton Ashley (1953-54), A. L. M. Co. shut down its hydroelectric system for a day so that Barbosa (1947-64), Aristides Noguiera da Cunha a critical water tunnel could be mapped. With one ex­ (1947-51), J. E. Gair (1954-56), P. W. Guild (1946- 51), Norman Herz (1956-61), E. F. Johnson (1954- ception, all companies freely provided all needed infor­ 56), C. H. Maxwell (1956-62), S. L. Moore (1955-57), mation on reserves, exploration results and analyses, J. E. O'Kourke (1952-54), J. B. Pomerene (1950-58). production, and other necessary data. Without this in­ E. G. Eeeves (1958-59), G. A. Eynearson (1951-57), formation the work would have taken much longer and G. C. Simmons (1957-62), E. M. Wallace (1956-61), would have been much less complete. It is hoped that and Amos M. White (1956-58). Their contributions are the results of the investigations will in some measure not to be measured alone by their reports and maps. repay the friendly open-handed cooperation, typical of The interchange of ideas and observations within the the country, which was received. PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A7

TRANSPORTATION panied by spectacular displays of lightning and very An asphalt-topped highway, BR-3, extends south­ occasional destructive windstorms, particularly in ward from Belo Horizonte to Rio de Janeiro, passing September and October. down the east flank of the Serra da Moeda for most of As can be imagined, the wide variation in precip^ta- its length and thence southeast out of the region (pi. 1). tion, both in the annual cycle and from year to year, Near Lagoa Grande the Estrada das Inconfidentes, a causes great fluctuations in the water table. This causes paved highway, branches east to Ouro Preto, passing considerable hardship for those dependent on surface through Itabirito and Cachoeira do Campo. A paved water and for hydroelectric plants. highway to Sao Paulo runs north of most of the area, Details of the local fluctuation in climate and in pre­ crossing the Serra do Curral in the Itatiaiugu quad­ cipitation and temperature in the area were given in rangle. Another leads east from Belo Horizonte, for reports by Gair (1962), Dorr and Barbosa (1963), and the most part north of the mapped area, to Sao Gon- C. H. Maxwell (written commun., 1964). galo de Eio Abaixo and Monlevade. This will eventually TOPOGRAPHY AND PHYSIOGRAPHY reach Vitoria. A branch leads northeast to Itabira. Other roads in the area were unpaved in 1962. They From a physiographic viewpoint the Quadrilatero are for the most part transitable by jeeps for the entire Ferrifero is one of the more scenic parts of Brazil. It year and by ordinary vehicles for most of the year. is marked by high and rugged peaks, long chains of Small back roads should be attempted only in special­ hogback mountains, high plateaus, some deep canyons, ized vehicles; local information should be obtained on and pleasant open valleys. It is an area of great vistas, their transitability, which changes from year to year. and, on a clear day, from the tops of the ridges and The railroad network is shown on plate 1. All rail­ mountains one can see at least 100 km in many roads except the main line from Rio de Janeiro to Belo directions. Horizonte, passing through Congonhas, Jeceaba, and The region is astride the watershed between two ma­ Fecho do Funil, are meter gage. The main line is broad jor river systems, the Sao Francisco, which flows about gage (1.62 m). Most passenger transportation within 1,500 km north to within 9° of the equator 'before t~irn- the region is by bus. The transportation network is dis­ ing southeast to reach the sea, and the Rio Doce system, cussed in greater detail in the forthcoming chapter on which flows 250 km northeastward and thence 200 km the economic geology of the region. Frequent airplane eastward to reach the sea. The maximum elevation in service to most of the major cities of Brazil is available the region is about 2,100 m, in the Serra do Caraga; the from Belo Horizonte. minimum is about 600 m, where the Rio Piracicaba leaves the region. CLIMATE DRAINAGE SYSTEMS The climate of the Quadrilatero Ferrif ero, which has had strong impact on the development of the landfonns The major river systems were superimposed on the and economic resources of the region, is semitropical, structure and thus are antecedent. These include the Rio and extremes of temperature are not known. Frost may Paraopeba in the west and the Rio das Velhas in the occur in isolated valleys, but snow and ice have not been central part of the region; both are major tributaries recorded. The maximum temperature is rarely over of the Sao Francisco, the first joining that river about 30 °C, and the mean annual temperature is slightly be­ 140 km north of the Quadrilatero Ferrifero, the other, low 20°C in much of the region. Mean annual rainfall about 300 km north. The Rio Piracicaba drains the varies with location from somewhat under 1,500 mm northeastern part of the region, and the Rios Gualaxo (millimeters) to somewhat over 2,000 mm, increasing and Carnio the southeastern part; both join the Rio to the southeast. Doce, the former about 75 km northeast by eas*. of Monlevade, the latter about 60 km east of Antonio The total rainfall varies greatly from one year to Pereira. another, and the maximum is well over twice the mini­ The smaller tributaries of these major rivers are gen­ mum recorded. The rainfall is very seasonal. At Itabira, erally well adjusted to the structure of the rocks, s^-ek- two-thirds of the animal precipitation occurs in 5 ing out the softer and less resistant formations, and, in months, and the months of June and July normally tne thicker and more uniform formations, follovdng pass without significant rains. In December, a week or faults, joint systems, and bedding. two of almost continual rain is common; this may also Many of the medium-sized tributaries of the major occur in January. Throughout most of the rest of the streams are well adjusted to rock structure in all but year rain normally falls in short hard showers, accom­ a few places. Stream piracy is still active; King (1956, A8 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL p. 197) remarked on the capture of the Rio Mata Pbrcos, In the Quadrilatero Ferrifero, significant quantities originally a tributary of the Paraopeba, by the Rio das of alluvium have been able to accumulate in only four Velhas system. Radical changes in drainage direction areas: one near Mariana, one just north of Serra do have also been noted in the Itabira district (Dorr, 1964, Caraga, one on the Piracicaba upstream from the town p. 1236) and observed in other localities, such as the of Rio Piracicaba, and one on the Rio das Velhas up­ northern part of the Moeda syncline. stream from Raposas. Despite the deep weathering, the The Rio Paraopeba and its major tributary the Rio gradient of major and minor streams is steep enough to Maranhao drain the Quadrilatero Ferrifero west of the remove all the detrital material brought to the streams. Serra da Moeda and the southwestern part to 4 km east A contributing factor is the fine particle size produced of Ouro Branco and as far north as the middle of the by the tropical weathering; relatively little coarse sand Dom Bosco syncline (pi. 1). The Maranhao cuts the is produced. North of the region, however, aggradation Serra da Moeda near Jeceaba, the only stream to do so is great; Burton (1869) reported that 100 years ago in the 60-km length of this range; the water gap is river boats of 40 tons burden came upstream from the narrow but not particularly steep. The Paraopeba in Sao Francisco River as far as Sabara; today a rowboat turn cuts the Serra do Curral at Fecho do Funil in a would have difficulty except in the flood season. Most narrow water gap, one of the two streams to breach reservoirs silt up rapidly; silt encroached on the Rio this range in about 100 km. de Pedras reservoir on the headwaters of the Rio das The Rio das Velhas rises in the Mariana anticline Velhas at least 1 kilometer during this worV. Those in (Antonio Pereira quadrangle) and flows over the south­ the Moeda syncline, however, where the lard has been east corner of the Bagao complex in an open valley be­ protected from overgrazing and destruction of the fore entering deep, steep, and narrow gorges in surface cover, did not silt up between 1946 and 1964. entrenched meanders in the constriction of the Rio das INFLUENCE OF THE ROCK TYPES ON TOPOGRAPHY Velhas uplift between Itabirito and Rio Acima. Be­ tween Rio Acima and the vicinity of Raposas the valley Ease of weathering and erosion of rocks and stability becomes broader, but the river again flows in deeply of their weathering products clearly are important con­ entrenched meanders to the north side of the Serra do trols in topographic evolution. In general, the more Curral, near General Carneiro, where the valley again complex the composition of the rocks and the finer opens. This river drains the Moeda syncline, much of grained the mineral constituents, the more easily the the Dom Bosco syncline, and all of the Rio das Velhas rocks are weathered in tropical and semitropical envi­ uplift as far east as the Serra das Oambotas. Deep and ronments. Another general control of weathering is the steep canyons in the Rio das Velhas Series rocks have degree of shearing and fracturing of the ro~,k. A third been excavated by the major tributaries. control is the rock isotropism. From the interplay of these major controls and other minor ones with larger East of the Serra das Cambotas and north of Bento rock structures and with earth movements and climatic Rodrigues, drainage is by the Rio Piracicaba, which changes, the topography and physiography of the re­ heads in the Conta Historia syncline (Capenema quad­ gion evolved. rangle) . The upper reaches of this river and the valley GRANITIC BOOKS near Monlevade are spectacular canyons, in the upper reaches largely following rock structure, in the lower The granitic rocks of the Quadrilatero Ferrifero are reaches crosscutting structure in entrenched meanders. in general the least resistant rocks of the region to The Gualaxo-Carmo system drains the southeastern weathering and thus to erosion. They form broad low­ part of the region and is characterized by deep narrow lands, low rolling hills, and relatively gentle slopes and valleys, in part controlled by structure, in part cross- are dissected by a dendritic drainage pattern. Weather­ cutting the structure. ing in granitic rock generally extends to a depth of 50m or more. Thus the master streams, most of which are en­ Most granitic rocks of the region break down to a trenched deeply in the harder rocks and less so in the clayey pink soil and saprolite. Locally the soil is highly softer rocks and all of which cross the ridge-forming aluminous and even somewhat bauxitic. Tv °- soil and rocks at some point in their courses, must have been out­ saprolite are subject to rapid and intense gullying lined at a time when the relief in the region was much where the protective surface of grass or forest is broken lower. The central and western parts of the region are by mule trails, overgrazing, improper cultivation, or part of the north-flowing drainage general through deforestation. The large area of badland topography in much of central Brazil; the eastern part is drained the southeastern part of the Bac.ao complex is but one across the geologic grain of the shield zone. of many examples of soil destruction. PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A9

111 many of the granitic areas resistant ribs and knobs and pocked by landslide scars. Because the phyllitic of unweathered granitic rock survive; these form bold saprolites are easy to excavate, they have been preferred smooth outcrops and waterfalls with very spectacular for highway alinements, but the poor mechanical quali­ pot holes in many localities, as along the Rio Arrudas ties of the fills and the unstable profiles of the cuts have between Belo Horizonte and Marzagania (Belo Hori- often made original savings illusory. Along the high­ zonte quadrangle). In some areas, such as the Caete way from Cachoeira do Campo to Mariana, some complex in the Antonio dos Santos quadrangle, only stretches have been abandoned entirely; others have not about 0.001 percent of the granitic terrain is fresh stabilized in 10 years. granitic outcrop; in others, as along the west flank of The argillaceous rocks in many places are slightly the Serra da Moeda in the Marinho da Serra quadran­ cemented or replaced at the surface by bauxite or gle, possibly as much as 25 percent of the granitic area limonite, depending 011 the composition of the rock. This is relatively fresh outcrop. cement serves to hold the surficial layer, from 1 to 10 cm Certain granitic rocks crop out boldly and are much (centimeters) in thickness, against mechanical erosion more resistant to erosion and weathering than the aver­ and permits the formation of steep and oversteep'3ned age. Such rocks as the Borrachudos Granite in the Ita­ slopes until this carapace is breached, after which bira district, the Floralia, and Santa Barbara quad­ erosion is rapid. rangles well exemplify the effect of lack of jointing, QUARTZOSE BOCKS coarse grain size, and isotropic structure on weathering. In the Itabira district the Borrachudos Granite stands The quartzose rocks of the region have several modes in bold relief above the foliated and jointed granitic of topographic expression, depending primarily on gneiss of somewhat similar composition. In the Monle- their composition and secondarily on stratigraphic vade area certain of the other coarser grained and more thickness and on the degree of shearing and fracture massive granitic rocks are also very resistant to erosion. to which the rocks were subjected. Graywacke inter­ The interface between fresh granitic rock and sapro- mediate in composition between argillaceous and lite varies radically; in some quarry faces a complete quartzose rocks weathers most rapidly and deeply, and transition can be seen in a few centimeters, but else­ fresh outcrops of such rocks are rare indeed. Subgray- where the transition is more gradual and occurs over a wacke and protoquartzite are more resistant to weather­ number of meters. Exfoliation is the rule on the ing and erosion but commonly do not crop out well interface. except where they occur as thick massive beds. Outcrops ARGILLACEOUS BOCKS of arkose are unknown. Outcrops of arkosic quar^zite and feldspathic quartzite are known in several forma­ The argillaceous rocks of the region are now largely tions, although the feldspar has been altered at the sur­ phyllite and schist. They are readily and deeply weath­ face in most places. ered. Saprolites are found to great depths; the writer Many quartzites contain notable quantities of mus- has been in adits more than 100 m below the surface in covite and sericite. Such rocks, found particularly in which weathered schistose rock can be smeared into the Maquine Group and in the Moeda Formation, are paste between the fingers. very resistant to erosion and in many places form bold In the Itabira district not one natural outcrop of ridges and outcrops, particularly where the rocks are fresh schistose rock was found. Few fresh outcrops massive. Where thin bedded or strongly sheared, they were found in other areas, and fresh argillaceous rocks weather and erode readily. With decreasing quantities are generally unknown except in deep gold mines and of white mica, the quartzites are increasingly resistant railroad tunnels deep in canyons. Because of their gen­ to mechanical and chemical erosion. The massive ortho- eral disaggregation, argillaceous rocks are nonresistant quartzites of the Cambotas and Moeda Formations are to mechanical erosion, and many of the deeper canyons among the most resistant rocks in the region, forming have been excavated in them, particularly in the Eio das great ridges and scarps. Weathering in such rock is by Velhas uplift. intergranular solution of quartz; where the rocks were Many of the fine-grained phyllites in the Piracicaba structurally crushed, very fine white sand, locally pure Group of the Minas Series and some of the phyllitic enough for use as glass or foundry sand, may accumu­ rocks of the Ita^olomi Series are particularly subject to late between outcrops. Sericite breaks down to clay, but landslides and soil creep where overgrazing, deforesta­ coarse muscovite is resistant to weathering. Soil derived tion, gold mining, highway construction, or other activi­ from such rock is, of course, sterile, and the rocks crop ties of man have disturbed the dynamic equilibrium out in small craggy patches and larger bare areas with between slope, climate, vegetation, and soil competence. a sparse and specialized flora where the general surface Many hillsides in such rocks are corrugated by creep is level. A10 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL One quartzite, the Tabooes, is exceptionally pure, con­ CANGA taining 98 percent silica. For some reason, this quartzite Canga is a weathering product of iron-formation and is commonly deeply weathered and disaggregated; per­ caps most of the major ridgelines in the region. It is haps the even and small grain size promotes inter- essentially iron hydroxide cementing larger or smaller granular leaching. quantities of detrital iron-formation, also generally DOLOMITE hydrated at the surface, and of hematite. This rock, inert to chemical weathering and very resirtant to me­ The carbonate rocks of the region vary greatly in chanical weathering, forms smooth slopes and plateaus. composition and therefore in reaction to erosion and Most of the higher old erosion surfaces are preserved weathering. Pure massive dolomite is in many places by canga, and the rock is now forming on loiter benches, fairly resistant to erosion and to solution and forms flat valleys, and pediments. scarps and a peculiar hummocky type of topography, as along the northwest side of the Rio Socorro (pi. 1) HEMATITE near its headwaters (Gandarela quadrangle). Soil Large masses of hard high-grade hematite are the formed on dolomite is clayey, and in most places dark most resistant rocks of the region to weathering and brown or black from small amounts of manganese and mechanical erosion. Certain ore deposits form spectacu­ iron commonly present; such soil is very fertile. One lar crags and scarped features visible for many kilo­ weathering product of dolomite was described at length meters. This rock may break down on weathering into by Guild (1957, p. 42) as a peculiar spongy material soft noncoherent hematite (Dorr, 1965, p. 42-43) in which the outlines of the original crystals are pre­ which crops out naturally in very few places. served by limonite and (or) manganese oxide. Phyllitic dolomite and dolomitic phyllite are among PHYSIOGRAPHIC DEVELOPMENT the least resistant rocks of the region to weathering. The physiographic development of the C'ladrilatero They form a deep saprolite and generally underlie gen­ Ferrifero is conditioned not only by the lithology and tle slopes. Quartzose dolomite is in many place less structure of its rocks but also by a complex history of resistant to chemical weathering than pure dolomite but epeirogenic uplift. Observers since Harder and Cham- may crop out prominently. berlin (1915, p. 370-378) have commented on the sev­ Many bauxite deposits in the region are closely re­ eral distinct levels of erosion preserved in the region. lated to argillaceous dolomite or dolomitic phyllite of Recently King (1956) attempted to relate these levels the Gandarela and Fecho do Funil Formations; they to others preserved in eastern Brazil, giving names to occur where removal of the weathering product has been four of them and assigning them ages on the basis of prevented by the physiographic environment of the fossiliferous rocks overlying erosion surfaces outside deposit or by the development of a hard aluminous and the Quadrilatero Ferrifero. ferruginous crust over the deposit. King's work consisted of a reconnaissance of eastern Sinkholes are known on the Gandarela Formation. Brazil from Bahia to the southern part of Sao Paulo All but one of the larger ones are in the Gandarela and much of central Brazil west of those limits. The syncline (pi. 1). One is on the west flank of the Serra do work was done in 2*4 months, during which time he Itabirito about west of the town of Itabirito, the Lagoa traveled 21,000 km by jeep, much of it over very ipoor das Cobras. roads, averaging 280 km per day, a feat of endurance IRON-FORMATION not likely soon to be equaled. Adequate topographic Four types of iron-formation crop out in the region: base maps in much of this part of Brazil were then non­ itabirite, dolomitic itabirite, amphibolitic itabirite, and existent; those then available in the Quadnlatero Fer­ carbonate-facies iron-formation. Except in canyons and rifero, at a contour interval of 100 m, were inaccurate, other small areas of accelerated mechanical erosion, only elevations being in error as much as 100 in. Under these normal itabirite, composed of hematite, very minor circumstances it could not be expected that the details magnetite, and quartz, crops out fresh. Fresh normal of correlation of erosion levels could everywhere be itabirite is also not a commonly occurring rock but is accurate or complete. known in canyons, in a few areas on ridgelines, in Despite these handicaps, King's work was a distinct monadnocks projecting above the ridgelines, and in contribution to the problem, representing as it did the landslide scars on oversteepened slopes. Some of the first attempt to relate the ancient surfaces of the Qua­ highest monadnocks, such as the Serras da Piedade and drilatero Ferrifero to those so well developed in other ItatiaiuQU, are composed of such rock, which there parts of Brazil. Elsewhere some of these ar°, fossil sur­ produces rugged and spectacular topographic forms. faces, being overlain by sedimentary rocks which have PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT All

been dated by their fuana and flora. Within the Quad­ region and the only one which, because it is a weather­ rilatero Ferrifero only one surface is overlain by ing product, regenerates itself (Dorr, 1964, p. 1234). suprajacent fossiliferous rocks; the age of the higher Because of this self-renewing feature, the rock surfaces and older surfaces is unknown. The staff of the present underlain by canga are probably the only ones which project had no opportunity to try to correlate the ero­ even approximate the original levels. Other surfaces sion levels within the area studied with those outside on quartzite and other rocks have been modified and their area of immediate interest, and thus the validity lowered by later erosion to such a degree that very little of King's interregional correlations cannot be con­ if any of the original surface remains, although the firmed. It is the writer's opinion, not shared by all mem­ topographic feature is preserved in more or less modi­ bers of the staff, that the ages assigned by King to the fied form. highest surfaces may be generally correct, although The highest erosion remnant is the relatively flat top certainly no proof of this now exists. of the Serra do Caraga, more than 2,000 m above the Table 1 lists eight major erosion levels observed in sea. Maxwell (written commun., 1964) pointed out that the Quadrilatero Ferrifero during this work. King this surface is locally overlain by deposits of well- specifically mentioned four of these surfaces and as­ rounded coarse gravel arid cobbles. Nowhere else in the signed definite ages to them; it was not possible to agree region is this level preserved by a relatively flat surface. with all his assignments, but the main scheme of King's More than 100 km to the north, the Pico de Itambe, 2,038 erosion levels and nomenclature has been followed. m above sea level, may also represent this surface. King Closer and specialized study undoubtedly will reveal and others have suggested that this high surface may new facets of the physiographic .problems not here con­ represent a Paleozoic erosional plane; 250 km to the sidered and will correct errors made herein. Such a west, Permian and Carboniferous sedimentary rocks study would be most useful from many viewpoints; it rest on a planed surface of Precambrian rocks (J. C. was outside the scope of the present project. Mendes, written commun., 1963). Other high peaks and surfaces such as those around TABLE 1. Erosion levels, Quadril&tero Ferrifero Pico de Itacolomi, Serra do Batatal, the Serra da Pie- [Modified from L. C. King (1956)] dade, and in the Serra do Caraga may indicate the pres­ Elevation Area of best development King's assigned King's name ence of another high erosion level between 1,700 and (meters) 1,850 m, but its age is unknown. Although King cor­ 2,000-2,100__ Serra do Caraca.-______.__.___------_ .-_-_---_-___.__ related both these high surfaces with his Gondwana 1,750-1,850-_ Serra do Caraca,-Itacolomi-...... _._...... _..__._.___ 1,500-1,650--.. ContaHistoria-Tutamea ----- Triassic_-._-__.- Gondwana. surface, the abrupt relief between such areas as the 1,250-1,400__ Serra de Rola Moca-Moeda Cretaceous.----- Post-Gondwana. syncline. Serra do Batatal and other nearby surfaces at more 1,100-1,200--.- Flanks of Serra do CurraL.--.------.-_--__ _._--_----.-__- 1,000-1,100-_ Lagoa Seca-Central Gandarela __-__.-..--..-. ....-.....--- than 1,800 m, the Serra do Ouro Fino at more than syncline. 850-950--.,- Rio das Velhas valley-Chapada Mid-Tertiary__ Sul-Americano. 1,600 m and the southwest end of the Gandarela syncline de Canga. 700-750 -.... Rio das Velhas valley. ------Late Tertiary... Velhas. at more than 1,600 m, all of which he also correlated with the Gondwana surface, makes such a correlation of the Five erosion cycles in eastern Brazil were named by higher surfaces questionable. King: the Gondwana, post-Gondwana, Sul-Americano, The highest surface generally present in the region Velhas, and Paragasu. Of these, only the first four are is that named the Gondwana surface by King. King represented in the Quadrilatero Ferrifero. The surfaces believed that the surface several hundred kilometers to destroyed by these erosion cycles, given the same names the west that he also called the Gondwana underlies the as the cycles, are naturally older than the time of their Botucatu Formation of Triassic age and the Upper destruction and were dated by the sediments deposited Jurassic plateau lavas; thus, it there must predate those on them. rocks. Within the Quadrilatero Ferrifero, no mode of King visualized that the destruction of the surfaces dating this surface is known, and continuity of the sur­ occurred by pediplanation and that the destruction of face to the areas of dating is not certain. If, as is con­ two or more surfaces may have been going on at the same ceivable, the mafic dikes of the Serra do Carac,a are of time; thus, since the oldest, no cycle has gone to comple­ the same age as the Upper Jurassic and Lower Creta­ tion throughout the area as a whole. There is reason to ceous plateau basalts of the Parana basin, the ages as­ suppose that still older surfaces existed at the time the signed by King to the older surfaces must be radically oldest named cycle began, a possibility admitted by revised and the surfaces will be much younger, for no King. remnant of the flows which should have been on these The best preserved old surfaces are all underlain by old surfaces now remains. If, on the other hand, the canga, the most evenly resistant widespread rock of the Caraga 'mafic dikes are much older than the basalts, 334-413 O 69 2 A12 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL King's age assignments may well be correct. As yet the necessary age determinations on the mafic dikes have not been made, and no outcrops of this rock fresh enough to permit accurate work are known. The Gondwana surface is preserved today as high areas of subdued relief or as ridgelines generally ac­ cordant with such areas. With the exception of the monadnocks mentioned, these are the highest parts of the region. King cited the ancient surfaces at the west end of the Gandarela syncline at elevations of about 1,500-1,650 m as belonging to this cycle. The Conta Historia area clearly belongs to the same cycle and lies at an elevation of about 1,550 m, as does the top of the Serra do Ouro Fino, at more than 1,600 m. To the west, on the western side of the Rio das Velhas, the highest peaks of the Serra da Moeda and the Serra do Itabirito FIGURB 2. Post-Gondwana surface preserved on canga sheet, Serra da Rola Moga. Note typical abrupt topographic uncon­ belong to this cycle. The ridgetop at Tutamea in the formity and monadnocks in the background. Lowland in the Serra da Moeda (Lagoa Grande quadrangle) is a broad background is on granitic gneiss; the ridgeline is on Caue area of rolling topography that is perhaps the largest Itabirite. Elevation about 1,350-1,370 m. View is to west- remnant of this surface. It lies between 1,500 and 1,550 southwest. m, and most of the sharp ridgeline of the Serra da nocks, such as Itatiaugu in the Igarape quadrangle. In Moeda is between those elevations, except at the south the west end of the Serra do Curral a widespread end. The Serra do Itabirito is generally somewhat remnant of a surface at about 1,300 m is preserved; lower, but the highest ridges are also above 1,500 m, this surface is somewhat higher to the east. and a few have flat tops as well. The Serra da Piedade Although King correlated the surfaces at about 1,100 has a prominent surface between 1,530 and 1,550 m. m in the Serra do Curral south of Belo Horizonte with According to King (1956, p. 200), the Goiidwana sur­ his post-Gondwana surface, the writer believes that this face lies at about 1,400 m in the Serra do Curral. This well-defined erosional level is better correlated with a writer doubts that the Gondwana surface is preserved younger surface not named by King. There is a very in the Serra do Curral south of Belo Horizoiite, except well developed higher surface, clearly younger than possibly in Pico de Belo Horizonte. At the west end the Gondwana surface, between 1,300 and 1,400 m in of that range in the Igarape quadrangle, two isolated much of the region south of Belo Horizonte which the flat-topped peaks above the general level of the ridge writer correlates with King's post-Gondwana level. are above 1,400 m; these might be relicts of the Gond­ Many broad, gently rolling surfaces, which the writer wana surface. refers to the post-Gondwana cycle, are found in the In the southern part of the region, many ridges lie Moeda syncline, dissected by younger deeply incised near 1,500 m in elevation in the Ouro Preto area, but streams (fig. 3). In the Marinho da Serra quadrangle, in the Serra do Itacolomi there is no apparent surface very well preserved canga-covered pediments are found at this level, although a well-defined level at about 1,450 on this surface (Wallace, 1965, plate 2); such canga- m and another at about 1,650 m are present. The Serra covered pediments were probably very common before do Ouro Branco, to the west, is capped by an ancient destruction by later erosion. In the eastern'part of the broad erosion surface between about 1,450 and 1,500 m Dom Bosco quadrangle, eight rounded hills culminate which may represent the Gondwana surface. between 1,400 and 1,460 m. According to King, the next younger surface, the On both the east and the west sides of the Rio das post-Gondwana, underlies lower and middle Cretaceous Velhas valley, broad dissected surfaces are developed sedimentary rocks outside this region; he said that the around 1,200-1,300 m, particularly on the east side, post-Gondwana cycle was incomplete and few large where the resistant quartzose rocks of the Maquine surfaces were produced. One of the best surfaces is Group have preserved the surface. (See pis. 5 and 6, the Serra da Rola Moga (elev 1, 350-1,450 m), part of this report, and Gair, 1962, pis. 1 and 2.) To the south, the Serra do Curral some 15 km southwest of Belo Hori­ much of the area in the west half of the Dom Bosco zonte (fig. 2). In the writer's interpretation, much of syncline may represent dissected remains of this sur­ the Serra do Curral west of Belo Horizonte is controlled face, also younger than the post-Gondwana. (See John­ by this surface, above which protrude some monad- son, 1962, pis 1 and 2.) PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A13

mation just northwest of Pico do Belo Horizonte, and many other local shoulders and breaks in slope at about this elevation both to the east and wTest of Belo Hori­ zonte, some of them canga covered, also represent this level. The common and locally large pediments both on the north and on the south sides of the Serra do Curral farther to the west (Pomerene, 1964, pi. 3) head at about this level. Many of the tops of rolling hills in the granitic and argillaceous areas of the Quadrilatero Ferrifero approximate this elevation, particularly in the Bacao complex. The south flank of the Gandarela syncline, an old flat surface pocked with sinkholes, ap­ FIGURE 3. Ancient surfaces on the Moeda Plateau, Lagoa proximates this elevation almost to Barao de Cocais. Grande quadrangle. Ridge in the far background is on Caue Johnson (1962, p. B30) pointed out that very well de­ Itabirite, elevation about 1,500 m, corresponding to the Gond- veloped terraces are to be found at elevations between wanda level. The rugged scarps in middle distance are on 1,100 and 1,200 m north of Morro do Caxambu (Dom the Cercadinho Formation, top about 1,350 m, corresponding Bosco quadrangle). Because of these features, the writer to the post-Gondwana surface. Other remnants of this surface are visible to the left. Dissected foreground surface is about believes two major surfaces developed between King's 1,250 m. View is to the northeast. post-Gondwana and Sul-Americano surfaces. No ele­ ments for dating them are known. To the east, correlation with King's older surfaces is The Sul-Americano erosion cycle is marked, according much more tenuous. The structure of the rocks is more to King, by broad surfaces and terraces at elevations complex, the country is more broken, and erosion more between 1,000 and 850 m. Such surfaces are clearly de­ complete. It is possible that the broad surface between fined along the Rio das Velhas, east of the Serra do 1,150 and 1,300 m on top of the Serra das Cambotas Caraga, and in the Paraopeba Valley. The largest rem­ might correspond to the post-Gondwana surface. South nant is probably the Chapada de Canga (Catas Altas of the Serra do Caraga the country is most broken and and Santa Rita Durao quadrangles), where an ancient the terrain is dominantly phyllitic. Conceivably the pediment carved on the Minas Series rocks at its west crest of the Alegria hogback in the Capanema and Santa end merges into an area covered by the Fonseca Forma­ Rita Durao quadrangles, about 1,375 m in elevation, tion (Maxwell, written commun., 1964), Tertiary fluvi- may correspond to the post-Gondwana surface,, but this atile and lacustrine sediments deposited on a planed is not certain. Another erosional remnant, Pico do granitic surface, at the east end. The Sul-Americano sur­ Frazao, is 1,350 m in elevation and has a large, relatively face is naturally below the sedimentary rocks; it might flat area above 1,200 m. In most of the southeastern represent an old wide river valley or might be a much and eastern Quadrilatero Ferrifero, there are no identi­ wider surface. About 10 km east of the east limit of the fiable remnants of the post-Gondwana surface. Santa Rita Durao quadrangle lies Fonseca, near which According to King, the post-Gondwana surface began town thin beds of lignitic material with plant fossils to be destroyed in the Late Cretaceous. The next major said to be late Tertiary in age underlie the canga cap of erosion surface cited by him was called the Sul-Ameri- the Chapada de Canga. This large canga surface very cano, supposedly formed during the middle Tertiary, as probably covered a much greater area in its original he believed that the surface is overlain by the Barreiras form; it is gradually being destroyed. The Piracicaba Formation of central and northern Brazil. This surface River has incised a canyon 100 m deep through the was stated to be wTell developed in Minas Gerais and in canga and Tertiary sediments into granitic rock. Be­ the Quadrilatero Ferrifero at an elevation of about 930 cause upper Tertiary sediments have accumulated on m near Itabirito and 850(?) m south of Belo Horizonte. the erosion surface 011 the granite, this surface is possibly However, in the Serra do Curral south of Belo Hori­ middle Tertiary, as stated by King. zonte a wTell-developed surface exists at about 1,100 m The youngest erosion cycle, the Velhas, is marked by below the post-Gondwana and above the Sul-Americano deep canyon cutting, and King referred the entrenched surface. The writer believes that King either missed an meanders of the Rio das Velhas and other rivers of the important surface in this area or confused two surfaces region to this cycle, supposed to have started in the late because of the lack of adequate maps. Lagoa Seca, about Tertiary. Some of these canyons are as much as 200 m 2 km south of the city on BR-3, is one remnant of this below the wide terraces marking the Sul-Americano intermediate surface. The flat on the Cercadinho For­ cycle. (See fig. 4.) A14 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL

Serra do Cipo Morro do Pires Serra das Cambotas (about 150 km distant, about 1300 m) (1375m) (1300 m, about 5p;^m distant) Pico de Belo Horizonte (1388 m) Serra da Piedade (1730 m) Moeda Quartzite Serra do Gurral Rio das Velhas watergap Q350m) (about 1300 m) (700 m) \ U , } (about 1300m)

FIGURE 4. Panoramic view from near N. 3,200 E. 6,550, Macacos quadrangle, looking from north to nearly east. Camera elevation is 1,400 m. Crags in the foreground are Moeda quartzite; dissected rocks of the Nova Lima Group are in the middle distance; ridgelines on the horizon are Minas Series rocks. Serra da Piedade is 40 km, and Pico de Belo Horizonte about 20 km, from camera. Rio das Arelhas, the master stream, flows in entrenched meanders behind Morro do Pires northward through the watergap in the Serra do Curral.

Long-continued weathering of the higher surfaces has (itabirites passing to pure iron ores), and limestones. All of had important economic consequences in the Quadrila- these rocks are greatly sheared, and characterized by a greater tero Ferrifero. If King's age assignments for these sur­ or less development of a micalike mineral (biotite, sericite, micaceous hematite, chlorite, talc, etc.) ; and, as in general they faces are correct, the Goiidwaiia and post-Gondwaiia are much decomposed, there is great difficulty in distinguishing surfaces have been exposed to weathering for more than the different members (except the quartzose and ferruginous 100 million years. As discussed in another paper (Dorr, ones), and thus far no successful attempt to work out the order 1964), these higher surfaces are the ones under which of succession in them has been made. It is tolerably certain that the major deposits of iron ore of supergene origin have the quartzose, ferruginous, and calcareous members are re­ peated at various horizons, and eventually they will serve as formed. Under the younger surfaces, some concentration reference lines for establishing the subdivisions of the series; of iron is known and locally thick beds of canga have but before this can be done the repetition due to folding and accumulated, but the major supergene iron ore deposits faulting must be determined and taken into account. Appar­ are below the older and higher surfaces. The accumula­ ently the whole series has been bent into closely appressed over­ tion of bauxite is less dependent on the age of the sur­ turned folds, and doubtless it has been much faulted. Until the region, or at least a typical portion of it, has been ac­ face, for such deposits are known on all except the curately mapped, any attempt to determine its detailed struc­ Velhas surface. Manganese deposits, concentrated to eco­ ture is hopeless. nomic grade by supergene action, are also less dependent on the age of the surface. The major manganese deposits Since Derby's work, continuing efforts have been all occur at higher elevations, and only one with com­ made by geologists to perfect the order of succession mercial production can be related to the Velhas cycle. and to clarify the structural and stratigraphic com­ None of the high-grade lode gold ores concentrated by plexities of the rocks and the region. The following supergene processes are related to low-level surfaces, pages are the latest, but certainly not the last, effort to although of course the lower grade primary sulfide gold bring order into the very complex relations seen and ores have no relation to present erosion surfaces. described by the great pioneer geologist. The present ef­ fort is, however, the first based on the careful regional STRATIGRAPHY mapping demanded by Derby more than half a century - ago. The Precambrian metasedimeiitary rocks of the As a result of this mapping, related as closely as pos­ Quadrilatero Ferrifero were first formally named by sible to earlier studies of many geologists, the Precam­ Derby (1906, p. 396-397) as the Minas Series: brian stratified rocks of the Quadrilatero Ferrifero The schistose series of the Serra do Espinhaco [this in­ have been divided into three series; one named forma­ cludes the Quadrilatetro Ferrifero J and adjacent regions, which tion of Tertiary age and several unnamed units of may conveniently be denominated the Minas Scries, consists of a great complex of predominantly argillaceous schists, with sub­ rocks of presumed Tertiary or Recent age have also ordinate masses of ordinary quartzites, ferruginous quartzites been recognized. Plate 12 gives the stratigraphic column PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A15 as used in this report and a summary of the lithologies rite crops out in many parts of the Brazilian Shield, of the formations, compares this column with several and in the past such rock has been correlated with the others worked out by earlier writers, and attempts a Minas Series of this area, it is now known that itabirite correlation between the various columns. Description of occurs in this region in each of two groups in two of the the rocks as here classified will precede discussion of series and in a rock of clastic origin generally similar in the development of nomenclature. lithology that is in the third series. All three series are In many Precambrian strata, metasedimentary rocks separated by profound unconformities indicating great that may be clearly divided into mappable stratigraphic time lapses. Clearly, 110 interregional correlation based units at one place may elsewhere be so metamorphosed solely on the presence of itabirite is well founded. that the original distinguishing characteristics cannot Similarly, although the orthoquartzite forming the be seen. All rocks may be gneissified in areas of greater Serra de Tiradentes at Sao Joao del Rei, some 100 km metamorphism, as in the Itabira district, or more sus­ to the southwest, is very similar in lithologic and sedi­ ceptible units may be so changed, as in the Monlevade mentary features to the main quartzite of the Tamandiia area, where certain lithologic units have been gneissi­ Group and the phyllitic rocks underlying it with angu­ fied and others remain identifiable. In this project, com­ lar unconformity to the Nova Lima Group, elements to pletely reconstituted gneissic rock in which all char­ make a secure correlation are not yet available, ^he acteristic features of original sedimentary units have writer believes, on the basis of discontinuous outcrop of been destroyed is considered new rock formed at the the characteristic manganese carbonate-silicate beds age of its metamorphism. In the Monlevade area, some associated with graphitic phyllite between the Lafaiete stratigraphic units can be correlated with confidence district and Sao Joao del Rei (Dorr and others, 1956, with named units in the rest of the region but are sepa­ p. 286; Ebert, 1963), that the phyllitic rocks there are rated by gneissified units whose, stratigraphic position Nova Lima Group; but until detailed mapping has corresponds to pelitic or dolomitic rocks in the normal proved this, it cannot be considered to have b0-en stratigraphic section. Names were assigned these established. gneisses (Reeves, 1966), and they are tentatively cor­ Correlation over wide areas by the aspect or mineral related with the less metamorphosed rocks known else­ composition of quartzites is equally perilous; detailed where in those intervals. Such gneissic rocks commonly tracing of beds in this region shows that facies, thick­ have distinctive features other than their stratigraphic ness, and mineral composition all change rapidly and position. that metamorphic and tectonic variation give the sp.me As in all Precambrian terrane, the lack of fossils re­ stratigraphic unit different aspects at different localities. quires that correlation be based on lithology. The vary­ The stratigraphic sequence presented in this paper is ing grades of metamorphism and degrees of structural the synthesis of the work of many geologists. Those v^ho disturbance combined with changing primary sedimen­ contributed most to the detailed subdivision of the Minas tary facies introduce uncertainty into the resolution of Series were G. A. Rynearson and J. B. Pomerene, v^ho several stratigraphic problems. investigated areas in which the stratigraphic sequence of In the Quadrilatero Ferrifero, certain distinctive se­ the Minas Series was not severely deformed, faulted, or quences of rock types are widespread and were used with highly metamorphosed. J. E. Gair and J. E. O'Rourke confidence in correlation. Sequence of rock types proved made the greatest contributions to the understanding of far more reliable than examination of individual rocks the Rio das Velhas Series, and G. C. Simmons and C. H. or specimens either in the field or, particularly, with Maxwell described and named the Tamandua Group. the microscope. A few rocks, such as itabirite and certain A. L. M. Barbosa, P. W. Guild, R. M. Wallace, and varieties of ferruginous quartzite, and specific sedimen­ C. H. Maxwell contributed most to a fuller understand­ tary features, such as characteristic pebbles, proved to ing of the Itacolomi Series. Naturally, the DNPM- be helpful guideposts in establishing positions in the USGS staff drew on the valuable stratigraphic informa­ stratigraphic section. On the other hand, the use of tion provided by predecessors in the region, particularly minerals which could have had either a sedimentary or Derby, Guimaraes, Harder and Chamberlin, Freyberg, a metamorphic origin, such as feldspar, proved unreli­ L. J. Moraes, F. Lacourt, and O. Barbosa. able as stratigraphic guides. For the reasons cited, correlation of the rocks within RIO DAS VELHAS SERIES the Quadrilatero Ferrifero with widespread areas of The Rio das Velhas Series was defined by Dorr, Gair, rocks of similar appearance and metamorphic grade out­ Pomerene, and Rynearson (1957) as comprising those side this region has not been attempted and is perilous schistose metasedimentary and metavolcanic rocks in except on the basis of detailed mapping. Although itabi­ the Quadrilatero Ferrifero older than the Minas Series. REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL

These rocks were then divided by those authors into two lowing Callaghan (private rept.), distinguished three groups: the Nova Lima Group, the older, and the stratigraphic zones in the Nova. Lima group in the Maquine Group, the younger. In 1961 Simmons and Raposas-Morro Velho area, two of which contain iron- Maxwell defined a still younger group which they in­ formation. He stated that it is impossible tc tell which cluded in the Rio das Velhas Series and named the zone is the oldest. Others who have mapped in the gen­ Tamaiidua Group, which did not crop out in the area eral region have expressed no opinion as to the divisi­ previously studied. The writer considers these rocks to bility of the rocks into formations, but most have be Minas Series, and they will be discussed under that mapped individual units, including iron-formations, heading. quartzite and conglomerate zones, graywr eke, gray- The Rio das Velhas Series, as shown on plate 1, is the wacke schist, and other rocks in the dominant chlorite- most widespread stratigraphic unit in the Quadrilatero sericite-quartz phyllite and schist. Ferrifero and occurs in generally anticlinal areas. These The writer believes that, when it proves economically rocks have locally been subjected to strong metamor- justifiable, the Nova Lima Group will be subdivided phism and metasomatism. It is possible that much of the by careful mapping into more than two formations and area shown on plate 1 as granite was derived through that much structural information can be gained in the ultrametamorphism or granitization from the Rio das process. Velhas Series, for locally the granitic gneiss grades into CONTACTS AND THICKNESS rocks of this series (Dorr and Barbosa, 1963, p. 41-42; The Nova Lima Group is apparently the oldest of the also fig. 24, this report). subdivisions of the Rio das Velhas Series. G:air showed The rocks of the Rio das Velhas Series can be traced that it is unconf ormably overlain by the Maquine Group out of the Quadrilatero Ferrifero to the south, to the (1962, p. 33) in the type localities of these formations west for an unknown distance, and to the east, where in the Rio das Velhas valley. To the east, the contact they soon give place to younger granitic rocks. They between these groups in the valley of the Rio Coiiceigao cannot be traced to the north, wThere Minas Series rocks (Capanema, Coiiceigao do Rio Acima, and Catas Altas are superposed and granitic rocks intervene. quadrangles) is indeterminate as to age relations be­ cause the contact is poorly exposed and may be faulted NOVA LIMA GROUP for some of its extent (Maxwell, written commun. 1962); Subdivision of the Nova Lima Group into named Moore, 1969). In the Antonio pereira and Santa Rita formations was not attempted during the mapping be­ Durao quadrangles the contact is also indeterminate. cause deep weathering, thick soil, and saprolite cover The simplest structural solutions in these p,reas would made such a separation impractical in the available suggest that the Nova Lima. Group is the older, as shown time. Rocks believed by the writer to be the equivalent by Maxwell in a series of block diagrams of the quad­ of the Nova Lima Group were subdivided into forma­ rangles mapped by him (written commun. 1962). tions in the area to the south and southwest of the The base of the Nova Lima Group has not been rec­ Quadrilatero Ferrifero by Ebert (1963). ognized anywhere in the mapped area. Th°/ rocks are As shown 011 plate 1, many of the more easily map- so strongly deformed and weathered that normal top pable lithologic units were traced in the Quadrilatero and bottom criteria are obliterated. Wher^ the Nova Ferrifero, giving a general idea of the structure of these Lima rocks are in contact with granitic rocks, the con­ rocks. A. F. Matheson (oral commun.) studied these tact is everywhere an intrusive one, as discussed below. rocks for the St. John del Rey Mining Co. and prepared Because the stratigraphic top and bottom of the Nova an unpublished map at a scale of 1: 60,000 of the region Lima Group are both unknown, the original thickness south of Belo Horizonte and north of Itabirito in which of the unit is also unknown. Structural complications he distinguished two major units in the Nova Lima make estimates of thickness hazardous, but, it is prob­ Group rocks, one domiiiantly quartzose, the other dom- able that not, less than 4,000 m of sedimentary rock is inantly phyllitic. The quartzose rocks, which crop out exposed in the Rio das Velhas valley. Exposures else­ in the axial area of an anticline, are the older. where are less complete, but an equal or greater thick­ Simmons (1968a, p. G9), as a result of mapping in the ness may be present in the Caete quadrangle and in the extreme western part of the Quadrilatero Ferrifero, dis­ Congonhas-Ouro Branco area. tinguished two stratigraphic facies in the Nova Lima Group by the presence or absence of iron-formation and LITHOLOGY ferruginous phyllite in those rocks. The ferruginous Because of inadequate outcrops, deep weathering, rocks there were believed to be younger than the non- complex structure, and, particularly, the iirture of the ferruginous phyllites. Tolbert (1964, p. 780-781), fol­ rocks themselves, the Nova Lima Group cannot be sys- PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A17

tematically described from a stratigrapliic viewpoint; (?) was also found; conglomerate was mapped in the only the lithology of the rocks studied can be discussed. Macacos quadrangle (Pomerene, 1964, p. D9) and re­ The rocks of the Nova Lima Group have been described ported in the Doin Bosco quadrangle (Johnson, 1962, p. in most detail by Gair (1962, p. 9-29), who was favored B7). Maxwell (written commun., 1964) and Simmons by many extensive underground workings and drill (1968a, p. G9) reported chlorite schist and phyllite, and holes in fresh rock in the vicinity of gold mines. His Simmons (1968a, p. G9) reported carbonaceous phyl­ work, carried on in the type locality of the group, is the lite. Wallace (1965) found a thin lens of garnetiferous most complete description of these rocks. Gair also limestone in a railroad tunnel. In zones of contact meta- studied the petrography of the Nova Lima Group as morphism, staurolite schist and garnet schist are co'ii- revealed in drill cores from the Itabira district (in Dorr mon. Thus, the Nova Lima Group contains the and Barbosa, 1963, p. 10-16). O. Barbosa (in Guima- metamorphic equivalent of almost every major type of raes, 1935) published petrographic descriptions of sedimentary rock except evaporites and organic fuels. many of the rock types in the Nova Lima area. (These Much of the widespread talc schist and talc phyllite in rocks were then included in the Minas Series because the the Nova Lima Group is probably metamorphosed and unconformity that is the present basis of the separation highly sheared ultrainafic intrusive rock, closely related of the Minas Series as now defined from the Rio das Vel- to soapstones and thus younger than the Nova Lima has rocks had not then been recognized.) Wallace (1965) Group.

was fortunate in having several railroad tunnels and a SCHISTS AND PHYLLITES few other fresh outcrops available for study in the Itabirito quadrangle; Guild (1957, p. 20-22) described The picture of extreme diversity given above by the rocks in the Congonhas area, now correlated with the recital of lithologies disappears when dominant litholo- Nova Lima, as the "greenschist sequence1' of the Minas gies are considered, for many of the rock types p-re Series from cores and a few fairly fresh outcrops. Be­ subordinate and known from only a few places. cause little fresh rock of this group has been found The dominant lithologies are chlorite phyllite and elsewhere in the region, detailed petrographic descrip­ schist, quartz-chlorite and quartz-chlorite-sericite phyl­ tions are elsewhere unavailable. Easily weathered rocks, lite and schist, and sericitic quartzite. Intergradational such as graywacke, are undoubtedly much more common ferruginous phyllite and iron-formation are character­ than reported. istic although volumetrically minor. All are fine grain ed Most rocks of the Nova Lima Group weather to buff, except in metamorphic aureoles. These rocks grade into red, pink, and purple saprolite, in many exposures re­ one another and undoubtedly represent a great mass of taining a strong banding which is clearly related to poorly sorted fine-grained argillaceous sediment, local foliation. Weathered chlorite is commonly the cause of accumulations of chemical sediments, and much inter­ the color; in certain metamorphic zones weathered bio- mixed material of volcanic origin. A. F. Matheson tite produces similar red and pink tones. In areas of (oral commun.) remarked that part of the Nova Lima deeply weathered rock in the medium and higher grades Group as exposed in the Gaia Tunnel (Rio Acima qurd- of metamorphism, it may be most difficult to distin­ rangle) had the aspect of the metamorphosed andesitic guish between biotite schist of the Nova Lima and bio- lavas of the Canadian shield. O. Barbosa (in Gni- tite gneiss of the granite gneisses. A useful but not maraes, 1935) and Guimaraes (1931, 1935, 1958) also infallible guide is the presence of small muscovite plates emphasized the volcanic contribution to much of this from small pegmatites, which are rare in the Nova Lima rock. rocks except at the contact with granitic rocks but quite Gair's description (1962, p. A9) of the dominrnt common in granitic rocks. schists and phyllites of the Nova Lima area is as In the type area, Gair distinguished the following follows: varieties of schists and phyllites: Q,uartz-sericite schists Schists and phyllites: Interbedded metasediinentary rnd and phyllites, quartz-chlorite and quartz-chlorite-seri- metavolcanic schists and phyllites form the bulk of the Nova cite 'schists and phyllites, carbonate-rich schists, Lima Group. Fresh rocks are mainly green or gray-green, but at the surface the schists and phyllites commonly are pinkish, quartz-biotite (or stilpnomelane) schists and phyllites, red, maroon, or buff owing to weathering. Secondary concentra­ quartzose and feldspathic schists, graphitic phyllites, tions of hydrated iron oxides at or near the surface have modi­ ferruginous schists, and metavolcanic schists. He also fied the rock to a dense brownish ferruginous schist in places. distinguished other rocks as follows: Carbonate-facies Natural exposures are generally small. Graphitic phyllite is iron-formation, quartzite, graywacke, quartz-dolomite widespread but not abundant. Except for graphitic phyllite, the weathered rocks generally cannot be related to specific un- and quartz-ankerite rock, sericitic. quartzite, and schist­ weathered schists and phyllites, making it difficult or impossible ose conglomerate. In the Itabira district, meta-arkose to map varieties of schist or phyllite. Although different type? of A18 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL schist and phyllite have been identified at individual locations, chloride schists and phyllites elsewhere in the region they have been mapped together as a single stratigraphic unit. strongly suggest that the volcanic fraction in the Nova Petrographic studies of more than 100 thin sections Lima Group is sizable and occurs everywhere, although of the schistose and phyllitic rocks of the Nova Lima rock fresh enough to permit petrographic confirmation Group by Gair (1962, p. 12-13) showed that quartz is was scarce. a major constituent in almost all types. Sericite and The widely varying composition of these rocks might muscovite are relatively common; biotite occurs in well indicate that the original volcanic contribution was some types; and fuchsite, the chromian mica, is found much mixed with sedimentary material before and dur­ in some. Chlorite is a major mineral in many rocks and ing deposition, locally being a rather pure tuffaceous subordinate or minor in others. Sodic plagioclase is not rock, locally flows, and elsewhere a siltstone mixed with rare but is a subordinate or minor mineral in most volcanic ash. rocks. Tourmaline in minor amounts was seen in a num­ Although the dominant rock types described above ber of thin sections. Epidote, graphite, magnetite, and account for the great bulk of the Nova L^ma Group, rutile are accessory minerals in many rock types. Car­ other rock types to be described below are important bonate appears in minor quantity in some rock types from a genetic and economic viewpoint. Furthermore, and as major mineral in the carbonate-rich schists. Sul- they occur in relatively narrow bands and thus are most fides, possibly introduced by hydrothermal fluids, pos­ useful in mapping the internal structure of the other­ sibly primary, were seen in some thin sections. wise nondescript rocks. In the Itabira district, similar rock types and also meta-arkose are found, all coarser grained because of IRON -FORMATION a higher grade of metamorphism; there, garnets are Thin discontinuous lenses of iron-formation charac­ abundant and andalusite has been found. Alkali meta­ terize the Nova Lima Group in this region. They are somatism clearly affected the rocks in that area (Dorr found from the Itabira district to the Sousas quadrangle and Barbosa, 1963, p. 11-12). In the Itabirito quad­ and from the Belo Horizonte to the Congonhas quad­ rangle, south of the area studied in detail by Gair, rangles and farther to the south in the Lafaiete area. Wallace found (1965) rocks similar to those reported Near Nova Lima, in part of the valley of tlx °, Rio Con- by Gair. ceigao, in the area north of Gandarela, and in the Rocks in the Gandarela quadrangle correlated with Sousas area, these lenses reach thicknesses of some tens the Nova Lima Group are so fine grained that the min­ of meters and can be traced continuously for several eral components cannot be identified in thin section. kilometers; many of these are shown on plate 1, and Herz (1962, p. C75) cited a chemical analysis of an more were shown on the individual quadrangle maps argillaceous rock collected by J. E. O'Rourke and cal­ by Gair (1962, pis. 1 and 2) and Pomerene (1964, pi. 2). culated (written commun. 1960) the probable mineral Elsewhere, as in the Itabira district, the Congonhas composition as follows: Quartz, 38 percent; sericite, 33 area, southwest and southeast of Caete, and east of Mari­ percent, chlorite, 28 percent; iron oxide, 1 percent, ana, they may range from a few centimeters or tens of Metavolcanic rocks described by Gair are largely centimeters to more than 10 m in thickness. Most thin chlorite schists, with some hornblende, epidote, clino- lenses are not shown on plate 1, and most were too thin zoisite, and subordinate sodic plagioclase and quartz, to show on the individual quadrangle maps. The num­ indicating an originally mafic rock. "Relicts of inter- ber of individual lenses in a ferruginous zone is difficult grown and crisscrossing plagioclase laths and possible to establish because of structural complexities, but Alves stretched quartzose amygdules in some of these rocks are (oral commun., 1954) found as many as seven in the the only direct evidence of volcanic origin found" (Gair Caete quadrangle; more may be present in the Nova 1962, p. 16). Dorr and Barbosa (1963, p. 13) described Lima and Macacos quadrangles, where they seem to be and figured rock from the Nova Lima Group of the most numerous. Itabira district believed to be a metatuff because of the All specimens of iron-formation from the Nova Lima embayed plagioclase crystals and grain-size differences Group obtained from below the surface zone of oxida­ between feldspars and the chlorite groundmass. Rocks tion proved to be either carbonate facies (James, 1954) near Congonhas described by Guild as part of his or carbonate-magnetite facies. Carbonate-facies iron- "greenschist sequence" (1957, p. 20-22), now correlated formation was first recognized in the reg; on by Gair with the Nova Lima Group, also are rich in chlorite, (oral commun., 1955; 1962, p. A16-A22) and Matheson epidote, zoisite, clinozoisite, and hornblende. They seem (1956). This discovery was so belated becp.use on out­ similar to the rocks described by Gair and very possibly crop it is most difficult to distinguish between oxidized represent metavolcanic rocks. The abundant chlorite and carbonate-facies iron-formation, characteristic of the PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A19

Nova Lima Group, and the better exposed oxide-facies most important gold mines with this host rock are the iron-formation, characteristic of the overlying Minas Raposas (Nova Lima quadrangle) (Gair, 1962; Tolb^rt, Series. 1964) and the Sao Bento mine near Barra Feliz (Santa Below the zone of oxidation most of the iron in car- Barbara quadrangle) (Simmons, 1968b). bonate-facies iron-formation (quartz plus siderite plus QUARTZ-DOLOMITE AND QUABTZ-ANKERITE ROCK magnetite) is in bivalent form. On the outcrop and in shallow openings, most of the iron in such rock has oxi­ In the Nova Lima and Rio Acima quadrangles Gair dized to the trivalent form, either as limonite or martite. described a peculiar rock as "a gray, massive qur.rtz- Fresh dolomitic oxide-facies iron-formation (quartz carbonate rock that resembles dolomite" associated with plus dolomite plus hematite plus varying quantities of 'schist of somewhat similar composition into which it magnetite) contains most of its iron in trivalent form grades (1962, p. A23). It consists of a mosaic irter- but on the outcrop appears very similar to oxidized growth of quartz and carbonate with minor plagioclase, carbonate-facies iron-formation, as the weathering chlorite, and graphite and occurs in discontinuous lenses. products of siderite and dolomite are similar and both The sedimentary origin of this rock type has not always rocks contain magnetite which alters to martite. Mobile been recognized, but both Gair (1962, p. A26) and limonite on the outcrop hides primary mineral identi­ Matheson (1956) showed clearly that it is an integral ties and relations in both f acies. part of the Nova Lima Group. In the Morro Velho mine Analyses and thin-section study of fresh rock from the rock is called "lapa seca." The premetamor^hic the Raposas and several other mines in the Nova Lima nature of this rock is not clearly apparent, nor is it and Rio Acima quadrangles and from the Sao Bento known why the iron carbonate should be ankerite rather mine, Santa Barbara quadrangle, show clearly that the than siderite or ferrodblomite; iron occurs both as magnetite and siderite, the two min­ This rock crops out poorly but has been mapped in the. erals varying widely in relative abundance (Gair, 1962, Rio Acima and Nova Lima quadrangles. It is the host p. A19; Simmons, 1968b, p. H9). Quartz content ranges rock for the gold ore in the Morro Velho, Bela Fr.ma, from 38 to 60 percent in the unweathered rock. Although and Bicalho mines and, judging from description1" of no other completely unweathered material is known, it mines now collapsed, may have been the host rocV in seems probable that most if not all of these Nova Lima other gold mines in the Caete area (Moraes and O. iron-formations were carbonate facies or carbonate- Barbosa, 1939) and at Bico de Pedra (Guimaraes, 1961, magnetite facies. There is no good evidence to the p.229). contrary. QUARTZITE Pomerene (1964, p. D7) listed useful criteria for dis­ Quartzite in the Nova Lima Group crops out widely, tinguishing the carbonate from the oxide facies on the but because it is commonly very impure, it weathers outcrops. Another criterion is that graphitic and carbo­ easily, breaks down to a saprolite, and is not ersily naceous phyllite is closely associated with carbonate- mapped as a lithologic unit. Where badly weathered, facies iron-formation in many localities, whereas these such rock has the aspect of quartz-sericite schist. This rocks are not known to be associated with dolomitic highly impure quartzose rock crops out near and south oxide-facies iron-formation, as predictable from the Eh of the highway between Sahara and Caete and is a environment of these facies (James, 1954; Castafio and prolongation of the quartzose zone mapped by Mathe­ Garrels,1950). son in the vicinity of Sabara. (See p. A16.) In places, bands of iron-formation grade, with an in­ Rather pure, locally resistant, massive quartzite was creasing content of aluminous material and decreasing mapped by Wallace in the Itabirito quadrangle. This quartz and iron, into ferruginous phyllite. Such rock green and gray slightly chloritic and sericitic quartzite is very common in the Sousas area (Simmons, 1968a) shows well-preserved crossbedding, locally highly intri­ and in the Santa Barbara (Simmons, 1968b), and cate, in roadcuts along the Rio Acima-Itabirito High­ Gongo Soco (Moore, 1969) quadrangles and is known in way. To the west this feature is obliterated by shearing, many other places. Gair described the carbonate-facies and the rock becomes schistose. Gair also mapped iron-formation as being closely related spatially with quartzite lenses in the Rio Acima and Nova Lima metavolcanic rocks (1962, p. 21 and 22). This is in strong quadrangles. contrast to the oxide-facies iron-formation found in the Many schistose rocks of the Nova Lima Group are Minas Series, which is not so associated. quartzose and grade into impure quartzites and Hck The carbonate-facies iron-formation of the Nova into quartz-mica schists both along and across the strike. Lima Group has been the host rock for widespread Systematic separation of such rocks must await more hypothermal gold mineralization in the region. The detailed mapping. A20 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL Metachert, now pure fine-grained white quartzite, bleached, and are normally very fine grained. In the grades laterally into iron-formation in many localities, lower levels of theiMorro Velho mine, which extends to notably in the Caete quadrangle. more than 2,500 m. below the surface, methr.ne has been reported, perhaps derived from the organic content of CONGLOMERATE such rocks. Simmons (written commun., 1962) reported Conglomerate is a very uncommon rock in the Nova the extraction of small quantities of orgi.nic matter Lima Group, but a large exposure of coarse conglom­ from carbonaceous phyllite in the Nova Lima Group. erate occurs in the Macacos quadrangle along the Kio de Janeiro-Belo Horizonte Highway (Pomerene, 1964, ENVIRONMENT OF DEPOSITION AND NATURE OF ORIGINAL SEDIMENTS pi. 2 and fig. 3). One contact of this rock is exposed in a roadcut, where there is an abrupt transition from The generally sparse outcrops of the Nova Lima weathered red phyllite to boulder conglomerate; here, Group and the pervading metamorphism and weather­ boulders of quartzite up to 50 cm in diameter are found ing cause great difficulty in interpreting the original in a phyllitic matrix. Away from the contact, which is nature and environment of deposition of these rocks. believed to be the stratigraphic base of the zone, the The Nova Lima Group consists of the following rock size of the detrital material grades through cobble size types: Preponderant fine-grained elastics, lesser vol- to pebble size. The zone can be traced for 3,000 m. with canics, some chemical precipitates, and rare coarse an outcrop width of 200-500 m. (Pomerene, 1964, DIG). elastics. The fine-grained clastic rocks were probably The coarse detrital material is well rounded and makes clay and silt, with admixtures of varying auantities of up 20-50 percent of the rock. volcanic ash. O'Kourke (written commun., 1956) was The coarse boulders are medium-grained quartzite the first to point out that these fine sediments were the and vein quartz; the smaller cobbles and pebbles higher result of mature weathering; he based his opinion on in the section consist of vein quartz, smoky quartz, the high alumina content of the rock. Herz (1962, p. quartzite, phyllite, schist, and highly weathered mate­ C75) recently reemphasized the point. rial that may have been granite or other feldspathic The volcanic rocks were probably volcanic ash, ap­ rock. Some schist and phyllite fragments are angular parently with some flows, as indicated by probable and probably fractured. amygdules seen in the Gaia tunnel. The source of the ash The smaller pebbles are oriented parallel to foliation and the flows is as yet unknown, for no feeders or and conceivably may have been somewhat squeezed, al­ ancient volcanoes have been found. though their axial ratios of 1:2-1: 3 are not unusual in The coarser grained sedimentary rocks range from stream or beach pebbles. The larger boulders are" well boulder conglomerate (tilloid) to quartzite and include rounded. meta-arkose and graywacke. At least some quartzite is The conglomerate is overlain by a sheared gray phyl­ crossbedded. All purer clastic quartzites are lenticular lite, typical of the Nova Lima Group, and a short but and few are as thick as 100 m. Kecrystallization of thick lens of pure white quartzite containing some quartz in micaceous quartzites and quartz-mica schist fuchsite. destroyed all evidence of original grain siz3 and nature This conglomerate corresponds closely to the material of original sediment. The presence of graywacke and the termed "paraconglomerate" or "tilloid" by Pettijohn tilloid strongly indicates the action of turbidity cur­ (1957, pp. 261-265) and attributed by him to subaqueous rents in the basin of deposition and thu

12 percent sericite. Original more homogenous quartzite the writer believes the formation thins stratigraph- was crushed during the tight folding. ically rather than tectonically.

Gair (1962, p. A30-A31) described the petrography LITHOLOGY of the Palmital Formation in some detail. He found by The Casa Forte Formation consists of schistose and thin-section study that at least one bed was gray wacke massive sericitic and chloritic quartzite, conglomerate, and suggested that closer study might reveal that this and lesser amounts of sericitic and chloritic schist and rock is more common than now known. Chloritoid in phyllite. quartzose rock is more abundant on the east side of the Quartzite of the Casa Forte Formation may be schist­ Vargem do Lima syncline than the west and occurs in ose to compact and massive. Locally it is in beds several rosettes and blades comprising as much as 30 percent of meters thick, some of which are conspicuously cross- the rock. The mineral is easily confused with kyanite in bedded. Massive beds crop out boldly and form peer liar the field. Small amounts of kyanite also have been haystack hills 20 m or more above surrounding terrain found. Gair also tentatively identified some fuchsite. (Gandarela N. 5,300, E. 3,350). (See pi. 5.) Where As originally mapped by O'Rourke, the Palmital schistose or more thinly bedded, the quartzite crops Formation contained a number of conglomeratic lenses out as low ridges scarred by solution pits, and these and beds, some traceable for many kilometers. Gair, ridges may look much like outcrops of Moeda quartzite subdividing the Maquine Group in the area immedi­ and Itacolomi quartzite. The rock ranges from white ately to the west of the area mapped by O'Rourke, by through buff and brown, depending on accessory and definition included all conglomeratic beds except the subordinate minerals. It may contain sparse well- basal conglomerate in the Casa Forte Formation, thus rounded pebbles of vein quartz and smaller angular redefining the then unpublished definition of the Pal­ quartz fragments. mital Formation. Gair's usage is here followed. The Quartzite of the Casa Forte Formation is characteris­ contact as mapped by O'Rourke in the Gandarela and tically sheared into a microaugen structure, as empha­ Rio de Pedras quadrangles was readjusted in the field sized by O'Rourke and Gair. Angular grains of quartz by the writer to fit Gair's usage, the contact being up to 1 cm across occur in a groundmass of fine-grained placed at the lowest conglomerate of the Casa Forte. comminuted quartz having a grain size of a few hun- dredths of a millimeter; folia of sericite cut across the CASA FORTE FORMATION groundmass and wrap around the larger fragments. The Casa Forte Formation as defined by Gair (1962, Conglomerate zones are lenticular but, particular'y in p. A31-A33) is separated from the Palmital Formation the Gandarela and Rio de Pedras quadrangles, can on the basis of the included conglomerate beds. It is be traced for several kilometers. They can be used for generally more quartzose and less sericitic and phyllitic deciphering structure if care is exercised not to con­ than the older rock. fuse different zones, for all are very similar litliologi- cally. Most conglomerate beds are dark brown to buff CONTACTS AND THICKNESS on the outcrop and form low scarps or linear ridges The Casa Forte Formation conformably overlies the clearly apparent in aerial photographs. The congi'om- Palmital Formation in the Nova Lima, Rio Acima, erate consists of rounded vein-quartz pebbles and Gandarela, and Rio de Pedras quadrangles. It has not elongate pebbles of phyllite, iron-formation, and fine­ been mapped separately elsewhere. grained white quartzite, probably metachert, in a The original thickness of the Casa Forte Formation matrix of quartzite or ferruginous quartzite. Pebbles is unknown, as it crops out only in the center of an of phyllite and metachert (?) are found throughout, but isoclinal syncline. If there is no repetition of beds by those of iron-formation are much more abundant ir the internal folding, the apparent remaining stratigraphic northern part of the belt than in the southern pazi,. thickness is about 600 m at maximum outcrop width. O'Rourke (written commim., 1956) stated that peb­ bles are elongated as much as 10:1. It seems certain Where the underlying Palmital Formation lenses out that some elongation is the result of tectonic stres^ on in the Rio de Pedras quadrangle, the Casa Forte For­ originally plastic pebbles and cobbles, but it is also mation overlies the Nova Lima Group without apparent probable that many fragments originally were platy discordance between the two units. The Casa Forte For­ rather than rounded, because most vein-quartz pebbles mation gradually narrows in outcrop width toward are undef ormed. the southeast, as the result either of stratigraphic thin­ Chloritoid is present in both the conglomerate and ning or of structural position in the axial part of an the quartzite; Gair stated that it occurs in fractures in isoclinal syncline. Because lineation plunges southeast, the conglomerates and both concordant to and cross- A24 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL cutting the foliation of the quartzite. Locally, the more poor exposures. The rocks are certainly in fault contact phyllitic quartzite may contain much chloritoid, which in some places and are apparently structurally and imparts a splintery fracture to the rock. stratigraphically conformable in others. Although Max­ The Casa Forte Formation apparently changes in well found in the Capanema quadrangle one locality composition along strike from north to south. Con­ showing strong angular discordance, he could not be glomerate lenses are more continuous but thinner, and certain whether this was a sedimentary or fault contact pebbles and cobbles smaller, toward the south. The because of poqr exposures; in view of the general con­ lenses are more continuous in the Gandarela and Rio cordance elsewhere, it may be considered a fault contact de Pedras quadrangles than in the Rio Acima quad­ until proved otherwise. Contacts with the overlying rangle. This lack of continuity may be due to increasing rocks in this area are probably unconformable. tectonic disturbance in the axial zone of the Vargem do In a small area in the southeastern part of the Santa Lima syncline, for the north end of this feature is Rita Durao and the northeastern part of the Antonio twisted to the east and cut by many more cross faults Pereira quadrangles, an isolated patch of gray quartzite than are apparent in the southern quadrangles (pi. 1). crops out boldly but with little relief. It was correlated with the Maquine Group undivided by Maxwell and GROUP UNDIVIDED Barbosa because of its lithologic similarity with the In the Conceigao do Rio Acima, Capanema, Sao upper part of the Maquine quartzites exposed along the Bartolomeu, Catas Altas, Santa Rita Durao, Mariana, road that climbs the north face of the Serra do Caraga. and Antonio Pereira quadrangles, the thick quartzite There is no way of deciphering the thickness or contact formation overlying the Nova Lima Group was mapped relations in this area. as Maquine Group undivided by Moore, Maxwell, and LITHOLOGY A. L. M. Barbosa. There the strata consist largely of The quartzose rocks in the southeast extension of the quartzite. The quartzite may be chloritic, pyritic, serici- Vargem do Lima syncline were reported by Maxwell as tic, or rather pure; locally it contains chloritoid. All essentially similar to the Casa Forte Formation to the these types may be conglomeratic. Quartz-chlorite schist, west and northwest. They include the same type of de­ quartz-sericite schist, and phyllite also occurred toward formed conglomerate beds, contain chloritoic1 and chlo- the bottom of the section. rite, and show microaugen structure and extreme inter­ The easternmost part of the Maquine Group undi­ nal crushing and shearing. With the exception of the vided in the southwestern part of the Capanema quad­ absence of iron-formation cobbles reported to the north rangle (the continuation of the Vargem do Lima syn­ by Gair and Dorr, the conglomerates seem entirely sim­ cline) is composed of schistose rocks that correspond ilar to those described above. Maxwell found subgray- in lithology to the Palmital Formation and are contigu­ wacke to be present in this area. ous with the formation as mapped in the Rio de Pedras Rocks mapped as Maquine Group undivided in the quadrangle. Maxwell (written commun., 1962) stated area of the Serra do Caraga differ somewhat hi lithol­ that these schistose rocks pinch out to the southeast. ogy, possibly in part because shearing was here less It will be remembered that the Palmital also wedged intense. Quartzites are more massively beddec1 where not out to the southeast in the adjacent Rio de Pedras quad­ schistose, locally contain much more chlorite, and are rangle on the west side of the syncline. The Maquine locally pyritic; true conglomerate is not common. Near Group in the type locality in the Vargem do Lima syn­ the base of the section, Moore and Maxwell found chlo­ cline is nowhere contiguous with the rocks mapped as ritic and sericitic quartzose schists, brown, gray, and Maquine Group undivided on the north and west flanks green on the outcrop, in some zones with scattered peb­ of the Serra do Caraca or to the southeast of that range, bles of vein quartz. Overlying is a sequence of chloritic and thus the correlation is not absolutely certain. and sericitic quartzite, some highly pyritized. Maxwell (written commun., 1962) also reported sub^raywacke CONTACTS AND THICKNESS and graywacke. Still higher in the section, th^ quartzite In the area east of the Vargem do Lima syncline, contains less micaceous minerals and is a gray to white deformation is strong and horizon markers absent; massive sericitic to nearly pure quartzite that crops out therefore, measurement of thickness of the stratigraphic in rugged scarps and crags. Scattered small pebbles also section is unreliable. Maxwell (written commun., 1962) occur in these strata. estimated that the stratigraphic thickness of the rocks may be about 1,400 in in the area mapped by him. ENVIRONMENT OF DEPOSITION The contact relations with the underlying Nova Lima Sedimentation in Maquine time evidently changed Group rocks are most obscure because of faulting and abruptly in environment and source of materials from PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A25 that which produced the much more argillaceous rocks cannot be considered a probable source of coarse­ of the underlying Nova Lima Group. This change is grained highly quartzose sediments because quartrose attested to by the more quartzose composition of the Nova Lima rocks are characteristically fine grair«*l. rocks, by the conglomeratic zones and beds, and by the Whether or not this other source rock was granitic is apparent lack of volcanic contributions or of chemical unknown. Although careful search was made by the sediments such as iron-formation or carbonate-rich staff, only Maxwell (oral commun., 1964) found any rocks. The Maquine Group evidently represents a evidence in the conglomeratic rocks for granitic peb­ nearshore environment. bles and this in only one locality. An erosional unconformity, perhaps with some struc­ The presence of chloritoid in very significant amounts tural disturbance of strata below the Maquine, is masked suggests a highly aluminous source, conceivably deeply by the intense deformation of the whole Rio das Velhas weathered Nova Lima Group rocks, for the sediments Series; it is indicated by the presence of a basal con­ metamorphosed into such rock. O'Rourke (written glomerate beneath the Palmital Formation on the west commun., 1957) was the first to call attention to the side of its outcrop. That erosion of the Nova Lima con­ evidence for mature weathering of some of the source tinued during deposition of the Casa Forte Formation rock. is suggested by pebbles and cobbles of iron-formation METAMORPHISM and metachert in the conglomerates of that formation. The Maquine rocks have generally been metairor- Both chert and iron-format ion are characteristic of the phosed to the greeiischist facies, and, except, for the Nova Lima Group. That these rocks were lithifiecl but presence of chloritoid and kyanite in areas of maximum not yet recrystallized when eroded is suggested by the stress discussed above, there is nothing else particularly latter extreme deformation of these pebbles in the same noteworthy about them. Their bulk composition, outcrops where relatively undeformecl vein-quartz peb­ dominantly quartzitic, precludes the formation of a bles are found, deformation of the same order of com­ wide suite of metainorphic minerals. plexity as that of the phyllite cobbles occurring with In the isoclinal Vargem do Lima syncline, the irost them. striking result of the dynamic metamorphism on the The wedgeout of the Pahnital Formation to the Maquine Group rocks is the tremendous degree of crush­ southeast, as shown on both the east and the west sides ing undergone by these rocks. The quartzose rocks were of the Vargem do Lima syncline by O'Rourke and Max­ brittle and therefore shattered; the more sericitic ro^ks well, is a strong indication of nearshore conditions. The sheared. In many places all original sedimentary struc­ location of the shoreline during Casa Forte time is not tures that might have been present, such as crossted- known, but the absence of the Maquine on the west side ding, have been obliterated, and metamorphic struc­ of the Rio das Velhas Valley suggests, but does not tures superimposed. Top and bottom criteria are prove, that it was somewhat to the west of present ex­ destroyed. Bedding has likewise been destroyed to a posures of those rocks. The change in type of sedimen­ considerable extent. tation and particularly of the distribution, size, and In the less deformed areas, as in the Serra do Car<\c.a frequency of pebbles in the rocks toward the east is also and southeast of that range, bedding has survived the suggestive of a source from the west. The changes in metamorphism in many localities, and crossbedd:ng rock type from east to west are expectable under near- can be seen in some rocks. shore conditions, as is the apparent thickening east­ ward in the Capanema and Conceigao do Rio Acima INTEEPEETATION OF THE BIO DAS VELHAS SEBIES quadrangles. The rocks thin abruptly eastward in the The major gaps in our knowledge of the Rio das Catas Altas quadrangle and were not found in the east­ Velhas Series are caused in large part by three factors: ern part of the Quaclrilatero Ferrifero. Whether this eastward thinning was caused by nondeposition or first, the type of lithology that results, under the cli­ post-Maquine erosion cannot be stated; the latter seems matic and physiographic conditions of the region, in the more probable because of the major unconformity poor exposures; secondly, the obliteration of sedimen­ above the Rio das Velhas Series to be discussed in later tary structures by the extreme deformation; ?nd pages. thirdly, the lack of mineral resources, except gold de­ Although the presence of phyllite, iron-formation, posits, which would justify expensive geologic explora­ and metachert cobbles in the Maquine clearly indicate tory effort. Until the present program of mapping, that some of this rock was derived by erosion of the these rocks had been studied only in the vicinity of g->ld Nova Lima Group, a second source of sediments, par­ deposits; they had not been mapped at detailed scries ticularly during deposition of the Casa Forte, seems except locally; and there was no appreciation of the re­ probable because the Nova Lima Group as now exposed lations between these rocks and other metasedimentary A26 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL units of the region, with which they had been confused. the lithic units; the other is a structural approach stress­ Because the work of the DNPM-USGS team was ing the nature of the contacts between the groups. All focused on the iron and manganese resources of the clastic sedimentary rocks are ultimately a result of dia- region, primary effort was not devoted to unraveling the strophism; therefore the lithic content of the formation, problems of the Kio das Velhas Series; information and particularly of conglomerates, affords much gained was in the normal course of regional mapping. information. A fertile field remains for future generations of The Nova Lima Group consists domiiiantly of chlorite geologists. schist and phyllite, sericitic phyllite, quartz-sericite The Eio das Velhas Series was separated from the schist and phyllite, graphitic phyllite, and graywacke Minas Series (Eynearson and others, 1954; Oliveira, and graywacke schist. It contains small amounts of 1956; Dorr and others, 1957) because a profound un­ quartzite and characteristic thin lenses of carbonate- conformity was discovered at the base of a persistent facies iron-formation and lesser amounts of tilloid con­ quartzite zone in rocks of the Miiias Series as originally glomerate and quartz-ankerite rock. Only one marble defined by Derby in 1906. Strata above that uncon­ bed was found in the whole region. A considerable ad­ formity fitted closely the lithology and succession of the mixture of tuffaceous material was found in the few subdivision of the Minas Series made by Harder and areas where fresh rock could be studied. Chamberlin in 1915. Below the unconformity the strata The total assemblage is similar to that described by form a suite which differs in lithology from the Minas Pettijohn as the graywacke (flysch) suite (1957, p. 615- Series as defined by Harder and Chamberlin and others. 616). Pettijohn stated that this suite is marked by great A remaining problem which we had to consider was thickness, by its predominantly argillaceou^ character, whether all the rocks here included in the Rio das by carbonaceous shales, by the relative absence of car­ Velhas Series are so related in time and environment of bonate rocks, and by subordinate quartzite. Bedded deposition that they should be included in one major cherts are said to be present, and waterlaid mafic tuffs unit or whether they should be divided into separate are said to grade into and resemble the coarser gray- units. wackes. The average ratio of sandstone to shale in the The use of radiometric dating can be quite deceptive flysch facies is about 1: 2 in sections cited by Pettijohn in this area (Hurley, 1961), but future refinements may and may be of this order of magnitude in the Nova Lima solve many age relations in the Eio das Velhas Series. Group as a whole, if the sericite-quartz phyllites are As of this writing, the only available determination on considered as impure sandstone. Although graywacke these rocks is a rubidium-strontium age of some 2,800 is found in few exposures, the rock type is particularly million years on mica from the Nova Lima Group just susceptible to deep weathering. Considering the abun­ west of the Eio cle Pedras reservoir, Eio de Pedras dance of other lithologies of the suite, it is probable quadrangle (Aldrich and others, 1964, p. 329). This that a much higher percentage of this rocV may exist specimen was taken in the contact aureole of the Bac.ao than appears at the surface. complex, indicating an older age for the Nova Lima Carbonate-facies iron-formation in the Nova Lima Group. It is quite possible that age determinations on Group indicates local reducing environments expectable rocks from other parts of the Eio das Velhas Series may in the type of sedimentary environment giving rise to show younger ages in other localities, for the rocks have the graywacke suite. The ubiquitous association of been affected by a more recent orogeny and metamor- graphitic phyllite with what were originally silty man­ phic events. These may have imposed their ages on the ganese carbonate beds (Dorr and others, 1956) about minerals in the rocks and probably obscured the effects 10 km south of the Quadrilatero Ferrif ero also indicates of the older metamorphism. The absolute age of the such an environment. Carbonaceous and graphitic phyl­ metasedimentary rocks cannot be established by avail­ lite are locally associated with the carbonate-f acies iron- able methods; only the age of their metamorphism can formation. For reasons still hotly debated, iron-forma­ be determined. tion is characteristic of Precambrian rocks; in the Nova In considering the relations between the two groups Lima Group that material, which grades into metachert, of rocks here included in the Eio das Velhas Series and may be equivalent to the bedded cherts cited by Petti­ in trying to decipher whether they represent a single john as typical of the flysch unit in younge^ rocks. evolutionary sedimentary unit or whether they repre­ Above the Nova Lima Group lies the Maquine Group, sent units widely separated in time and genetically un­ consisting of quartzose rocks, subgraywacl~e, and con­ related, two major lines of approach seem most likely glomerates. Pettijohn said (1957, p. 616) : to provide reliable information. The first is the. nature the typical graywacke (flysch) assemblage commonly increases of the sediments themselves and the. relations between in coarseness upward. Conglomerates, if any, appear high in PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A27 the section. The upward increase in coarseness is associated berlin in 1915, was restricted at the top by Gruimara°ss with an increase in the proportion and coarseness of the sands. in 1931, was restricted at the base by Oliveira in 19£6, The sands, moreover, become cleaner and are subgraywackes was divided into three groups by Dorr and colleagues and even protoquartaites rather than graywackes * * *. The typical flyseh facies has been displaced by the molasse facies in 1957, and was subdivided into nine formations 'by a product of paralic sedimentation. Wallace (1958), Maxwell (1958), Dorr (1958a, b), This statement well describes the Maquine Group, which Pomerene (1958a, b, c), Simmons (1958), and Gair is characterized by, in ascending order, a local basal (1958). The Tamandua Group, here included as a fourth conglomerate, phyllite, phyllitic and schistose rocks group of the Minas Series, was first described by Sirn- with increasing quartz, subgraywacke, conglomerates, mons and Maxwell (1961), who assigned it to the Rio sericitic and cliloritic quartzite, and clean quartzites das Velhas Series. and grits. Despite the masking effects of metamorphism As a whole, the Minas Series is quite different in na­ and strong deformation, there is little possible question ture from the underlying Rio das Velhas Series, for it of the nature of these rocks or of the fact that they form consists largely of blanket and tabular formations a molasse suite coherent with the underlying flyseh fa­ (Pettijohn, 1957, p. 618) relatively consistent in lith­ cies of the Nova Lima Group. ology for great lateral extent, a type of formation not The contacts between the two groups of rocks forming found in the Rio das Velhas Series. The Minas sedi­ the Rio das Velhas Series are of significance in this mentary rocks are well differentiated, except the highest analysis. On the west side of the Vargem do Lima syn- formation, also in contrast to most of the Rio das Ve'- cline the contact between the Nova Lima and Maquine has rocks. The uppermost formation of the Minas Se­ rocks is essentially concordant, perhaps with a slight ries, the Sahara Formation, is poorly sorted and con­ angular and erosional discordance. On the northeast tains much graywacke and tuffaceous rock. side of the Vargem do Lima syncline the contact is gra- Thus, as illustrated in the stratigraphic column pre­ dational. Further southeast it is a fault contact or a sented on plate 12, the series is here divided into four suspected small angular discordance. groups, composed respectively of two lower groups of Fragments of rock believed to have come from the two clastic formations each, separated by a probable Nova Lima Group have been reported in the Casa Forte erosional unconformity; a gradationally overlying Formation in the Vargem do Lima syncline. The Pal- group composed of two formations of dominantly chem­ mital Formation wedges out in the southern part of that ical or biochemical origin; and an upper group com­ syncline. Thus, it seems certain that locally an erosional posed of five clastic formations, the uppermost quite unconformity exists between the Maquine and Nova different in general environment from the other forma­ Lima Group, but elsewhere sedimentation may have tions of the series. been essentially continuous. Given the environment of deposition of these rocks the upper unit, a nearshore CONTACT WITH THE BIO DAS VELHAS SERIES or continental suite, and the lower, a geosynclinal basin The contact of the Minas Series with the underlying suite local erosional unconformities and small angular Rio das Velhas Series is structurally and stratigraph- unconformities are to be expected. The gradation be­ ically unconformable, but the extent and degree of the tween the phyllitic rocks and the quartzose rocks, so unconformity is difficult to establish with precision in well exposed at the northeast corner of the Vargem do all localities. That unconformity was first demonstrated Lima syncline, and the wedgeout of the Palmital For­ in the Moeda Plateau and along the south side of the mation and overlap of the quartzose Casa Forte rocks Serra do Curral by Rynearson, Pomerene, and Dorr in on the argillaceous rocks of the Nova Lima Group are 1954. Plate 1 shows that mappable lithologic zones in the Nova Lima Group strike nearly at right angles to also congruent with the inferred environment. the Minas Series in the Itabirito quadrangle. Lenses of The lithology of strata here included in the Rio das iron-formation in the Nova Lima Group strike at high Velhas Series, and evidence provided by contacts be­ angles into the trend of the Minas Series in the Ser~a tween its two groups, indicate that the Rio das Velhas do Curral both to the west in the Sousas quadrangle and Series rocks seem to form a single evolutionary strati- in the Macacos quadrangle in the central part of the graphic entity reflecting a progression from eugeosyn- range. The angle of intersection between iron-formation clinal to paralic sedimentation. of the Nova Lima Group and the Minas Series in the Gandarela and Itabira quadrangles is lower. Reeves MINAS SERIES (1966) discussed evidence pointing toward an angular The Minas Series was first named by Derby in 1906, unconformity between the Minas and Rio das Velhas was divided into five formations by Harder and Cham- Series rocks in the Monlevade district. 334-413 O <69 3 A28 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL

Along the Serra do Itabirito and in other localities nitic rocks to the north and west also fcorned or were where a strongly angular unconformity can be proved, emplaced during this interval. the two series are also locally conformable. This sug­ gests that the pre-Minas folding there was rather tight. DISTRIBUTION, WEATHERING, AND On the other hand, the apparent conformity between METAMORPHISM the Minas Series and the underlying rocks in that part The Minas Series is preserved only in synclinal struc­ of the Gandarela syncline where rocks trend northwest tures in the Quadrilatero Ferrifero, in the complex mass by north (Gandarela quadrangle), contrasted with the of the Serra do Caraga, and in the Serra das Cambotas. 30° angular unconformity where the Minas beds turn The Minas rocks commonly lie in orderly sequence, al­ toward the northeast (pi. 1), indicates that the older though in many of their exposures they are inverted, for beds had been strongly tilted in th'.s area before Minas most of the syiiclines are overturned to the north and sedimentation started. The tight folds found in these west. Many relatively thin lithologic units of the Minas older rocks 20 km to the west are not here present. Series can be traced for tens of kilometers along strike In other areas, as along the south flank and eastern without change in lithology except that resulting from part of the Gandarela syncline, near the axis of the variation in metamorphic grade. Mariana anticline (Antonio Pereira quadrangle), and Thickness of some of the units varies radically; in south of Serra da Piedade (Serra da Piedade quad­ most places this is clearly caused by tectonic deforma­ rangle), bedding is obscure in the Nova Lima Group, tion of the rocks. Formations characteristically thicken and foliation is essentially parallel to the bedding in in the axial zones of folds and thin on the flanks within the Minas Series. Angular relations are strongly indi­ a few hundred meters to a few kilometers; the Caue cated by the more complex folding of the Nova Lima Itabirite varies in apparent thickness from a few meters Group rocks in some of those areas, and Simmons (writ­ to more than 1,-iOO m. Such variations are not confined ten commun., 196-1) stated that there is an angular dis­ to the plastic iron-formation and argillaceous rocks but cordance of about 20° in the Santa Barbara area. are also found in more massive quartzites. Naturally this Profound erosion of the Rio das Velhas Series before tectonic deformation obscures the originrl sedimentary Minas sedimentation began is certain. The only place features of the rock in many areas. where a minimum absolute value of such erosion could Most strata of the Minas Series are deeply weathered; be calculated is in the northwest corner of the Gan­ exceptions are the massive quartzites, locally the itabi­ darela syncline; there marker beds in the Nova Lima rite, and still more locally the dolomite lenses. The only Group show that not less than 1,000 meters of sediments rocks which could be satisfactorily studied petrograph- was removed. More than 1,000 meters of quartzose rock ically are the monomineralic rocks, some ferruginous of the Maquine Group must have been eroded between rocks, and a few phyllites and schists incised by recent the Vargem do Lima syncline and the Serra da Ouro rapid mechanical erosion. Thus, our petrographic knowl­ Fino; very probably all of the Maquine Group was edge of the Minas Series rocks is uneven; some we know removed in extensive parts of the region before Minas well, but most we know poorly. sedimentation began. The metamorphic grade of the Minas Series ranges Wherever the contact of the two series can be traced, from the greeiischist facies to the amphibole-almandine it is evident that the pre-Minas surface was one of very facies (Herz, written commun., 1965) and increases gen­ low relief. The maximum local relief on the old erosion erally toward the southeast and east. Variation depends surface of perhaps 50 m was produced on a hard quartz- upon the regional position and proximity to younger ite of the Nova Lima Group (fig. 5). The irregular granitic bodies. Certain formations have been metamor­ phosed to granitic gneisses in the Monlevade and Itabira accumulation of polymictic basal conglomerate in the areas. Shearing was strong in places. Both shearing and Minas Series is an indication of old valleys, a few of metamorphism obscure sedimentary features, and the which were several tens of meters deep but most of greater the degree of metamorphism and shearing, the which were much shallower. Presumably the surface deeper the weathering. had been peneplaned, as suggested by Harder and Chamberlin (1915). TAMANDUA GROUP Regional evidence shows that a great time interval The Tamandua Group was defined by Simmons and elapsed between Rio das Velhas and Minas Series sedi­ Maxwell (1961) as comprising the quartsdtes, quartzose mentation to permit folding of the older rocks and and argillaceous schists, and phylitic and dolomitic beveling of them to a nearly plane surface. Part of the itabirite stratigraphically between the Maquine Group Ba^ao granitic complex and probably some of the gra­ of the Rio das Velhas Series and the Caraga Group of PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A29

FIGURE 5. Vertical aerial photograph showing the unconformity between the Minas Series (left) and the Rio das Velhas Series (right) in the Serra do Itabirito near N. 8,000, B. 3,000, Itabirito quadrangle. The Minas Series is here folded past the vertical and dips steeply east. The relief on the ancient pre-Minas erosion surface, formed on the hard quartzite in the Rio das Velhas Series, is somewhat exaggerated by the angle of view. Approximate scale, 1:10,000. A30 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL

the Minas Series. The group was named from exposures and 7), not only where underlain by the relatively non- in the Serra do Tamandua (Santa Barbara and Cocais resistant granitic gneiss, as in the Serras do Tamandua quadrangles); it also crops out in the Antonio dos San- and das Cambotas, but also where it lies adjacent to the tos, Gongo Soco, Capanema, and Santa Rita Durao generally less resistant rocks of the Minas Series, as in quadrangles. Rocks tentatively correlated with the Cam­ the Serra do Caraga. It is one of the most resistant botas Quartzite of this group crop out in the Floralia, formations in the region to chemical and mechanical Dom Bosco, Ouro Branco, and Santa .Rita do Ouro weathering. Preto quadrangles. DISTRIBUTION AND THICKNESS The Tamandua Group was originally divided into one Rocks mapped as Cambotas Quartzite crop out in the named and three unnamed formations by Simmons and Serra do Tamandua, the Serra das Cambotas, the east Maxwell. Simmons (1968b) subsequently grouped the end of the Serra Geral, and the Serra do Caraga. Sim­ three unnamed units into a single formation, and this mons and Maxwell (1961) divided the Cambotas into restriction is followed in this paper. The Cambotas two units at the type locality near the junction of the Quartzite crops out extensively and well; the overlying Serra Geral and the Serra das Cambotas a lower unit strata, composed of phyllite and iron-formation, crop 92 m thick composed of phyllitic schist, quartzitic phyl­ out poorly and are mappable in only a few restricted lite schist, conglomeratic phyllite schist, and quartzite,1 areas. and an upper unit 638 m thick composed almost entirely The rocks of the Tamandua Group are generally of quartzite. The lower unit cannot be traced laterally similar in lithologic aspect to many other rocks of the with any consistency because of poor exposures and region and cannot be traced continuously far from the migmatization by the underlying granitic gneiss. type locality. Stratigraphic relations between them and In the Serra do Caraga, Simmons and Maxwell others are obscured by metamorphism, faulting, ero­ divided the Cambotas Quartzite into four units, consist­ sion, and overlap of younger formations. Hence, the ing from top to bottom of coarse-grained quartzite, 100 group is identified only tentatively in parts of the m; crossbedded fine- to coarse-grained sericitic quartz­ region. ite, 580 m.; fine-grained sericitic quartzite, 70 m.; and The descriptions below relate only to the areas definitely identified as Tamandua Group on plate 1. Other areas will be discussed separately.

CAMBOTAS QUARTZITE The Cambotas Quartzite was defined by Simmons and Maxwell as including the quartzites and phyllitic and quartzitic schists at the base of the Tamandua Group. The unit characteristically forms bold scarps (figs. 6

FIGURE 7. Typical cliffy outcrops of the Cambotas Quartzite in the Serra das Cambotas, Santa Barbara quadrangle.

1 Although Simmons includes this lower unit in the Nova Lima Group in his most recent report (Simmons, 1968b, p. H10), this writer agrees with the original assignment of these rocks to the Cambotas Quartzite (Simmons and Maxwell, 1961). This opinion is based on the following criteria : (1) The high quartz content of the rocks, typical of the Cambotas but atypical of the Nova Lima Group, (2) the presence of conglomerate beds, regionally most uncommon in the Nova Lima Group and not reported in that group elsewhere in this vicinity, (3) the absence of chlorite, a characteristic mineral of the Nova Lima Group but not of the Cambotas Quartzite, and (4) the fact that the original description did not mention an important structural or Stratigraphic break between the lower and upper units (although a bedding-plane fault was noted). The original description FIGURE 6. Typical bare outcrops and cliffs of the Cambotas states that near the type locality the Cambotas as there denned rests Quartzite in the Serra do Caraga. The chapel of Colegio do discordantly on the Nova Lima Group. If the lower unit of the original description were Nova Lima rather than Cambotas, an interval equiva­ Caraga in the foreground is at an elevation of about 1,300 m; lent to the whole Maquing Group must be present between these con­ the peaks, over 1,750 m. Photograph by Norman Herz. cordant units. PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A31 conglomerate and very coarse grained quartzite, 40 m. west of the point where the road to Gongo Soco from the Maxwell, who mapped the Serra do Caraga, elsewhere Caete-Barao de Cocais highway crosses the serra. stated (written commun., 1962) that there Moore, who mapped this area, stated (written commun., the thickness of the Cambotas Quartzite is variable and no­ 1962) that the formation is cut out by a thrust fault; where certain * * *. The only locality in the Alegria district Bolau (1952, p. 482) considered that the formation which has younger pre-Minas rocks overlying the Cambotas wedges out to the west. The writer, after a careful field Quartzite is in a zone of complex faulting, so complex that meas­ check in 1966, believes that a sedimentary wedgeout is urement of thickness is impossible. There are no definitive the more probable, for the formation does not reappear marker beds in the quartzite * * *. The maximum apparent thickness is about 1,000 m. west of this area and the postulated thrust fault was not found. In the Serra do Tamandua the outcrop width ranges LITHOLOGY from 0 to 1,900 m. The thickness is at least 200 m. in The Cambotas Quartzite is largely an orthoquartr.ite. much of the range, but the rocks disappear abruptly to Maxwell stated that in the Serra do Carac.a the rock is the east. relatively pure quartz with minor amounts of sericite- CONTACTS muscovite, chlorite, clinozoisite, and kyanite. Kyanite is The Cambotas Quartzite was said by Simmons present only in zones of intense deformation (Maxwell, (1968b) to overlie granitic gneiss with metasomatic con­ written commun., 1962). Rock types present are pel ble tact wherever the contact between those rocks is ex­ conglomerate, grit, and fine- to coarse-grained quart­ posed in the Cocais and Santa Barbara quadrangles. zite; some layers are crossbedded. A few thin zcnes To the west, Moore mapped a normal discordant sedi­ consist of quartzose sericite phyllite. mentary contact between the Cambotas Quartzite and In the Serra das Cambotas the unit is lithologically the Nova Lima Group in the Gongo Soco quadrangle; similar, although Simmons reported no chlorite or clino­ the quartzose grit and conglomeratic grit are there zoisite. There, the most distinctive feature of the rock apparently concordant in strike with underlying rocks, is the preservation of original sedimentary structures, but dips may or may not be concordant. An erosional including ripple marks, scour and fill, and crossbedding. contact between the two groups is proved by the pres­ These are preserved despite the fact that the strata have ence of fragments of a distinctive thin-bedded Nova been so crushed that a microaugen texture has resulted Lima iron-formation in the basal conglomerate of the and, on weathering, the rock breaks down to silt-sized Cambotas at Gongo Soco N. 8,100, E. 6,800. The iron- particles (Simmons, 1968b). In the Serra do Tamandua formation crops out about 50 meters to the north. Max­ the lithology is apparently similar to that in the Serra well (written commun., 1962) stated that in the Catas das Cambotas. Altas quadrangle the Cambotas may overlie the Nova Simmons found that the Cambotas Quartzite in the Lima Group with unconformable contact. Simmons and Cocais area was strongly metasomatized. The partial Maxwell (1961, p. 9) stated that the Tamandua Group alteration of the quartzite to granitic rock can be sf,en near its type locality discordantly overlies the Nova a few hundred meters west of the Barao de Cocris- Lima Group. Cocais road, just north of the Serra do Tamancfua, The Cambotas Quartzite overlies the Maquine Group where quartzite occurs as isolated remnants in the gneiss. with apparently concordant contact in most places in Bolau (1952, p. 488) determined that feldspar occurs the Catas Altas and Capanema quadrangles, but the locally in rocks now called Cambotas Quartzite, which regional outcrop pattern was said by Maxwell (written he correlated with the basal Minas Series. Simmons p.nd commun., 1962) to show a discordance which could be Maxwell also mentioned the presence of this mineral; in either erosional or structural. At Capanema N. 5,400, E. some rocks the feldspar is sedimentary in origin, but in 6,600, the contact between these rocks is said to be an others it is a metasomatic product. Its presence or ab­ angular unconformity of 60°. Elsewhere in these quad­ sence has no diagnostic value from a stratigraphic rangles there are suggestions of a fault contact. Sim­ standpoint. mons and Maxwell (1961, p. 10) stated that the UNNAMED FORMATION Tamandua Group discordantly overlies the Maquine Group in the western part of the Serra do Carac,a. DISTRIBUTION, LITHOLOGY, AND THICKNESS Moore (1969) stated that in the Coiiceigao do Rio Acima The unnamed formations of the Tamandua Group quadrangle there is a discordance in dip of 5°-30° and in (Simmons and Maxwell, 1961) overlie the Cambcftas strike of 10°-80°. Quartzite in the type locality in the Gongo Soco ?,nd The Cambotas Quartzite thins abruptly to the west Santa Barbara quadrangles, in the Cocais quadrangle, in the Serra Geral and disappears a few hundred meters and at all observed localities in the Santa Rita Durao A32 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL

and Catas Altas quadrangles. Simmons originally sepa­ but the twTo units are conformable (Simmons and Max­ rated these formations in the Santa Barbara and Cocais well, 1961). Simmons cited no locality wl Qre known quadrangles, but they have not been separated else­ upper Tamadua is in angular sedimentary contact with where ; they crop out only in roadcuts or other artificial known Minas Series rocks. The persistence of this thin exposures or, rarely, in gullies. For this reason, Simmons and relatively soft formation for some 35 km across later (1968b, p. H11-H13) considered the three forma­ the Santa Barbara and Cocais quadrangles indicates tions as members of a single formation, despite their that if the contact is an angular unconformity, the angle varying lithology. must be less than 3°, hardly noticeable witl such poor As described in the type locality, the upper unit is exposures, and of minor significance. According to Sim­ phyllitic schist, 35 m thick; the middle unit is dolomitic mons (oral commun., 1964), not only the upper forma­ and phyllitic iron-formation, 41 m thick; and the lowest tion of his then threefold division, but all the formations unit is phyllitic and quartzitic schist, 211 m thick. Sim­ thin to the east, which suggests to the writer a wedgeout mons stated (written commun., 1964) that the upper rather than an angular or erosional unconf ormity. Max­ formation may be divided into nine lithologic units 2.5 well (written commun., 1962) implied that the uncon­ km west of the type locality, but that none of the units formity was an erosional one in the quadrangles mapped could be correlated with those of the type locality. From by him, but did not advance specific evidence. The per­ this it would appear that the lithology of the formation sistence of thin soft rocks along the east side of the Serra is highly variable. The thickness of these rocks in the do Caraga indicates that erosion there mus4: have been Catas Altas and Santa Rita Durao quadrangles was not minor if it occurred. mentioned by Maxwell (written commun., 1962), but the ROCKS TENTATIVELY CORRELATED WITH THE TAMANDUA maximum outcrop width is perhaps 200 m in a place GROUP where dips are steep. The Serra do Ouro Branco in the southwestern part In these rocks white mica is second in abundance to of the region is an isolated large mass of quartzite which quartz; small quantities of chlorite are also present. The Simmons and Maxwell (1961) and this writer tenta­ iron content of the iron-formation is not known; some tively correlate with the Cambotas Quartzite of the 28 in of the total of 42 m of this formation at the type Tamandua Group. This most impressive mountain is locality is dolomitic and phyllitic itabirite, and the rest composed of white and gray sericitic quartzite with an is dolomitic phyllite. The writer believes, but cannot apparent maximum thickness of about 1,400 m (John­ prove, that the iron-formation was oxide facies. A single son, 1962, p. 10 and fig. 2), bounded on the north by a analysis presented by Simmons (1968b, p. H13) showed major fault. To the south, the quartzite is1 in contact 71 percent iron; this was from a small soft high-grade with gneiss at least in part formed by metasomatism of iron ore body in the Serra do Tamandua. Such high- the Nova Lima Group and, with angular unconformity, grade iron deposits are known only from oxide-facies the Nova Lima Group. This great quartzite mass wedges iron-formation elsewhere in the region. The rapidly out gradually to the east, as in the Serra do Tamandua, changing lithology as most recently described by Sim­ and abruptly to the west, as in the Serra Geral. mons might indicate nearshore sedimentation. The quartzite of the Serra do Ouro Branco is com­

CONTACTS posed, according to Johnson, of gray to white sericitic coarse-grained quartzite and grit. The coarse quartz The contact between the unnamed formation and the grains, some well rounded, others angular because of Cambotas Quartzite was seen in only two places. In crushing, are in a matrix of fine-grained crushed quartz one, at the type locality, it was interpreted as a bedding and sericite. Hematite and magnetite, zircon, and locally fault (Simmons, 1968b, p. H13). At the other, in the feldspar are accessory minerals. Well-rounded quartz Santa Rita Durao quadrangle near Morro da Mina, the and quartzite pebbles as much as 10 cm long are reported relations are obscured by severe shearing and faulting, but sparsely scattered pebbles of vein quartz and quartz­ and the fundamental nature of the contact could not be ite about 1 centimeter in maximum dimension are more deciphered (Maxwell, written commun., 1962). Sim­ common. Ferruginous quartzite and phyllite pebbles mons (1968b, p. H13) believed that the contact is prob­ also are present. One conglomerate bed 5 m thick com­ ably essentially conformable. posed of 50-70 percent rounded pebbles wr,s found by The contact between the unnamed formation and the Johnson, who suggested that the quartzite pebbles may overlying rock of the Caraga Group of the Minas Series be recrystallized chert. Pebbles in this bed are squeezed was believed by Simmons (1969b, p. H14) to be struc­ and alined parallel to schistosity rather than to bedding. turally conformable where observed. At the type locality Bedding in the quartzite is nearly obliterated by the contact is marked by several centimeters of quartz schistosity, which may cross the bedding at a large PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A33

angle. Compositional banding, reflecting original bed­ probably the Maquine rocks are too fine grained to sup­ ding, is locally several meters or more across. Because of ply coarse sediments of this nature. The source of the the strong foliation in this area, crossbedding is difficult sediments is not clear, but distribution of the coarser to identify with confidence; many structures in the sediments might suggest a source to the west. rock which appear to be crossbedding are due to tec­ tonic causes. Johnson did not definitely identify cross- CABAQA GROUP bedding and the writer believes that, although the na­ The Caraga Group was established by Dorr, Gair, ture of the rock indicates that the beds were probably Pomerene, and Rynearson (1957) to include the clastic crossbedded originally, this feature cannot now be rocks above the major regional unconformity exposed identified. in the western Quadrilatero Ferrifero and below the The isolated quartzite mass of Ouro Branco was cor­ chemical sediments of the overlying Itabira Group. As related with the basal Minas Series by Harder and originally defined, it included the Caraga Quartzite and Chamberlin (1915) and by Guild (1957). Barbosa (pi. the Batatal Schist of Harder and Chamberlin. The 9) mapped it as Maquine Group, queried. Guimaraes Batatal Schist was renamed the Batatal Formation by (1931) correlated it with the Itacolomi Series. As Maxwell in 1958 because it is largely phyllite and con­ pointed out by Johnson (1962, p. 10-12), who tenta­ tains other lithologies. The basal quartzite formation tively correlated it with the basal Minas Series, the as then known, called the Caraga Quartzite by Harder stratigraphic position of this rock is indeterminate. It and Chamberlin, was renamed the Moeda Formation is here correlated with the Cambotas Quartzite on litho- from the Serra da Moeda (Wallace, 1958), where the logic grounds; it can hardly be Itacolomi because of structure is simple and where there can be no question stratigraphic relationships. The wedgeouts to the east as to its stratigraphic position and its relation to other and west are more characteristic of the Cambotas than rocks of the Minas Series and the Nova Lima Group. In of the Moeda Formation quartzites. The lack of chlo- 1957 the quartzite making up the main mass of the Serrr rite and the relative purity of the quartzite strongly sug­ do Caraga was still unmapped and its age was moot, as gest that the rock is not Maquine Group, for it is a Guimaraes (1931) considered it to be Itacolomi. paralic orthoquartzite rather than a molasse In the course of mapping the Serra do Caraga, Max­ protoquartzite. well perceived that the quartzites making up the main Two arcuate masses of quartzite isolated in granitic mass of the mountain were somewhat different from the gneiss in the Floralia quadrangle were mapped by Herz, quartzite immediately below the Batatal Formation in who correlated them with the Cambotas Quartzite. that area. At the same time, Simmons, mapping in the These are said to lie with angular discordance on the Santa Barbara and Cocais quadrangles to the north, Nova Lima Group of the Rio das Velhas Series (Herz, found that the principal quartzite in the Serra do Ca­ oral commun., 1965). raga was lithologically similar to that in the Serra clas Cambotas and the Serra do Tamandua and that in the ENVIRONMENT OF DEPOSITION latter ranges two phyllite and one iron-formation zones The rocks of the Tamandua Group differ from those separated that quartzite from thin quartzite and phyl­ of the underlying Maquine and Nova Lima Groups in lite zones which he correlated with the Moeda and Ba­ that they are shelf deposits dominantly composed of tatal Formations. The major quartzite underlying the orthoquartzite, quartzose phyllite, and phyllite. The Serra do Tamandua, the Serra das Cambotas, and the Cambotas Quartzite is a prismatic paralic deposit; the Serra do Caraga and the phyllitic and iron-formation overlying phyllitic and dolomitic rocks might indicate zones were grouped by Simmons and Maxwell (1961) deeper water, although the rapidly varying lithology as the Tamandua Group, and the major quartzite called indicates nearshore deposition also. The group overlies the Cambotas Quartzite. the older eugeosynclinal and molasse rocks, certainly The Caraga Group must be redefined because the origi­ with deep erosional unconformity and probably with nal definition was made before the presence of older an angular unconformity. Wherever observed, it under­ rocks above the regional unconformity and below the lies the Minas Series with angular confonnity; no sig­ type Caraga Group had been established. The Caraga nificant erosional unconformity has been certainly Group is here redefined as comprising the metasedimen- demonstrated although it may well exist. tary quartzitic and phyllitic formations conformably Minor chlorite present gives no hint of the volcanic underlying the chemical sediments of the Itabira Group activity which characterizes the underlying sediments, of the Minas Series and overlying with essential con­ and the coarse quartzites must have been derived from cordance either the quartzose and phyllitic rocks of the the erosion of a granitic terrane, for the Nova Lima and Tamandua Group or, commonly with angular and ero- A34 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL sional discordance, the rocks of the Rio das Velhas this area the formation averages more than 300 m in Series. thickness, with the maximum apparent thickness greater The type locality of the group was originally re­ than 1,000 m (See fig. 8.) The coarse-grained facies also stricted (Dorr and others, 1957) to the south flank of occurs in the Monlevade area; there, the marimum thick­ the Serra do Caraga in the Capanema quadrangle, fol­ ness of some 500 m is attributable to both tectonic and lowing the definition by Harder and Chamberlin (1915) original stratigraphic thickening. as closely as possible, because of the uncertainty of the The thinner, fine-grained facies is peripheral to the age assignment of the main mass of quartzite in that coarser facies and has been mapped in the eastern, west­ mountain. Subsequent mapping proved the wisdom of ern, northern, and southern extremities of the region, this restriction, for the main mass of the Serra do as illustrated in figure 8. The fine-grained facies is less Caraga is now assigned to the Tamandua Group. The than 100 m thick but, because of the obscure outcrops, Caraga Group is present in the area cited in 1957, and could not be located with certainty in many places and the type locality may be maintained there. The best, may average less than 50 m in thickness. Ir the western most typical, and most accessible exposures of this part of the Serra do Curral, it is between 30 and 70 m group are in the Serra da Moeda and in the thinner, thick according to Simmons. In the eastern part of the platform fades to be described, in the Ouro Preto region. Serra do Curral, it cannot be more than a few tens of At least two geologists on this project believe that the meters thick and locally may be absent. In the east end rocks of the Tamandua Group should be included in the of the Gandarela syncline it was mapped with the Ba­ Caraga Group. As will be demonstrated, the rocks of tatal Formation and may be a few tens of meters thick or the Caraga Group as here defined are blanket sediments, less. In the Ouro Preto area, it averages perhaps 50 m whereas the rocks included in the Tamandua Group are thick and may be missing in part of the Dom Bosco lenticular prismatic sediments in the Quadrilatero Fer- quadrangle. However, it thickens to the nor^h in the Sao rifero. The different environment of the two divisions Bartolomeu quadrangle. Around the Serr^, do Caraga and the fact that an erosional interval possibly separates the fine-grained facies crops out only sporadically and them seems to the writer to justify the preservation of is only 10 m to a few tens of meters thick. the divisions as two groups. LITHOLOGY

MOEDA FORMATION The two facies of the Moeda Formation have in com­ mon a dominantly quartzose composition. The lithology In the Quadrilatero Ferrifero the Moeda Formation varies with the thickness, and the two facies will be occurs in two intergrading facies. The more prominent described separately. They are intergradational along is a thick coarse-grained quartzite that crops out strike. boldly; a thin fine-grained facies that crops out ob­ The coarse-grained facies is composed of quartzite, scurely is at least as extensive. In part of the region it grit, conglomerate, and phyllite. The rock types are was impossible to separate the thin facies from the over­ commonly well differentiated, and some of the quartzite lying Batatal Formation because of poor exposures; in approaches orthoquartzite, although nearly all contains such areas the two formations were mapped together as varying and locally large quantities of ser;oite or mus- Caraga Group undivided. To the west, north, and east covite. Both quartzite and grit have been sheared and the transition from the thick coarse-grained to the thin crushed to a microaugen texture in many localities, and fine-grained facies is relatively abrupt, occurring over the larger grains are there very angular. W ere original 3-10 km. This is perhaps best illustrated in the Ibirite constituents of the rock can be distinguished, grain size quadrangle, where the transition from a thickness of can be seen to change rapidly and widely both along perhaps 500 in to about 100 m or less takes place in and across the strike of the beds. about 3 km. In the Bagao quadrangle the diminution is The grit and quartzite both contain widely varying slightly less abrupt but better exposed. At the south end of the Serra da Moeda the thinning is much more quantities of iron oxides, in many places concentrated gradual, however. on shear planes and bedding planes; it is impossible to be certain how much of the iron has beer moved and DISTRIBUTION AND THICKNESS recrystallized during the metamorphism and local meta­ The coarse-grained facies of the Moeda Formation somatism to which the rocks have been subjected. For crops out in a roughly triangular area with the base in the same reason, it is most difficult to be certain how the Serra da Moeda and adjacent parts of the Serra do much of the feldspar found in the rock is original detri- Curral and the apex to the east in the east end of the tal feldspar; this mineral is most abundant in proximity Gandarela syncline and the Serra do Ouro Fino. Within to granitic rocks, as is the locally abundant tourmalinite PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A35 r

5

00 ft 00 O> >a ^s

(/) O

3(U ,_,8 O2 ro§ E, V 2 A 8 i c ! o I

\ V A36 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL and tourmaline (Guild, 1957). Quartzite and grit are by far the most common rock types in the coarse-grained f acies of the Moeda Formation. Conglomerate is a characteristic rock of the coarse­ grained f acies. Basal conglomerate containing angular and squeezed fragments of the underlying Nova Lima Group phyllites together with well-rounded pebbles, frt cobbles, and boulders of vein quartz and quartzite of unidentifiable source, and, rarely, amphibolite, is excel­ lent evidence for the unconf ormable nature of the con­ tact of the Minas Series with the Rio das Velhas Series. The phyllite fragments occur not only at the contact but as much as 10 m above it. They are not found in the many conglomerate lenses higher in the section, and it is significant that no granite, amphibolite, gabbro, or other similar rocks have been found as cobbles in these higher conglomerates. The basal conglomerate occurs in irreg­ ular lenses ranging from a few centimeters to perhaps 30 m thick and from 10 m to a few hundred meters in outcrop length; most exposures show a thickness of only a few meters. In all the conglomerate beds, the quartz and quartzite cobbles vary widely in dimension (figs. 9 and 10); gen­ erally they are larger where the conglomerate lenses are thicker. The largest boulder seen was 40 cm in maximum dimension; this was in the basal conglomerate in the Serra da Moeda near Macacos N. 1,000, E. 500. Most cobbles are 5-10 cm in length or smaller. The size of the cobbles gradually and irregularly diminishes from west FIGURE 9. Cobbles in basal conglomerate of the Moeda Forma­ tion near N. 1,000, E. 500, Macacos quadrangle. Colonial fort to east across the approximately 40-km-wide area of oc­ in the background was built to protect gold mine in the basal currence of the coarse-grained f acies. conglomerate. With one exception, the conglomerate lenses are ran­ dom in distribution throughout the Moeda Formation. boulder conglomerates are not found; pebble conglom­ In the Serra da Moeda a persistent zone of conglomerate erate is relatively rare, although in some places isolated lenses occurs over a strike length of at least 10 km at the pebbles occur in quartzite. Grit is also an uncommon base of a quartzite zone overlying a very persistent rock. Quartzite is more evenly grained and finer phyllite member in about the middle of the formation. grained, and in many localities is very phyllitic. Phyl­ Phyllite occurs in the coarse-grained facies of the lite in separate beds and members was not found, Moeda Formation in lenses of varying dimension. Where although exposures of this facies are so bad that it may the formation is thickest, in the Serra da Moeda and be present but unseen. In general, the rock types are Serra do Itabirito, these lenses are mappable members less differentiated, but in some areas, as near Ouro as much as 166 m thick (Wallace, 1965), one of which Preto, the quartzite is evenly grained and relatively is known to be continuous over four quadrangles. Where pure. the facies is thinner, as in the east end of the Gandarela PHYSIOGRAPHIC EXPRESSION syncline, the lenses are a few meters to hundreds of The physiographic expression of the Moeda Forma­ meters long. This sericitic phyllite is only slightly tion depends upon its thickness. Where 200 m or more quartzose. thick, it forms scarps with very spectacular and The lithology of the fine-grained facies of the Moeda craggy outcrops (fig. 11). West of Itabirito the forma­ Formation is quite different from that of the coarse­ tion is more than 400 m thick and makes an impressive grained facies although the bulk chemical composition scarp above the adjacent Nova Lima Group. Farther of the rocks may be very similar. Except for the south along the Serra do Itabirito, the Moeda Forma­ presence of basal conglomerate lenses indistinguishable tion thins to less than 100 m; the scarp disappears, and from those in the coarse-grained facies, cobble and the highest rocks along the range are the Nova Lima PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A37

bedding, and foliation all control the detailed erosional features of the outcrops.

BATATAL FORMATION The Batatal Formation (Maxwell, 1958) is the sedi­ mentary unit described and named the Batatal Schist by Harder and Chamberlin (1915, p. 356-357). The formation consists largely of phyllite and includes minor metachert, iron-formation, and graphitic phyl­ lite. It occurs widely but rarely crops out.

CONTACTS In most places the Batatal overlies the Moeda For­ mation with a generally sharp contact, but in the Moeda Plateau the rocks may interfinger locally on a small FIGURE 10. Puddingstone conglomerate in the Moeda Forma­ tion, Serra da Moeda. The conglomerate is typical of discon­ scale (Wallace, 1965) where both formations are well tinuous lenses in the body of the formation. developed. Where both formations are thin and the Moeda Formation is argillaceous, the two are inter- Group, there indurated by contact metamorphism. gradational and may not be separable. The contact is Spectacular scarps 100-300 m high and some 40 km conformable wherever seen. The Batatal Formation is long in the Serra da Moeda bordering the valley of the in conformable and sharply gradational contact with Paraopeba result from the relatively greater resistance rocks of the overlying Itabira Group. to erosion of the Moeda Formation compared with that of the subjacent granitic rocks. DISTRIBUTION AND THICKNESS Where the Moeda Formation is thin and phyllitic, as The Batatal Formation is easily weathered and non- along the north side of the Gandarela syncline, it crops resistant and lies between two resistant ridge-forming out very poorly, if at all. Where thin but consisting of formations; therefore, it is not well exposed. Its well-indurated and relatively pure quartzite, as along presence is revealed by subsurface workings, by gullies, the Serra do Ouro Preto, it forms ridges and exten­ and by smooth, normally convex, slopes which support sive dip slopes from which overlying units have been a depauperate vegetation of stunted trees and sparse stripped (pis. 7 and 9). grass. It weathers to a gray to buff or brown soft sapro- The quartzite weathers into bold slabby outcrops lite which tends to be somewhat limonitized at the pitted and scarred by solution cavities, as is typical of surface. all the more massive quartzites of the region. Jointing, The Batatal was identified throughout the central part of the Quadrilatero Ferrifero and is present, but was not mapped separately from phyllites in the Moeda Formation in the Congonhas area. Along the south bor­ der of the region it does not seem to be present in some localities east of Dom Bosco. Good exposures in canyons and gullies have not revealed a mappable thickness in the Ouro Preto area. In the western Serra do Curral, Simmons (1968a, p. G13) reported that its thickness is consistently about 30 m. The formation is traceable eastward to the Nova Lima quadrangle; thence inter­ mittently eastward to the Serra da Piedade quadrangle. Outcrops are poor in the east end of the Serra do Curral, and the formation may well be essentially con­ tinuous although thin. In the central part of the Serra do Curral, it thickens to more than 100 m in some areas, FIGURE 11. Typical topographic expression of the coarse­ notably to the southeast and east of Pico Belo grained facies of the Moeda Formation in the Serra das Horizonte. Gaivotas, Ibirite quadrangle. The even ridge in background is the Serra da Moeda. showing the Gondwana surface. Low­ Moore, O'Rourke, and Simmons traced the formation land in the foreground is carved on Nova Lima Group rocks. in the Gandarela syncline and found it to thin eastward. A38 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MESTAS GERAIS, BRAZIL Simmons (1968b, p. H15) stated that the formation Breger (Simmons, 1968a, p. G13) extracted soluble is present but thin except along the south limb of the organic material from this rock. Limonite pseudo- Gandarela syncline at its east end; on the north limb morphs after pyrite are seen in graphitic phyllite in it is 9 m thick near the west border of the Santa Bar­ places. Thin discontinuous lenses of iron-formation also bara quadrangle. In much of the Monlevade area it is are known, and the phyllite may be very hematitic 20-50 m thick, but in the southern part it could not be in the upper few meters of the formation. Magnetite separated from the Moeda (Eeeves, 1966). Maxwell crystals are common in these uppermost beds. mapped the formation along the east side of the Serra do Caraga and intermittently along the south side, but ENVIRONMENT OF DEPOSITION it was not found farther south in the Santa Rita Durao The Caraga Group consists of the Moeda Formation : quadrangle. an extensive quartzose formation, and the Batatal For­ Thus, the Batatal Formation ranges in thickness mation, an argillaceous formation. The Moeda Forma­ from a few meters to more than 200 m. It is thickest tion occurs in two facies: a thick coarse-grained in the Serra da Moeda. It is persistent across the central orthoquartzite and protoquartzite, and a thinner, part of the region but disappears or thins markedly fine-grained facies. The transition between the two to the northeast and to the southeast. facies is in most places fairly abrupt. The thicker facies, averaging perhaps 300 m in thick­ MTHOLOGY ness but much thicker in some parts of its outcrop, is The Batatal Formation is sericitic phyllite. Particu­ characterized by coarse grain size, abundant conglom­ larly in the central part of the region, a field test for erate lenses, and well-differentiated phyllite lenses and the rock is the absence of grittiness between the teeth. members. This facies is typically developed in the east Where both the Batatal and the underlying Moeda For­ and west sides of the Moeda Plateau, where it has a mation are thin and poorly differentiated, the Batatal north-south extension of some 40 km. Eastward the may also contain significant amounts of quartz. Locally coarse-grained facies narrows in outcrop width and the formation contains significant amounts of chlorite, thins stratigraphically. East of the westernmost parts graphite, and carbonaceous material. of the Gandarela syncline and of the Serra do Ouro Lenses of pure white fine-grained equigranular quart- Fino, it is found only in a small part of the Monlevade zite are known in an adit in the Serra do Curral near quadrangle. Jangada (Ibirite quadrangle); at about N. 500, E. The thinner facies, which averages less than 100 m 4,500, in the Belo Horizonte quadrangle; at N. 8,950, in thickness, is characterized by finer and more even E. 1,900, in the Lagoa Grande quadrangle; in tunnel F grain size, by mixing of the argillaceous fraction with in the Rio do Peixe canyon, Lagoa Grande quadrangle; the quartzose fraction of the formation, and by the at N. 12,550, E. 11,700, Lagoa Grande quadrangle; and lack of conglomerate lenses. This facies has been mapped in an exploratory adit in Conceic.ao Peak, Itabira dis­ in most of the region peripheral to the area underlain trict. The occurrence in the west side of the Lagoa by the coarse-grained facies. Grande quadrangle, which is on the road to Piedade de Both facies are characterized in many localities by Paraopeba, is the most accessible. Such lenses, inter­ a basal conglomerate made up of fragments of phyllite preted as lenses of metachert, are fortuitously revealed derived from the underlying Nova Lima Group phyl- by cuts, gullies, and exploratory workings; they may be lites and well-rounded vein-quartz and quartzite peb­ more common than now known, as they crop out only bles and cobbles. The lenticular basal conglomerate in unusual topographic situations. probably represents filling of ancient shallow valleys carved in the older rocks. The distribution of the coarse In tunnel F, Lagoa Grande quadrangle, such meta­ facies, and particularly the size distribution of the cob­ chert grades upward into iron-formation (Wallace, bles and boulders in the conglomerate lenses in that 1965). Along the road from Gongo Soco to the Caete- facies, strongly indicates that the source of the material Barao de Cocais highway, a thick zone consists of in- was to the west and that the coarse facies was a deltaic terbedded metachert and phyllite; bands are a centi­ deposit; the finer facies represent poorly sorted silts meter or so thick, are fairly well separated, and alter­ deposited in quieter water around the delta. nate from one rock type to the other. Phyllitic bands The Batatal Formation is nearly coextensive with the contain some hematite, and the sequence grades upward Moeda. In the Quadrilatero Ferrifero it ranges from a into the Caue Itabirite. quartz-free argillaceous rock where it overlies the thick­ The Batatal Formation also contains dark-gray to er facies of the Moeda Formation, to an argillaceous nearly black graphitic and carbonaceous phyllite. I. A. rock with small amounts of free quartz where that for- PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A39 mation is thin and itself somewhat argillaceous. The ITABIRA GROUP Batatal varies directly in thickness with the Moeda, The Itabira Group was so named by Dorr, Gair, being thinnest to the south and east. Pomerene, and Kynearson (1957) and comprises: the Harder and Chamberlin (1915) were the first to point itabirite included in the Itabira Formation of Hrrder out that the quartzose sediments are transgressive over and Chamberlin (1915), as well as closely related dolo- an old peneplaned surface; it seems probable that the mitic rocks in part included by Harder and Chamberlin transgression was from west to east. The Moeda Forma­ in their Itabira Formation and in part probably in­ tion preserves the transition from paralic to stable shelf cluded in their Piracicaba Formation. The older de­ environments in its transition from coarse to fine quartz­ scriptions are rather vague as to exactly what rocks were ose sediments. The argillaceous Bafcatal may be at least included in specific formations. The Itabira Group is in part contemporaneous with part of the Moeda, an separated from the Caraga Group because it is domi- offshore sediment deposited on a slowly sinking wide nantly composed of chemical sediments whereas the shelf or platform. The presence of lenticular masses of Caraga Group is dominantly composed of clastic sedi­ metachert and the gradational contact with the overly­ ments. In 1958 the Itabira Group was divided by the ing chemical sediments of the Itabira Group also point writer into two intergradational units: the Caue toward this environment. Both formations are blanket Itabirite, a metamorphosed oxide-facies iron-formation, formations, having great lateral extent in comparison and the Gandarela Formation, in large part composed with their average thickness. of carbonate rocks of various types. Because the two formations are intergradational, and METAMORPHISM particularly because the weathering products hide, the The Moeda Formation, having a bulk composition contact zone in most places, it is difficult or impossible not sensitive to metamorphic effects, contains few and in certain areas to separate the two formations. In 18 simple minerals and does not vary widely with changing quadrangles they were separated; in 10 they were metamorphic environment. Argillaceous minerals were mapped together as Itabira Group undivided. This dif­ changed to sericite and muscovite, their grain size vary­ ficulty led Simmons (1968b, p. H15) to suggest that the ing directly with the intensity of metamorphism, which rocks should be considered a single formation; other increases to the east. Locally, biotite is formed at the geologists do not concur because the units can be sepa­ appropriate metamorphic grade. rated in most places, because the economic products The pelitic Batatal Formation, however, is more com­ from each of the units are different and it is therefore plex in its metamorphic reactions. In the Monlevade important to make the separation where possible, and district, much of the rock was converted to a laminated because the difficulty in separation is caused moro by quartz-muscovite schist containing as much as 10 per­ weathering products than by lack of distinguishing cent almandine garnet (Beeves, 1966). In the Itabira features in the fresh rocks. district garnets also are found in this unit, although The Itabira Group is the most important in the region the rock is a phyllite rather than a schist. Further to the from an economic point of view. It contains great re­ west this mineral has not been observed in the Batatal. serves of high-grade iron ore of metasomatic origin and Graphite and carbonaceous matter in certain zones of weathering of the Caue Itabirite has produced enormous the rock inhibited recrystallization. Such zones look like reserves of lower grade but potentially very useful iron argillite where adjacent formations show the grade of ore. Much manganese ore has been produced from both regional metamorphism to be well up in the greenschist formations of this group, different in nature and ure in facies. each formation. Tens of tons of gold were extracted Guild (1957) and O'Rourke (written commun., 1954) from ores in the Caue Itabirite during the 19th century. noted the presence of kyanite in this formation as well The element palladium was first isolated by W. H. Trol- as in the Moeda Formation in some localities. Both laston from palladic gold from the Gongo Soco deposit Guild (1957) and Dorr and Barbosa (1963) cited the (Henwood, 1871). Bauxite is derived from the Gan­ locally abundant tourmaline in the Batatal Formation darela Formation in many localities. Building stone, and, in the Congonhas area, in the Moeda Formation. road metal, paving stone, and fine ornamental stone, as Guild (1957, p. 12) attributed the tourmaline to meta- well as dolomite for metallurgical and refractory pur­ somatic addition of boron from emanations or solutions poses, are extracted from the Gandarela Formation. In from post-Minas granitic rocks; Dorr and Barbosa many localities the Caue Itabirite can be an ample envisaged the same source and related the mineral to the source of underground water, an economic asset as yet fluids that produced hematite ore bodies in the overlying practically untouched. The economic geology of the Ita­ Itabira Group. bira Group will be discussed in another paper; in this A40 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL report the rocks are considered only from a strati- area during early Minas sedimentation^ and the forma­ graphic, structural, and physiographic standpoint. tion may never have been deposited there; both the Moeda and Batatal Formations are also thin r,nd locally CATJE ITABIEITE missing in that area. The type locality of the Caue Itabirite is in Caue Comparison of figure 12 with figure 8 shovs that the Peak in the Itabira district, where the formation is thick areas of maximum thickness of the Moeda Formation and is well exposed by surface and underground work­ and of the Caue Itabirite to a large extent coincide. ings. Elsewhere the formation is poorly exposed except Despite the very different nature of the two formations, in readouts, tunnels, adits, and a few canyons, because this may well indicate the zones of maximum subsidence canga, the weathering product of itabirite, conceals its during early and middle Minas time. lithology and contacts. There is no way of knowing what the original strati- graphic thickness of the Caue Itabirite may have been. DISTRIBUTION AND THICKNESS Because the rock was originally a blanketing oxide- The Caue Itabirite crops out almost everywhere in facies iron-formation, a chemical sediment (Tyler, the Quadrilatero Ferrifero where structurally expect­ 1948) underlain by clay stone and overlain by dolomitic able and was mapped along some 540 linear kilometers rocks also of great lateral extent, wide prirrary varia­ of outcrop. Iron-formation believed by most geologists, tion in thickness is not to be expected, and therefore the including the writer, to correlate with the Caue Itabirite general thickness of between 200-400 m in relatively also has been traced for tens of kilometers to the north undisturbed areas may be considered to approximate on the east flank of the Serra do Espinhac,o, where the original stratigraphic thickness. The only observable typical quartzite-phyllite-itabirite sequence is found. systematic thinning is at the west end of tin Serra do Isolated occurrences of gneissic iron-formation, simi­ Curral, the south end of the Serra da Moeda, along the lar to the Caue where that formation has been highly south margin of the B'agao Complex, and possibly at the metamorphosed, are found in the granitic gneisses for east end of the Gandarela syncline. tens of kilometers to the east and northeast of the Quad­ rilatero Ferrifero. (See map in Harder and Chamber- CONTACTS lin, 1915, p. 347.) Similar rocks are known in the grani­ The contact between the Caue Itabirite and the Bata­ tic gneisses some tens of kilometers west of the region tal Formation is gradational over a few centimeters or, mapped, near Oliveira. Iron-formation has also been in very few places, 1 or 2 m. Commonly the gradation reported some tens of kilometers to the southeast. It is is marked by the appearance of hematite in the Batatal, not known whether these isolated occurrences of highly gradually increasing in quantity toward the Caue. The metamorphosed iron-formation are correlative with the rock becomes banded by the appearance of quartz lay­ Caue Itabirite. ers; and phyllite disappears, generally abruptly. In The Caue Itabirite is quite variable in thickness (fig. some places there may be one or more bands of iron- 12). Where the formation is essentially undisturbed by formation up to tens of centimeters thick, overlain by local deformation, as on the limbs of the Moeda syncline thinner bands of phyllite. Locally the iron mineral near and on the upright limb of the Gandarela syncline, it the contact is coarse euhedral magnetite, particularly ranges from 300 to 500 m in stratigraphic thickness. In where the Batatal is graphitic. Where the Caue Itabirite axial areas of folds such as the west end of the Gan­ overlies quartzose rocks the contact is abrupt. darela syncline, in the Alegria syncline, in the Conta The Caue Itabirite grades upward into the Gandar­ Historia area, in the vicinity of Morro do Chapeu, and ela Formation. This contact is described in more detail at the north end of the Serra da Moeda it may have an in later pages. The contact is in many places indefinite, apparent thickness of more than 1,000 m. In contrast, as the transitional zone may be tens of meters wide. the formation may thin radically on the flanks of tight folds. Near the main church in Itabira, the formation LITHOLOGY AND COMPOSITION on the overturned limb of the major syncline is only The Caue Itabirite is composed of itabirite, dolomitic about 15 in thick, whereas on the upright limb of the itabirite, and ainphibolitic itabirite, with very small syncline near the axial zone in Caue Peak it is appar­ lenses of phyllite and marble. The rock itabirite has ently almost 500 m thick. Clearly the formation was been defined (Dorr and Barbosa, 1963, p. 18) as follows: highly plastic during deformation. The term itabirite denotes a laminated, metamorphosed, oxide- The f onnation does not crop out in its expectable posi­ faeies formation, in which the original chert or jasper bands tion for several kilometers along the south edge of the have been recrystallized into granular quartz and in which the Bagao complex. This area may well have been a positive iron is present as hematite, magnetite, or martite. The quartz PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A41

OS O I

A42 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL bands contain varied but generally minor quantities of iron- As first pointed out by Guild (1957, p. 15), dolomitic oxide, the iron-oxide bands may contain varied but generally itabirite in many places contains magnetite or mixtures minor quantities of quartz. The term should not include quartz- of that mineral and hematite. This is thought to be a ite of clastic origin with iron-oxide cement even though such rocks are sometimes grossly banded. It should include only primary rather than a metamorphic feature of the rock. rocks in which the quartz is niegascopically recognizable as Guild also noted that talc may accompany dolomitic crystalline, in order to differentiate it from unmetamorphosed itabirite; it is believed to be a metamorphic product. oxide-facies iron formation. A certain amount of impurity in As shown by thin-section study, dolomite substitutes in the form of dolomite or calcite, clay, and the metamorphic min­ part for quartz in the quartz bands of the rock and less erals derived from these materials may be included, but these may never be dominant constituents over any notable thickness. generally for the iron minerals, and in soma. places oc­ Where they are, the rock term must be qualified by the use of curs as separate thin bands, forming a triply layered the appropriate mineral name as a qualifier (for example, dolo­ rock. mitic itabirite, a rock in which the dolomite largely takes the A third general type of itabirite is amphibolitic. Field place of the quartz). Rarely itabirite grades into ferruginous chert which, when recrystallized, may look like low-grade ita­ evidence indicates that this type forms by ccntact meta- birite, although commonly it is finer grained and whiter. To pre­ morphism of dolomitic itabirite by nearly granitic vent confusion, a cutoff point of about 25 percent iron should be rocks and thus it does not correspond to the silicate- established. This figure is a practical one, as few itabirites are facies iron-formation of James (1954). Typical local­ so lean in iron and most ferruginous cherts do not contain so much iron. ities may be observed where the Itatiaiugu-Itauna road crosses the west end of the Serra do Curral (Itatiaiugu, Thus, the normal fresh itabirite of the Caue Itabirite is a very simple rock from a petrographic viewpoint, N. 500, E. 1,600), in roadcuts on the east ban]- of the Bio being composed almost entirely of quartz and hematite, Paraopeba (Fecho do Funil quadrangle), and along and locally, magnetite. Such rock makes up by far the Br-31 north of Serra de Piedade (Serra da Piedade, major part of the Caue Itabirite in the Quadrilatero N. 9,500, E. 5,750). At other localities, such as south of Ferrifero. The chemical composition of unweathered Sao Juliao in the quadrangle of that name, such a rela­ normal itabirite averaged (Dorr, 1964, p. 1212) : iron, tion cannot be demonstrated, but at what depth the 37.9 percent; SiO2, 44.7 percent; A12O3, 0.5 percent; nearby intrusive granitic rock may lie is, of course, phosphorus, 0.05 percent; H2O + , 0.3 percent. Assum­ unknown. ing all the iron to be present as Fe2O3 and all the phos­ The amphiboles which have been identified are cum- phorus as P2O5, the total is 99.7 percent, leaving 0.3 mingtonite, tremolite, and actinolite. Most ar? so weath­ percent for other elements, probably largely calcium ered that identification is not possible, but cummington- and magnesium. An average of 11 random samples in ite apparently is the most common. The iror mineral of which FeO was determined, taken in widely separated amphibolitic itabirite is generally magnetite. Five parts of the region by the writer, showed 0.54 percent samples were taken which were fresh enough to make FeO and 55.0 percent Fe2O3, indicating the great pre­ chemical analyses worthwhile. The average composi­ ponderance of hematite. The range of the individual tion, in percent, was: iron, 31.9; SiO2, 50.8; A12O3, 0.7; analyses was not great. CaO, 2.2; MgO, 2.0; phosphorus, 0.08; FeC was deter­ Dolomitic itabirite is widely present in the Caue mined as 13.2 percent. Itabirite, particularly in the upper third of the forma­ It should be emphasized that none of th-^se samples tion. Fresh rock is exposed in few places for it weathers were of completely unweathered rock and tl°,t original to great depth. The variations in chemcial composition magnetite probably was in part oxidized to martite. are still unknown; most of the analyses available were of samples taken in areas of metasomatic enrichment It is believed that only dolomitic itabirite gives rise and therefore cannot be assumed to be of unaltered rock. to amphibolitic itabirite upon metamorphism. Where Moore (1969) cited the following analysis of a sample normal itabirite was intruded by granitic rocks, it did taken in a deep canyon away from areas of known not form significant amounts of amphibole. Grunerite, metasomatism in the Conceigao de Rio Acima quad­ a common amphibole in the Lake Superior area there rangle: Iron, 29.4 percent; SiO2, 47.8 percent; A12O3, thought by some to be a metamorphic product, by 0.5 percent; CaO, 3.2 percent; and MgO, 1.6 percent. others to be primary has not been found deroite search. This is the only analysis of fresh dolomitic itabirite Other iron silicate minerals, such as riebecHte, chamo- sampled away from areas of metasomatic enrichment site, and greenalite, known in iron-formatiors elsewhere known to the writer, despite the many references to in the world, are unknown in this region. Stilpnomelane this rock in recent literature (Pomerene, 1964; Guild, was tentatively identified in two areas but is not 1957; Simmons, 1968a, b). abundant. PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A43

Few other minerals are known in fresh itabirite. Sulfides have been reported only in the vicinity of gold mineralization and were introduced (Dorr, 1965, p. 25) . Small amounts of chlorite have been reported in a num­ ber of localities; the variety penninite (Herz, oral com- mun., 1958) is particularly well developed in itabirite where the Rio Acima-Gandarela road climbs the scarp at the west end of the Gandarela syncline. It occurs in dark-green plates several millimeters long oriented across the bedding of the coarsely laminated itabirite. .Chlorite is more common in dolomitic itabirite. Asbestiform amphibole, too weathered for precise identification, occurs in several localities. Such ma­ FIGURE 13. Sill in itabirite exposed in exploratory adit (mine terial was attributed to the metamorphism of volcanic 1) in Itabira district. bombs by Guimaraes (1953), but we found no evidence of any volcanic material in the Caue Itabirite. Although analyses of fresh itabirite reveal less than crushed vein quartz concordant with the bedding, rather a tenth of a percent manganese in most samples, man­ than a primary sediment. Localities cited elsewhere in ganese oxide is not rare in weathered itabirite and many the Quadrilatero Ferrifero are now interpreted as be­ commercially important deposits of manganese ore are longing to the clastic rocks of the Itacolomi Series known closely associated with itabirite, the largest total­ rather than to the Caue Itabirite (Guild, 1957; Wallace, ing more than 5 million tons. The larger deposits are in written commun., 1961). dolomitic itabirite, but many are in normal itabirite; Weathered beds thought to have been marble are no significant deposits have yet been found in amphib- known in the Caue Itabirite in many areas. Fresh mar­ olitic itabirite. The mode of occurrence of manganese ble occurs in one lens near the Jangada iron-ore deposit in fresh itabirite is not certainly known; presumably (Coelho, oral commun., 1962) ; the thickness exposed it occurs as the oxide but in dolomitic itabirite may well in an adit was about 20 m. In most other known local­ occur as manganoan calcite or manganoan dolomite. No­ ities, such as in the Lagoa Grande manganese mine where was the rock explored deeply enough to reach (Lagoa Grande quadrangle), in the Itabira district, and unweathered material. The manganiferous itabirite oc­ in the Congonhas district, lenses are much thinner, curs in lenses up to several meters thick; individual measured in tens of centimeters, and all are weathered. lenses may be more than 1,000 m long; and, in the Serra Most are in the upper part of the Caue Itabirite. da Moeda, zones containing manganiferous lenses are Thus, the Caue Itabirite consists largely of inter- found for a strike length of more than 10 km. Manganif­ bedded laminae of quartz and hematite, in places con­ erous itabirite most commonly occurs in or close to the taining some dolomite, magnetite, and amphibole and transition zone into the Gandarela Formation but is not occasional lenses of marble and phyllite. The rock is confined to this part of the stratigraphic section. entirely similar in lithology to most major oxide-facies Thin concordant schistose, phyllitic, and clayey lay­ iron-formations known throughout the world. ers in itabirite are not unusual in some parts of the PHYSIOGRAPHIC EXPRESSION Quadrilatero Ferrifero. In many exploratory adits and roadcuts, these layers can be observed to connect with The Caue Itabirite is the major ridge-forming forma­ crosscutting altered dikes and thus are sills emplaced tion in the Quadrilatero Ferrifero. This is primarily before the close of the metamorphic period (fig. 13) . In because the weathering product, canga, (page A73) is inert to chemical weathering and very resistant to me­ other places such concordant layers are very probably chanical weathering. Canga overlies weathered itabirite, sedimentary layers; near the surface they are so thor­ a disaggregated rock not resistant to mechanical ero­ oughly altered and weathered that the original nature sion. (See fig. 2.) Fresh hard itabirite forms a few of the rocks cannot be deciphered. monadnocks projecting above ridgelines, such as Serra Despite a statement to the contrary by Dorr and Bar- da Piedade and Itatiaiugu, and in such places makes bosa (1963, p. C22), coarse clastic material is not cer­ blocky and craggy mountains. Normally the canga- tainly known in the Caue Itabirite. Further study in the covered ridgelines are smooth and rounded, forming Itabira district made clear that the material is probably ranges continuous for many kilometers and standing a 334-413 O- A44 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL

hundred to several hundred meters above the surround­ variable in detailed lithology throughout the region and ing terrain. Where the parent itabirite dips between is gradational with the Caue Itabirite over a wide zone; 20° and 40°, hogbacks with sharp inface scraps are in some quadrangles it could not be separated from that formed, as in the Serra do Curral; where dips are formation or was not present. steeper, more symmetrical ridges are formed except where the canga cap is broken by erosion. In these places DISTRIBUTION AND THICKNESS blowouts caused by rapid removal of the softened The Gandarela Formation was mapped in most of itabirite leave craggy needles of hard itabirite project­ the Serra da Moeda, in the central part of the Serra ing above the steep surface, as in the Grota de Esmeril do Curral, and in the Gandarela syncline. An amphi- in the Itabira District. Figure 14 illustrates typical land- bolite believed to be the metamorphic equivalent of the forms developed where the formation is excep­ formation occurs in the Monlevade area and was named tionally thick. by Reeves (1966) as the Sitio Largo Amphibolite. The Where the formation is thick or the dip is very low, Congonhas district was mapped before the formation broad ridges result. Where the formation is thin, canga had been recognized, but it is separable in much of that may not form and the formation may be hidden for area. It is not separable in most of the area immediately considerable distances by soil and slope wash. In such east and south of the Serra do Caraga (Maxwell, written places the soil gives no hint of the rock below; the commun., 1962) or in the Itabira district (Dorr and formation can be located by using a magnetometer or Barbosa, 1963). In some of these places the rocks are dip needle. known to contain lithologies typical of the Gandarela GANDARELA FORMATION Formation in the expected stratigraphic position, but The Gandarela Formation was first clearly distin­ they are so thin or so poorly exposed that they cannot guished by O'Rourke (written commun., 1954) and by be mapped. Pomerene (1964, p. D15) and was named by Dorr The thickness of the Gandarela Formation is variable. (1958b) from the type locality in the vicinity of In the type locality, the formation seems to be about Fazenda Gandarela in the quadrangle of that name. 750 m thick. In the Serra da Moeda, it is about 500 m In that area the formation is thick, and deep erosion has thick at the north end and about 200 m at the south produced excellent outcrops. Elsewhere the formation edge of the Marinho da Serra quadrangle. In the Serra is very poorly exposed and was mapped on the basis of do Itabirito, the formation thins from about 400 m its weathering products and landforms. These are fairly near the Codornos Dam in the Lagoa Grande quad­ distinctive to geologists with some mapping experience rangle to a knife edge in the Bagao quadrangle at N. in the region, but it may be difficult for others to delimit 1,800, E. 300 (Wallace, 1965). Pomerene (1964, p. D16) the formation with confidence. The formation is quite measured a thickness of 400 m in the Serra do Curral at Acaba Mundo and believed the formation is of about the same thickness in the Macacos quadrangle. The formation thins rapidly to the east and was not found in the Nova Lima quadrangle and only sporadically in small discontinuous outcrops to the east in the Serra do Curral; the abruptness of this thinning is in part tectonic. To the west along the Serra do Curral, the Gandarela thins to about 150 m in the Rio Paraopeba water gap (Fecho do Funil quadrangle). The forma­ tion wedges out complete!}- 1.4 km west of the east boundary of the Igarape quadrangle. Perhaps the formation is thickest in the west end and north limb of the Gandarela syncline in the Gandarela, Caete, and Gongo Soco quadrangles. In the western part of the Gongo Soco quadrangle an apparent thickness of 900 m was calculated from mapping by Moore. In the central part of the same quadrangle near the present FIGURE 14. Caue Itabirite between Alegria and Conta Historia, village of Gongo Soco, the formation was calculated to Capanema quadrangle. Gentle slopes below are protected by a be about 600 m thick. One kilometer north of Barao do thin canga cover, breached by the small stream. The formation here has an apparent thickness of more than 1,000 m. Photo­ Cocais, Simmons (1968b, p. H18) measured a section graph by Norman Herz. about 70 m thick. PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A45

The south limb of the Gandarela syncline is strongly lows. The weathering product of dolomitic itabirite, faulted in many places, and the true thickness of the where not hidden by canga, is more hydrated than that formation is there unknown. In the Gongo Soco quad­ of normal itabirite. It is ocherous, tends to be yellow rangle a broad plateau marked by three large sinkhole instead of dark brown, may be porcellaneous witt an lakes has developed on the outcrop belt. Farther east in aluminous aspect although chemical analyses (Pomer- the Santa Barbara and Cocais quadrangles the forma­ ene, 1964, p. D20) showed little more A12O3 than the tion was estimated by Simmons to range between 40-70 normal weathered itabirite in such material and may m in thickness. be magnetic from the higher original magnetite content In the Ouro Preto-Mariana area, as calculated from of dolomitic itabirite. Magnetite octahedra may be visi­ mapping by A. L. M. Barbosa, the Ganda>rela Formation ble. The presence of amphibole replaced by limonite is is on the order of 75 m in thickness; to the north in the a sure sign of dolomitic itabirite. The canga overlying region of Antonio Pereira it may be on the order of dolomitic itabirite may be denser and may contain less 200-500 m. Westward along the south edge of the Quacl- quartz and commonly contains more alumina than that rilatero Ferrifero the structure is so complex that no over normal itabirite. Where slopes are gentle, the firm estimate of thickness is possible; Johnson (1962) variety "chemical canga" (Pomerene, 1964, p. D30; estimated a thickness of perhaps 200 m in the northern Dorr, 1964, p. 1224) is common over dolomitic itabirite. exposures in the Dom Bosco quadrangle and more than So many exceptions to all these guides are known that 400 m in the southern. no single one can be blindly followed; field experience is the best guide in separating the Caue Itabirite from CONTACTS the Gandarela Formation. The Gandarela Formation lies with gradational con­ LITHOLOGY AND COMPOSITION tact on the Caue Itabirite and is overlain with local disconformable contact by the Piracicaba Group of the The Gandarela Formation consists largely of dolo­ Minas Series. The disconformity is an erosion surface; mitic and calcitic strata. The most conspicuous rock in no angular discordance is known. the type locality is marble, for the most part, dolomitic. Elsewhere in the region, dolomitic phyllite, dolomitic The contact with the Caue Itabirite is an arbitrary iron-formation, and phyllite may be dominant rock one. The original definition stated (Dorr, 1958b, p. 63) : types. "The contact of the Gandarela Formation with the underlying Caue Itabirite is commonly gradational over In the type locality the marble is red, white, and a few meters to tens of meters. The boundary should gray; O'Rourke (written commun., 1954, and pi. 5) be drawn whe,re dolomitic itabirite or dolomite becomes divided the formation in the Gandarela quadrangle into predominant over normal itabirite."1 It is now known three members on the basis of color. These memHrs that dolomitic phyllite of the Gandarela Formation may were separated one from another by thin but extensive lie directly on Caue Itabirite, although this is not lenses of itabirite, in part dolomitic. In other quadran­ common. gles the marble cannot be so separated, and elsewhere color has no discernible stratigraphic significance; the As stated in the description of the Caue Itabirite, that itabirite lenses are absent or too random in distribution formation may be dolomitic in part, and, where it is dolomitic, the dolomitic itabirite is more common in the to correlate with those in the type locality. upper part of the formation. Similarly, the Gandarela The red marble is generally somewhat higher in iron Formation in many localities contains lenses and lentils content than the other colors, but all types may contain of dolomitic itabirite and normal itabirite. Such lenses several percent of iron or manganese. The iron is pres­ may occur anywhere in the formation but are more com­ ent both as ferroan dolomite and calcite and as firmly mon in the lower part. Were the rocks everywhere visible divided hematite and magnetite. Manganese minerals in fresh outcrops there probably would be little difficulty are not megascopically or microscopically visible in the in seperating them, but unfortunately they are com­ marble, and the element undoubtedly occurs as man- monly covered by canga and are thoroughly leached and ganoan dolomite or calcite (O'Eourke, written commun., hydrated; thus, much uncertainty is introduced. Where 1954; Dorr and others, 1956, p. 303). marble or dolomitic phyllite overlies itabirite, contacts The marble is fine grained and dense, in many locali­ are drawn with little question, for marble weathers to a ties showing strong flow structures and healed breccias. characteristic soil, and dolomitic phyllite is covered by Bedding is obscure in many outcrops. Hundreds of soil quite different from that on most normal itabirite. chemical analyses of these marbles have been publish ed; Where the contact between normal itabirite and these indicate that for the most part the rock is dolo­ dolomitic itabirite is gradational, much difficulty fol­ mite or dolomitic marble but that locally it is pure cal- A46 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL cite marble. Whether the original sediments were related only on the basis of the general stratigraphic dolomite or calcite is not known; the irregular distri­ succession. bution of calcite marble would suggest that dolomitiza- Johnson (1962, p. B15) described fresh grayish-green tion occurred. If so, this must have taken place before dolomitic phyllite in the Dom Bosco quadrangle as or during rnetamorphism, as the rock is now dense and being composed of sericite or chlorite with some dolo­ equigraiiular. The widespread presence of talc also sug­ mite, enclosing lenses and streaks of granular quartz, gests that magnesium-bearing solutions once permeated with much magnetite and some epidote. Bedding is dis­ the rock. No vestige of organic life now remains, tinguished by color banding and may be either essen­ although the abruptly lenticular nature of some marble tially parallel to or may crosscut foliation at a high lenses suggests either giant boudinage or original angle. He also described greenschist of possible vol­ bioherms. canic origin in two outcrops in the same area. These are An unusual lithologic zone in the Gandarela Forma­ the only known occurrences of such rock in the Gan­ tion marble is a sharpstone intraformational conglom­ darela Formation in the region and wer*. also men­ erate composed of tabular fragments of metachert and tioned by Guild (1957, p. 48-49). This material occurs dolomite in a dolomite matrix. This zone ranges from in bands a few tens of centimeters thick and consists a few centimeters to more than 1 meter in thickness dominantly of chlorite, which constitutes as much as and has been located in the Sao Juliao, Dom Bosco, 80 percent of the rock, with quartz, biotite, gud euhedral Lagoa Grande, Belo Horizonte, Gandarela, and Gongo magnetite. Sooo quadrangles. The rock was well described and il­ The basal beds of the Gandarela Formation are dolo­ lustrated by Guild (1957, p. 16). Outcrops are too scat­ mitic itabirite in much of the region. In some parts of tered to indicate whether the sharpstone conglomerate the region, notably in the Serra do Curral in the Nova occurs at one horizon or at several horizons in the Lima, Ibirite, and the eastern part of the Fe-Vho do Funil marble beds; if further work proves it to be confined quadrangles and in the quadrangles immediately south to a single zone, it would be a most useful marker in of the Serra do Caraga, substantial parts of the whole regional correlation. Wherever observed, it was in the formation may be composed of dolomitic itabirite, and upper part of the Gandarela Formation. the marble facies of the formation may be very sub­ Pyrite is a widespread but sparse constituent of the ordinate or absent. Elsewhere the Gandarela Formation marble and is associated with sparse chlorite in some characteristically contains lenses and bands of dolomitic exposures. Talc is also a common associated mineral; it itabirite and, more rarely, normal itabirit"}. Most such occurs on shear and bedding planes and, to a lesser ex­ lenses are small, ranging in thickness from a few centi­ tent, disseminated in the marble. Phlogopite was re­ meters to a few meters and in length from a few meters ported by Moore (1969) in the Gongo Soco quadrangle, to a few hundred meters; but at the west end of the and amphibole was noted by Guild in the Congonhas Gandarela syncline, O'Rourke (written commun., 1954) area. The marble is host for various introduced sulfied mapped two outcrop bands which he traced for more minerals, notably the copper sulfieds, but these have than 10 km with few interruptions (pi. 5). not been found in commercial quantities. A small de­ Simmoiis (1968b) pointed out that in the Barao de posit of stibnite and its oxidation products was found Cocais area, dolomitic itabirite is much more coarsely near Belo Horizonte, presumably derived from an epi- banded in the Gandarela Formation than in the Caue genetic deposit in marble (I. S. Coelho and Luis de Itabirite. Guild (1957, p. 49) earlier reported a similar Castro, oral commun., 1965). Fluorite lenses were re­ relation in the Congonhas district. ported by Pomerene (1964, p. D18) near Acaba Mundo Pomerene (1964, p. D20) described a "dolomite iron­ in the Belo Horizonte quadrangle and by Johnson stone" in the lower part of the Gandarela Formation (1962, p. B15) near Dom Bosco. that consists of grossly banded siliceous limonite and In much of the region, particularly in the southern some hematite and magnetite. He believed it represents part, phyllite and dolomitic phyllite 2 are important limonitized and silicified weathered dolomite. rock types in the Gandarela Formation. Dolomitic PHYSIOGRAPHIC EXPRESSION phyllite weathers rapidly and deeply to a saprolite simi­ lar in appearance to, but somewhat more spongy than, The Gandarela Formation stratigraphically overlies that formed from other phyllites. Such rock can be cor- the most resistant formation in the region and underlies a ridge-forming formation. Thus, its physiographic ex­ 2 In this weathered terrain, identification of carbonate minerals in the outcrop of complex rocks is usually impossible. Because dolomite is pression is to some extent controlled not only by its own the most common carbonate in those few places where determination is lithology but also by the structure, thickness, and atti­ possible, geologists of the DNPM-USGS team arbitrarily considered all phyllite containing carbonate to be dolomitic phyllite. tude of adjacent units. Almost everywhere it lies at PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A 47

lower elevations than the stratigraphically underlying itabirite on trend with the south end of the Serra da Caue Itabirite, and, in such localities as the west end of Moeda some 40 km southwest of the Quadrilatero Ferri­ the Gandarela syncline, drainage follows the contact as fero. The formation thins radically to the northwest at a preferred zone. Where the marble is thick and over­ the west end of the Serra do Curral. On the east snd lying units thin or soft, drainage follows the upper north sides of the Quadrilatero Ferrifero, however, no contact, as the Rio Socorro in the central part of the signs of a wedgeout are known. The structure of the Gandarela syncline. Where marble predominates and the south side is too complex to permit judgment as to sequence is upright, the land surface is hummocky and probable continuity southward; in the Ouro Preto^- marked by isolated craggy outcrops. Where the rock Mariana area the formation is prominent and persistent. sequence was rotated well beyond the vertical, as in part Thus, the Caue Itabirite can be classified as a sheet or of the Serra do Curral, the topographically overlying blanket deposit (Pettijohn, 1957, p. 618). Caue Itabirite forms great scarps which, except in a The Gandarela Formation is not known to be as few localities of rapid mechanical erosion, supply debris extensive as the Caue Itabirite; however, it is kno^n that conceals the topographically lower rocks. Where for a distance of 130 km from east to west (assuming the formation is nearly vertical, scattered outcrops may the equivalence of the Sitio Largo Amphibolite with be found in stream valleys and at the base of cliffs in the Gandarela Formation) and 65 km from north to the Caue Itabirite. Along the inf ace slope of the Moeda south; it, too, fulfills Pettijohn's definition of a blanket Plateau, the Gandarela Formation is covered by soil deposit. The formation has not yet been recognized and canga and is marked by a gentle slope and small outside the Quadrilatero Ferrifero, but, because the local relief. Where the formation is dominantly phyl- outcrops are obscure and the formation has been recog­ litic, as in the Dom Bosco quadrangle, it is indistin­ nized as a unit within the Quadrilatero Ferrifero only guishable topographically from the other phyllitic in recent years, it may be present. Simmons (1968a) formations of the region. considered the erosional disconf ormity at the top of the Sinkholes develop on the marble, and locally lakes Gandarela to be responsible for its absence at the w^st of some size are found in them. The largest are the end of the Serra do Curral; this might be the situation Lagoa das Antas, on the south limb of the Gandarela elsewhere. syncline, which is some 500 m in diameter, and the All formations of the Carac,a and Itabira Groups r.re nearby Lagoa das Coutas, which is more than 600 m intergradational blanket deposits, and the two forming long Goiigo Soco quadrangle). Other smaller lakes oc­ the Caraga Group are normal transgressive stable-shelf cur near Pico de Itabirito and near Fazenda Gandarela. deposits. The Gandarela Formation is also a typical off­ Shallow undrained depressions without permanent shore stable-shelf deposit, and thus it seems probable water are known elsewhere, as at Lagoa Seca south of that the Caue Itabirite, too, was deposited on a staH« Belo Horizonte. shelf. Thus, the topographic expression of the Gandarela The worldwide occurrence of banded iron-formation Formation is distinctive and may be used with confi­ in the Precambrian and the almost complete abseroe dence for mapping, although the characeristically poor of this rock type in post-Algonkian sediments has long exposures prevent accurate location of its contacts and been known. No completely satisfactory hypothesis of soil and canga conceal the greater part of the unit. the genesis of this rock has yet been advanced, and the

ENVIRONMENT OF DEPOSITION problem of source of the enormous quantities of iron involved (for the Caue Itabirite, some 5X1012 tons of The Caue Itabirite once covered almost completely elemental iron) remains enigmatic. As shown above, the 7,000 square kilometers of the Quadrilatero Ferri- there is no internal evidence that the Caue Itabirite and fero, 155 km from east to west and 100 km from north the Gandarela Formation were deposited in other than to south. If itabirite near Serro and Morro do Pilar a stable-shelf environment or that they were related (towns north of Itabira) was once continuous with that of the Quadrilatero Ferrifero, as believed by the writer to island arcs or volcanism, were estuarine or fresh­ and other geologists (Harder and Chamberlin, 1915; water deposits, or, in fact, were other than nonral O. Barbosa, 1954), the maximum known north-south epicontinental marine deposits. Thus, it is here assumed extent was another 100 km. There is good reason to that they were deposited as chemical or perhaps bio­ think the formation may have continued at least 200 km chemical deposits in a normal marine environment. more to the north. The formation seems to thin to the The origin of iron-formation is discussed in more detril southwest near Jeceaba, but John Forman (oral elsewhere, where the various types present in tl is commun., 1965) reported highly metamorphosed region are examined as a group. A48 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL The Gandarela Formation, with its intercalations of PIRACICABA GROUP original claystone, now phyllite, and intraformational The Piracicaba Group was named by Dorr, Gair, sharpstone conglomerate in the dominantly carbonate Pomerene, and Rynearson (1957) and is roughly equiv­ sediments, appears to have been deposited in epiconti- alent to the Piracicaba Formation as defined by nental seas which in the central part of the region were Harder and Chamberlin (1915, p. 362-363). The latter, free from fine detrital sediments. In the southern part however, may have included part of what is now called of the area considerable quantities of fine detrital the Gandarela Formation in their Piracic^ba Forma­ material were deposited. The single, very local occur­ tion, for at that time the erosional disconformity at the rence of rock of possible volcanic origin emphasises that top of the Gandarela Formation had not been discov­ strong contemporaneous volcanism was not prevalent ered ; moreover, it was not known that units with sim­ in the region. ilar lithologies occur in both the Itabira and the Contemporaneous sedimentation in an environment Piracicaba Groups. The type locality of the Piracicaba of overall marine transgression of the Caraga and Formation chosen by Harder and Chamberlin is in Itabira strata in different parts of the area is possible. the headwaters of the Rio Piracicaba soutl west of the The Moeda Formation would represent nearshore Serra do Caraga, where it is now known that much of deposition, the Batatal Formation offshore deposition, the group has been removed by erosion. The varied and and the Caue Itabirite and Gandarela Formation persistent lithologic units comprising the Piracicaba deposition well offshore. The transgressive sea would Group, so well exposed in the Serra da Moeda and on produce an on-lap sequence. Oscillations and regres­ the north side of the Serra do Curral, were not studied sions would be indicated by phyllitic zones in the in detail by Harder and Chamberlin. dominantly quartzose Moeda Formation, by metachert It seems very probable that Harder and Chamberlin and minor hematite in the dominantly phyllitic Batatal considered the ferruginous quartzites here included in Formation, by rare shale beds and dolomite lenses in the Cercadinho Formation of the Piracicaba Group to the Caue Itabirite, and by itabirite lenses and zones in be part of their Itabira iron-formation, for they refer the dolomitic Gandarela Formation. Elements to prove to rounded quartz grains in itabirite. Careful search or disprove this possible contemporaneity are missing during this mapping program revealed no single in­ because of the absence of throughgoing marker beds stance of rounded quartz grains in the rocl- here called or fossils; in any one locality, the formations are the Caue Itabirite, whereas such material is very com­ systematically superposed, and the lower formations mon in, if not characteristic of, the ferruginous quartz- are there older than the overlying formations. ite of the Cercadinho Formation. Because Harder and All formations of the Carac,a and Itabira Groups in Chamberlin considered the iron-formation to be a clastic the Quadrilatero Ferrifero are thickest in a zone extend­ deltaic formation rather than a chemical precipitate, ing across the middle of the region in a roughly east- the presence of rounded quartz grains in iron-formation west direction and curving northward on the east side. seemed a congruent rather than exotic occurrence, and This may indicate that the area was the site of major they had no reason to reexamine their stratigraphic con­ introduction of sedimentary materials and of sub­ clusions because of the presence of clastic auartz. sidence, particularly in Caraga time. The Piracicaba Group overlies the Itr.bira Group METAMORPHISM with structural conformity but erosional disconformity Metamorphism of the Caue Itabirite produced a in the west side of the Quadrilatero Ferrifero. This radical increase in grain size with increasing meta- erosional disconformity has not been found east of the morphic grade. Specific data on this variation through­ Moeda Plateau and the vicinity of Sabara in the Serra out the region was summarized by Dorr (1964, table 2) do Curral. It is quite possible that in the east side of and Herz (written commun., 1964) and show that such the region the group may be in conformable and gra- variation, although qualitatively similar to that dational contact with the older rocks, for in the one described by James (1955) in the Lake Superior region good exposure in the Itabira district the contact was of the United States, is anomalously greater. The gradational, and the few other exposures of the actual increase in grain size has important economic effects; contact known in the eastern part of the region do not supergene enrichment of the iron-formation is rather give critical evidence. closely controlled by this phenomenon. The formations of the Piracicaba Group consist Amphibolitic itabirite is believed to be a product of largely of clastic sedimentary rocks, ranging from fine the contact metamorphism of dolomitic itabirite. Meta­ conglomerate through quartzite to phyllite and gra­ morphism completely recrystallized the carbonate rocks phitic phyllite; chemical (or biochemical) precipitates of the Gandarela Formation. included in the sequence are sporadic beds of dolomite PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A49 and small lenses of iron-formation. Most formations cause the formation had not yet been identified at the contain relatively simple and well-differentiated mature time mapping was done there. The distinctive litholo- sediments. All are now metamorphosed, most to the gies are present in all three of these quadrangles, how­ greenschist f acies and in the east side of the region to ever. Thus, the formation extends for a minimum of the almandine-amphibole f acies. about 100 km from north to south and 150 km from east The formation at the top of the group consists of to west. The formation has not yet been surely identified graywacke, chloritic .schist and phyllite, tilloid, and outside the Quadrilatero Ferrifero although rock of sandy schists, with small lenses of iron-formation and similar lithology crops out above itabirite about 10 km possibly some chert. The lithology of this sequence dif­ east of Morro do Pilar, about 75 km north of the Quad­ fers substantially from that of the underlying forma­ rilatero Ferrifero. tions, and its thickness is greater than that of all the The Cercadinho Formation varies widely in thick­ others combined. ness. At the west end of the region, it is invaded by In the Monlevade area, all rocks are more highly granite and is only about 30 m thick. It thickens east­ metamorphosed than elsewhere in the Quadrilatero Fer- ward to about 80-120 m in the Fecho do Funil quad­ rifero, and stratigraphic units believed correlative with rangle and, at the type locality in the Belo Horizonte the Piracicaba Group are altered to gneiss named by quadrangle, Pomerene (1964, p. D20) measured a sec­ Reeves (1966) the Elefante Formation. It contains a tion 317 m thick; 3 km to the southwest the formation Pantame Member believed to correlate with the Cer- is only 80 m thick (Pomerene, 1964, p. D20). Gair cadinho Formation, and a Bicas Gneiss Member pos­ (1962, p. 38) found the thickness in the western part sibly the equivalent of the Barreiro or Fecho do Funil of the Nova Lima quadrangle to be about 100 m, but Formation or both. the formation thickens to about 300 m in the rest of that The original thickness of the Piracicaba Group is quadrangle. Farther to the east on the north flanks of unknown because the group is always limited upward the Serra da Piedade, specific data are not available, by intrusive granitic rocks, the present erosion surface, but the formation is very thick. or an unconformity with the overlying Itacolomi Series. To the south, in the Moeda Plateau, the formation The maximum known thickness of the Piracicaba Group varies between 400 and 900 m according to Wallace is in the order of 4,800 m. (1965), who attributed the maximum thickness to

CERCADINHO FORMATION tectonic thickening. In the Gandarela syncline O'Rourke (written commun., 1954) estimated a stratigraphic The Cercadinho Formation, the lowest unit of the thickness of about 300 m at the west end. Farther east Piracicaba Group, consists of interbedded ferruginous Moore estimated (1969) a thickness of 400 m at the quartzite, grit, quartzite, ferruginous phyllite, and phyl­ west side of the Gongo Soco quadrangle; this thins lite. A few small lenses of dolomite are known. The for­ eastward to about 200 m. Still farther to the east the mation was described by Pomerene (1958a) from the formation is squeezed in the syncline, and the upper sequence exposed in the Corrego de Cercadinho in the contact is not known (fig. 15); Simmons found a maxi­ Belo Horizonte quadrangle. The formation is distinc­ mum of 225 m near the east end of this structure. To the tive in its lithology and outcrop pattern, the quartzose south of the Serra do Caraga, Maxwell (written com­ beds are ridge formers, and the formation is a most mun., 1964) estimated the maximum thickness of the useful horizon marker in geologic mapping. formation as about 225 m. In the northern part of the DISTRIBUTION AND THICKNESS Antonio Pereira quadrangle the outcrop of the Cerca­ The Cercadinho Formation was mapped from the dinho Formation is several kilometers wide; cross fold­ westernmost quadrangles along the full length of the ing and relatively low dips make exact measurements Serra do Curral for 100 km in a northeasterly direction. difficult, but the stratigraphic thickness cannot be less It was mapped the full length of the Gandarela syn- than 300 m. cline except at the east end, where it is faulted and CONTACTS squeezed out. Rocks of the same lithology are present Pomerene and Ashley, working in the Belo Horizonte in the east end of the Quadrilatero Ferrifero in the and Nova Lima quadrangles, were the first to find the Itabira district but were not mapped separately owing erosional unconformity between the Cercadinho and to poor exposures. The formation is known from the Gandarela Formation. There it is marked by a basal north side of the Serra do Curral to south of Ouro conglomerate and grit containing fragments of the Preto; it was not mapped separately in the Dom Bosco Gandarela Formation, small hematite chips, and iron- quadrangle because exposures were too discontinuous, rich concretions (Pomerene, 1964, p. D22). The discon- or in the Sao Juliao or Casa da Pedra quadrangles be­ formity was found still farther west in the Serra do A50 REGIONAL, GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL the overlying Fecho do Funil Formation is everywhere concordant and gradational.

UTHOLOGY Ferruginous and nonferruginous quartzite, pebble conglomerate, grit, ferruginous phyllite, "silver phyl- lite," dolomitic phyllite, dolomite, and purple phyllite compose the Cercadinho Formation. These lithologies are interbedded and strongly lenticular ; quartzose rocks in some places are crossbedded. Diversity of sequence and the presence of ferruginous quartzite and "silver phyllite" are the distinguishing features of the formation. Basal beds FIGURE 15. Cercadinho Formation in the Gandarela syncline, The basal beds are quartzite in most places, but grit, Gongo Soco quadrangle. Rocks in the bottom of the valley pebble conglomerate, or phyllite are found elsewhere. and at the top of the ridge are Gandarela Formation. Cliffy Basal conglomerate and conglomeratic beds contain peb­ outcrops on the far side of the valley are Cercadinho Forma­ bles of itabirite and pink metachert derived from the tion, repeated by isoclinal folding. The Serra do Caraga, Gandarela Formation or possibly the Caue Itabirite, formed of Cambotas Quartzite, is in the distance. Photograph by Norman Herz. as well as pebbles of vein quartz and quartzite of in­ determinate origin. Small elongated chips of hematite, with maximum dimensions of 8 by 3 by 2 mm, were Curral by Simmons (1968a), who determined that the reported by Pomerene (1964, p. D22). No dolomite Cercadinho Formation gradually transects the Gan­ pebbles have yet been reported. This material derived darela Formation and the Caue Itabirite. This could be from the Itabira Group is coarser and more abundant caused either by an angular discordance on the order in the west side of the area and was not found on the of less than 3°, or the Gandarela and Caue might wedge east side. out to the west and the Cercadinho Formation overlap. Pomerene (1964, fig. 9) illustrated concretions of The extent of the erosional unconformity so well ex­ specular hematite that occur at the base of the forma­ posed in the Serra do Curral is uncertain because the tion in the Belo Horizonte quadrangle. They are al­ actual contact is visible in few places in the rest of the mond-shaped ovoids as much as 4 cm long, with con­ region. In the Itabira district, a gradational contact centric structure, and commonly contain as a nucleus was observed between the underlying Itabira Group and an angular quartz fragment. All are oriented with their conglomeratic quartzite interpreted by Dorr and Bar- long axes parallel to regional lineation and are thought bosa (1963, p. 22-23) as basal Cercadinho Formation. to represent deformed and metamorphosed limonite con­ Between that area and the Serra do Curral, no strati- cretions formed on a pre-Cercadinho erosion surface, graphic contacts between these rocks have been observed. similar to those now forming on some ferruginous phyl- The units are everywhere structurally concordant, but lites and quartzites in the region (p. A73). in the Gandarela quadrangle and south of the Serra do Quartzite Caraga float similar to the basal conglomerate of the The quartzite beds of the Cercadinho are both ferru­ Cercadinho in the Serra do Curral has been found, in­ ginous and nonferruginous. They range in grain size dicating a probable erosional unconformity in those from grit to fine-grained quartzite, but medium- to areas. On the other hand, no such basal conglomerate coarse-grained quartzite is the most common rock. The was found in the Gongo Soco or Santa Barbara quad­ coarser sizes are most abundant on the west side of the rangles in exposures of the basal beds of the Cercadinho region. Quartz grains within a single bed are generally Formation. The writer believes that epeirogenic uplift well sorted, and the relatively high degree of rounding took place without deformation of the earlier sediments of the grains is one of the diagnostic characteristics of but with a radical change in the sedimentary environ­ this formation. ment. The upper part of the Itabira Group was re­ Ferruginous quartzite crops out more widely in the worked to a varying but generally minor degree, and west side of the region but is known everywhere. The the detritus was incorporated in part in the Cercadinho iron occurs as hematite or, less commonly, as magnetite Formation. or martite, forming the matrix for the quartz grains. Thb contact between the Cercadinho Formation and The iron content varies abruptly in individual beds PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A51 across the bedding and to a lesser degree along strike. grains are alined parallel to the regional metamonhic This variation imparts a grossly banded aspect to the lineation; the intermediate axis is generally parallel to rock in many places. The more ferruginous bands may foliation. contain as much as 60 percent iron, but the average is The origin of these spindle-shaped quartz griins, much less, perhaps in the order of 15-20 percent by found only in the Cercadinho Formation, has provoked weight. The crystalline matrix imparts a dark-gray to much speculation. Pomerene attributed them to stretch­ black color to the rock. For some reason unknown to the ing of original subspherical quartz grains, Simnons writer, the iron minerals of the Cercadinho ferruginous (1968a, p. G17) to shearing of quartz granules. Because quartzite do not hydrate as readily as the same minerals the phenomenon is confined to few and thin lenticular in the other iron-bearing formations. beds in a single formation that is but a small part of a The hematite in the ferruginous quartzite was ap­ stratigraphic section with many beds of coarse quartz parently deposited as very fine grained detrital material grains and granules, all of which have been subjected to derived from erosion of the Itabira Group to the west similar deformation, a primary lithologic control may and, locally, as black sand deposits, because some of the be indicated instead. It is here suggested that the "rice ferruginous quartzite beds are crossbedded and con­ grains" were originally small pellets or concretion^ of glomeratic. The rapid variation in iron content of the silica gel or cryptocrystalline quartz which deformed ferruginous quartzite both normal to and parallel to into uniform shapes and orientation and recrystallized bedding and the gradual diminution in iron content to during orogeny and metamorphism. Conceivably the the east suggest that the chemical or biochemical pre­ uniform size is due to winnowing action of the currents cipitation of iron minerals which was the rule in the when the original pellets were deposited. underlying Itabira Group rocks had ceased during the Phyllite deposition of the sediments under discussion. The most abundant rock in the Cercadinho Formation, Normal quartzite is also a widespread rock in the particularly in the eastern part of the region, is phyllite. Cercadinho Formation. It is interbedded with the fer­ Some of the phyllite, like some of the quartzite, is very ruginous quartzite and is more common than ferrugin­ distinctive and cannot be confused with phyllite from ous quartzite in the eastern part of the region. To the other formations. Like all phyllites in the region, 1 ow- east the percentage of sericite in the normal quartzite ever, outcrops are sparse, and the unit is seen only in increases, the grain size decreases, and the relative quan­ artificial openings and in localities where mechanical tity of quartzite in the formation as a whole decreases erosion has been unusually rapid. markedly. The most distinctive phyllite was given the field rame Many quartzite beds in the Cercadinho Formation are "silver phyllite" because of its silvery silky sheen. The crossbedded, generally at low angles with the true bed­ rock consists of sericite and of hematite in the form of ding but in some places at high angles. Well-preserved very fine grained plates of specularite. Both platy min­ ripple marks have been found in two places, one in the erals are oriented parallel to foliation and the specularite Belo Horizonte, the other in the Lagoa Grande quad­ adds its high reflectivity to the sericite, producing the rangle (Wallace, 1965, fig. 9). Ripple marks are closely distinctive sheen. Hematite is present in varying per­ spaced and indicate fairly shallow water. Graded centages in the phyllite, ranging up to perhaps 50 per­ bedding on a small scale has been seen in several cent; in the higher ranges the rock makes a red streak. quadrangles. Quartz is very subordinate, and the rock was originally A peculiar variety of quartzite of wide distribution a hematitic claystone. The "silver phyllite" is inter- in the Cercadinho Formation was given the field name bedded with both ferruginous and nonferruginous of "rice grit" by Pomerene (1964, p. D23). The material quartzite, forms partings between beds of ferruginous occurs in lenses as much as 1 meter thick but rarely quartzite, and occurs without quartzite in zones as much traceable for as much as 50 m along strike in any one as several meters thick. zone. It is not known whether these lenses occur at one Normal phyllite is also found in the Cercadinho For­ or several stratigraphic horizons; all are toward the mation but in most places is indistinguishable from the top of the Cercadinho Formation. The "rice grit" con­ phyllites of other formations. Much of it is quar^zose sists of spindle-shaped grains of quartz with axial ratios and some is dolomitic. In the foothills north of the $;Q.rra of 1:2 :&-A ; maximum length of grains is about 6 mm. do Curral a considerable thickness of poorly foliated The grains are similar in size and shape to rice grains weathered purple phyllite is exposed in roadcuts. This and make up as much as 50 percent of the rock. Their material, probably originally dolomitic, weathers to a matrix is fine-grained quartzite, locally ferruginous, very weak saprolite that produces pestilential quanti­ locally only slightly ferruginous. All long axes of the ties of impalpable dust on unpaved roads. A52 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL

As well as the minerals cited above, the phyllite of the Formation. These are best developed in kymite-quartz Cercadinho Formation may contain chlorite, biotite, and veins in the phyllite and occur as blue to gray radiating chloritoid. This formation seems to be the only one in crystals as much as 10-15 cm long normal to the walls. the Piracicaba Group that contains chloritoid, accord­ Pyrophyllite is a common accessory miner?,!. Wallace ing to Herz (written commun., 1965). (1965) believed that the alumina in the kyanite was de­ Dolomite rived from the phyllite of the wallrock, as the mineral Small dolomite lenses have been reported in the Cerca­ also occurs as scattered discrete crystals in the country dinho Formation in the Ibirite, Belo Horizonte, and rock. Throughout most of this region, coarse blue kya­ Nova Lima quadrangles, the largest some 20 m thick nite crystals characterize the formation and are not al­ and perhaps 200 m long. Elsewhere this rock has not ways associated with quartz veins. been seen; the writer believes that it was not deposited Fine gem-quality imperial topaz in euhedral crystals outside the areas mentioned. as much as several centimeters long is found in sapro- Strata-bound epigenetic minerals lite over Cercadinho phyllite in a linear zone several Related to the stratigraphy of the Cercadinho For­ kilometers long in the Ouro Preto and Dom Bosco quad­ mation is the occurrence of several epigenetic minerals, rangles, associated in some places with large terminated the presence of which seems closely related to the com­ quartz crystals. The mineral may well have formed by position of the rocks composing the formation. These the reaction of fluorine-bearing hydrothermal solutions include manganese oxide and notable euhedral hematite, with the highly aluminous phyllite. kyanite, and imperial topaz crystals. The occurrence of unusually fine crystals of these In the Ibirite and Belo Horizonte quadrangles and aluminous minerals in the Cercadinho phyllite would in Pogo Fundo in the Casa da Pedra quadrangle, many seem to indicate some special property of these rocks. small prospect pits and a few larger ones have explored Because of the difficulty in finding fresh rock to study, supergene concentrations of manganese oxide to shallow no contribution to this problem was made in the present depth in the phyllitic rocks of the Cercadinho Forma­ study other than the recognition of the virtual restric­ tion. At the time of the mapping, these openings had tion of the topaz and the best development of the collapsed and slumped to such an extent that little could kyanite in the cited host; it is hoped that others will be seen of the environment of the mineralization and carry forward the study. less of the host rocks. The linear distribution of the Weathering properties workings and the presence of what appeared to be The quartzite of the Cercadinho Formation is resist­ weathered manganiferous dolomitic phyllite suggest ant to weathering where thick and massivo. It forms that the manganese oxide represents surficial concentra­ the ridgeline of the Serra do Curral for several kilo­ tion of unidentified manganese minerals deposited syn- meters east of Pico do Belo Horizonte and a prominent genetically in the rock. Elsewhere the formation is not ridge north of that Serra in part of the Belo Horizonte, manganiferous. Ibirite, and Fecho do Funil quadrangles. Characteristic of the Cercadinho Formation ferrugi­ Weathered phyllitic rocks of the formation are soft nous quartzite is the development of large euhedral and nonresistant and most unstable in artificial ex­ hematite crystals in localities where hydrothermal al­ posures. Roadcuts, particularly 011 dip slopes, are teration has been strong. Such crystals, in many locali­ nearly impossible to hold, and inordinate maintenance ties near and in quartz veins, are known from many expense is incurred in road alinements wr.ich ignore parts of the region, notably just east of the Pico de this fact. Near Ouro Preto large retaining walls have Frazao, Antonio Pereira quadrangle and near N. 7,800, been swept away by the inexorable slow creep of these E. 12,400, Lagoa Grande quadrangle (Wallace, 1965). deeply weathered rocks. At the latter locality Eynearson found a group of crystals as much as 50 cm across and weighing as much FECHO DO FUNIL FORMATION as 150 kg (kilograms) each; one was sent to the National The Fecho do Funil Formation was nam°-d and de­ Museum in Washington, others are in the DNPM mu­ scribed by Simmons (1958, p. 65-66) from a sequence in seum in Belo Horizonte. Smaller tabular crystals which the quadrangle of the same name. In the type locality, can be cut and polished for jewelry are found in a the formation consists of brown and dark-gray dolo­ number of localities. mitic phyllite; argillaceous and siliceous dolomite; and Kyanite occurs in many of the aluminous formations pink, gray, and tan phyllite, locally silty. It is in gra- in the region, but, with a few exceptions, large-bladed dational contact with the underlying Cercadinho For­ crystals are characteristic only of the Cercadinho mation and the overlying Tabooes Quartzite. PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A53

DISTRIBUTION AND THICKNESS LITHOLOGY The Fecho do Funil Formation has been mapped in The Fecho do Funil Formation is composed of doTo- the Serra do Curral eastward from the Igarape quad­ mitic phyllite, phyllite, siltstone, and impure dolomite. rangle to the Belo Horizonte quadrangle. It was not The dolomitic phyllite has nowhere been observed fresh, recognized in the Nova Lima quadrangle or to the but the spongy texture of the rock in weathered ex­ east. It has also been mapped in the Moeda Plateau in posure is typical of such rock. As pointed out by Sim- the Lagoa Grande, Itabirito, and Marinho de Serra mons (1968a), in the type locality the formation consists quadrangles. Although it is present in the Sao Juliao of rocks containing differing quantities of three con­ and Casa da Pedra quadrangles, the formation was stituents dolomite, quartz, and sericite. Dolomite may not mapped because it had not yet been distinguished at be substituted by calcite and quartz and is almost the time of the mapping; in the Dom Bosco quadrangle entirely absent from some phyllite. Some phyllite is it was not mapped because of poor exposures. It is pres­ slightly ferruginous but nowhere contains as imi°,h ent and thick in the Ouro Preto district and was also hematite as phyllite of the Cercadinho Formation. mapped in the Santa Kita Durao quadrangle to the Siltstone is very fine grained and in most places sericitic. north, but it was not identified in the Capanema or Lenses of impure marble are common in the forma­ Catas Altas quadrangles. It does not occur in the Gan- tion. This rock at the type locality contains 36 percent darela syncline and may be presumed not to have been SiO2, 4 percent A12O3, 16 percent CaO, and 11 percent deposited there. In the Itabira district, thin lenticular MgO, with 23 percent loss on ignition (Simmons, 1968a). zones of maiiganiferous saprolite derived from dolo- Wallace (1965) reported seven analyses of marble from mitic rocks above typical Cercadinho beds may repre­ the Moeda Plateau which are much purer; SiO2 ranges sent the formation. In the Monlevade area the zone is from 1 to 7 percent, A12O3 from 0.1 to nearly 2 percent, now gneiss and no rock corresponding to the Fecho do CaO from 28 to 51 percent, and MgO from 1.4 to 19 Funil lithology can be identified. Thus, the formation percent, with the loss on ignition ranging from 41 to 48 is everywhere present except in the eastern Serra do percent. Four of the seven samples approach the end Curral, in the Gandarela syncline, and east of the Serra members of the limestone-dolomite series. The iron con­ do Caraga. tent ranged from 0.6 to 4.2 percent, and one sample con­ The formation varies gradually in thickness. In the tained 1.8 percent manganese. Some of the iron is present type locality Simmoiis (1968a) measures a thickness of as hematite and magnetite, but some is probably fer- 410 m. It thins to 100 in to the westward before being roan calcite and dolomite. All the manganese probably cut out by intrusive granite >and thins to the eastward occurs as manganoan calcite or dolomite, as no primary in the Belo Horizonte quadrangle to perhaps 100 m. manganese oxide minerals are visible. Johnson (1962, Good exposures in the Rio das Velhas canyon north B14) reported nearly pure dolomite from the Cum be of Sabara do not reveal rocks of the lithology of the quarry, Dom Bosco quadrangle. The dolomite lenses Fecho do Funil Formation, and it may be presumed may be as much as 30 m or more thick and as much as that this formation wedges out in the western part of several hundred meters long. Although in the Moeia the Nova Lima quadrangle, as it is not present in the Plateau these lenses tend to be near the base of the Gandarela syncline either. formation, they may occur anywhere in the formation. To the south of Belo Horizonte in the Moeda Plateau, These lenses are bluntly lenticular. A well-exposed the formation was estimated by Wallace (1965) to example can be seen in the valley to the south of Ouro average 300 m in thickness. In the Ouro Preto area the Preto-Belo Horizonte Highway in the Ouro Preto quad­ outcrop width suggests a thickness of 200 m or more. rangle near Botafogo, where the phyllite wraps around In the Santa Rita Durao quadrangle, Maxwell (written the end of a dolomite lens that has been quarried. It is commun., 1963) estimated a thickness of about 175 m. possible that the lenticularity may be due to a gia,nt The Fecho do Funil Formation is one of the least boudinage structure, or the lenses may be bioherms. resistant to weathering in the region and forms low Naturally, in completely recrystallized Precambrian relatively flat or smoothly rolling slopes. In the Moeda marble the lack of fossils is not critical. Plateau the Rio Mata Porcos has carved a deep and steep canyon in this formation, providing intermittently This marble has locally been quarried for metallur­ good outcrops of this generally nonresistant sequence. gical purposes and for building and paving stone. Near Commonly only saprolite is seen, even in deeply incised the Corrego do Eixo in the Marinho da Serra qur.d- canyons, but included lenses of dolomite crop out boldly. rangle, nonecoiiomic occurrences of tetrahedrite and The formation is subject to landslide and creep where other copper minerals occur in several separate lenses it underlies steep slopes. (Rynearson, written commun. 1951). Small concentra- A54 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL tions of manganese oxide occur in the saprolite derived voids about 1 mm in diameter stained with brown limo- from the rock throughout the area. nite, which give the rock a curiously specl'led appear­ ance. Their origin is unknown. TABOOES QUARTZITE Although the fresh quartzite is gray, vitreous, and The Tabooes Quartzite was named by Pomerene hard, weathering dissolves the quartz cement and the (1958b) from the type locality in Corrego Tabooes in the rock becomes noncoherent, although it str.nds well in Ibirite quadrangle. It is composed of fine-grained cuts because of the interlocking grains. It is stained quartzite. brown in some localities and is white in others. The wea­ DISTBIBUTION AND THICKNESS thered rock is a favorite source of sand for cleaning The Tabooes Quartzite is found only in the Igarape, and scouring for the local rural inhabitants p,nd roadcuts Fecho do Funil, Ibirite, Belo Horizonte, and Lagoa in many places are marked by small excavations where Grande quadrangles an east-west distance of 45 km sand has been removed for that purpose. It has also and a north-south distance of about 22 km. At the type been used as a source of high-quality silica sand for locality the Tabooes Quartzite is 121 m thick; elsewhere refractories and as a dressing on blacktop pavement. In in the Ibirite quadrangle it may be thicker. In the Belo one locality in the X/agoa Grande quadrangle, the wea­ Horizonte quadrangle it thins to the east and is only thered quartzite was infiltrated and locally replaced by locally present in the east half of that quadrangle. West manganese oxide derived from the topographically of the type locality the formation continues across the higher Fecho do Funil Formation; a few tons of com­ Fecho do Funil quadrangle, where Simmons (1968a) mercial ore were removed. measured two sections, 75 and 60 m thick, respectively. BARREIRO FORMATION It extends some distance into the Igarape quadrangle, where it is about 30 m thick and is there cut out by The Barreiro Formation 3 was named by Pomerene intrusive granite. In the Moeda Plateau the Tabooes (1958c, p. 68-69) for the well-exposed sequence of schist, Quartzite is much thinner and crops out only at the phyllite, and graphitic phyllite along Corrego do Bar­ north end, where it ranges from 40 m in the north to 2 m reiro in the northeast corner of the Ibirite quadrangle. near the south boundary of the Lagoa Grande quad­ DISTRIBUTION AND THICKNESS rangle (Wallace, 1965). Where not deeply weathered and disaggregated, the The Barreiro Formation crops out on the north flanks Tabooes Quartzite forms a low ridge. Where strongly of the Serra do Curral from the Igarape quadrangle in weathered and completely disaggregated, it forms de­ the west to the Belo Horizonte quadrangle, in the east; pressions instead of ridges and is in many places marked the formation wedges out in both directiors. It is pres­ by deep gullies. Where the formation is thin, it is diffi­ ent in the Moeda Plateau and along the south border cult to locate, as the normal topographic expressions of the region in the Ouro Preto and Mr.riana quad­ cannot develop. rangles and, to the east, in the Santa Rita Durao quad­ CONTACTS rangle. For structural reasons, it does not crop out in The Tabooes Quartzite grades downward within a few the Catas Altas quadrangle and, like the underlying centimeters to a meter into the underlying Fecho do Tabooes Quartzite, was not deposited in th e Gandarela Funil Formation, which is normally phyllitic at the syncline or the east end of the Serra do Curral. It is not top. It grades upward into the Barreiro Formation, known in the Monlevade area, but rocks of this lithology also phyllitic, within a few centimeters. In the western are present in the Itabira District. Thus, it is known for part of the Serra do Curral it is overlain by the Sabara a maximum distance of 110 km in an eas*-west direc­ Formation with a rather abrupt contact. tion and for 90 km in a north-south direcHon but was not deposited everywhere in the region. Ro?k of similar LITHOLOGY lithology has been seen east of Morro do Pilar, some 75 The Tabooes Quartzite is an equigranular fine-grained km north of Itabira. massive orthoquartzite; chemical analyses show that the The formation never crops out well, and estimates of rock contains 98.5 percent SiO2 (Pomerene, 1964, p. thickness are uncertain at best. It is about 124 m thick D26; Simmons, 1968a). Its grain size is in the order of in the type locality and about the same thickness in the 0.05 mm, and the grains are subrounded to subangular. Moeda Plateau. It is perhaps 15 m thick in the Itabira The rock is completely massive and bedding planes are District. not seen in outcropJ Both on weathered and on unweath- 8 English speaking readers should be careful not to confuse this ered outcrops the quartzite is distinguished from all name or formation with the widespread Baxreiras Formation of mid- Tertiary age found in Amazonia. No confusion wiT be possible in other quartzites in the region by the presence of small Portuguese as the two words have quite different meanings. PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A55

CONTACTS found. Outcrops of the saprolite are scattered and in­ The Barreiro Formation overlies the Tabooes Quartz- conspicuous. The complex bulk composition of the for­ ite with conformable and gradational contact. The mation makes it more sensitive to-contact and regional contact is gradational over a few centimeters and is metamorphism than other formations of the Minas marked by an abrupt change from pure quartzite to Series. Because of these facts and the great thickr^es phyllite. It is structurally conformable with the Sabara of the formation, there is much uncertainty as to the Formation, and in some areas the contact is gradational detailed lithology and as to the correlation of particu­ (Pomerene, 1964, p. D27), in others, marked by an lar horizons or zones from one part of the region to erosional unconformity (Simmons, 1968a). another. DISTRIBUTION AND THICKNESS LITHOLOGY The Sabara Formation has been mapped continuc'iis- The Barreiro Formation is dominatly composed of ly from the west side of the Itatiaiugu quadrangle near phyllite and graphitic phyllite. The nongraphitic phyl­ the west end of the Serra do Curral to the east side of lite is pink and light gray; the graphitic phyllite is the Santa Luzia quadrangle near the east end of the black to dark gray. Serra do Curral, a distance of about 65 km. It crops In the type locality and many other places, the lower­ out in part of the Gandarela syncline but is squeezed most beds of the Barreiro Formation are nongraphitic out east of the Gongo Soco quadrangle. Kocksof similar phyllite. These lower beds, some 10-20 m thick, are lithology were seen in the Itabira district but not overlain by graphitic phyllite, in the type locality in mapped individually owing to the poor exposures. Rocks two zones totaling about 100 m in thickness separated at this stratigraphic position in the Monlevade area by a 4-meter break of nongraphitic phyllite. In much are all transformed to gneiss. In the Ouro Preto area of the region, the graphitic phyllite is the key lithology the Sabara Formation is widespread, and it has b^en of this formation, for it is not easily confused with mapped in the southern part of the Santa Rita Durao strata of other formations. quadrangle; it is not found in most of the Serra do Ca- A sample of the graphitic phyllite collected by raga area, as only older rocks crop out there. In the Pomerene contained 4.4 percent carbon, Simmons Dom Bosco area and probably in the Congonhas area, (1968a, p. G20) reported that Breger extracted small rocks of similar lithology are known but are not mapned quantities of soluble organic material from this gra­ separately. The formation is not found in the Mo^-da phitic phyllite and that analysis of the sample collected syncline, probably having been eroded off in pre- by Simmons showed that the rock contained 1.5 percent Itacolomi time. Thus, the Sabara Formation is wide­ soluble and insoluble carbon. spread in the Quadrilatero Ferrif ero. The graphitic phyllite locally contains cubic casts The maximum apparent thickness is at the type lo­ which were probably pyrite. In a relatively small area cality and in the Belo Horizonte quadrangle, where the near Ouro Preto, the rock contains as much as 20 per­ formation is about 3,000-3,500 m thick. Maxwell (writ­ cent pyrite and has been mined on a small scale for the ten commun., 1963) stated that it is about 1,000 m thick pyrite content, used for making sulfuric acid. This in the Santa Rita Durao quadrangle. Moore (19^9) pyrite is very probably syngenetic pyrite, as first sug­ stated that it is more than 1,000 m thick in the Gongo gested by O. Barbosa (1954). Soco quadrangle. The rock weathers easily, crops out only in artificial The formation is always limited upward either by an openings and gullies, and has poor mechanical charac­ ancient or modern erosion surface or by intrusive gra­ teristics for highway and other engineering works. nitic rocks. Along the north side of the Serra do Curral SABARA FORMATION it is everywhere in contact with younger granitic rocks The Sabara Formation, a eugeosynclinal assemblage, that have strongly metamorphosed the formation. In was named by Gair (1958) for the excellent exposures the Ouro Preto area and in the Santa Rita Durao area in the valley of the Rio das Velhas just north of the it is unconformably overlain by the Itacolomi Series. city of Sabara, where that river cuts the Serra do Cur­ In the Gandarela syncline and the Itabira district it is ral in a long canyon. The highway on the east bank and squeezed into the cores of synclines. Its original thick­ the railroad on the west bank provide the most continu­ ness is therefore unknown. ous exposures yet discovered across the strike of this formation. CONTACTS Throughout the region, the Sabara Formation is The upper limit of the formation is an erosion surface deeply weathered to saprolite, and only in a few locali­ or a contact with intrusive granite. The lower contact ties can rock fresh enough for thin-section study be varies in nature within the region, and the Sabara For- A56 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL mat ion directly overlies the Barreiro Formation, the A. L. M. Barbosa and J. H. Grossi Sad (oral commun., Tabooes Quartzite, the Fecho do Funil Formation, and 1966) found in a new roadcut near Saramenlia (Ouro the Cercadinho Formation at one place or another. The Preto quadrangle) that there the Sabara rests on an Barreiro and the Tabooes are formations of limited erosional surface cut on rocks of the Fecho do Funil regional extent; it is not known whether this is the re­ Formation. Thin quartz veins in the lower rocks are sult of nondeposition or pre-Sabara erosion. Structual- said to be truncated by this surface. The rocks above ly, the Sabara Formation is everywhere concordant and below this surface are concordant, and no basal with those underlying formations. In the Belo Hori- conglomerate is present. zonte quadrangle the contact is gradational (Pomerene, Thus, the contact appears to be transitional with no 1964, p. D27); farther west it is abrupt (Simmons, significant pre-Sabara erosion in some parts of the re­ 1968a). In the Serro do Curral the base is marked by gion but abrupt and with significant pre-Sabara erosion a conglomeratic zone a few centimeters thick, consist­ in other parts. The structural concordance of the Sabara ing of fragments of phyllite in a phyllite matrix. In with all directly underlying formations in all parts of the Antonio Pereira quadrangle A. L. M. Barbosa dis­ the region shows that no orogeny occurred in the time covered a boulder conglomerate composed largely of interval represented. dolomite boulders and cobbles overlying with structur­ al conformity the Fecho do Funil Formation (fig. 16). LITHOLOGY The dominant rocks of the Sabara Formation at the type locality are mica and chlorite schistf with inter- bedded graywacke, subgraywacke, and quartzite. The chlorite schists grade upward into biotite-muscovite schists with porphyroblasts of garnet and staurolite as the contact with the granitic rocks is approached. Gray­ wacke and subgraywacke occur in lenticular beds and zones as much as 70 m thick; these make up about 10 percent of the lower half of the formation, whereas the upper two-fifths of the formation is almost entirely schist (Gair, 1962, p. A40-A41). In the type locality, and everywhere else along the north flank of the Serra do Curral, the upper part of the formation has been thoroughly metamorphosed by the intrusive granitic rocks to the staurolite isograd and it is difficult to decipher the exact original nature of the rocks. Abundant chlorite and biotite schist in the type lo­ cality of the Sabara Formation may well represent meta­ morphosed tuffaceous rock. Although criteria to defi­ nitely prove such an origin have been largely destroyed there by contact metamorphism and strong shearing, in other parts of the region much of the rock has a defi­ nitely tuffaceous aspect; in places the material appears to be a volcanic agglomerate. Such material has been reported from many parts of the region. Although deeply weathered, much of such rock wo^ild probably best be termed greenstone where it is not highly metamorphosed. Fresh graywacke and subgraywacke crop out well only in the type locality, where active canyon cutting by the Rio das Velhas has removed most saprolite. Else­ where these rock types have not been widely observed FIGURE 16. Basal conglomerate of the Sabara Formation near but are believed to be present under the soil and sapro­ N. 11,800, E. 7,100, Antonio Pereira quadrangle. The cobbles and boulders are almost all dolomite, presumably from the lite cover. Fecho do Funil Formation. In the valley of the Rio Socorro, outcrops of quartzite PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A57 and feldspathic quartzite were reported by Moore same area. Roundstone conglomerate is found in two (1969). The grain size of the quartz and feldspar ranges exposures north of the Serra do Curral in the Igarape up to 1.0 mm, and the feldspar makes up as much as 15 quadrangle (Simmons, 1968a) in which pebbles and cob­ percent of the rock. Epidote and sericite or chlorite are bles of granite, quartzite, quartz, and phyllite occur in other constituents of this rock. a phyllite matrix. The writer observed granite cobbles In the upper part of the formation in the type section as much as 20 cm in maximum dimension in the outcrop a thin poorly laminated lenticular "iron-formation" oc­ along the Belo Horizonte-Sao Paulo highway.) curs, interpreted by Gair as a ferruginous nietachert. Tilloid (fig. 17) composed of widely scattered large The rocks here are highly metamorphosed, and it is not granite boulders as much as 1 m in maximum dimen­ possible to distinguish with certainty ferruginous meta- sion in a matrix of chlorite schist or phyllite is revealed chert and iron-formation in such an environment. Thin by a deep roadcut just south of Mariana; smaller scat­ lenticular beds of iron-formation are known in the Con- tered cobbles and boulders in chloride phyllite are also gonhas area and in the Dom Bosco quadrangle in rocks found near the base of the Sabara Formation near of lower metamorphic grade that are believed to cor­ Fazenda Mirandinho at the south edge of the Santa Rita relate with the Sahara. Barbosa mapped (pis. 9 and 10) Durao quadrangle. Such rock is revealed only fortui­ extensive ferruginous quartzite and itabirite beds in the tously by readouts and is possibly much more common Sabara Formation in the Ouro Preto and Mariana than now known. quadrangles. Thus, in contrast to the underlying Minas £'?,ries Two persistent thin beds of white pure very fine­ stable-shelf sediments, the Sabara Formation indicates grained quartzite also occur in the upper part, of the a repetition of the eugeosynclinal environment of the Sabara Formation, interpreted by Gair as metachert. Nova Lima Group: volcanism again contributed sedi­ They are clearly part of the Sabara Formation in the ments to the region, many of the rocks are turbidites, vicinity of Marzagao. (See Gair, 1962, pi. 1.) Pomerene and the presence of granitic boulders indicates nearby (1964, p. D28) found similar rock in the Ibirite quad­ crustal instability and the unroofing of batholithic rangle entirely surrounded by granitic rocks; he noted masses. The concordant and gradational contacts with that it had the appearance of Arkansas novaculite. Pom­ the shelf sediments indicate that this particular area was erene suggested that this material might correlate with not itself strongly deformed at this time, but clearly rock other than the Sabara Formation. The present the long period of crustal quiescence indicated by the writer believes it more reasonable to correlate the Ibi­ mature and well-differentiated underlying sediments rite material with that near Marzagao, as such a min- came to an end at the beginning of Sabara time. erologically simple rock would not easily be altered by granitizing processes or invading granitic magmas and could be expected to remain relatively unchanged. Three kinds of conglomeratic rock occur in the Sa­ bara Formation a sharpstone phyllitic conglomerate, true boulder and cobble roundstone conglomerate, and tilloid. The sharpstone conglomerate is at the base of the formation in contact with the Barreiro Formation north of the Serra do Curral (Pomerene, 1964; Sim- mons 1968a). The angular detrital material is phyllite in a matrix of phyllite; the fragments are small, and the bed a few centimeters thick. Roundstone basal conglom­ erate several meters thick in the Antonio Pereira quad­ rangle is composed of dolomite cobbles and boulders (fig. 16). Higher in the section a few zones of scattered well- rounded quartz pebbles in a phyllite or quartzose phyl­ lite matrix are found in the western Serra do Curral. Maxwell (written commim. 1964) referred to scattered small pebbles in 'a phyllite matrix in the Santa Rita FIGURE 17. Boulder of granite in phyllite matrix in tilloid of the Durao quadrangle and to scattered pebbles and cobbles Sabara Formation, near N. 16,600, E. 9,100, Mariana of granite, quartz, and quartzite in graywacke in the quadrangle. A58 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL

ENVIRONMENT OF DEPOSITION starved section the euxinic facies, mainly black shales with, The rocks here included in the Piracicaba Group are, perhaps, very thin and evenbedded siltstones, siliceous limestone or chert. This facies is succeeded by the normal fl^sch facies, with the exception of the Sahara Formation, all blanket the rhythmically bedded dark shales and graywackes. shelf deposits. The Sahara Formation represents a This description of normal geosynclinal sediments flysch suite deposited in a eugeosynclinal environment. suggests that the Piracicaba Group of the Minas Series The erosion interval at the end of Itabira time marked may represent this normal development. The Cerca­ a withdrawal of the sea or an epeirogenic uplift of the dinho Formation was the first blanket sediment in this land surface. In this region, major folding is not evi­ group, a shelf sediment, deposited on a flr.t surface. dent; at most, a very minor warping occurred. The The Fecho do Funil may represent the first down- amount of erosion that took place in this region was not warping of the miogeosyncline, with siltstones and great. Whether the Gandarela Formation was com­ siliceous dolomite common. The Tabooes and Barreiro pletely removed in those parts of the region where it is Formations correspond to the thin, starved, and, in the missing or whether it never had been deposited could Barreiro, the euxinic section, overlain by the thick not be determined. In the eastern part of the region eugeosynclinal flysch suite of the Sahara Formation. nondeposition is indicated, but relations in the western The sequence represents a single geosynclinal sedimen­ Serra do Curral suggest that the formation may have tary cycle. The rocks were slightly deforned after been removed by erosion there. The presence of coarse Sahara sedimentation and before Itacolomi time. pebbles, cobbles, and finer fragmental material from the The sudden and drastic change of sedimentation that Itabira Group in the basal conglomerate of the Cer- began with the Sahara Formation is thought by some cadinho Formation in the Serra da Moeda and the west­ to justify setting that formation apart in a separate ern Serra do Curral, and the generally coarser sediments group in the Minas Series (A. L. M. Barbosa, written in the Cercadinho in the west-central part of the region, commun., 1966) or as a lower part of the overlying indicates extensive erosion of the Itabira Group in an Itacolomi Series (Teixeira da Costa, 1961, p. 60). How­ area west of the Quadrilatero Ferrifero. The less ever, there is no compelling evidence any wl Q.re in the abundant hematite in the Cercadinho Formation to the Quadrilatero Ferrifero of a significant angular discord­ east also suggests an origin to the west. ance in the sedimentary sequence now included in the The alternating beds of fine, medium, and coarse Piracicaba Group. In certain parts of the region, as in lenticular crossbedded and ripple-marked quartzite and the Ibirite quadrangle, sedimentation seems to have of phyllite which compose the Cercadinho Formation been uninterupted, and the Sahara is gradational with are typical of nearshore or deltaic sediments deposited the underlying Barreiro Formation (Pomerene, 1964, on a rather flat surface. The impure dolomite and dolo- p. D27). In other areas the Sahara Formation overlies mitic phyllite of the overlying Fecho do Funil Forma­ the Cercadinho and the Fecho do Funil Formations. tion indicate gradual encroachment and deepening of The absence of the Tabooes and Barreiro Formations the sea. On these sediments, regional in extent, were can as well be attributed to nondeposition as to erosion; deposited the more restricted quartzite and phyllite of the former formation certainly lenses out southward in the Tabooes and Barreiro Formations, which are local the Lagoa Grande quadrangle (Wallace, 1965). Except accumulations of very well differentiated and mature for the cited localities near Fazenda Gu^laxo and sediments. The high graphite and locally high pyrite Saramenha, there is no sign of significant subaerial contents of the Barreiro Formation indicate an euxinic erosion in these areas. The thin basal conglomerate environment, the result of local downwarping. found in some localities at the base of the Sahara For­ The graywacke-subgraywacke-metatuff-tilloid-proto- mation contains no exotic rocks, merely those, from the quartzite-metachert-phyllite suite of the Sahara Forma­ underlying formation, plus some quartz grains and tion clearly marks a sudden and complete change both pebbles. The angular phyllite fragments ir this con­ in depositional environment and source of sediments glomerate could not have been transported guy signifi­ from those which produced the well differentiated sedi­ cant distance; even if the material now pi yllite had ments of the rest of the Minas Series. already been lithified, for the shale would be very soft. Pettijohn (1957, p. 640) stated the following on the Further evidence on the erosional discordance at the normal course of development of geosynclinal base of the Sahara should be sought to supplement sediments: present inadequate information caused by lack of out­ There is what may be called the normal geosynclinal cycle. The crops. New roadcuts and other excavations in scattered geosyncline is initiated, filled, and commonly, though not always, parts of the region will in the course of time provide deformed. The sediments which accumulate during the different stages are more or less petrographically distinct. The early critical evidence now lacking as to the laters 1 extent of stages of the geosynclinal cycle are commonly marked by a the erosional interval now definitely known only in the PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A59

southern part. If this irosion surface proves wide­ litic facies was named the Santo Antonio Formation spread, and if the conformity of the beds below and by O. Barbosa (1949, p. 7), who assigned the unit to tlie above this surface proves to be a general rule through­ Minas Series. Guild (1957, p. 22) established this se­ out the region, very possiUy the surface should be used quence as equivalent to the Itacolomi Series and re­ to separate the Sabara from the rest of the Piracicaba named it the Santo Antonio facies (p. 25). Quartzite Group. If subdivided into formations by more detailed of the type locality will herein be referred to informally mapping, the Sabara coutl then be raised to group as type Itacolomi, and the phyllitic sequence of O. Bar­ status. The conformity of these rocks with the under­ bosa and Guild will be referred to as the Santo An­ lying Minas Series rocks, ii contrast with the angular tonio facies. unconformity of the Sabaia with the overlying Ita- TYPE ITACOLOMI colomi Series rocks, indicates that the Sabara should The quartzites grouped informally as type Itacolomi remain in the Minas Series. are those described by Harder and Chamberlin (1915), METAMOEPHISM Freyberg (1932), Lacourt (1935), A. L. M. Barbora Most formations of the Piiacicaba Group are found (in Dorr and others, 1961, p. 20-22), and others. They in the axes of synclines, wheie they were subjected to crop out boldly and well in the type locality, the Pico strong deformation but were protected from the contact de Itacolomi and vicinity, south of Ouro Preto. Tl Q- nietamorphism produced by invading granitic rocks, type locality is isolated by deep canyons and, the farther which occur largely in anticlinal areas. Thus, they tend from the type locality, the less certain the correlation to show a relatively low grade of regional meta- (A. L. M. Barbosa, in Dorr and others, 1961, p. 20). morphism but may be strongly folded and sheared. Harder and Chamberlin (1915, p. 364) pointed out that The major exception to the above generalization is the relatively clean quartzite of the type locality grades found on the north side of the Serra do Curral, where laterally into phyllitic and schistose rock of quite dif­ granitic rocks transgress all the formations of the ferent aspect. Piracicaba Group except the Tabooes Quartzite, which Until the mapping by A. L. M. Barbosa (pi. 10), wedges out before contact with the granite. Here, the geologists considered the thick section of phyllite in tl Q- effect of the bulk composition of the formations on Pico de Itacolomi to be an integral part of the Itacolomi their nietamorphism is clearly apparent. The hetero­ Series (Lacourt, 1935). Barbosa, however, determined geneous complex sediments of the Sabara Formation that the phyllite is equivalent to the Sabara Formation were strongly affected, for staurolite and garnet zones of the Minas Series, thrust over Itacolomi quartzite arc! are strikingly developed parallel to the contact in bands in turn overlain by Itacolomi quartzite (A. L. M. Bar­ more than 1 kilometer across. The better differentiated bosa, in Dorr and others, 1961, p. 20-21). sediments of the lower formations were not so strongly DISTRIBUTION AND THICKNESS affected. Quartzites were metasomatized for only a few Quartzite of the type Itacolomi Series occurs in the tens of centimeters at the granite contact, and phyllites Ouro Preto and adjacent Mariana quadrangles. Quart­ were recrystallized into somewhat coarser grained mica­ zite of similar appearance imconfonnably overlyirg ceous rocks for distances measured in meters or tens of the Minas Series also occurs in the Pico de Frazao in meters. ITACOLOMI SERIES the Antonio Pereira quadrangle. In the type area, intense shearing and particularly Strata now included in the Itacolomi Series were thrust faulting make any estimate of thickness perilous1 ; discriminated by Harder and Chamberlin (1915, p. A. L. M. Barbosa (written commun., 1965) suggested 364-368) as the Itacolomi Quartzite,4 considered by that the series may be as much as 2,000 m thick. The them to be part of the Minas Series. In 1931, Guimaraes formation is limited upward by the present erosion raised the unit to the rank of a series on the basis of the surface, and we have no evidence as to the original thicV- angular unconformity between the quartzite and the ness, which may have been much greater. South of Pas- underlying rocks of the Minas Series. The Itacolomi sagem de Mariana and elsewhere, the type Itacolomi Series is now considered to consist of two facies, one of locally contains abundant garnet; the presence of this quartzite and one containing much phyllite. The phyl- relatively high temperature metamorphic mineral sug­ 4 Readers should not confuse the stratigraphic unit with the rock gests that a great thickness of younger but premetamor- known as itacolomite, a flexible micaceous quartzite named originally phic rock has been eroded since nietamorphism. from the Quadrilatero Ferrifero but occurring in many parts of the world. This rock has no stratigraphic significance whatsoever, being LITHOLOGY found in the Rio das Velhas, Minas, and Itacolomi Series. Many geologists haye confused the two terms ; for this reason most geologists The type Itacolomi is dominantly quartzite and grit in Brazil since Derby and Harder and Chamberlin have avoided using the term "itacolomite." and contains varying quantities of sericite. The quartz- 334-413 O <6I9 5 A60 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL ite is conglomeratic and also contains many lenses of of the Itacolomi strata in the type locality consists of conglomerate. Thin and very lenticular interbeds of medium-grained quartzite and. grit, about 30 percent of phyllite are present in some areas. conglomeratic quartzite, and about 20 percent of Most strata consist of coarse quartz grains in a matrix conglomerate. of fine-grained quartz, sericite, and muscovite. Feldspar CONTESTS may be either detrital or a metasomatic product; A. L. The Itacolomi Series in ttie type locality rests with M. Barbosa (written commun., 1965) emphasized that angular unconformity on :he Sabara, Barreiro, and the presence or absence of this mineral is of no strati- Fecho do Funil Formations of the Minas Series. The graphic significance. It forms a very small percentage maximum angular unconformity found was about 12°, of the rock. Hematite and martite are common inter­ just to the south of Ouro Preto. stitial constituents. On the main mule trail south from The polymictic conglomerates found in the Itacolomi Mariana, a zone containing much iron is found near the Series (Freyberg, 1932, p. 105) prove major erosion of base of the sequence, and similar zones occur elsewhere. the Minas Series before and during Itacolomi sedimen­ In many places, hematite and martite are disposed on tation, and the unconformable relation is beyond ques­ shear planes cutting the bedding at low angles. The tion. There is no way of evaluating the time interval quartzite is crossbedded at many localities. involved, but it was great enough to permit removal of The conglomerates consist of rounded pebbles, cob­ at least 1,000 m of Minas sediments before local accumu­ bles, and, in one locality, boulders as much as 0.5 m in lation of Itacolomi sediments began. maximum dimension of vein quartz, quartzite, itabirite, SANTO AJTT6NIO FACIES and, sparsely, phyllite; A. L. M. Barbosa confirmed the presence of one cobble well enough preserved to be cer­ The rocks originally named the Santo Antonio For­ tainly identified as a granitic rock. Many pebbles and mation were described by O. Barbosa (1919) from the cobbles are somewhat squeezed and discoid; their long Morro Santo Antonio in the Casa da Pedra and Sao dimensions are in the plane of foliation and may be at Juliao quadrangles. The rocks that crop out there are right angles to bedding. Phyllite fragments in such lo­ phyllitic quartzite, roundstone and sharpstone conglom­ calities are sheared out into long wisps. erate, phyllite, and ferruginous and nonferruginous L. J. de Moraes pointed out to the writer that itabirite quartzite. Probably, both O. Barbosa and Guild in­ cobbles show that the iron-formation had been slightly cluded in their Santo Antonio some rocks which are now folded before erosion, a fact first noted by Lacourt assigned to the Cercadinho Formation; at the time of (1935). Despite careful search, nowhere were cobbles or their work this formation had not yet been distinguished boulders found with the characteristic sharp and intri­ as a separate unit and exposures in the area are obscure. cate creiiulation and folding now found in that rock, The Santo Antonio sequence in most localities is easily although in modern gravel deposits cobbles and weathered and only locally forms ridges or hills; where boulders with such folding can be found. This indi­ quite phyllitic, it underlies low valleys and gentle slopes. cates that the most intense folding occurred after DISTRIBUTION AND THICKNESS Itacolomi sedimentation but that some folding had The Santo Antonio facies of the Itaco^mi Series is occurred before. the youngest Precambriaii metasedimentary rock wher­ All conglomerate beds in the type Itacolomi are lentic­ ever it crops out and is always limited upward by an ular. Some are more than a few meters thick, and a erosion surface. It rests upon an erosion surface carved few are more than 10 m thick. The lenticularity of the into the Piracicaba Group of the Minas Series in all conglomerate beds suggests scour and fill. No persistent known exposures. In the type locality it rests on rocks basal conglomerate has been reported in the type Ita­ which the writer would correlate with the Cercadinho colomi, but exposures of the bazal zone are rare indeed. Formation, elsewhere on all the other formations of the Phyllite in the type Itacolomi consists largely of seri- Piracicaba Group expect the Tabooes Quartzite. citic lenses not more than a few tens of centimeters thick (A. L. M. Barbosa, written commun., 1965). Kyanite The facies is known to crop out in the Lagoa Grande, is found both in phyllite and quartzite. Casa da Pedra, Sao Juliao, Antonio Pereira, Capanema, The Itacolomi quartzites are crossbedded, and La- and Santa Kita Durao quadrangles (pi. ]). Individual court (1935) mentioned the occurrence of ripple marks areas of outcrop are not large and are isolated one from in the quartzite. Bedding is coarse to massive in the the other in three general areas, the largest of which is coarser grits and quartzites. The finer sediments are 9 km long. It survives only in downfaulted and down- more thinly bedded. folded areas. A. L. M. Barbosa (written commun., 1965) stated The original thickness is, of course, unknown, as the that, ignoring the thin phyllitic lenses, about 50 percent top is nowhere exposed. The apparent thickness in the PIYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A61

Lagoa Grande area is aout 150 m, in the Congonhas mentioned above, in contrast with the Caue Itabirite, area in the order of 1,00 m, and in the Santa Eita which never contains coarse, and rarely contains fine, Durao area in the order Barbacena Series, much rock of post-Minas Caraga between the basal Minas Series as it occurs in age and included in the Minas Series stratigraphic units the Moecla Plateau and the Rio das Velhas Series, sub­ of pre-Minas age.5 In 1956, Oliveira restricted the Minas Series to the strata above the major angular un­ dividing this group into four formations. They as­ conformity demonstrated by Rynearson and colleagues signed this new group, the Tamandua Group, to the and below an angular unconformity with the Itacolomi Rio das Velhas Series. Simmons (1968b) later com­ Series. bined the top three formations into one formation. An erosional disconformity between the Itabira and There is some question as to the proper series assign­ Piracicaba Formations of Harder and Chamberlin, dis­ ment of the Tamandua Group. To evaluate the as­ covered by Pomerene and Ashley in 1952, was taken as signment, a recapitulation of the development of the line of separation between those two units, a separa­ nomenclature of the rocks assigned tc the Caraga tion indefinitely defined by Harder and Chamberlin. Quartzite by Harder and Chamberlin is necessary. Those In 1957 Dorr, Gair, Pomerene, and Kynearson formally geologists included in their Carac,a Quartzite all the proposed the name Rio das Velhas Series for the strati­ quartzite in the Serras do Caraga, Tamandua, and das fied rocks below the angular unconformity at the base Cambotas, assigning the unit to the bottom of the Minas of the Minas Series as restricted by Oliveira. They also Series. Later Guimaraes assigned the qug.rtzites of the divided these two major series into groups so that fur­ Serras do Carac,a and das Cambotas to the Itacolomi ther subdivision would be possible, for mapping had by Series; Bolau (1952), agreeing with Harder and

5 For these reasons, as well as the lack of an adequate definition, the 8 The subdivision was then made because in that year the Brazilian term Barbacena Series is not used in reports resulting from this project. Geological Society held its annual meeting in the Quadrilatero Ferrifero. Some of the rocks included by O. Barbosa in his Barbacena Series are It was deemed desirable to presemt the results of 11 ymrs work because undoubtedly the stratigraphic equivalents of some rocks in the Rio das the Minas Series was the focus of scientific and economic interest at that Velhas Series of these reports. meeting. PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A65

Chamberlin, believed tint they belonged to the basal concordantly overlie the Tamandua Group with nor­ Minas Series. mal stratigraphic succession and are not overturned. At the time members of this team named the various There are no apparent signs of major faulting between formations of the Minas Series (1958), the Serra do the units. The structure in the Serra do Caraga is mc^e Caraga had not been mayped and, because the strati- complex and the rocks more deformed, but the Cam­ graphic position of all qiurtzites in that range had not botas Quartzite there also underlies the unquestioned yet been established and because structure was there Minas Series in normal stratigraphic sequence. Thus, most complex, the basal qiartzites of the Minas Series the Tamandua Group and therefore the Cambotas as restricted by Oliveira (1956) were named from the Quartzite must be older than the overlying units of the Moeda Plateau region, wlere the structure was rela­ Minas Series and cannot be correlated with the Ita­ tively simple and the stratigraphic succession unequivo­ colomi Series. cal. It was then known at least part of the Serra do 2. No locality was found by Simmons, Moore, or Caraga was made up of quirtzite correlative with the Maxwell in areas mapped by them where the Tamandua Moeda quartzite, and thereiore Harder and Chamber- Group underlies the unquestioned Minas Series w;th lin's name for the basal Minss Series rocks was retained demonstrable angular relation. A number of localities with group status to include both the Moeda Forma­ are cited in reports by these geologists where the rocks tion and the Batatal Fornation, which also occurs are in structural concordance. An erosional unconfor'n- there, to be consistent with tie upgrading of the other ity may or may not exist between the Moeda Forma­ original formation names to group status. This proved tion and the Tamandua Group; the writer believes, but wise, for further mapping proved that the Serra do cannot prove, that an erosional unconformity exists. Caraga is made up of quartzites of three different ages That it is not great is shown by the lateral persister^e and lithologies, the Maquine, the Cambotas, and the for some 25 km of the soft relatively thin phyllitic beds Moeda, and that the Cambotas forms the bulk of the of the upper Tamandua Group. mountain range. 3. The Cambotas Quartzite rests with certain e**o- Moore was the first to suspect that quartzite in the sional unconformity and probable angular unconform­ Serra das Cambotas and its lithologic equivalent in the ity on the Nova, Lima Group in the Gongo Soco quad­ Serra do Caraga might be older than that mapped as rangle. Rocks now tentatively correlated with the Cam­ the Moeda Formation. Later detailed mapping by botas Quartzite rest with angular unconformity on the Simmons and Maxwell showed that this was indeed Nova Lima Group in the Floralia quadrangle (Herz, true and that the quartzite of the Serras das Cambotas oral commun., 1965), in the Serra do Ouro Branco and do Caraga was lithologically somewhat different (Barbosa, 1949; Guild, 1957; Johnson, 1962), and in from the Moeda Formation and underlay that forma­ the Serra do Espinhago (Pflug, 1965). No unequivocal tion. These geologists assigned these older quartzites evidence for or against an angular unconformity be­ and overlying phyllitic beds to the Eio das Velhas tween these units is known elsewhere. Series on the supposition that a major unconformity 4. In the part of the region in question, as in other existed between them and the Moeda quartzite. places where the Tamandua and Maquine Groups are Exposures of the critical zone of contact in the areas missing, the Caraga Group of the Minas Series rests un- of outcrop of the Tamandua Group are uniformly conformably on the Nova Lima Group. Simmons poor and, although the zone of contact between the (1968b, p. H8) stated that, in the south limb of the Gan- Moeda Formation and the Tamandua Group and be­ darela syncline in the Cocais and Santa Barbara quad­ tween the Tamandua Group and underlying units rangles the average angular discordance between these total many tens of kilometers in length, only a few rocks is 20°. In the northeastern part of the Conceigao significant data on the contacts have been found. do Rio Acima and the southeastern part of the Gongo These incomplete data lead the writer to believe Soco quadrangles, Moore (1969) described the dis­ that the Tamandua Group is better assigned to the cordance between Minas Series and Nova Lima Group Minas Series than the Itacolomi Series (Guimaraes) rocks as "slight to moderate"; at Conceigao do Rio or to the Rio das Velhas Series (Simmons, Maxwell, Acima N. 5,300, E. 1,400, there is a difference in both and Moore) for the following reasons. strike and dip of 40°. At the west side of the Conceigao 1. In the Serras do Tamandua and das Cambotas, the do Rio Acima quadrangle, the basal conglomerate of the Cambotas Quartzite is shown by ripple marks, cross- Moeda Formation contains angular fragments of Nova bedding, scour-and-fill structures, and other sedimen­ Lima Group phyllite. tary features to be upright, not overturned (Simmons, 5. The contact between the Cambotas Quartzite r,nd 1968b). The unquestioned units of the Minas Series the underlying Maquine Group of the Rio das Velhas A66 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MESTAS GERAK, BRAZIL

Series was said by Maxwell (written commun., 1962) formation. He believes that tie presence cf large rela­ to be sedimentary but discordant. Maxwell cited a low tively isotropic and rigid leiticular masres of ortho- angle structural unconformity at Capanema N. 13,600, quartzite in dominantly argillaceous terrain localized E. 8,300, and a 60° angular unconformity at Capanema the synclines during the epoch of deformation and that N. 5,400, E. 6,600. In the Conceicao do Rio Acima quad­ localization of the quartzites on one side of synclines rangle, Moore considered this contact to be a fault con­ is expectable rather than anomalous. tact. As shown on plate 1, this writer interprets it to Disagreement on correlation of a band of phyllitic be a sedimentary contact. The angular discordance as rocks resulted from poor exposures alor? the Serra mapped by Moore is as much as 30° between the two Geral in the Caete, Gongo Soco, and Santa Barbara units. quadrangles. Thus scattered field evidence shows that a definite an­ Chloritic phyllites with iron- formation lenses uncon- gular and erosional unconformity exists between the formably underlying the Moeda Formation in the Tamandua Group and both groups of the Rio das Vel- Caete and Gandarela quadrangles have been traced has Series and, where the Tamandua Group is missing westward without structural or stratigraplic break into in this area, between the rest of the Minas Series and the type locality of the Nova Lima Group and eastward the Rio das Yelhas Series. No angular unconformity into the western part of the Gongo Socc quadrangle. has anywhere been demonstrated in the Quadrilatero Moore mapped the band in the Gongo SOCT quadrangle Ferrifero between the Tamandua Group and the rest as Nova Lima Group, lying with fault contact on the of the Minas Series. Because the angular unconformity Cambotas quartzite in the eastern half. Simmons between the Tamandua Group and the Rio das Yelhas (1968b), however, interpreted this phyllitic zone above Series is very probably the same unconformity found the Cambotas Quartzite. in the Santa Barbara quad­ between the Minas Series and the Rio das Yelhas Series rangle as the upper formation of the Tamandua Group from the westernmost to the easternmost quadrangles, and projected it into the Gongo Soco quadrangle. it seems quite probable that the Tamandua Group should The writer, in reinterpreting the stratigraphy and be assigned to the Minas Series, for this regional un­ structure for plate 1, assigned the rock in the eastern conformity was cited by Oliveira as the basis for the part of the Gongo Soco quadrangle to the upper forma­ separation of the Minas and the Rio das Yelhas Series. tion of the Tamandua Group, in agreement with Sim­ Because the upper Tamandua Group is not known mons1 assignment, and the western part to the Nova west of a line between Henrique Fleuiss and the south­ Lima Group, in agreement with MooreV assignment. east corner of the Serra do Caraga, there is no reason The two units are thought to be separated by an east- to suppose that it was ever deposited west of this line. trending fault. The reasons for this tentative assignment The Cambotas Quartzite thins abruptly toward the are that the phyllitic rocks in this band contain several west and disappears in the middle of the Gongo Soco thin iron-formation zones and are chloritic in the west quadrangle and is also relatively thin northeast of the side of the Gongo Soco quadrangle, typical of the Nova Conta Historia sync-line (Capanema quadrangle). It Lima Group, and are sericitic on the east side and con­ seems reasonable to suppose that there it wedged out tain one relatively thick iron-formation zcne, typical of abruptly westward. The quartzite of the Serra do Ouro the Tamandua Group. The writer considers that both Branco, tentatively correlated with the Cambotas, units of the Tamandua Group wedge out abruptly to wedges out abruptly to the west the west, because neither is found to the west of the mid­ Simmons (personal commun., 1964) emphasized the dle of the Gongo Soco quadrangle. When better expo­ sporadic distribution of the Tamandua Group within sures of the rocks in the area are made by roadbuilcling the Quadrilatero Ferrifero as evidence of a strong un­ or mining activity, the area should be restudied, as pres­ conformity between these rocks and those of the Caraga ent exposures do not permit certain resolution of the Group and cited the fact that the Tamandua is found problems. only on the north and west limbs of synclines as difficult After this paper had been prepared in essentially final to explain. It is quite conceivable that the Cambotas form, a report resulting from a reconnaissance of the was deposited as beaches in bays between headlands, in part of the Serra do Espinhago between 19°30' S. and a manner similar to the present beaches being deposited 17° S. appeared in Germany and Brazil (Pflug, 1965). along the Sao Paulo-Rio de Janeiro littoral; a discon­ The Brazilian version of this report will be cited here. tinuous highly lenticular formation would ensue. The The region discussed comprises some 30,000 square kil­ writer believes, but cannot prove, that some erosion took ometers immediately north of the easteri part of the place between Tamandua and Caraga deposition; this Quadrilatero Ferrifero; a planimetric sketch map of does not imply either a long time break or a major de­ the area and part of the Quadrilatero Ferrifero at a PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A67

scale of about 1:1,000,000 was included, showing distri­ The writer therefore believes, with Pflug (1965), bution of stratigraphic units and of lithology. Harder and Chamberlin (1915), and O. Barbosa (1954), Pflug advocated that tie Miiias Series should be di­ that the Minas Series as exposed in the region to th^ vided into three interfingering lithologic f acies, depos­ north of the Quadrilatero Ferrifero is basically similar ited synchronously. These f acies were called by him the in stratigraphy to rocks included in the Minas Series in "Diamantina f acies" (quartzite), the "Itdbira f acies" the Quadrilatero Ferrifero. Detailed mapping will prob­ (itabirite), and the "Guanhaes f acies" (eugeosynclinal ably show that some formations distinguished to tin sediments). North of the Quadrilatero Ferrifero, the south will not be present to the north, and some nev f acies crop out in three north-trending belts distributed formations may well be distinguished, as all the forma­ from west to east, respectively: the first, 40-100 km tions are not continuous everywhere even in the Quad­ wide, said to represent miogeosynclinal sedimentation; rilatero Ferrifero. the second, 10-15 km wide, said to represent sedimenta­ The distribution in the Quadrilatero Ferrifero of tin tion at the transition between the miogeosyncline and eugeosynclinal Sahara Formation, similar to the "Guan- the eugeosyncline; the third, over 100 km wide, said to haes facies" and shown as "Guanhaes facies" in Pflug's represent sedimentation in the eugeosyncline (Pflug, figure 13, proves definitely that it is much younger thar, 1965, p. 31-33 and fig. 14). This interpretation contrasts rather than being a synchronous and interfingering fa­ with the subdivision of the Miiias Series into four super­ cies of, the basal quartzites of the Minas Series. Not only posed groups, subdivided into 11 formations and several is the Sahara coextensive with those rocks where it has members, made by the DNPM-USGS geologists in the not been removed by later erosion, but is separated fror* Quadrilatero Ferrifero. them by one and possibly two erosional unconformities The basal formation of the Miiias Series in the Quad­ as well as by about 1,000 meters of well-differentiated rilatero Ferrifero, the Cambotas Quartzite, can be sediments, some of which have been recognized to tin traced from its type locality northward, with a few north. Instead of being confined to the area east of tin breaks of a kilometer or so where it has been eroded off, quartzite outcrops, the Sahara continues to the west with

into the "Diamantina f acies" of Pflug.~ The correlation full thickness about 100 km beyond the westernmost out­ is secure and is followed in Pflug's map, which unfortu­ crops shown on Pflug's map, where it is cut out by nately omitted several large areas of the quartzite be­ younger granite. Thus, neither to the east, nor to tin tween the Serra das Cambotas and the area to the north. west can it be considered to be synchronous or inter- Pflug found that an angular unconformity exists be­ fingering with the quartzite rocks traceable from th°i tween his "Diamantina facies" and rocks similar litho- area mapped by Pflug into the Quadrilatero Ferrifero. logically to the Rio das Velhas Series, that the Minas Superposition of the Caue Itabirite above the quart­ Series sediments had come from the west, and that the zite which Pflug considered to be a time equivalent of grade of metaniorphism systematically rose from the the itabirite is equally clear in the Quadrilatero Fer­ west to east, all facts consistent with those reported from rifero from the easternmost to the westernmost expc- the Quadrilatero Ferrifero and tending to make more sures, or about 140 km in an east-west direction and for certain the interregional correlation. Pflug did not find a some 100 km in a north-south direction. break between the Tamandua and Caraga Groups and Thus, the Quadrilatero Ferrifero offers no support did not subdivide the basal Minas Series quartzites, al­ for the relatively stable shoreline during Minas Series though he did refer to local unconformities in the sedimentation demanded by the interfingering facies quartzite. concept but offers much evidence of systematically This writer has not systematically mapped in the area transgressing shorelines resulting in distinctive super­ discussed by Pflug but has crisscrossed it numerous times posed blanket formations. in geological reconnaissance while trying to decipher The fact that the sediments of the Piracicaba and the geological frame around the Quadrilatero Ferrifero. Itabira Groups are at least as well developed in tin In that area he has recognized ferruginous quartzites Serra do Curral, their present northernmost limit in the lumped with true itabirite by Pflug in his "Itabira fa­ Quadrilatero Ferrifero, as anywhere else in the region cies." Two true itabirate zones occur in the northern area suggests strongly that these rocks once extended far with outcrops measured in tens of kilometers in length, to the north of their present limit. Available evidence but the rock type is there in thinner zones and is not indicates that their present distribution is controlled everywhere continuous. Rocks of the eugeosynclinal more by postsedimentation tectonics, granite forms,- suite, called the uGuaiihaes facies" by Pflug, were ap­ tion, and erosional events than by the original limits parently similar in lithology and stratigraphic position of the sedimentary basin in which they were deposited. to the Sahara Formation of the Quadrilatero Ferrifero. It is most difficult to imagine that a series of marine A68 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINTAS GERAIS, BRAZIL sediments more than 4000 m thick, which can be traced type locality near Fonseca, a town east of Agua Quente south from the Serra do Curral for 100 km there and outside the Quadrilatero Ferrifero. being limited by tectonic features and for 140 km in At the base of the unit is a conglomerate about 7 an east-west direction, should pinch out abruptly to the m thick consisting of pebbles and cobbler: of granitic north. If, as supposed, the Minas Series once extended gneiss, phyllite, and quartzite. Siltstoiie, claystone, and far to the north of the Serra do Curral, Pflug's facies sandstone overlie this basal conglomerate. About 27 m concept evidently could not be defended, for that con­ above the base of the unit, lignite lenses ranging in cept demands a land area to the north of the Serra do thickness from 2 to about 80 cm appear in the other Curral during deposition of the whole Minas Series. He rocks; an old analysis indicates the following composi­ named this hypothetical land area the San Francisco tion in percent: ash, 17.95; volatiles and hygroscopic Massif (Pflug, 1965, fig. 14). water, 44.91; fixed carbon, 37.14 (Gorceix, 1884, p. 88). The interfingering of phyllite and locally ferrugi­ The formation lies on granitic gneiss and is overlain nous quartzite with the purer quartzite of the central by canga and ferruginous laterite. Serra do Espinliac,o that gave rise to the interfingering The Fonseca Formation underlies the Fonseca basin, facies concept is also clearly evident in the Moeda For­ about 35 square kilometers in extent. The top is a rela­ mation in the Quadrilatero Ferrifero. In the Serras da tively plane surface and the bottom a Tertiary erosion Moeda and do Itabirito, where this formation is thick, surface described earlier in this paper (p. ^ 13). Gorceix well exposed, and clearly underlying rather than inter­ was the first to describe these strata; the abundant fingering with the Itabira and Piracicaba Groups, the plant fossils were assigned to the Pliocene or upper quartzite contains many lenses of phyllite, some of Miocene (Gorceix, 1884, p. 88-89). More recently Som- which were mapped separately as members of the Moeda mers (written conimmi., 1962) judged that the fossils Formation (Guild, 1957; Pomerene, 1964; Wallace, could not be dated more closely than late Tertiary. 1965). This lithologic succession is a natural one in OTHER TERTIARY ROCKS transgressive nearshore formations and can be matched Several small accumulations of sedimentary rock of today in many localities, particularly those with ag­ Tertiary age are known in the Gaiidarela, Macacos, and grading barrier beaches and lagoons. Nova Lima quadrangles, and undoubtedly others will The writer believes that careful mapping at adequate be found elsewhere beneath canga, bauxite, or soil. One scales of the central Serra do Espinhago may reveal of these deposits, in the Gaiidarela basin, yielded com­ some of the same formations found in the Quadrilatero mercial lignite during the Second World War. Ferrifero, arched into a great anticline that was much The Gaiidarela basin, so named and described by disturbed structurally, sliced by north-trending, west- Gorceix (1884), is about 750 m long and 250 m wide. moving thrust faults, probably with imbrication, bent In this basin claystone, sandstone, and lignite overlie the by cross folds, and cut by younger granitic and gab- Gaiidarela Formation in a roughly elliptical pattern broic rocks. Before erosion, rocks included in the "Ita­ (pi. 5). A drill hole near the south margin of the basin bira" and "Guaiihaes facies" probably extended over showed some 60 m of semiconsolidated sediments, which the top of the "Diamantina facies" that now fonns the could not be successfully cored, overlying ferruginous central part of the Serra do Espinhago, as postulated rocks interpreted as canga. The drill did not reach un- by Harder and Chamberlin (1915) and O. Barbosa weathered bedrock. Dense vegetation and poor outcrops (1954), who called them the Itabira and Piracicaba prevented close study of the Tertiary rocks. Two adits formations. Itabirite cobbles in ancient (Cretaceous?) driven on lignite beds were found, but the timber at the high-level gravels west of the Rio das Velhas, more portals was crushed and a respectable f ow of water than 200 km north of the Serra do Curral, and well west emerged; it was not considered wise to enter them and of the Serra do Espinhago suggest that iron-formation their extent and the thickness of the lignite beds are once covered that range to its west side and possibly unknown. At least two lignite beds were reportedly beyond. found, one more than 1 meter thick. The lignite seen at the portal and in an old bin contained many frag­ UNMETAMORPHOSED SEDIMENTARY ROCKS mentary leaf impressions and much pyrite, largely FONSECA FORMATION smeared on fractures and bedding planes. The mine Underlying the "Chapada de Canga" in and east of water was very acid, and limonite was being deposited the Santa Rita Durao and Catas Altas quadrangles is near the portals. an area of flat-lying poorly lithified sedimentary rocks In the eastern part of the area occupied by the Ter­ named the Fonseca Formation by Maxwell (written tiary sedimentary rocks, the beds are nearly flat lying, comniun., 1965). The unit is about 85 m thick at the but near the western margin they steepen to an atti- PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A6D

tude of about 50° E. This steep cip led Brajnikov material, and, where found, they are small angulr.r (1947) to postulate a fault and an important Tertiary fragments of rocks cropping out near the occurrence. tectonism; careful search for evidence of faulting in The basal parts of the masses are exposed in only a few the adjacent rocks and tie Tertiary rocks by several places, but there the clay may contain a basal con­ geologists of our staff was, however, fruitless. It is glomerate composed of angular or rounded fragments rather believed that the steep dips are caused by slump­ of the underlying rocks and vein quartz ranging up to ing and consolidation of sediments deposited in a sink­ 40 cm in maximum dimension. (See fig. 19.) In one hole in the carbonate rocks of the underlying Gandarela body overlying iron ore, the basal conglomerate is com­ Formation. A small sinkhole lake choked with vege­ posed of hematite blocks. The bodies are not cut by tation, the Lagoa de Metro, still exists a few hundred quartz veins or by dikes. meters northeast of the Gardarela basin but will soon On the outcrop the "mudstone" presents a rugose sur­ be destroyed, as it is perched on a steep hillside caused face barren of vegetation, even of moss or lichen, except by deepening of the principal valley to a level about 50 where soil has washed or blown over the rock (fig. 20). m below the lake. A canga lip preserves the topographi­ Many outcrops are mottled white and brick red, per­ cally anomalous feature. haps more are brick red, and a few are white or choco­ Pomerene found thin Tertiary deposits in the valley late brown mottled with white or light tan. No vestige of the Rio do Peixe near the Miguelao dam, Macacos of bedding survives, but some bodies are cut by joints or quadrangle. Little is known of their extent or thick­ shrinkage cracks. In exposures where the third dimen­ ness, for they are exposed only when the reservoir is sion can be seen, it is apparent that the mottling is empty and are indistinguishable from recent sediments caused by bleaching of the colored clay in vertical or except for the fact that they contain a thin bed of fos­ nearly vertical cylindrical or elongate forms (Guild, silized leaves, identified by Roland Brown as Tertiary 1957, fig. 22). The white clay is formed by complete (Pomerene, 1964, p. D29). Similar deposits of clay bleaching of red or brown clay (Guild, 1957) caused b^ containing fossilized leaves were reported by Gair leaching of iron. (See table 2, analyses 1 and 2.) In a (1962, p. A41-A42) in the Nova Lima quadrangle; un­ few exposures the white clay has been bauxitized along doubtedly other occurrences of such material are still joints. to be found. A number of chemical analyses have been made of MUDSTONE the material from various bodies, and typical analyses Isolated masses of imstratified nonplastic laterized of the white clay are presented in table 2. Although clay, as much as 50 m thick and 300 m wide and as long most analyses available did not determine TiO2, those as 2 km, are found in much of the Quadrilatero Ferri- that did show an abnormally high content of titania, fero. They occur within relatively limited ranges of locally more than 5 percent. Such a high percentage is elevation and seem to be related to the post-Gondwana characteristic of laterites formed from mafic (Gold- erosion surface (p. A12). Although unfossiliferous, the schmidt, 1954, p. 419) or alkalic rocks. material is judged to be Tertiary or perhaps older. Ge­ ologists of the staff applied the field name "mudstone" to this material, the origin of which is puzzling. Although mottled clay is a common result of the laterization of granitic and other rocks in Brazil and other tropical countries, the material here described is different from such weathering products in that it is nonplastic and much more indurated. It never grades into saprolite and thus is not the result of weathering of rock in place but of some transported material, differ­ entiating it from the type laterite of India which it superficially resembles. The rock is composed of semi-indurated nonplastic clay containing in many localities randomly distributed sparse frosted grains of quartz 1-3 mm in diameter. The quartz grains probably make up less than one-half percent of the rock. Except in the basal part of the FIGURE 19. Basal conglomerate in "mudstone" near N,. 6,700 bodies, in few places are rock fragments found in the E. 12,900, Lagoa Grande quadrangle A70 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL Perhaps the most significant feature of tr 3 occurrence of the mudstone is its virtual restriction to elevations between 1,200 and 1,350 m in the Moeda Plateau, the Serra do Curral, and the southern part of the region, where it is most abundant. A small body near Santa Kita Durao and the larger body overlying the Chacrinha hematite ore body above the city of Itabira are some­ what lower, but perhaps 85 percent of the known oc­ currences, and all the larger ones, are between the cited elevations. On the Moeda Plateau and in many other occurrences the "mudstone" appears to be plastered on one or both sides of ancient high-level valleys, the floors of which have since been lowered below the bottom of the clay deposits. In other places the base level of local drainage is still above the base of the "mudstone1' deposits, which are very resistant to erosion because of their homogene­ ity and lack of bedding planes or other points of attack by running water. In areas where the general stream level is below 1,200 m, the clay may form topographic noses and isolated rounded hills at about that elevation. Thus, it seems very probable that the rraterial from which the "mudstone'' formed is related to the erosion surface between 1,200 and 1,350 m, called the post- Gondwana surface by King (p. All). The spatial relations of the "mudstone" with the underlying Precambrian rocks and with the old erosion surfaces have led to several interpretations of the origin of this rock. Guild (1957, p. 43), who was the first to FIGUEE 20. Outcrop of "mudstone" near N. 5,250, E. 4,200; Ibir- describe the "mudstone", believed it to be a weathering ite quadrangle. The rugose and pitted surface and absence of product of dolomite. It is quite true that much of the vegetation and soil are typical. material overlies dolomitic rocks, particularly the Gandarela Formation, but since the time of Guild's field- TABLE 2. Analyses of "mudstone" work a number of occurrences have been found which overlie phyllite, quartzite, and other rocks. Further­ more, the presence of basal conglomerates between the SiO2_.--.-.-.._.-______-. 36.2 44.1 46.4 42.0 45.6 43.1 42.2 "mudstone" and the underlying rocks, described by Ti0 8 .-...... (i) (i) 4.3 5.2 4.2 4.7 4.5 AUOs------32.9 38.9 32.8 36.4 33.2 35.3 36.9 Johnson (1962), Pomerene (1964), Dorr and Barbosa FejOs------.- 2 17.6 22.6 1.8 1.5 1.5 1.5 1.6 P. .-.-.---..--.., -.....-. 0.05 0.04 (3) (3) (3) (3) (3) (1963), and Wallace (1965), unknown at the time of Loss on ignition----____.__- 12.4 13.5 11.7 14.1 12.8 14.0 14.5 Guild's work, proves that the mudstone is not a simple Total...... 99.6 99.5 97.0 99. 2 97.3 9.7 decomposition product, but rather that the original 1 TiOa not determined; probably reported with AhOa. material from which the rock formed had been trans­ 2 Calculated from determination of Fe. 3 Not determined. ported. The rounded frosted quartz grains in the "mud- 1. Dark-red clay. Analyst, Dr. Cassio Pinto, DNPM (Guild, 1957, p. 43). stones11 have not been observed in the Gand arela Forma­ 2. White clay, same locality. Analyst, Dr. Cassio Pinto (Guild, 1957, p. 43). 3-7. Clays. Analyses by Cia. Magnesita, S. A., Belo Horizonte (Wallace, 1965, p. F30). tion, a chemical and fine clastic formation, although they could have come from the quartzite? of the Cer- "Mudstone11 is most abundant in the west-central part cadinho Formation. of the Quadrilatero Ferrifero and particularly on the Pomerene (1964, p. D31) believed the material Moeda Plateau. Several large bodies are known in the formed by laterization of siliceous ferrugirous dolomite, southern part, particularly near Dom Bosco, and on the modified by soil creep and slump. Gair (1962, p. A30) north slope of the Serra do Curral east of Jangada, but believed it may have formed from reworked ocherous in the easternmost part of the region it occurs only in itabirite and emphasized the resemblances to mudflows. the Itabira district. Johnson (1962, p. C25) believed that the material was PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A71

not derived from dolomite but that it had been trans­ by rapidity of accumulation of fine homogenous material ported, filling old channels. Wallace (1965, p. F29), who and the effects of later laterization. mapped the area in which such deposits are most wide­ Neither the hypothesis of origin put forth above nor spread, showed that their distribution is not closely any other yet advanced satisfactorily accounts for the related to bedrock lithology. He attributed the material presence and even distribution of the rounded fronted to further weathering of preexisting bauxite or lateritic quartz grains found so ubiquitously in this material in soils. Wallace was the first to note the close relation of the region. the "mudstone" to a restricted elevation range in the An hypothesis of origin which would explain the pre­ area mapped by him; the writer found that this was true sence and distribution of quartz grains in the rock and on a regional basis, with a few exceptions. the lack of stratification would be that the "mudstone" The writer believes that the close spatial association is thoroughly laterized windblown dust, a type of lo°ss, with the Gandarela Formation in many parts of the dating from an ancient arid cycle. By this hypoth°!sis region is due to the normal topographic expression of it would be difficult to explain the abnormally high ti- that formation, which provided an ideal setting for the tania in the rock and the close relationship of the "mud- accumulation of the material, the laterization of which stone''' to a certain range in elevation, unless it were plso gave rise to the "mudstone/' postulated that the material forming the loess wa? in The abnormally high titanium content of the rock fact derived from volcanic ash during early Tertiary or may provide a clue to the original material from which late Cretaceous time. The origin of this distinctive rock it was derived, for none of the other rocks of the region must be considered to be still open to question. contain abnormal quantities of this element. Chemical analyses of the iron-formation and dolomites show very WEATHERING PRODUCTS low titanium. Herz (1962, p. C75) has shown that the Several materials formed by weathering processes phyllitic rocks of the Minas Series are low in titanium. from the underlying bedrock are sufficiently widespread The bauxites in the region, commonly the site of con­ to be mappable or of geologic interest. Those of economic centration of titanium, are not abnormally high in ti­ value, such as bauxite and canga, will be discussed li ere tanium. It therefore is probable that the material from as geologic rather than economic features. which the "mudstone''' formed may have been introduced from sources outside the region. SAPROLITE Such a source might be found in windblown volcanic The most common weathering product in the region is ash derived from the volcanoes thought to have accom­ saprolite, a. term applied by Becker (1895) to thorough­ panied the emplacement of the titanium-rich alkaline ly decomposed, earthy, but untransported rock. In this intrusive rocks a few hundred kilometers to the west and report, the term is restricted to weathered rock whose south of the Quadrilatero Ferrif ero. Large complexes of original structure and texture is preserved but in which such volcanic rocks are found in the vicinity of Araxa most original mineral constituents other than resist?,tes and POQOS de Caldas in Minas Gerais, in Sao Paulo, and have been so altered by weathering processes that they Rio de Janeiro. Those in Araxa and POC.OS de Caldas are have been destroyed. Most polyniineralic rocks, such thought, on the basis of age determinations, to have as gneiss, phyllite, graywacke, and other rocks, were been active between 60 and 100 million years ago, or in altered to saprolites almost everywhere and completely the early Tertiary to middle or Late Cretaceous (Herz, decomposed; their feldspars and much of the mica were written commun., 1965). This age assignment is not radi­ altered to clay minerals. Pure quartzites were disaggre­ cally different from that made, on completely different gated through solution of cementing material. After grounds, by King (1956) for the post-Gondwana surface some experience, one can with confidence broadly clas­ on which most of the "mudstone'" deposits are found. sify rock on the basis of textures and relict minerals Possibly windblown ash from explosive eruptions in the saprolite. Whether particular granitic rocks were blanketed the ancient surface and was concentrated in granites, quartz monzonite, or quartz diorite cannot be the valleys in that surface by slope wash and local mud- established with certainty except in fresh rock. Struc­ flows. The deeper valleys would have been choked and tural features in weathered rock tend to be preserved. filled with such ash, which would have been very fine to Decomposition of feldspars and micas results in im­ have traveled so far. Basal conglomerates are to be ex­ poverishment of the more mobile elements, particularly pected in such an environment, as preexisting cobbles the alkalis, calcium and magnesium, with a coordinate and boulders would be surrounded by the fine introduced increase in the more stable elements such as aluminum, material and would be easily transported by mudflows. and, under some circumstances, iron. The stability of The complete lack of bedding perhaps can be explained quartz is believed by the writer to depend primarily on A72 REGIONAL GEOLOGY OF THE QUADRIL ..\TERO FERRIFERO MINAS GERAIS, BRAZIL

the grain size of the mineral (Dorr, 1964, p. 1230-1231) ; (1958, p. 71-73) re,·iewed the evoh ing usage of the where finely cli,·ided as in the Tabooes quartzite, some term since its introduction by Buchanan in 1807 to de­ itabirite, some phyllites, and some schists, it is fairly scribe certain alumino-ferruginous material in India fugiti,·e. \Yhere quartz is coarse grained, as in many derived from the weathering of basalt. Recanse of its sedimentary grits and granitic gneisses, it is quite stable. lack of specificity, the term as here used is defined, not In itabirite, leaching of quartz is much reduced where with the thought that ·this is the only correct usage or its grain size is grea,ter than 0.1 mm. Mafic dikes and the that the definition includes all conceivable laterites, but gabbros do not form deep saprolites because the complex rather to help the reader relate the material called later­ silicates alter readily and contain no primary resistates ite in the QuadrihHero Ferrifero to that of other areas to presen e orig-inal textures. in which the term may be used somewhat differently. Laterite is used herein for the st11tctureless material, SPLASH ROCK commonly highly aluminous and in many phtces ferru­ Carbonate rocks are exceedin(Yly soluble in this region ginous, formed by residual weathering of any rock ex­ and weather deeply. 1\"'here protected from mechan­ cept orthoquartzite or quartz n~ ins. 1~Teathering, in this ical erosion, the weathering product is an unusual ma­ climate, results in the breakdown of silicates, in decrease terial consisting of a tine cellular boxwork of limonite of the silica content, and in concentration of hydroxides and, locally, manganese oxide. These insoluble minerals, of iron and, particularly, a.lnminum. Laterite is dis­ deri,·ed either from the carbonate rock itself or nearby tinguished from saprolite on the basis of the lack of rocks, are deposited around the indi,·idual crystals of residual structure, from most soil on the basis of its carbonate. The latter are subsequently dissolved, leav­ residual nature and Ia ·k of organic material. Certain ing a microscopic boxwork of limonite or manganese residual ·oils are al o laterite, and such soil is called oxide around ,-oids now filled with water. The material latet·it ic soil. This writer does not include ·anga as a has se,·eral identifying characteristics : when struck laterite, although it could be argued that it is a special with a hammer, the hammer sinks into the material with product of laterization. a splash; when rolled between the fingers, small frag­ Laterization of most rocks in this region results in ment ha,·e some initial strength but collapse suddenly, a. high per ·entage of bauxite minerals, other hydroly­ lea\ ing a moist or wet fine paste which roughens the zat-es, and resistates in the laterite or lateritic soil and skin; and when dried, a chunk of the material will a gradual impO\·eri ·hment in iron oxide and hydroxide. usually float in water before disintegrating as capillarity As shown by Cruild (1957, p. 43), clay co ntaining more sucks the water into the cells with such violence that the than 1:2 percent iron became virtually iron free on con­ walls break. tinned weatherinu, and the alumina content rose accord­ Guild (1957, p. 42--:!-H) and Pomerene (196-!, p. D18- ingly. Some buff, pink, or white lateritic soils derived D19) were the first to demonstrate that such material is from granitic rocks in this area contain significant, the weathering product of dolomite. The material had though noneconomic, quantities of bauxite minerals as been discussed by earlier ~eologists as earthy iron oxide established by difl'erential thermal analysis (S. ~.Gold­ or earthy man~aniferou s rock (Henwood, 1871 ; Scott, ieh, oral co mmun., 1!)5;3). 1900). Soft clayey manganiferous and limonitic rock Although much of the material considered as laterite propel'ly described as earthy is also abundant but should or lateritic is red, much of it is light pink, light yellow, be cliff •rentiated from the material here described as buff, or '"hite, owing to la k of iron. Some of the com­ "splash rock,:' a.s it is not necessarily deri,·ed from car­ mercial bauxite in the region is red because of the high bonate rocks. iron content; other bauxite in commercial depo its is "Splash rock ' is an excellent guide t.o underlying light buff to nearly white (fig. 21). Thus, the definition dolomite or limestone in deeply weathered terrain. Much of laterite as "red residual so iF: gi,·en in the glo sary of of the rock mapped as clo1olllitc in the Quadril(ttero Fer­ the American Geological Institute ( 1960, supplement, rifero wa s on the basi. of the presence of "splash rock" p. 37) is incomplete, for the bauxite is an end product at the outcrop or in the ndits and cuts in the zone of of laterization. weatherin~. No other rock yields this weathering Laterization depends upon good subsurface drainage product. to ca rry away the soluble co nstituents of the material LATERITE being leached. The best de,·eloped laterites are on hjgh The term ' laterite" ha s been used in many senses in plateaus, terraces, or shoulders or flanks of ridges where the geologic literature and must now be considered a the strong subsurface drainage is stimulated by favor­ very genera 1 term indeed. Gordon, Tracey, and Ell is able topography. PHYSIOGRAPHIC STRATIGRAPHIC, AND TRUCTURAL DEVELOPMENT A73

lar to the material in Inrlia for which the term wa coined. It charncteri tic texture and red or brown color i an excell nt O'uirle to the underlyin(Y rock. f..Alteritic oil OY~r granitic rock i thick and lo ·ally hi(J'h in alumina. Q,·er phyllite mo t lateriti soil is thin and . terile, the in oluble hydrolyzat forming a liO"htly re i tant ru t a few millim t r or centimeter thi k. Tan limoniti nodule , red or black inside, and om­ monly with a quartz (Yrain a a m1 1 u · and lo all ' magnetic, form in place OY r rtain phyllite zon . Protoquartzite may form a lat~riti soil b) r mond of ilica and breakdmYn of mi a and other alnminon. mineral . Ferruginou I i olitic con r tion ar not rare on protoquartzit that contain appre iabl hematit or mag-n tite. Fo. ~ il oil and cn nCYa mn) be ome later­ ized ex mplified in the Itabira di tri t (Dorr and Bar­ bo. a, 1963, p. 3:..- 33), as may other unco n o1idat d or semi on olidat d ed iment protect d from orr si tnnt to ero ion, u h a the ' mud ton :· depo it d ribed abO\ e. arbon a te-ri h ro k '"hen nbject to in ten lat rizn­ tion, produce c rtain clHnn ct ri ti £ atureL \lumi­ non carbonat rock is the be t our e ro k for t h omm rcial bauxite l po ·it . The pur r rlolomit and lime tone. ma ' form a. d ep porou chocolnt -brown to reel lateritic oil wh i 'h if t h originn 1 ro k wn man(Yanoan or f rroan, contnin abundant pi ·olite or nodule of mangane e oxide or 1imonit (D IT and other , 1956, p. 30 .... -305). Lo ally, n on th border of the Gnndar la and Gon rization profile, .Mut uen bauxite mine. :\. ,000, · La.t~rite and lat riti oil chnract ri . ti nll) are K 2,:JOO. ~I atato~ (]mHlrnngle. 'l'be uppe rmost zon is rhemi a l poron and oak np water like a pong . Deep cut in ·an~a (c). The nt:'xt zont:' i · light-brown to tan highly a lumi­ om lo alitie. re,·ea l th·tt interconnect ing ·hnnnelwa. .,. non. .,; and fNrnginon~ hard <"H il ga (ch), with strongly mark d n •rti<-al lintin~ ::; . Hen<'ath tht:> <·a n~a i ·a zone of pink to bnff de,·elop in th material, for, luring hard rain. , water <·ommerda l ba uxit ~ (b) 1 -~ m thick. ThE' hamnwr i in ba nxit . l urt: from . tnall tubular openin(Y · n if from an open fk low thP hauxite i:-; a Ynrying- hut u~unlly thick hig-IHllnmina p1pe. elay, a t lenst 10m thiek in this area. Her' the c: hemi a l cn nga CANGA at tlw top may w ell be :-;ompwhat younger than the unde rlying Canga i: a rock type originally de:cribed four zones and unrE> ln ter it wa . Guild. H 1 fin d laterite n un­ t it ore. The ro k i important from a. phy ioo-raphi ·<:>m<:>ntecl iron-formal ion d bri talu · and deep-re 1 Yi ewpoint and a 1 o from an e onomi Yiewpoint, as it f rrnginon: .oil" 1957 p. +1). To the writ r: know]- ha b n u eel a an ore of iron. clg<:> tlw fir. t two cia . . ifi cation would b ex ·luded from Ca.no·a ha been defined (Dorr, 196±, p. 1219) a a la,terite ttnd er any om monl) n c pted definition· th lithifiecl ·m·ficinl or 11 ea r- urface rock formed of widel) la :t is here ·on ·i lrred lnterit . Th fir. t two ro k typ ,·nrying quantiti s of detrital frngm nt , ommonly were mapped by other geologi. t of th t·eam a. collu,·ial high-oTade hemntit or iron-formation, cern ented b) ore and a: taltt: or rnbbl ore · they ar ,. r) wide pread. limonite into a blanketing body. Thi ' i e entially the rabbro, liaba. atHl amphibolite w ath r to a nm mea ning gi,·en to th term by Derby (1910, p. )1 , lat ritr high in iron an l alumina that i probnbl 1m1- :20- ' 1) and by Iln rder and hamb rlin ( 101.), A74 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MEXTAS GERAIS, BRAZIL p. 392). Gorceix (1884, p. 80-83) defined the rock as a cal weathering and tough and resistant to mechanical ferruginous conglomerate cemented with an argillaceous weathering. Canga is effectively attacked by mechanical and ferruginous cement. Gorceix defined the rock as it erosion only by undercutting of the edges of the sheets, appeared in a deposit now known to be a ferruginous where removal of the deeply weathered anc1 soft under­ bauxite (I. S. Coelho, oral commun., 1964) ; clay or lying rocks is easy (fig. 2). Data on the thickness of the aluminous material is not ordinarily a major constituent canga sheets are scanty. Scarps on the edges of the sheets, of most canga in the Quadrilatero Ferrifero. The term and subsurface exploration show that the maximum "ferruginous conglomerate'' is, strictly speaking, cor­ thickness is more than 30 m; in many places it is be­ rect, as the rock is composed of cemented coarse detrital tween 1 and 5 in and is commonly more than 50 cm. material; it is, however, misleading in that conglomer­ The surface of the canga-capped plateaus underlain ate is commonly a cemented waterlaid gravel, whereas by the Caue Itabirite and of the canga-covered slopes canga is quite different in appearance, mode of origin, is bare in part,; in part it is covered with a layer of iron- and composition from waterlaid gravels. In the region rich, generally highly porous soil containing coarse fer­ much ferruginous conglomerate that is cemented with a ruginous detritus a few centimeters to a few tens of centi­ ferruginous or argillaceous cement is not canga. meters thick. In few places has canga been found under The detrital fragments in canga are derived from any great thickness of soil, although it has been found immediately adjacent or underlying rocks and were under laterite, mudstone, and lake beds(?) at depths up transported by slope wash or creep rather than by to 100 m. The vegetation on canga is xerophytic and stream action. The itabirite, quartz, and phyllite frag­ orchids, cacti, and widely spaced low stunted bushes ments are angular to subrounded; those composed of and gnarled trees are typical. high-grade hematite may be angular or may be rounded Canga is most common above the Caue Itabirite and by weathering (Dorr and Barbosa, 1963, p. C76). They the Gaiidarela Formation and laps out over adjacent vary in size from 1 mm to more than a meter but in most formations. No large bodies of canga and only a few places range between 1 and 10 cm. small ones are known over the iron-formations of the Fragments of hematite in the rock are hard and Nova Lima Group, although boulder trains of canga- unhydrated, but itabirite fragments are more commonly like material are not uncommon over those rocks. Prob­ thoroughly hydrated with the quartz leached com­ ably the pre-Minas iron-formations are too thin to give pletely or in part, being partially replaced by limonite. the proper physiographic setting for the development of Phyllite fragments are not common except over or near canga; these rocks crop out as rather narrow ridges phyllitic rocks and are replaced to greater or lesser rather than flat-topped plateaus. It is also quite possible degree by limonite. Vein-quartz fragments are angular, that the fact that the pre-Minas iron-formations are car­ but most of them break down into smaller pieces with a bonate facies rather than oxide facies has inhibited the light blow; leaching along crystal interfaces has development of canga, for ferrous iron is much more weakened them. soluble than ferric iron under near-surface conditions, The cement in canga is generically called limonite. and the iron may have been to a considerable extent It is composed of several varieties of hydrous iron oxide, dispersed from the site during weathering. very locally with significant admixture of hematite or Canga has been subdivided into many types, but a of AIO(OH), which may be in the crystal lattice of gross division into three types is adequate for most goethite (Correns, 1952). The cementing limonite ranges purposes. These subdivisions are normal canga, chemical from yellow through tan, reddish tan, and brown to canga, and canga rica. nearly black, depending on the minerals, the amount of Normal canga contains between perhaps 20 and 80 clay or alumina, the degree of crystallization of goethite, percent detrital fragments of hematite or itabirite, lo­ and the porosity. The porosity ranges from very low cally with a little phyllite and quartz. It forms primarily in certain types of canga to perhaps as much as 50 per­ on the old erosion surfaces and on slopes up to 20° on cent in other types. Permeability is low except in one Caue Itabirite but may be found on other rocks, even type. The amount of cementing limonite in the rock granite, where topographic conditions favor supply of may range from perhaps 5 percent in certain types to the proper detritus and dissolved iron to cement the more than 95 percent in others. detritus. Normal canga contains between 64 and about Canga forms large blanketing sheets on high and in­ 50 percent iron depending on the degree of hydration termediate surfaces; some sheets are measured in tens and the amount of quartz. ALO3 is rarely as much as 6 of square kilometers. As discussed on page All, these percent; commonly it is less than half that amount. surfaces defy erosion because the canga is inert to chemi­ Figure 22 illustrates rather rich normal canga. PHYSIQGRAPHK, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A75

FIGURE 22. Broken fragments of rich normal canga, about 62 FIGURE 23. Canga rica, at the Aguas Claras deposit, near N. percent iron, lying on a typical canga surfac*. Note botryoiclal 5,500, E. 10,200, Belo Horizonte quadrangle. The material here forms and concentric- growths. Lagoa Seca, near N. 3,800, E. will average over 67 percent iron. The blocks and fragments 6,100, Belo Horizonte quadrangle. of hard hematite are firmly cemented by limonite.

"Chemical canga" is the name applied to canga very publication (Dorr, 1964, p. 1219-1223, 1232-1236). low in detrital material and high in limonite. It is most Clearly the detrital part of the rock is derived in large abundant 011 the gentle slopes in the valleys and swales part from the Caue Itabirite and the hematite lenses on the Gandarela Formation several hundred meters therein included. It is believed that iron in the cement­ or more away from the Caue Itabirite, although it too ing limonite was dissolved from the formation during may form on any rock if the topographic position cuts the weathering process, transported in the ferrous state, off coarse iron-rich detritus and supplies solutions rich and redeposited at or near the surface, where the trans­ in iron to an environment where it may be precipitated porting waters became oxygenated and changed the at or near the surface. This variety is generally light soluble ferrous iron to insoluble ferric iron. Some iron brown to tan, contains 6 to more than 10 percent alu­ is probably carried as a colloid and some as ferric iron; mina, rarely more than 55 percent and in many examples both would be deposited by evaporation of ground water less than 50 percent iron, and is dense and brittle with near the surface during the dry season. very low porosity and permeability. The detrital frac­ tion is commonly about 5-10 percent of the rock, and it ALLUVIUM is angular and relatively fine, with fragments commonly The fact that the region is now being actively eroded smaller than 1 cm. Beds are usually thin and covered and that most of the streams are well above base level, with soil; the rock type forms extensive surface crusts combined with the fact that weathering in this region in few places. This type of caiiga is similar in aspect to breaks most of the rocks down to clay and silt-sized ironstone concretions, although no iron carbonate is particles, which are easily removed during the heavy known from such bodies. freshets of the rainy season, make major accumulations Canga rica (rich canga) is hematite rubble ore ce­ of alluvium rather uncommon in the Quadrilatero Fer- mented by limonite. The cement may fill all the inter­ rifero. The few places where such alluvium does occur stices between the hematite pebbles and cobbles or only are just above canyons where major streams cut through enough may be present to hold the pebbles and cobbles resistant rocks, as near Honore Bicalho and Sahara on together, with most of the interspaces empty. The iron the Kio das Velhas; where stream gradients abruptly content is commonly about 64 percent, in many cases flatten, as on the Rio ConceiQao southwest of Barra more than 66 percent. Quartz and alumina are very low, Feliz; and where mining activity or rapid erosion of commonly less than 2 percent. The hematite fragments soft itabirite has dumped large amounts of sand and may be angular or rounded and range from small pebble gravel into a stream too small or with too low a gradient to boulder size. Figure 23 illustrates a typical outcrop to carry it all away, as between Catas Altas and Agua of canga rica. Quente and below Itabira. Various hypotheses of the genesis of canga have been In the smaller streams narrow zones of discontinuous advanced and were discussed at some length in a recent alluvium are found. East of the Moeda Plateau and 334-413 O 69- A76 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, UNAS GERAIS, BRAZIL Congonhas do Campo, many of these were worked for foundry sand. Tie purity is such that much qualifies as gold during the 18th century, and both these deposits glass sand (Walace, 1965, p. 46). and some larger ones are marked by irregular heaps of LANDSLIDES coarse gravel and unnatural drainage patterns, attest­ ing to the thoroughness of the old placering operations. The breakdowi through deep weathering of much of In few places is there enough sand and gravel to sup­ the micaceous aid feldspathic rock to clay, coupled ply more than local needs for these commodities; with the deforestation of the region since the Euro­ crushed rock is commonly used for concrete aggregate in peans moved inx> the area much of it in the 19th and place of gravel, and sand is dredged out of the rivers 20th centuries and the practice of construction of long themselves and trucked for considerable distances to canals to bring water to the site of gold washings dur­ supply the building industry of Belo Horizonte and ing the 18th anl 19th centuries, plus severe overgrazing other cities. by domestic animals all have combined to cause many Where soft itabirite has been rapidly eroded or ex­ landslides in tie region. Most are small, a few hectares tracted on a large scale during mining operations, as or less in extent, but some are larger. The soft phyllites just south of Cat as Altas and below the Grot a de Es- of the Minas Series, and particularly the phyllitic phase meril in the Itabira district, the alluvium contains high of the Cercadmho Formation, are preferred sites for percentages of hematite in sand sizes and as gravel. such slides. The construction and maintenance of high­ Locally, this material has been so sorted by the streams ways has been seriously hindered by landslides caused that a banded unconsolidated rock composed of hema­ by location of the roadbed without regard to the in­ tite and quartz is formed; it is very similar in appear­ herently unstable condition of the subsoil in certain for­ ance to some of the ferruginous quartzite of the Cerca- mations and under certain topographic situations. dinho Formation or the conglomeratic " iron-forma­ IGNEOUS AND GRANITIC ROCKS tion1'' in the Itacolomi Series near Fabrica in the Congonhas district, which are believed to have formed The igneous and granitic rocks of the Quadrilatero in an essentially similar manner. Large quantities of Ferrifero consist of large masses of granitic rocks and hematite are contained in such alluvium and some day smaller bodies of gabbro and metagabbro, diabase dikes, may well prove to be of economic importance. Dredg­ hypabyssal porphyritic dikes, and many small bodies of ing operations for gold below Mariana on the Rio chlorite schist, talc schist, soapstone, and serpentinite Carmo, during the decade of the 1950's, recovered sig­ that are alteration products of igneous rocks. Metatuffs nificant quantities of very high grade magnetite which in the Rio das Velhas and Minas Series were considered could not be sold because of the cost of transportation as sedimentary rocks because it is believed that they from that locality. were waterlaid and mixed with varying quantities of normal clastic sediments in a eugeosynclinal TALUS AND COLLUVIUM environment. Erosion of the high itabirite ridges and hematite de­ Because Norman Herz has prepared f- separate chap­ posits has given rise to major accumulations of talus ter of this professional paper on the geology, petrogra­ and colluvium. Because during the weathering of ita­ phy, and petrology of the igneous and granitic rocks of birite most of the quartz is removed, these accumula­ the region, these rocks are not discussed in any detail tions are of economic interest and have provided in this chapter. This report presents only those generali­ millions of tons of high-grade and medium-high -grade zations needed to understand the stratigraphic and iron ore for export and for domestic use. Such deposits structural development of the region; detailed evidence are found on the flanks of ridges; mining has shown and information will be found in the cl apter by Herz. some of the colluvial accumulations to be as much as 30 m thick. They tend to accumulate most extensively GRANITIC ROCKS in old valleys and indentations in the ridge side, and The generalized geologic map of the region, plate 1, most are not continuous for more than a few hundred illustrates the distribution of the granitic rocks within meters, although some are much longer. the area under discussion, but to fully understand the Of the other rocks in the region, only quartzite sup­ granitic rocks within the region and their relation to plies much talus, and this tends to consist of large iso­ the overall geologic setting, a short description of a lated blocks which break down to sand size on weather­ wider setting is essential. To the east and southeast of ing. Accumulations of sand derived from serrate the mapped area, granitic rocks and gneisses are con­ quartzite outcrops in the Serra do Itabirito, a type of tinuous to the Atlantic Ocean, with a few small islands colluvium, are locally very pure and have been used for of recognizable metasedimentary rocks, generallv at a PHYSIOGRAPHY, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A77 very high grade of metamorphism, Although these ite through mafic granodiorite. Most are foliated to rocks have not been studied in any d'tail, it is certain greater or lesser degree, being properly called granitic that they comprise both ultrametamo-phosed sedimen­ gneisses. The largest bodies of nonfoliated granite, the tary and intrusive rocks. A long belt »f charnockites is Borrachudos and Peti granites, are characterized by known from Sao Paulo through southrn Minas Gerais strong linear structure caused by alinement of biotite to Espirito Santo, and it probably continues to the blebs. Some of the granitic gneisses are clearly ortho- north into Bahia. It is believed that ihe area between gneiss, others are clearly paragneiss (fig. 24), and many the Quadrilatero Ferrifero and the sea was the focus are of uncertain origin. Unquestionably there has been of at least one post-Minos orogeny. much remobilization and anatexis (Guimaraes, 1961). To the north, the granitic rocks shown on plate 1 con­ The task of unraveling the relations between all these tinue for some tens of kilometers in most places, before rocks is much complicated by the prevailing deep being overlain disconformably by the Sao Francisco weathering of granitic rocks, for it is difficult to solve Series. The Sao Francisco Series is shown on most maps petrologic problems from saprolite. as Silurian, but many Brazilian geologists now believe Radiogenic age determinations on rocks in and near that the rocks are at least in part late Precambrian. Al­ the Quadrilatero Ferrifero have shown that there were though the granitic rocks adjacent to the Minas Series two major epochs of granite formation or emplacement: north of the Serra do Curral are everywhere younger one about 2,700 million years ago, the other about 1,350 than the Minas Series, a few kilometers farther to the million years ago (Herz, written commun., 1965). An north older granites are present in the central and west­ event resulting in widespread ages of biotite of about ern parts of the region. North of the eastern part of 500 million years, in rocks the feldspar of which is much the Quadrilatero Ferrifero, the rocks and structure are older, also occurred; with the exception of a pegmatite most complex; younger gabbroic rocks are widespread, in the extreme northeastern part of the region, no gra­ and the Serra do Cipo, made up of quartzites at least in nitic rock of that age is yet known within the Quadri­ large part correlative to the Cambotas Quartzite, forms latero Ferrifero. Radiogenic ages of minerals in granitic a broad north-trending arch for several hundred kilo­ rocks which fall between the cited ages are known but meters, the east flank of which is apparently overlain by are presently interpreted as recording modification of higher rocks in the Minas Series intruded by post-Minas the minerals by later events rather than as representing granite. true rock ages. Much more detailed mapping of the To the west, granitic rocks and rocks believed to be granitic rocks and many more radiogenic age deter­ equivalent to the Rio das Velhas Series are known for minations will be needed before the absolute ages of all tens of kilometers. The granitic rocks are believed to these rocks can be determined with confidence. be in large part pre-Minas, although post-Minas gran­ ites almost certainly occur, as attested by some metasedi- mentary units included in them which, on stratigraphic grounds, seem best related to the Minas Series. To the south, a complex mixture of Rio das Velhas Series metasedimentary rocks and post-Rio das Velhas granitic rocks extends for at least tens of kilometers. Although part of this area has been studied in recon­ naissance, more detailed studies and much age work will be needed before it is fully understood. Minas Series rocks end a short distance southwest of the mapped area but probably reappear farther to the southwest. The Serra do Tiradentes in Sao Joao del Rei, 100 km to the southwest, is composed of quartzite lithologically similar to the Cambotas Quartzite, and the unconformably underlying metasedimentary rocks FIGURE 24. Paragneiss forming from schist. About 20 km are lithologically very similar to the Nova Lima Group. southwest of Itabira on the new road from Belo Horizonte to Granitic rocks there are younger than the eugeosyn- Itabira. The light material is faintly foliated granitic gneiss; clinal sediments, and it is most probable that granitic the dark is biotite (?) schist. Concordance of foliation from one block of schist to another is apparent. This material is rocks of several ages exist. saprolite; vertical lines are the result of rill erosion in the The granitic rocks within the Quadrilatero Ferrifero bank and show the greater resistance of schist and phyllite to form a complex suite ranging in composition from gran­ mechanical erosion. A78 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, UNAS GERAIS, BRAZIL

From the stratigraphic and structural viewpoint of Pre-Minas graiite is not known in the eastern part this paper, the most important results of the age work of the region. It may have once existed there and been are that field evidence for two or more orogenic events remobilized completely, as that area if one of high is strongly fortified and that there is definite evidence metamorphic grade relative to the rest of the region, that granitic masses of considerable size existed before and the focus of the post-Itacolomi orogeny was to the the orogeny which deformed the Minas Series. Such east and southeast of the region. The vriter believes, masses inevitably affected tectonic style and oriented on the basis of iie tectonic style of the post-Itacolomi strain during that orogeny. Such confirmation of the orogeny in tha: area, that any preexisting granitic field evidence is most welcome because the older gra­ rocks there were deeply buried beneath the Rio das Vel- nitic masses have in places been immobilized at their con­ has sedimentary rocks, which were in lar^e part altered tacts with younger rocks and elsewhere younger gra­ in that orogeny to the post-Itacolomi granitic rocks now nitic rocks have been emplaced along their borders, as at exposed. the south edge of the Bacjio complex and just north of There is no reason to believe that the older granitic the Serra do Curral. (See fig. 25.) It is most difficult, if rocks exposed in and immediately adjacent to the region not impossible, to prove on direct field evidence that contributed sediments to the formation of the Minas any granitic rock in the region is older than the Minas Series rocks. With the exception of some granite boul­ Series, although there is much indirect evidence for the ders in tilloid of the Sahara Formation near Mariana existence of the older granite. and in the western Serra do Curral, no granitic pebbles Herz showed that the known distribution of the older or cobbles are known in the Minas Series, not even in granitic rocks is not uniform in the region; such masses the basal rocks immediately overlying granite in the occur in the Bac.ao complex, in the Bonfim dome to Serra do Moeda. Except for the Sahara Formation, the west of the Serra do Moeda, and to the north of the most Minas Series sediments are well differentiated. It Serra do Curral, north of a younger granite which has therefore seems probable that the abundant quartzite metamorphosed the Minas Series. It is believed, but in the Minas Series \vas derived from the granitic rocks not certain, that the granitic rocks intruded into and formed from the Nova Lima Group along the south now exposed well to the west and southwest of the re­ edge of the Quadrilatero Ferrifero also belong at least gion and was transported somo distance, to the site of in large part to the older group, although thin stringers sedimentation. Only one granitic pebble has been re­ of granitic material are found in the quartzite of the ported from the Itacolomi Series conglomerates, and Serra do Ouro Branco at one place. it is thought by this writer that the quai*tzose material in that unit was also transported some distance to the site of sedimentation.

ULTRAMAFIC ROCKS Throughout the Quadrilatero Ferrifero, many bodies of ultramafic rock are now altered to serpentinite, soap- stone, and talc-chlorite schist. The most common rock is soapstone, locally steatite; serpentinite is much less common. At contacts, the material is sheared to talc schist and smaller bodies are entirely sheared to talc schist (Guild, 1957; Gair, 1962; Dorr and Barbosa, 1963; and Pomerene, 1964). All workers since O. Bar­ bosa (1949) agree that the soapstone was derived from the alteration of dunite and periodotite. Serpentinite is commonly farther from contacts with granitic rock, and steatite is closer to such rock; thus, the steatization FIGURE 25. Granite intrusive into Piraeieaba Group schist north may well be related to the granitic intrusion (Dorr and of Serra do Curral near N. 7,000, E. 4,300, Serra da Piedade Barbosa, 1963, p. 48). Petrographic features of these quadrangle. Note alteration of schist at contact. The sapro- rocks were described by Herz (written co'ttmun., 1966). lite surface shown is nearly horizontal. Scale is given by a low bush in the granite and the hammer handle in rill in granite, In this report, certain problems concerning age and center. stratigraphic relations are discussed. PHYSIQGRAPHIt, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A79

DISTRIBUTION in the Floralia, Santa Barbara, and Catas Altas quac1 - The altered ultramafic rocks are jound in greatest rangles. This body is about 30 km in mapped length and abundance and in largest masses in tht Congonhas area, for most of its outcrop it is less than TOO m wide, al­ in the Itabira district, and astride the louiidary between though the maximum outcrop width is about 2,500 m. the Rio Acima and Macacos quadrangles. Similar but The rock consists for most of its outcrop of chlorite- more tabular bodies were mapped neir Caete. East of antigorite-talc schist, but at the south end, particularly the Serra do Caraca in the Catas ^\ltas quadrangle, between Agua Quente and Catas Altas, it contains inter- serpentinite and steatitic rock occur; ihey are mapped grading large bodies of serpentinite and steatitic rock together on plate 1 with greenstone scaist because dur­ apparently identical in aspect with that in many of tr?. ing mapping the poor exposures prevented separation bodies believed to be pre-Minas in age. Near Quebra of these rock types. In this zone, serpentinite and steatit­ Ossos, the schistose phase contains abundant discrete ic rocks are more common between Agua Qnente and euhedral crystals of magnetite as much as 2.5 cm on a Catas Alt as. South of Ouro Preto in the Santa Rita de side and euhedral limonite pseudomorphs after pyrite Ouro Preto and the Ouro Preto quadrangles, small bod­ as much as 6 cm on a side (Maxwell, written commm:., ies of such rock are also found but were not mapped, 1964). Unmapped talc schist occurs in the granitic rock south This complex body lies on the contact betiveen posV of the village of Suzana in the Marinho da Serra quad­ Minas granitic gneiss and the Rio das Velhas Series rangle, west of the Serra da Moeda. North of the Serra for much of its length, but at the south end it cuts tH da Piedade, small bodies of altered ultramafic rocks Minas Series and is in contact with the Tamandua, containing asbestos are entirely surrounded by post- Caraca, and Itabira Groups. Locally it may contain Minas granite. mappable slivers of itabirite believed to be derived from the Caue Itabirite. Thus, this rock too is post-Minas. AGE It has been cut by post-Minas faults, unlike the dikes re­ Two ages of ultramafic rock seem to be indicated by ferred to above which occupy post-Minas faults. This the occurrences in the Quadrilatero Ferrifero. With is the rock referred to by Harder and Chamberlin three known exceptions, all the bodies are in the Nova (1915, p. 356-357), who, not having mapped it in detail, Lima Group of the Rio das Velhas Series or in granitic considered it to be a metamorphosed lava flow that oc­ rocks derived from those rocks. In only one locality, curred during sedimentation of the principal Minas that astride the Rio Acima-Nova Lima quadrangle Series iron-formation and therefore attributed some of boundary, is the Minas Series in apparent sedimentary the iron in that iron-formation to a volcanic origin. contact with such rock; elsewhere direct evidence is not available. Since 1954 (O. Barbosa, 1954, p. 21), GABBROIC ROCKS AND DIABASE most geologists have agreed that these ultramafic bodies Gabbro, amphibolite gneiss considered to be meta- are pre-Minas in age, for ultramafic intrusive rocks are gabbro, and diabase probably were intruded at two or expectable in the eugeosynclinal sediments of the Nova more epochs. The oldest is amphibolite or amphibolite Lima Group and there is no indication that most of the gneiss, some so closely associated spatially with soap- bodies are post-Minas. stone in the Itabira district that Dorr and Barbosa Altered ultramafic rocks which the writer believes to (1963) considered that the original rock might be ci- be of a different age are known in three places. Two of magmatic Avith the dunite there that was thought to these one in the Itabira District, the other in the Ta- be the parent rock for the soapstone and serpentinite. mandua iron-ore deposit in the Macacos quadrangle Other considerable masses of such rock, not associated consist of talc and talc-chlorite schist. These are postore with the soapstones, occur in the granitic gneiss in that dikes and sills containing fragments of iron ore and district and are locally in lit-par-lit relation to Nova are foliated parallel to the foliation in the wallrocks Lima Group schists. In the Monlevade quadrang1^, (Dorr, 1965, p. 31; DOIT and Barbosa, 1963, p. C45, Reeves mapped a long thin body of similar rock in the C70, and fig. 34; fig. 13, this report). Given the present Monlevade Gneiss that he considered to be a metamor­ composition of these dikes, it seems reasonable to sup­ phosed mafic flow or dike (1966, p. E13). Comparable pose that they \vere originally ultramafic in composi­ rock is not uncommon elsewhere in the Quadrilatero tion. The rocks are post-Minas in age and were em- Ferrifero in granitic gneiss or the Nova Lima Group. placed before the end of the latest metamorphism and The material is similar in appearance and mineralogic after the high-grade hematite ore was formed. makeup to amphibolite in the higher metamorphic The third and most complex occurrence is the long facies of the Minas Series in the Monlevade district, septum of rock mapped as greenschist shown on plate 1 where it may represent metamorphosed impure dolomi- A80 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, 1VINAS GERAIS, BRAZIL tic (Reeves, 1966). This problem was considered by mentioned their presence south of the Ouro Branco Herz (written commun., 1966) and will be discussed in quadrangle. his forthcoming chapter of this professional-paper Simmons (196Ba) described the occurrence in the series. Fecho do Funil quadrangle (N. 2,200, E. 2,550) as a Unfoliated gabbro, rock intermediate in texture be­ weathered mafic dike, which he related to the diabasic tween gabbro and diabase, and diabase are found intrusions discussed above. This rock is1 characterized throughout the region as bodies ranging from a few by large (20 mrn) oriented feldspar ph^nocrysts, now meters to one mass nearly 2 km wide and 6 km long. all altered to clay, in a fine-grained brown weathered Even larger bodies of gabbro are known to exist north groundmass. The rock cuts the Piracicab^, Group of the of the area mapped. Their shape tends to be tabular, and Minas Series. Fresh rock of very similar aspect was they are controlled in many places by faults, joints, and found by Pomerene (1964) near the Fazenda Rosario, bedding. Most are dikes or sills, many are irregular in Ibirite quadrangle; Pomerene also related this rock to outline, and several are small plugs. Most thick bodies the diabase dikes. It is here in the staurolite zone of the exhibit chilled walls, some flow structure along walls, Sahara Formation. In the same quadrangle, where the and cores of unoriented more coarsely crystalline rock. road to Fazenda Jangada crosses the Se^ra do Curral, These features may be seen a few kilometers west of Sao weathered rock with relations and aspect similar to GoiiQalo on the Monlevade-Belo Horizonte highway those of the weathered rock in the Fecho do Funil quad­ and in the Itabira district at the east end of the Bor- rangle is found in the Gandarela Formation. In the rachudos water tunnel. Thin dikes have diabasic texture. Catas Altas quadrangle, Maxwell mapped a small body In the Serra do Caraga an intricate system of dikes of similar rock intrusive into the Cambotas Quartzite occupies thrust and normal faults. Fresh dike rock has at N. 7,650, E. 2,550. never been seen, and its petrographic nature is unknown. PEGMATITE DIKES Moore (1969) stated that the rock is mafic, shows thin chilled zones, and is generally unsheared, except in some Although Minas Gerais is famous for the wide variety fault zones. Maxwell (written commun., 1962) stated of minerals occurring in commercial quantity in peg­ that the rock has ophitic texture. Such dikes are prob­ matite dikes, the area under discussion is notably poor ably related to the less weathered diabasic and gab- in such minerals, and only in the MomSvade area has broic rocks to the north in the Serra das Cambotas, the production of valuable pegmatite minerals been pos­ Itabira district, and elsewhere in the region. sible. These deposits are in gneiss thought to have been The dike swarm in the Serra do Caraga is largely derived from the Piracicaba Group of the Minas Series; confined to the Cambotas Quartzite but also cuts the Aldrich, and others (1964, p. 329) dated feldspar from Maquine Quartzite and the Nova Lima Group. Prob­ one of them as 545 million years old by the rubidium- ably this distribution is related to the brittleness of the strontium method. The main area of commercially host rocks, for orthoquartzite of the Cambotas contains valuable pegmatites begins a few kilometers to the east the most dikes, protoquartzite and subgraywacke of the and southeast. Probably these and the pegmatite in the Maquine fewer, and the argillaceous rock of the Nova Monlevade area are of different magmatic affiliation Lima Group least. Similar dikes cut the Minas Series and age than those to be described below. elsewhere in the region and probably the Itacolomi Elsewhere in the region most pegmatite dikes are Series south of Ouro Preto. Such mafic rock thus is one thin a few centimeters to a meter or, very rarely, more of the youngest of the region, a fact also indicated by than a meter in width discontinuous, and are confined to granitic gneiss and stratified rock adjacent to or the unfoliated nature of dikes and their characteristic fairly close to the contact with granitic rock. A few ex­ chilled walls, indicating- that at the time of intrusion the ceptions are known: Guild (1957, p. 57) found a small wallrocks were cool. pegmatite body in the Caue Itabirite in the Casa da PORPHYRY DIKES Pedra area some 3 km from the nearest outcropping granite. Dorr and Barbosa (1963, p. 46) mapped peg­ Herz (written commun., 1964) described under the matite bodies perhaps 800 m from the nearest known name porphyry dikes, intrusive rocks which have been granite in the Itabira district. variously described by other geologists on the DNPM- Within granitic rock, pegmatite dikes and pegma- USGS team. These rocks are commonly weathered, and toid bodies are much more common, although also gen­ their relations and extent are obscure. They are known erally small. In the Rio de Pedras quadrangle, much in the Catas Altas, Ibirite, Belo Horizonte, and Fecho black tourmaline and tourmalinite is found in the Ba§ao do Funil quadrangles. Johnson (1962, p. B24) and Herz complex near the Rio de Pedras reservoir. Poor expo- PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A81

sures conceal the exact mode of occuirence of the min­ "cleavage'' quartz. It is closely related to and inter- eral, which is found as float on the surface of the sapro- graclational with fibrous quartz. lite and soil. Nearby, on the road to Tazenda Ajuda, a "Cleavage1' quartz was described by Guild (1957, p. pegmatite dike perhaps 10 m thick coirains coarse mica. 30), as follows: In granitic gneiss, several generations of pegmatite Coarse-grained vein quartz with several sets of closely spaced dikes are seen in quarry faces. Some cccupy ptygmatic fracture planes reflect light evenly from large surfaces mu"li folds, others have indefinite walls and grade into the as coarse-grained pegmatitic feldspar does from cleavage plants. surrounding granitic rock, and still others are straight, Crystal faces have not formed in most of this "cleavage" quartz, cut the other pegmatites, and have sharp walls. All are so that the relationship between these planes and the crystal- lographic directions cannot be determined in hand specimen. unf oliated and many have quartz cores. In those crystals in which the faces can be identified, however, The Borrachndos Granite has no pegmatite dikes as­ the fracture planes are at angles that do not correspond to the sociated with it, although this very coarse grained gran­ principal planes. Statistical studies with oriented thin sections ite itself approaches a pegmatite in its texture and con­ would aid in solving the problems presented by this quartz, tains abundant fluorite. but they have not as yet been made. In many localities the "cleavage" surfaces, which are QUARTZ VEINS rhombohedral, are curved and look like large curved Throughout the Quadrilatero Ferrifero quartz veins dolomite crystals. "Cleavage" surfaces may be several abound. They include quartz associated with steatitic centimeters across. The material commonly has a silhy bodies, "cleavage1" quartz, quartz veins that may to vary­ sheen or luster and is semitransparent. ing degree be regarded as metamorphic segregations, This type of quartz occurs in segregations and crors- and gold-bearing hydrothermial quartz veins. cutting short veinlike bodies most commonly associated with the Caue Itabirite. In many places it is found QUABTZ ASSOCIATED WITH STEATITIC BODIES in dolomitic itabirite, particularly where that rock is Steatite or soapstone is in many places associated with manganiferous, but it has not been seen in place in large masses of ''bull quartz/' Such masses can be seen dolomite. in the Itabira district (Dorr and Barbosa, 1963, p. 47) Fibrous quartz is most common in the same environ­ and in the Congonhas district (Guild, 1957, p. 29). The ment and is intergradational with "cleavage" quarfz. same relation occurs in a talc schist and soapstone body This type of quartz grows from the walls of the en­ near Suzana, Marinho da Serra quadrangle, and else­ closing rock toward the center of veins. It occurs as thin where. Such large quartz masses have not been observed (1 mm or less) to thick (5 mm or more) fibers or col­ near serpentinite, nor are they found with all steatite umns, some separated by hydrous iron or manganese bodies. The largest masses are found in the Itabira dis­ oxides, some only slightly stained by such material. T' a trict, where they are as long as 100 m and 20 or more fibers or columns may be as long as 15 cm, although most meters wide on the outcrop. Quartz there seems very are less. Maxwell (written comniun., 1962) showed that pure and contains no silicates, carbonates, or sulfides; the fibers were single crystals. In those examined by him, the material may be of economic interest. fine hairlike inclusions tentatively identified as rutile In the Itabira district these quartz bodies are believed were found oriented obliquely to the quartz fibers. If the to have been emplaced during the formation of talc inclusions are rutile, the quartz is probably not supe^- from serjDentine derived from the original ultramafic gene in origin as has been suggested. rock (Dorr and Barbosa, 1963, p. 47-48). The regional absence of quartz around the serpentine bodies lends QUARTZ AS METAMORPHIC SEGREGATIONS credence to the idea. The steatitization is thought to have Material in a vein or irregular mass is considered to been effected by hydrothermal solutions rich in CO2, be a metamorphic segregation if it was derived from also accounting for the local presence of much carbon­ the wallrock under metamorphic conditions with minor ate magnesite (Itabira) or ankerite (Congonhas) in transportation. Much quartz in the metamorphic rook the steatitic rocks. Because the steatitic bodies are close- of the region must have been formed as segregations to or in the granitic rocks and the serpentine bodies are during the inetamorphic epochs, as the bodies are form­ further removed, it seems reasonable to suppose that less pods or pockets, and their abundance is closely re­ the source of the fluids involved was the granitic rock. lated to the abundance of quartz in the enclosing rock. Thus, in such formations as the Moeda Formation, the CLEAVAGE AND FIBROUS QUABTZ Cambotas Quartzite, or the Maquine quartzose forma­ Aii unusual type of quartz is very common in the tions, quartz veins are very common indeed, whereas in Quadrilatero Ferrifero and was given the field name of such rocks as the Barreiro Formation and the more ar- A82 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MNAS GERAIS, BRAZIL gillaceous parts of the Nova Lima Group, they are rela­ GOLD-BEARING HYDROTHERMAL QUARTZ VEIN'S tively rare. With exceptions, such bodies are not The Quadrilataro Ferrifero has produced in the persistent along strike for many meters; more common­ order of 2,000 tor.s of gold since mining activity began ly they are podlike and seem to have no connection with in the early 18th century. Much of this was won from throughgoing structures or veins, although in many placers, more of it from hypothermal replacement places they are parallel to bedding or in shear zones or deposits in the Nova Lima Group rocks, r,nd much from in minor fold axes. supergene enriched deposits in the Caue Itabirite and A number of these quartz veins in the quartzite are the hematite ores associated therewith (Dorr, 1965). vuggy and may contain water-clear euheclral quartz Gold-bearing quartz veins have been mined in many crystals of a size usable for their piezoelectric properties. localities but generally speaking, do not seem to have Also to be regarded as a type of metamorphic segre­ been either rich, persistent, or very productive. No gation are the quartz-kyanite veins so common in lode quartz mines are now in operatior, and the sur­ the Cercadinho Formation. Such veins are found in face manifestations are obscure and hard to find, quartzose rocks also containing high percentages of owing to the generally deep weathering. Few through- aluminous material. Guild (1957, p. 29) was the first going quartz veins occur in the region and still fewer to suggest that the alumina of the kyanite had been contain the limonite-stained holes and casts of sulfide derived from the wallrocks, a suggestion concurred crystals that might be expected to be the surface mani­ with by all later workers in the area. festation of sulfide-bearing mineralized veins. Perhaps The quartz-kyanite veins are continuous for meters the most accessible occurrence of such material is in to hundreds of meters and ordinarily are concordant the Eio Acima quadrangle on the road from Eio Acima with the bedding of the host rock. They are normally to Ganclarela, near N. 3,400, E. 12,800, where a breccia composed of a quartz core, with kyanite growing in zone in quartzite is mineralized with quartz, pyrite and blue-gray or greenish-gray radiating crystals between arsenopyrite, and gold in low values. the core and the walls, oriented in many cases normal The primary gold mineralization in the region is to the walls. Such crystals may be as long as 15 cm; in hypothermal where the environment can be deciphered some localities they are in part replaced by pyrophyl- (Gair, 1962; Dorr, 1965). In the Santana mine near lite. Locally kyanite may be in the quartz core, but this Mariana (pi. 10), a system of short gash veins is found is not common. The largest and best exposures of such in the Batatal Formation and both the schistose rock quartz-kyanite veins are to be seen in the Itabirito and the quartz veins are mineralized with arsenopyrite quadrangle near N. 5,000, E. 500, in excavations made and some gold. Probably most of the gold occurrences in the. Botica kyanite mine, from which 1,450 tons of in quartz veins in the region were generally similar kyanite was produced (Wallace, 1965). Herz and Dutra to this occurrence short, nonpersistent, low grade, (1964) described the geochemistry of some kyanite and hypothermal in environment. Reconcentration by occurrences and came to essentially the same conclu­ weathering and in stream gravels probably accounts sions as Guild with respect to the type here discussed. for the originally rich placer and elluvial deposits. Other vein deposits which might be considered essen­ The major lode-gold deposits of the area are replace­ tially metamorphic segregations are the quartz-hema­ ment deposits rather than gold-bearing auartz veins. tite veins in the Caue Itabirite and the Cercadinho For­ mation. Such veins have been described by Maxwell REGIONAL STRUCTUEE (written common., 1964) and Guild (1957, p. 29-30) and are also known to occur elsewhere. In some veins, The Quadrilatero Ferrifero is an inland of meta­ many of which follow faults or fractures in itabirite, morphosed strata in a sea of granitic rocks. It is a quartz forms a core, and hematite is hard, pure, fine region in which only the roots of g^eat structures grained, and unfoliated, identical with unfoliated high- remain. A large but unknown thickness of younger grade hematite ore. Such hematite grades more or less strata once overlying the region has been removed, for abruptly into itabirite. In many localities both hema­ the youngest remaining sedimentary rocks do not sig­ tite and quartz contain, or are covered by, coarse nificantly differ in metamorphic grade from the oldest. specularite plates up to 2 cm across and by coarse talc. During major deformation, the rocks flowed plastically, From visual inspection, the bulk composition of the as shown by their internal structures and by the abrupt veins approximates the bulk composition of the ita­ thinning and thickening of formations which must birite; that is, about half the material by weight is originally have been relatively constant in thickness. hematite and half is quartz. Such veins may be from 5 Three major periods of deformation occurred in the to 50 cm thick. region. The first two, one between the Pio das Velhas PHYSIOGRAPHC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A8?

and Minas Series and the other between the Minas and of which was very severe, obscured and complicated the Itacolomi Series, are obscure, ard it is not possible earlier deformation and rendered interpretation of pre- to affirm either their degree or ther extent. The last Minas events most hazardous. The nature of the pre- and the strongest involved all Precambrian sedimen­ Minas sediments, eugeosynclinal and molasse suiter tary rocks; it is post-Itacolomi in ags. without throughgoing marker horizons which can be Epeirogeny and warping occurred within the named correlated throughout the region, further complicate? sedimentary cycles but cannot be proved to represent the evaluation of early structural events. major deformation or erosion within this area, although The Rio das Velhas rocks were folded before Minor undoubtedly significant orogeny and erosion occurred sediments were deposited, for the younger rocks rest in these intervals in the borderlands oitside this region, with locally profound angular unconformity upon the as attested by the sedimentary rocks themselves. Epeiro­ older. Regional evidence also shows that the pre-Mina? geny also affected the region in post-Cambrian time deformation was stronger in the west than in the east. and major Tertiary uplift resulted in the present moun­ Near the west end of the Serra do Curral, Simmonr tainous topography. (1968a) mapped local iron-formations in the Nova Lima, At the beginning of the major post-Itacolomi orogeny, Group in strong angular unconf ormity with the Mina? pelitic and chemical sedimentary rocks were fine grained Series. Although these iron-formations are now nearly and relatively weak, for they failed by folding and flat lying and arched, the overlying Minas Series har plastic flow. During the progress of this main erogenic been rotated through about 120° in later deformations; period, all those rocks recrystallized and grew in grain therefore, at the time of Minas sedimentation the pre- size, micaceous minerals formed from clays, cherty iron- Minas iron-formations must have been steeply dipping- formation became itabirite, sandstones recrystallized and strongly folded. South of the Serra do Curral in the into quartzites, siltstones became resistant quartzose Belo Horizonte quadrangle, the angular discordance be­ phyllites and schists, and the relative competency of the tween local iron-formations in the Nova Lima Group individual formations changed with time throughout and the overturned Minas Series proves strong pre- the whole erogenic process. Because of the changing Minas deformation in that area. To the south, in the nature of the rocks and secular variation in confining Serra do Moeda, the Minas Series is strongly discordant and shearing stress, zones of easy failure became rigid, with quartzite lenses in granitic gneiss thought to have buttresses formed, and, in the latter stages of the orog­ been derived from the Nova Lima Group (Marinho da eny, rocks fractured instead of folding and yielding by Serra quadrangle). Along the Serra do Itabirito, folds plastic flow. Granitic rocks were em placed and were hundreds of meters in amplitude with east-trending formed from preexisting sedimentary rocks, substitut­ axes in the Nova Lima Group are truncated by the reg­ ing large masses having isotropic structural characteris­ ular, generally north-northwest, trace of the basal Min­ tics for somewhat anisotropic bodies. as outcrops. A major quartzite lens in the Nova Lima This region is but a small fraction of the whole oro- Group in the Itabirito quadrangle is at right angles to genic belt and is on the northwest margin of that belt. the strike of the Minas Series (fig. 5). Thus, in the Intensity of deformation and metamorphism diminishes west-central and western parts of the region, strong pre- from southeast to northwest; therefore different parts Minas folding must have occurred. of the region have different structural patterns and dif­ To the east, this pre-Minas folding became less in­ ferent tectonic styles. Reports prepared by members tense. In the northwest end of the Gandarela quadran­ of the DNPM-USGS team detail the structural devel­ gle a 30° discordance in strike and dip between marker opment of specific quadrangles; in this report major beds in the Nova Lima Group and the overlying Minas structures are described and their interrelations dis­ Series is apparent. Farther to the east, in the Santa cussed on a regional basis. Barbara quadrangle, Simmons (1968b) found that the angular discrepancy between the Nova Lima Group and PRE-MINAS STRUCTURE the Minas Series was about 20°. An angular discord­ The only reference plane by \diich the nature and ance also exists in the Monlevade and Itabira districts extent of the pre-Minas deformation can be definitely in the extreme eastern and northeastern parts of the measured is the base of the Minas Series, which was region. Gentle fold structures in the Nova Lima Group very probably a nearly plane erosion surface with a are suggested by outcrop patterns in the Caete quad­ slight inclination to the east and without major relief rangle and probably in the western part of the Gongo at the time Minas sedimentation began. That reference Soco quadrangle (Moore, 1969), but farther to the east plane has been deformed at least twice since the begin­ these disappear and a moderate tilting of the rocks ning of Minas sedimentation. These deformations, one below the reference plane is the only clearly pre-Minas A84 REGIONAL GEOLOGY OF THE QTJADRILATERO FERRIFERO, MNAS GERAIS, BRAZIL structure. In the southeastern part of the region, post- complex. There, the presence of higher Minas forma­ Minas deformation was so complex that the degree and tions indicates that the Caue Itabirite is absent because nature of the pre-Minas deformation cannot be of nondeposition not erosion. The source of the coarse deciphered. detrital material was probably to the west and south­ Thus, the structural evidence suggests that the focus west, suggesting stronger tectonic activity in that direc­ of the pre-Minas deformation was to the west or south­ tion during the interval between Minas and Itacolomi west. Also, pre-Minas granitic rock is known in the sedimentation. western and northwestern parts of the region and is POST-ITACOLOMI OROGENY suspected to the southwest. Stratigraphic evidence within the Minas Series indicates that the sediments The major orogeny of the region occurred after dep­ making up those rocks came from the west or south­ osition of the Itacolomi Series. However, the relatively west. Therefore, it seems reasonable to conclude that the systematic sequence of alined anticlines: and synclines focus of the pre-Minas orogeny was to the west and found in many younger orogens cannot be found in the southwest and that this orogeny resulted in the forma­ Quadrilatero Ferrifero. Although many major struc­ tion of a major orogen which influenced later geologic tural features strike northeast, several strike northwest events and processes. The area making up the Quad­ and two of the largest strike north and east, respectively. rilatero Ferrifero must have been on the east side of Most of the synclinal folds are overturned toward the this erogenic belt; the extent and trend of this belt are west and northwest at least in part but they are sepa­ still undefined. rated by broad clomal areas which acted as disruptive buttresses in the course of orogenesis. The structural PRE-ITACOLOMI FAULTING pattern inherited from earlier deformation and the Faulting which occurred either in pre-Minas or pre- post-folding formation of granitic rocl" that literally Itacolomi time has not been documented in the region, engulfed preexisting strutures and Stratigraphic se­ although several of the geologists of the DNPM-USGS quences complicate the picture still more. team have stated their belief that such faults must exist Complex fault patterns further obscure the analysis, and that some faults that were active in post-It acolmi particularly in the eastern and southeastern parts of time may have been originally formed in pre-Minas the Quadrilatero Ferrifero, where thrusting occurred on time. This seems reasonable, but is not proved. Simi­ a major scale. They, too, reflect rock transfer toward the larly, no post-Minas-pre-Itacolomi faulting has yet west and northwest. been proved, although it may have occurred. Elements The description of the individual IT ajor structures for such proof are lacking because the Itacolomi Series rests upon this complex history and upon remaining and is absent in most of the region. identifiable Stratigraphic units. For practical purposes, POST-MINAS-PRE-ITACOLOMI DEFORMATION these include of certain units in the Rio das Velhas Series and the Minas and Itacolomi Series. On plate 1, The Mimas Series was warped and uplifted in the the major structures involving the two upper series are interval between its sedimentation and Itacolomi sedi­ best outlined by the distribution of the Itabira Group. mentation ; within this region the Itacolomi Series rests Orogenic forces thus came from the east and south­ on almost every formation of the Piracicaba Group, in­ east. In the southeastern and eastern parts of the re­ dicating not less than 1,000 m of pre-Itacolomi erosion, gion, both folding and fracturing deformed the rocks; possibly much more in some localities. The angular un­ in the central and western parts of the region, deforma­ conformity is nowhere great; the maximum observed tion was clominantly by folding except along the south was 12°. In the Quadrilatero Ferrifero the interval thus boundary. Thrust faulting followed folding, and nor­ represents mild diastrophism and warping rather than mal faulting followed thrust faulting. Granite was orogeny and may perhaps be best regarded as the first mapped as being thrust faulted only in the Mariana stirrings of the great paroxysm that followed Itacolomi time. quadrangle; here, the thrust fault might be pregranite Warped cobbles of itabirite probably derived from in age, as the rocks involved are thought to be grani- Caue Itabirite are found in Itacolomi conglomerates. tized Nova Lima Group. The granitic rocks are cut by Unless removed from the anticlinal areas within the many normal faults. region, these cobbles must have come from some area Before the folding and faulting of this orogeny, outside the Quadrilatero Ferrifero because the Caue the rocks now compressed into the Quadrilatero Ferri­ Itabirite is everywhere present within this region where fero must have occupied hundreds of square kilometers expectable except in a small area south of the Bagao to the east of their present location, for the total crustal PHYSIOGRAPHC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT shortening amounts to tens of kilome;ers in an east-west south side of the Serra do Caraga. The north end was direction. greatly compressed against the Caraga quartzite mass MAJOR FOLDS and dismembered by thrust faulting. The cross folds The Quadrilatero Ferrifero is boinded on the east are cut by a system of northwest-trending faults and hy by the Santa Rita syncline, on the \\est by the Moeda high-angle thrust or reverse faults, trending north- syncline, and on the south by the Dom Bosco syncline. south. The first is grossly overturned and sliced by thrust The cross folding is best developed in the Santa Rita faults; the second is overturned in part; and the third Durao and Capanema quadrangles but is also present in is generally upright but telescoped by many thrust the Antonio Pereira and Mariana quadrangles to the faults. To the north, the region is limited for 100 km south there best shown, perhaps, in the convolutions of by the generally overturned north sides of the Rio das quartzites included in the Rio das Velhas Series and Velhas uplift and the Bonfim dome, forming the Serra strata of the Minas Series on the east flank of the syn­ do Curral. Farther to the east, it is limited to the north cline. North-trending high-angle reverse faults slice the by the Serra do Tamandua, the eastern extremity of the syncline into many parts, involving the youngest known Gandarela syncline, which forms the east and southeast beds of the Itacolomi Series as well as the Rio das Vel­ flank of the Rio das Velhas uplift. The Gandarela syn­ has sequence. Granites intruded the faults in many cline is isoclinal and overturned at the east end but at places and have formed from the pelitic rocks of the the west end, near the center of the Quadrilatero Ferri­ Rio das Velhas Series on the east flank; quartzites cor­ fero, it is open and upright. The eastern outliers in the related by A. L. M. Barbosa and Maxwell with the Mr- Itabira district and the Monlevade district are respec­ quine Group are relatively unaltered. tively composed of a complex syncline and a system of The major syncline is strongly overturned to the wert synclines and anticlines, both trending N". 45°-50° E. in the Antonio Pereira and in the Santa Rita Durao The Rio das Velhas uplift is the master structure quadrangles. The overturned east flank dips east from within the region, for most of the other major structures 25° to 45°, the wast flank also dips east at similar angles. are peripheral to it at least in part. It occupies the south- The structure is so disturbed by cross folds and thrurt central part of the Quadrilatero Ferrifero, trends north faults that it is impossible to locate its axial trace. To by west and then to the northeast and disappears in the the north, on the east flank of the Serra do Caraga, the sea of granitic rock to the northeast. The Caete uplift syncline is abruptly and radically compressed and is of Bolau (1952) is the northeast extension of this fea­ nearly vertical. There, its west limb is in part cut out ture. Another major anticlinal feature largely within by a steep reverse fault in the southern portion of the the Quadrilatero Ferrifero is the Conceic.ao-Carac.a up­ Catas Altas quadrangle. In the central and northeri lift, lying between the, Gandarela syncline cand the Santa parts of this quadrangle, the west limb is missing. Th^i Rita syncline. east limb continues north in an ever narrowing sliver Superimposed on all these major folds are many of quartzite and itabirite bounded on the west by post- smaller folds and on those still smaller ones ranging Minas greenstone and on the east by post-Minas granitic down to crenulations. Some are parallel to major struc­ rock in metasomatic contact with the Minas Series tures ; others are cross folds. (Maxwell, written commim., 1962). The cross folds deforming the Santa Rita syncline are SANTA RITA SYNCLINE larger than those found in other major synclines and on°- This syncline was described in detail by Maxwell of them, the Conta Historia syncline, develops to th°i (written commim., 1962), who discovered and mapped northwest into an important fold that underlies thQ, the northern two-thirds of the structure. Deciphering Serra do Batatal and the Serra do Ouro Fino (Maxwell, this structure was a notable feat of geologic mapping; a written commim., 1962). The writer concurs with th°. reasonable interpretation of the rock relations on the suggestion by Maxwell that this fold once was continu­ east flank of the Serra do Caraca had never before been ous with part of the complex west end of the Gandarela advanced despite the fact that it has been the site of syncline. Because both limbs of the Santa Rita syn­ mining activity for 150 years. cline are cross folded, those structures probably formed The Santa Rita syncline is perhaps the most complex after the major syncline had developed, possibly as a major fold in the region. The major north-south struc­ result of a change in direction of regional stresses, pos­ ture, the earliest to form, was itself later folded by sibly as a result of new couples set up by the earlier de­ northwest-trending open cross folds expressed on plate formation. The cross folds plunge to the southeast. It 1 by salients and reentrants in the contact between the is clear from the outcrop pattern that cross folding Cambotas Qiiartzite and the younger rocks along the preceded thrust faulting of the Minas and Itacolomi A86 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL Series. The northwest-trending faults may have devel­ thrust fault that brings the Nova Lima Group over oped during the cross folding and perhaps precede strata of the Itacolomi Series. The north limb is folded thrust faulting. The cross folds are the sites of impor­ sharply around the Mariana anticline to continue north­ tant bodies of supergene iron and manganese ore in the west as the west limb of the Santa Rita syncline. Itabira Group, as described by Buchi (1961). Maxwell The syncline may be divided into two main parts: (written commun., 1962), and Dorr (1964). an eastern part characterized by a series of arcuate The tight and dismembered north end of the Santa thrust and tear faults, and a western par4" in which tear Rita syncline contrasts with the wider and nearly recum­ faulting is more important. In the eastern segment suc­ bent main body of the fold south of the Serra do Caraga. cessive thrust faults have pushed Rio das Velhas rocks Evidently major westward and northwestward dis­ over Itacolomi and Minas Series rocks,. Minas Series placement of the rocks occurred in the central part of rocks over Itacolomi Series rocks, Minas Series over the syncline, and the massive monolithic body forming Minas Series rocks, and Itacolomi Series over Rio das the Serra do Carac,a formed a buttress against which Velhas Series rocks in a great succession of imbricate the eastward-trending forces wrapped the Minas Series thrust plates. Movement was from east to west. into a tight fold. Individual thrust plates preserve the synclinal struc­ The structural evolution of the central part of the ture, with the apparent fold axes in each individual syncline may have been defined either by the original plate somewhat offset, indicating that the thrust faults sedimentary limits of the Maquine and Cambotas are younger than the folds. The north limb of the syn­ Quartzites or by erosion which may have removed cline is not involved in the thrust faulting except at the much of these rocks south of the Serra do Carac,a before extreme east end. A high order of field mapping was the deposition of the Minas Series. The Cambotas does demonstrated by A. L. M. Barbosa in unraveling this not occur south of the central part of the Capanema complex structure. quadrangle except, questionably, in the Serra do Ouro In the central and western parts, the structure of the Branco- Dom Bosco syncline changes from dominant thrust The southern part of the Santa Rita syncline was faults in west-moving slices to gradually increasing held to the east by the Mariana anticline. The writer be­ complexity of internal folds; the axes of the minor folds lieves that pre-Minas granitic rocks that form an unex- trend parallel to the axis of the major fold. Thrust posed part of the Bagao complex underlie the fold at faults are prominent in the southern pprt of the syn­ no great depth and formed a keystone around which the cline. At its west end, the north limb of the syncline is younger Minas Series rocks were forced and compressed a simple fold, although its south limb is complexly during post-Itacolomi deformation. The contact be­ faulted (Johnson, 1962; Guild, 1957). tween the Minas Series and the Rio das Velhas Series In the eastern, central, and western parts of the Dom on the west limb of the Santa Rita syncline is a thrust Bosco structure the strata are generally upright. How­ fault in much of the axial zone of the Mariana anticline, ever, the western part of the fold is locally overturned with the Minas overriding to the northwest. In the east toward the south. limb, however, Minas Series and Itacolomi Series rocks The southern part of the Dom Bosco syncline in the were overridden by the Rio das Velhas Series. The de­ Dom Bosco, Sao Juliao, and Casa da Pedra quadrangles formation was similar to, but less intense than, that on is most complex. It is a westward continuation of the the east side of the Serra do Caraga, probably because zone of thrust faulting that sliced the east end of the the buttress is deeper in relation to the present surface. structure. The relatively regular and even thrust fault­ ing found to the east breaks up into a large number of DOM BOSCO SYNCLINE individual thrust plates, seemingly bounded to the north The Dom Bosco syncline is the second most compli­ and south by tear faults and by east-west high-angle cated major structure in the region. Its south side is faults with considerable vertical displacement. The area enigmatic for most of its extent because it is sliced and is one of greatly confused structure, which only becomes torn by thrust and tear faults of unknown but probably somewhat more simple west of Sao Juliao. The large considerable displacement. The axial plane cannot be block of quartzite which forms the £?rra do Ouro accurately located, and its trace as shown on plate 1 Branco may have caused finer fracturing of the thrust may well be much in error. plates to the north of it as they slid westward past this The syncline lies south of the Rio das Velhas uplift, monolith, which itself is strongly sheared in a general and extends eastward from its junction with the Moeda N. 70° W. direction. syncline in the Marinho da Serra quadrangle to the In the Casa da Pedra-Fabrica area a series of im­ Mariana quadrangle, where it is obliterated by a major bricate thrust plates is piled up against the Serra da PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A87

Moecla, overriding the east side of tie Moeda syncline MOEDA SYNCLINE and thus possibly originating in ai expectable anti­ The Moeda syncline borders the west edge of the clinal fold at the junction of the Dom Bosco and Moeda Quadrilatero Ferrifero and is one of the larger struc­ synclines, a fold possibly destroyed by the faulting. tures in the region (pi. 1). The gross structure is simple, Guild, who studied the area before adjacent areas had but many cross folds exist within the central area; this been mapped, was the first to demonstrate major is best seen on the more detailed maps accompanying thrusting in the Quadrilatero Ferrifero and made a the report by Wallace (1965). Those maps illustrate fundamental contribution to the structural geology of well the greater power of structural portrayal and the region. Three and possibly five thrust plates are here analysis made possible by detailed stratigraphic sub­ exposed (Guild, 1957, p. 35). Some are folded into division, for 10 units were mapped in the Minas Seriee wrinkles and subsidiary folds with orientations mainly instead of the three shown on plate 1. in the eastern quadrants and plunging to the east and The Moeda syncline can be traced for about -10 km southeast. A remnant of Itacolomi rocks is preserved south from the Serra do Curral to its junction with the, in one of these folds, and it is locally isoclinally folded Dom Bosco syncline. The west limb can be traced for and overturned. nearly 20 km more to the southwest corner of the The Dom Bosco syncline plunges gently east at the Quadrilatero Ferrifero and then continues a few west end; the subsidiary folds are canoe shaped, as kilometers southwesterly from there before losing its pointed out by Johnson (1962, p. B28). identity in post-Minas granitic and gneissic rocks. The Most of the south side of the east-west belt of meta- east limb merges into the north limb of the Dom Bosco sedimentary rocks folded into the Dom Bosco syncline syncline in a gentle curve which contrasts with the junc­ is bounded by the Engenho fault (Guild, 1957, p. 36), tion between the Dom Bosco and Santa Rita synclines to which is upthrown to the south. Johnson considered that the east. a branch of the fault was the locus of the large mafic Taking the base of the Minas Series as the datum, th°, dike in the Serra do Ouro Branco. Farther to the east Moeda syncline is about 16 km across at its widest point the fault separates granitic rocks to the south from the but narrows to about 5 km at its north and south end?. Itacolomi Series. The west flank, known as the Serra da Moeda, dip^ The structural evolution of the Dom Bosco syncline regularly to the east at 30°-45° and only locally dip« was possibly initiated by uplift of the Rio clas Velhas more steeply. The trend is even and regular, running block to the north and an unnamed large block to the about, north-south to the Casa da Pedra quadrangle, south. Many subsidiary small east-west folds formed where it flexes to the southwest. The limb is cut by few during that period of folding, as wrinkles on the major cross folds or cross faults. The east limb, known as the structure, but a major north-south shortening of the Serra do Itabirito, is much more complex, bowing out magnitude of the east-west compression shown in the from its junction with the Dom Bosco syncline and Santa Rita syncline did not occur. Later, major west- trending northeast to the Itabirito quadrangle, thence directed stresses produced the imbricate thrust plates warping again to the northwest across the Itabirito and along the south limb of the fold and also its major tear Macacos quadrangles, and gradually turning more to faults. the north at the north end. That limb is cut by many The distribution of the thrust faulting along the cross faults and warped by several cross folds. For much south limb of the syncline suggests that the south block of its extent, the east limb is either nearly vertical or was the active one and itself had a westward component, overturned toward the west, dipping as low as 45° E. as is also suggested by the westward bend at the south in a few areas. end of the Serra cla Moeda (pi. 1). Very probably the The axial trace of the syncline, in part located by presence of the large granitic mass of the Bagao com­ infolded Itacolomi Series strata, is generally north- plex channeled the stresses and movement into a west­ south but flexes far to the east in the Lagoa Grande erly direction instead of to the northwest as is common quadrangle. The north end is complicated by a thrust elsewhere in the region. Similarly, the imbrication of and tear fault and by a saddle and cross fold; the thrust plates against the south end of the Serra da nature of the junction with the Serra do Curral is not Moeda indicates that a buttress existed there, as sug­ clearly understood. gested by Guild, indicating the presence of a pre-Minas As demonstrated by Wallace and indicated by aero- western granitic mass. The last event was the movement magnetic surveys (W. B. Agocs, oral commun.), the on the Engenho tear fault. This had an upward com­ syncline is not a deep one. It seems to flatten abruptly ponent in the south block and possibly a western com­ and has a W-shaped cross section. Maximum depth to ponent as well. the base of the Minas Series may be in the order of A88 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL 2,000 m. Pomerene (1964, fig. 25, p. 39) illustrated well ernmost tip of the Serra do Itabirito is cut by a thrust the supposed cross section by a photograph of an actual and tear fault that is well revealed in surface and sub­ small fold found just to the north. surface exposures in the Mutuca II mine and environs. Mapping by Wallace showed that two systems of To the west, in the Serra das Gaivotas, a south-trending cross folds traverse the structure, one trending north­ septum of Caraga Group strata that is essentially anti­ west, the other northeast. The sharpest of these folds clinal in structure is in thrust contact on its east side is reflected in the nose of Itabira Group trending south­ with the Nova Lima Group. east from the Serra do Moeda in the Lagoa Grande Pomerene (1964, p. D39) believed that this western quadrangle (pi. 1). Most of these folds are not clearly thrust fault connects with that exposed in the northeast reflected in the Caraca and Itabira Groups but were corner of the Moeda syncline, assuming that it is the found by careful mapping of the individual formations major plane of movement. The writer suggests that the of the Piracicaba Group. In the center of the syncline, thrust fault 011 the east side of the Serra das Gaivotas the two systems intersect to form a domal structure may swing to the northeast rather than to the southeast (Wallace, 1965). (fig. 26) and that the well-exposed fault on the east side Extremely tight internal folding in the Caue Forma­ of the syncline is a separate structure. The writer would tion produces radical thickening of that formation in interpret this postulated fault as cutting across the soil- places. The more massive Moeda Formation is also thick­ aiid canga-covered area at the junction of the structures, ened by internal folding and by plastic flow, for its swinging gradually to the northeast, and finally running bedding is almost completely obliterated in some areas nearly parallel to the Caraca-Nova Limp, contact and such as near Morro do Chapeu, Macacos quadrangle, dying out several kilometers to the northeast in the older where strong schistosity and flow cleavage is normal to rocks. A zone of intense brecciation is revealed by the bedding. Just northwest of that locality, Pieter van large readouts on the north and south sides of the Via- Leuwen (oral commun., 1962), mapping an iron-ore de­ duto de Mutuca, about Macacos N. 13,OCO, E. 3,200), posit, believed he found a five-fold repetition of one hori­ where the Moeda Formation is thin and mylonitized. If zon in the iron-formation. Such structural complexities this interpretation is correct, the northern part of the would never be suspected on most maps and undoubted­ Moeda syncline was thrust northwestward for some un­ ly are much more common than now known, for internal known, but not necessarily great, distance against, and folding in evenly trending formations is hard to prove to some extent over, the overturned limb of the pre­ without distinctive horizon markers and good outcrops. existing homoclinal structure now forming the Serra The junction of the Moeda syncline with the Serra do do Curral. Curral is enigmatic; the critical areas for deciphering SERRA BO CURRAL it are covered by soil and canga. We have here a major syncline oriented north-south butting against a major The Serra do Curral is a throughgoiiig linear moun­ tain some 100 km long extending from the Serra da homoclinal structure striking about N. 60° E. for about Piedade quadrangle on the east to a short distance 100 km. The junction is in about the middle of the N. 60° E. structure. The puzzling aspect is the uninter­ west of the Itatiaiucu quadrangle. The range has been given many local names along its extent, and the name rupted and undeviated sequence of the Piracicaba and Serra do Curral is commonly applied only in the area Itabira Groups on the north side of the Serra do Curral just south of Belo Horizonte. Its geologic structure and also the apparently uninterrupted trend of the Ita­ is homogeneous, however, so this one name was applied bira Group across the junction and into the Moeda syn­ to the whole range by the USGS-DNPM staff in the cline. If the synclinal fold merely opened out on its interest of simplicity. flanks and merged into the transecting structure, a hiatus and deflection would be expected in the trend of The homogeneous nature of the structure is clear on the Itabira and Piracicaba Groups in the Serra do Cur­ plate 1. The same stratigraphic units dip steeply or are ral. Obviously a more complicated structure exists. overturned throughout the whole Serra do Curral. The About 5 km south of the junction, the Moeda syncline most complex part of the range from th°, structural is very narrow. This construction is interpreted as a viewpoint is in the Belo Horizonte and Ibirte quadran­ saddle fold, a type of fold found elsewhere in the region. gles (Pomerene, 1964, p. 37), where it is cut by a number The syncline opens again to the north but is cross folded of small thrust faults which swing into the plane of the 011 the east side, revealing an unconforinable contact bedding and disappear, by cross faults, and particular­ between the Caraca and Nova Lima Groups. The Caraca ly by subsidiary folds, the axes of which parrllel the axis Group is much sheared in this area and thins radically of the range in a manner similar to that in the Dom along the south flank of the Serra do Curral. The north­ Bosco syncline to the south. PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT

EXPLANATION

C, conglomerate in me Nova Lima Group

6 KILOMETERS

FIGURE 26. Alternative explanation of structure at junction of the Moeda syncline with the Serra do Curral, to be compared with plate 1. See plate 1 for structural details, topography, and drainage.

Such subsidiary folds are also present, farther west in north limb of the syncline is cut out by a cross fr»ult the Fecho do Funil quadrangle, as mapped by Sim- which puts granite against Cercadinho Formation ro".ks. mons (1968a). Mine exploration carried on after Pome- If this contact is followed to the west about 1 km, an­ rene completed mapping showed that in the Ibirite other cross fault can be found, west of which Alves quadrangle a subsidiary fold, not visible on the outcrop, mapped a sharp and small eastward-plunging tight occurs in the Samambaia iron-ore deposit. Other folds anticline exposing Itabira Group and Caraga Group probably occur in the Belo Horizonte quadrangle, and rocks at its core and Piracicaba Group rocks on both the writer believes, but cannot prove, that the sudden the north and the south flanks. These rocks are in con­ thickening of the Caue Itabirite in the great Aguas tact to the north and probably to the west with intrusive Claras iron-ore deposit is caused by such a fold. Sharp granite; a small thrust fault may be present to the flexures or cross folds occur in a few places, such as that in the southeastern part of the Santa Luzia quad­ west. Isolated in the granite to the north is a bloch of rangle and on the west boundary of the Fecho do Funil metasedimentary rock some 3 km long and more than quadrangle; such cross folding is not common. 1 km wide consisting of a thin remnant of the Nova The east end of the Serra do Curral, locally called Lima Group, overlain by the Caraga, Itabira, and part the Serra da Piedade, swings abruptly to the north and of the Piracicaba Groups of the Minas Series. This block continues a kilometer or more back in a westerly direc­ and the rocks in it strike about N. 15° E., and the Hds tion, reflecting the structure of the rock, here folded dip -iO°-600 NW. and make a sharp angle with the gen­ into a sharp syncline named by Alves, who discovered eral trend of the Serra do Curral and the adjacent m°ta- it, the Piedade syncline (Alves, 1961). (See pi. 1.) The sedimentary rocks to the south. The nearest rocks of the A90 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MTNTAS GERAIS, BRAZIL north limb of the Piedade syncline strike about east the region (Guild, 1957; Johnson, 1962; Reeves, 1966; and dip south. Barbosa, pis. 9 and 10 of this report). Cercadinho and Interpretation of this enigmatic structure at the east Caue iron-rich rocks are typically continuous along end of the Serra da Piedade, dismembered by faulting strike. Regardless, if the subsidiary longitudinal anti­ and largely engulfed in younger granite, is tenuous. W- cline within the Piedade syncline were ar persistent as shaped synclines (larger synclines with smaller longi­ some in the region, the north limb of the Piedade syn­ tudinal anticlines in the middle) are not uncommon cline would have been far to the north before granitic features in the region; numerous examples are found in emplacement, and any repetition would be caused by the the Itabira district, in the Moeda syncline, and else­ subsidiary anticline. The problem is an excellent ex­ where. The small anticline just north of the Serra da ample of Dor's second law of scientific investigation: Piedade might be the beginning of such a subsidiary "The less the evidence, the greater the number of opin­ longitudinal anticline within the Piedade syncline. ions and the more firmly held they are."

Were this so, the anomalous northeast-trending block GANDARELA SYNCLINE now surrounded by granite could be explained as being the remains of the south limb of the subsidiary anticline, The major structure lying to the southeast and east folded back by a cross fold. A suggestion of such a cross of the northern part of the Rio das Velhas uplift is fold is to be seen roughly on strike to the southwest in known as the Gandarela syncline. It is in part open and the Serra do Curral (pi. 1). The north limbs of that upright and in part isoclinal and overturned toward the anticline and of the Piedade syncline would then be northwest. He who walks the 5-kilometer ridge at the somewhere to the northeast, completely destroyed by west end of the syncline,7 observing the parallelism of the invading granite. From this area to the westernmost the structure beneath his feet to that in the valley of the extension of the Serra do Curral, no sign of the north Rio das Velhas below to the west, is immediately aware limb of the Piedade syncline can be found, although that the blunt end of the Gandarela syncline belongs to such a limb must have once existed. To the west, post- the Rio das Velhas uplift rather than to the tight iso­ Minas granite is in intrusive contact with the Minas clinal syncline stretching to the northeast. To the west Series along the entire north flank of the Serra do Cur­ lies one structural style, to the east, another. ral. causing a notable contact aureole. The Gandarela syncline splits at its Tvest end, as Conceivably, the north limb of the Piedade syn­ shown by O'Rourke's careful mapping (pi. 5), one axis cline veered sharply to the north, paralleling the struc­ veering to the west and north of west, from the generally ture in the Serra das Cambotas, for the prevailing southwest trend of the main fold, the other axis veering structural trend north of the mapped area is north for almost south, with the blunt end between them belong­ hundreds of kilometers. ing to the Rio das Velhas structure. The two limbs are Bearing on this problem is the great thickness of the roughly parallel for about 10 km to the northeast of the Sabara Formation at its type locality along the Kio das west end, but beyond the main axis swings to the east- Velhas, some 10 km to the southwest. Gair (1962) northeast, and the limbs converge as the fold becomes mapped this 3,500-meter-thick section of the formation tighter, isoclinal, and overturned. The south limb is as deformed but not repeated by folding back upon it­ largely cut out, by the great, Fundao fault in the vicinity self. Teixeira da Costa (1961) suggested that some of the of and east of Barao de Cocais, and the Piracicaba area mapped as Sabara Formation along the Rio das Group is largely squeezed out of the center of the fold. Velhas belong to the north limb of the Piedade syncline, In the Gongo Soco and Conceicao do Rio / cima quad­ because of the presence of quartzite and "iron-forma­ rangles, both limbs dip 40°-60° S., and at the east end tion." Simmons (1968a, p. G29) agreed with this view of the fold both limbs dip 45°-60° S. The fold disap­ and cited the presence of lenticular quartzite bodies in the granite as evidence. He, too, suggested (oral com- pears on the east bank of the Santa Barbara river in the mun., 1964) that a highly metamorphosed lenticular Sao Gon^alo quadrangle, cut off by the Fundao fault. "iron-formation" in the rocks mapped as Sabara might 7 Surely the view from this ridge is one of the mos^: impressive in be better correlated with the Cercadinho ferruginous the whole region. To the east, sliced by deep rugged canyons, soars the quartzite or the Caue Itabirite. The writer, agreeing high massif of the Serra do Caraga, with a gothic skyl'ne and water­ falls several hundred meters in height making white ribb>ns down sheer with Gair, thinks Teixeira da Costa's and Simmons' evi­ quartzite scarps; to the west lies the great stepped chf sm of the Rio dence is not critical, because the quartzite zones (meta- das Velhas valley, bounded to the west by the scarps ani ridges of the Serra do Itabirito. Forty-five kilometers beyond that Serra protrude the chert ?) are not stratigraphically identifiable and lentic­ castellated crags of the Serra do Itatiaiugu, and 100 km to the north lies ular iron-formations and ferruginous quartzite in the the broad high arch of the Serra do Cipo. The area has no permanent in­ habitants except rabbits, foxes, anteaters, armadillos, and an occasional Sabara Formation have been mapped in many parts of wolf and jaguar. Silence is tangible. PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A91

In the Gandarela syncline, axes of subsidiary folds corner of the structure the Moeda is piled in imbricate are parallel to the main synclinal axis, particularly at thrust slices branching from the Fundao fault, vdth the east end. The syncline is cross folded at two places spectacular scarps as much as 150 m high, but tJ-°se north of Barao de Cocais, attributed by Simmons slices merge and die out to the south, where the forma­ (1968b) to doming by gabbroic rocks exposed in a tion is locally faulted out. stocklike body to the north. Because the cross folds are As illustrated on plate 1, the Fundao fault charges related in direction to the other cross folds and because direction about 90° just north of the syncline, changing elsewhere the gabbro is controlled by postfolding faults, from a northeast-trending thrust fault to a southeT,st- it may well be that the folds controlled the emplacement trending tear fault parallel to and limiting the Rio das of the posttectonic gabbro rather than the converse. Velhas uplift and cutting off the Conceic,ao anticl;ne. The Gandarela syncline plunges northeast at its west As also shown on plate 1 and in more detail on plates end and southwest at its east end at low angles, locally 5 and 6, a long projection of the Moeda Formation reversing. The width of the outcrop of the Piracicaba extends southward from the Gandarela syncline, and is Group is a very rough measure of the tightness of truncated by the Fundao fault; the dip of these rocks folding and the depth of the synclinal keel. The width reverses from 20° to 30° N. near the contact with the of outcrop of the Itabira Group may also be a rough Itabira Group to 10°-25° SE. near the fault. Thus, measure of the degree of compression; at the west end the fault cut off a saddle fold or transverse anticl'ne, the Caue Itabirite is grossly thickened by internal overriding the Moeda Quartzite with Nova Lima Gnup crenulations and folds, indicating flowage of material rocks, an extension of the Conceic,ao anticline, from the into the axial zone. To the northeast (except at the ex­ southeast. treme east end) such crenulation and internal folding In the central part of the Ouro Fino syncline in the are less pronounced or absent in large areas; probably, zone where the strike of the axial plane abruptly swings material has been squeezed out of the limbs. In the about 90° from northeast to southeast, Maxwell (writ­ southern part of the west, blunt end, many small folds ten commun., 1962) noted east-trending minor folds transect the trend of the formation, indicating compres­ and lineation. These structures, in conjunction vith sion from the south and southeast. O'Rourke found a the outcrop pattern and the relationship to the mrjor shear fold in the Gandarela Formation just west of the Fundao fault, indicate that the whole syncline orig­ Fundao fault (pi. '5), and this fold indicates the same inally was oriented northwest parallel to the Rio das type of deformation in the axial area. Velhas uplift but that the north end was refolded to

OTTRO FINO SYNCLINE its present northeast trend by movement of the over­ riding sheet of the Fundao fault, as first pointed out The Ouro Fino syncline is the small refolded syncline by O'Rourke (written commun., 1956). The syncline lying immediately to the south of the west end of the was originally separated from the Gandarela syncline Gandarela syncline and separated from it by the Fundao by a broad saddle fold, now truncated by the fault. The fault and the west end of the Conceigao anticline. At its Ouro Fino syncline is a rootless structure at its no^th- south end, this structure is parallel to the Rio das Velhas east and north ends, as inferred by O'Rourke. uplift and trends northwest, but the north end is paral­ The excellent mapping of the Gandarela and Rio de lel to the Conceigao anticline and the Gandarela syn­ Pedras quadrangles by O'Rourke and the Capanema cline and trends northeast. The flexure is abrupt and quadrangle by Maxwell and the structural analyser by more than 90°. The fold is closely related to the Gan­ these men are basic to the understanding of this compli­ darela syncline to the north and to the Conta Historia cated structural development. syncline, a cross fold on the Santa Rita syncline, to the south. The refolding of the structure was controlled by STRUCTURES IN THE ITABIRA AND MONLEVADE DISTRICTS the Fundao fault, which limits it to the west and north­ The Itabira and Monlevade districts have certain west (pi. 6). structural features in common and some different f^om The Caraga Group and the Caue Itabirite are the only each other and from those in the rest of the Quadrilatero units of the Minas Series still preserved in the Ouro Ferrifero. Both are separated from the main body of Fino syncline. The Moeda Formation, which rims the the region by some 15 km of paragneiss and orthogr. eiss syncline except where faulted out, overlies the Nova and contain downf olded remnants of Minas Series rocks Lima Group with sedimentary contact north and east that have been eroded off or destroyed in the interven­ of the fold but is in fault relation to those strata on the ing space. As indicated by the prevailing higher grade northwest and south sides. The Moeda is faulted against of metamorphism, the rocks were subjected to higher Maquine Group rocks to the west. In the northwest temperatures, probably because more deeply folded into 334-413 O 69 7 A92 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL the substratum. Both areas show moderate-sized fold RIO DAS VELHAS UPLIFT structures in contrast to the large ones preserved to the The master structure in the Quadrilatero Ferrifero west. Thrust faults are minor and rare. However, the is the Rio das Velhas uplift (Gair, 1962, p. A8). All largest normal faults known in the region are to be major synclinal structures except those in the Itabira found in the Monlevade district. and Monlevade districts are at least to some extent The Itabira district consists of one northeast-trend­ conjugate to this great structure, which is traceable for ing syncline of Rio das Velhas and Minas Series strata 100 km north from the south edge of the Bagao com­ surrounded by metasomatic and igneous granite. As plex to near Nova Lima and thence nortlsast, where defined by the base of the Minas Series, the syncline is it disappears into the granitic rocks of the Caete com­ about 10 km long, is isoclinal for much of its length, and plex. This central uplift is the key to tli3 structural is locally overturned toward the northwest, locally development of most of the region. nearly vertical. The syncline ends to the northeast The Rio das Velhas uplift is about 19 km wide at the against a fault which may be pregranitization, for narrowest point and widens to about 47 km at the south vestiges of iron-formation are found in the granitic end, taking the base of the Minas Series as datum, and is gneiss east of the fault. even wider at the north end (fig. 4). At tl °- northeast The southeast limb of the syncline, which is poorly end and the south end, large masses of granitic rock exposed, is sinuous but not complexly folded; and the have invaded the general anticlinal structure; at the Caue Itabirite, which is very strongly lineated, is much south end the masses are of two ages, and possibly they thinner on the southeast limb than on the northeast are of two ages at the north end. In the main body of the limb. The northeast limb has been folded into two very uplift, diabase and metadiabase dikes are krown but are marked subsidiary synclines and their conjugate anti­ not abundant; significant postmetamorphic gabbro clines, those in the central part of the district tight and masses are not known. overturned and the most easterly syncline open and Because of the base of the Minas Series is arbitrarily upright. In these subsidiary folds and in the axial zone taken as the edge of the uplift, the metasedimentary of the main fold, the Caue Itabirite is much thickened. rocks of the uplift are Rio das Velhas Ser^s. The Rio The larger synclines are marked by subsidiary longi­ das Velhas Series shows at least two ages of folding: tudinal anticlines within them (Dorr and Barbosa, the pre-Minas folding discussed on pages A83-A84, and 1963, pi. 8) similar to those in the Moeda and Piedade folding which this writer believes is related to the main synclines to the east. Although the major syncline post-Itacolomi orogeny. As brought out in some detail developed as a result of compressive forces from the in the section on stratigraphy, the Rio das Velhas Series southeast, the marked development of cross folds in the for the most part lacks mappable stratigraphic zones northwest limb and the marked lineation parallel to and thus intricacies of the structure can be deciphered their axes suggests that a left-lateral shear couple in few places. Included distinctive mappabb lithologies operated during the folding. are lenticular; the only mappable throughgoing contact The Monlevade area is unique in the region for its is the base of the Maquine Group. Most primary sedi­ series of well-defined medium-scale conjugate synclines mentary features were destroyed by metamorphism; and anticlines; all plunge to the northeast or east. In hence, tops and bottoms of stratigraphic sections cannot this area the metamorphic grade is so high that most be established. of the Piracicaba Group pelitic rocks have been trans­ The foliation in the Nova Lima Group is roughly formed into gneiss, and much of the Caue Itabirite is parallel to the contact with granitic rock around the highly magnetitic. Some of the synclines, such as that Bagao complex except at the north end, wh^.re it strikes culminating in Morro Agudo, are closely appressed and across the contact. A band of iron-formation, which isoclinal in the keel zone; others flare out as relatively crosses the old road from Itabirito to Pico de Itabirito in the northern part of the Bac.ao quadrrngle, shows gentle folds. Intensity of shearing and tightness of fold­ that foliation and bedding are parallel ther^. Quartzite ing increases to the south-west, and structures swing to beds found by A. L. M. Barbosa in the Sao Bartolomeu an east-west orientation in the same direction. The folds quadrangle, on the other side of the Bac.ao complex, are broken by two major cross faults, the largest known indicate the same relation. Probably, the southern part in the region. Thrust faulting is minor except in the of the uplift was bulged upward by the Bagao complex, southwest corner, where one limb of the Agua Limpa and the bedding is largely concordant witl the contact syncline is overridden by gneissic rock. in this part of the uplift. PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT ^93

In contrast, the foliation of the Nova Lima Group described by Maxwell (written commun., 1962) as is strongly discordant with much of the contact between cataclasites. In the Capanema quadrangle, shearing is the Caete complex and the Nova Lima Group south of parallel to the many strike faults in the Minas Series the town of Caete, but from there northward the folia­ in the Conta Historia syncline and, in part, in the Ouro tion is closely or completely concordant with the granite Fino syncline to the east. These strike faults are also contact. West of Caete an unmapped quartzite body in parallel to the fault between the Vargem do Lima syn­ the Nova Lima Group is parallel to the contact and cline and the Conta Historia syncline part of the sys­ concordant with the foliation; there it trends northeast. tem called the Fundao fault to the north. Here, the The structure on the west side of the Bio das Velhas fault is a tear rather than a thrust fault, and its e^st uplift north of Itabirito is complex indeed as the result side has moved northwest. One branch of this fault of strong pre-Minas and post-Itacolomi folding. In the system was mapped by Maxwell entirely within the vicinity of Nova Lima, Honore Bicalho, and Raposas, Maquine Group and joins the Fundao fault at the abundant bands of iron-formation and other mappable most westerly part of the Serra do Ouro Fino. rocks trend northwest and are complexly crumpled; the The north end of the Vargem do Lima syncline warps crumpling is thought to be related to the post-Itacolomi sharply to the northeast, and the axial trace trends orogeny, as it can be best explained by movement against parallel to that of the Rio das Velhas uplift and to a buttress to the northwest. Lineation and axial planes iron-formation beds in the Nova Lima Group mapped of the folds in the Nova Lima Group are generally by Alves and O'Rourke. One may presume that the parallel to the same structures in the nearby Minas fold continues on to the northeast in the Nova Lima Series. These vermiform structures give some hint of the Group rocks, to disappear in the Caete granitic complex complexities that may be hidden in areas of uniform near Caete. Foliation in the Nova Lima Group is parallel lithology. to this trend, but the lack of throughgoing mappable The Mariana anticline, the southeasternmost exten­ horizons precludes tracing of the structure. sion of the Rio das Velhas uplift, is in part a post- The Vargen do Lima syncline is not deflected by the Itacolomi structure. However, the writer suspects that folding which produced east-west structures on the west pre-Minas granitic rocks of the Bagao complex may ex­ limb of the Rio das Velhas uplift, folding that iras tend eastward from the outcrop of that complex, hidden clearly pre-Minas. It is thought that the present form under the broad area of Nova Lima Group forming the of the Vargem do Lima structure was largely produced axis of the Mariana anticline. Such a mass, if continuous by the post-Itacolomi folding, probably contemporane­ with the Bac,ao complex, would have formed a rigid ous with the northwest-trending fold of the Conta east-west buttress. Such an east-trending granitic mass Historia and Ouro Fino synclines. A pre-Minas fold, would explain the sharp bend in the Dom Bosco-Santa much more open, existed here, because the Minas ro?ks Rita synclinal system and the channeling of the major just to the east overlie Nova Lima Group rocks, not westerly trending thrust faults in the Dom Bosco syn- Maquine rocks. cline. Unless such a buried pre-Minas granitic mass exists, the above structures are most difficult to resolve. CONCEigAO-CARAgA UPLIFT The Conceigao-Caraga uplift lies in the eastern part, of VAKGEM DO LIMA SYNCLINE the Quadrilatero Ferrifero (pi. 1). Except at its extreme On the east flank of the Rio das Velhas uplift lies northeast end, where it runs into the Santa Barbara the isoclinal overturned east-dipping Vargem do Lima Gneiss, no granitic rocks are known in this uplift, end syncline (Gair, 1962, p. A-49). In this structure, the in this it is unique in the region. It is bounded to the Maquine Group crops out for about 53 km. This syn- north by the Gandarela syncline and locally by the cline generally trends northwest and swings sharply to Fundao thrust fault, 011 the east and south by the Santa the northeast at its north end (pi. 4). The north end Rita syncline and its subsidiary folds, and on the west plunges south, flattening in the Nova Lima quadrangle. by the Conta Historia syncline. The Nova Lima Group The plunge of the syncline at its south end is not deter- crops out on its north flank and the Maquine r.nd minable, owing to great shearing in that area and the Tamandua Groups in the central and southern parts of presence of the overlying fault plate of Minas Series the structure. Deformation within this uplift is v?ry rocks. complex. Throughout its whole extent, the Vargem do Lima The north side of the uplift is occupied by the Nova syncline is strongly sheared parallel to its axial plane. Lima Group; these rocks trend northeast subparallel Such shearing seems to diminish in intensity to the to the Gandarela syncline and the structure was named northwest; at the southeast end, the quartzose rocks were the Conceigao anticline by Moore (1969). This struc- A94 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MESTAS GERAIS, BRAZIL ture can be traced to the northeast to the Santa Barbara interpretation was followed on plate 1. Kelations be­ Gneiss and the Peti phase of the Borrachudos Granite, tween the Nova Lima, Maquine, and Tamandua Groups both of which have discordant contacts with the Nova in the central part of the Capanema quadrangle suggest Lima Group. Long lenses of iron-formation in the an anticlinal rather than a synclinal structure; conceiv­ Nova Lima, mapped by Moore, Simmons, and Maxwell, ably the axis of the Conceigao anticline swings south to showed that these rocks probably are folded into tight this area rather than continuing to the southwest as isoclinal folds. Because all top and bottom criteria were shown on plate 1. It should be made clear that this area destroyed by metamorphism and tight folding, it is not is the least accessible and among the most rugged in possible to ascertain which of the two folds indicated on the whole region and that exposures are poor. There­ plate 1 is the anticline, but it seems more logical to sup­ fore, it is not surprising that the complex structures re­ pose that it is the more northerly, which would make sulted in some disagreement. the Nova Lima Group subparallel to the Minas Series In the main mass of the Serra do Carag-r,, the Cam- in the Gandarela syncline. Simmons indicated that a betas Quartzite was folded into tight isoclinal struc­ low-angle unconformity between the Minas and Rio das tures, some recumbent, and was sliced intajnany imbri­ Velhas Series rocks is present in the area mapped by cate thrust plates, probably individually small. It is him. If the south fold were the anticlinal fold, tight also cut by many steep faults occupied by mafic dikes. pre-Minas folding and a sharp unconformity between The area is one of great tectonic disturbance. As is evi­ the Minas and Eio das Velhas Series would be indi­ dent from the fault pattern, rock failure was primarily cated. As shown above, the Minas-Rio das Velhas un­ toward the west and, in the southwestern part of the conformity seems to be characteristically low angle in structure, toward the northwest (pi. 1). ThQ, northwest- the eastern part of the region, and therefore it is believed trending cross folds on the Santa Rita syncline are that the northern fold is the anticline. responsible for the sinuous south border of the The lack of mappable zones in the Nova Lima Group Conceigao-Caraga uplift. The west half of the south west of the end of the iron-formation exposure makes it border is a magnificent dip slope on bare quartzite, some impossible to decipher the detailed structure of the Con- 600-800 m in height, crenulated by minc^ folds and ceicao anticline in that area. Possibly the isoclinal fold­ faults. ing in the Nova Lima continues southwestward, but this The east border of the Conceigao-Caraga uplift, one of cannot be proved, and the writer thinks that it may well the more spectacular sights in the Quadrilatero Fer- die out to the southwest, as it did in the Gandarela syn­ rifero, is a wall of bare quartzite, sloping more than cline, the conjugate fold just to the north. 50° and more than 1,000 m high, cut by crevasses and The northeast end of the Conceigao anticline is in con­ punctuated by castellated crags (fig. 15). Bedding of the tact with posttectonic Borrachudos Granite and foliated strata is obscured by the intense folding, faulting, and granitic rocks of the Santa Barbara Gneiss. The contact foliation of the quartzite but trends nearly north and with the latter rock is marked by a long foliated green­ dips steeply east at the north end and also probably stone schist intrusive of post-Minas age. (See page A79.) farther south, where the mountain is to-> rugged to The southwest end of the Conceigao anticline butts with climb. On the east side, the overlying formations are fault contact against the Vargem do Lima syncline in tightly wrapped around the quartzite anc1 are faulted the Rio das Velhas uplift. with strike faults and tightly folded. In the Conceigao do Rio Acima quadrangle, Moore The extensive dike system in the Serra do Caraga and (1969) named a fold that lies just south of the Conceigao the many late tensional faults may well indicate that a anticline the Capivara syncline. Maquine Group rocks large gabbroic body similar in size or larger than that were thought to have been isoclinally folded upon unroofed north of the Serra do Tamandua may under­ themselves in this syncline. Because the Cambotas lie the Serra do Caraga, accounting in part for its Quartzite of the Tamandua Group lies immediately to uplift. the southeast of the Maquine Group, in space that OTHER UPLIFTS should be occupied by older rather than younger rocks Because most uplifts around the periphery of the if a major syncline exists, this writer believes that this Quadrilatero Ferrifero consist largely of granitic rocks syncline must be a local wrinkle rather than a signifi­ that crop out poorly, lie in great part outside the cant feature. A major thrust fault was postulated by mapped area, and are of slight economic interest, they Moore to explain the anomalous relations but this fault were not mapped in detail. Similarities with gneiss was not mapped in adjacent quadrangles, where the con­ domes in Rhodesia and other parts of the v^orld suggest tact between the Maquine and Tamandua Group rocks that these gneissic domes grew outward in successive was deemed by Maxwell to be sedimentary. Maxwell's periods of accretion. The areas merit further study. PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A95 The uplift west and southwest of the Serra da Moeda the Serra do Carac,a may have a cumulative displace­ may have been similar to the Eio das Velhas uplift. ment of not less than several kilometers. Part of the Two sides of the uplift are delineated within the westward movement of the Itacolomi and Minas Series mapped area, but the other two sides were not found was caused by the nearly recumbent overturned folding in reconnaissance and may have been destroyed by in that direction, but part was due to transport we^t- granitization or intrusion, as on the north side of the ward in thrust slices. Serra do Curral. Or, as seems more likely from the The great Fundao fault system is probably the long­ stratigraphic evidence, the area may have been one of est single fault in the region. It has been traced south­ major erosion during Itacolomi time, and the Miiias west from it-he Floralia quadrangle, where it disappears Series and perhaps the underlying metasedimentary in granitic rocks, to the Eio de Pedras quadrangle, a rocks may have been removed then. Alternatively, the distance of 48 km. There it veers southeast as a system keels of the bounding synclinal folds may have been of tear faults and thrust faults for another 30 km to eroded away in post-Itacolomi erosion. the axis of the Mariana anticline 8 and thence, as thrust faults, another 10 km to the west, a total extent of 88 km. FAULTS It very probably joins the major thrust-fault system Three main types of faults developed in the region in the Ouro Preto and Mariana quadrangles, which during post-Itacolomi deformation: (1) thrust faults, swings south and again to the east, bounds to the north some intimately related to (2) tear faults parallel to the the largest block of Itacolomi Series in those quadran­ strike of the rocks, and (3) cross faults. Of these, the gles, and continues out of the region to the eastward as first two seem closely related to the earlier compressive a tear fault. Thus, the system may be more than 110 km stage of orogeny, and the last, to the end stages. long. The Fundao fault system marks the boundary be­ THRUST FAULTS tween the highly faulted area which lies to its east and The south and east sides of the Quadrilatero Ferrif ero the relatively uiif aulted area that lies to its west. It is are intricately cut by thrust and tear faults, in several thus both a major structural boundary and a major areas by imbricate sets. Such faults are superposed on structure in its own right. the earlier folds and thus indicate their later age. Other The displacement on the Fundao fault itself cannot thrust faults and fault systems of similar age are found be proved to be great. Simmons (1968b) stated that a elsewhere in the region but are not nearly as complex minimum displacement of 500 m would explain all the or as extensive. observed features in the Santa Barbara and Cocais Displacement along thrust faults has nowhere been quadrangles. In the Conceigao do Eio Acima quadran­ measured and, because of the lack of known matching gle, Moore calculated the minimum strike-slip displace­ elements, cannot be calculated without much more ment along this fault, which there splits into three detailed field mapping. Facies changes in stratigraphic slices, as 2.4 km. If, as the writer believes, the northeast­ units are gradual or erratic and, without fossils, cannot ward-trending part of the Ouro Fino syncline is the be used with any confidence. At the junction of the result of a 90° bend of this part of the syncline by Serra do Curral and the Moeda syncline, an abrupt movement of the upper plate of the Fundao fault facies change in the Caue Itabirite was noted by Pom- toward the west-northwest, the displacement must be in erene (1964) and this feature has been seen elsewhere. the order of 5 km. Too little is known of the nature and rapidity of such No horizon markers are known to measure displace­ changes to use them as a measure of the distance thrust ment along the part of the Fundao fault system between plates may have moved. the Serra do Ouro Fino and the axis of the Marirna Cumulative westward movement along the south edge anticline. The fault is a tear parallel to the strike of of the region perhaps is on the order of tens of kilo­ the wallrocks in the Eio de Pedras and Capanema quad­ meters. If the most westerly thrust plate in the Jeceaba rangles. The Maquiiie Group is faulted against the Ita- and Casa da Pedra quadrangles was torn from the bira Group for much of this distance. The Maquine south limb of the Dom Bosco syncline near the west end rocks to the southwest of the fault trace are very of that, feature., movement there may have been on the strongly sheared parallel to the fault zone, and much order of 7 km. The west- and northwest-trending subsid­ displacement is represented in the shearing. Maxwell iary folds in that thrust plate would be expectable (written conimun., 1962) remarked on the extreme elon- where the Dom Bosco syncline swung to the southwest 8 Although plates 7 and 8 by A. L. M. Barbosa do not show this f'nilt to join the missing limb of the Moeda syncline. as continuous across the Sao Bartolomeu and AntSnio Pereira quad­ rangles, Barbosa said (oral commun., 1964) that he thinks it may well Similarly, the north-striking thrust faults south of be present where not mapped.

334-431 O 69 8 A96 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL gation of many cobbles and pebbles in the Maquine beds, 2,000 m, but it could be more than 4,000 m. The north­ with axial ratios of as much as 20:1. This northwest- ern extension is called the Agua Quente fault and is trending shearing diminishes to the northwest. On the traceable for a distance of 50 km (Maxwell, written northeast side of the tear fault, close northwest-trend­ commun., 1964). At its north end the Agua Quente fault ing faults in the end of the Conta Historia syncline dis­ disappears in the Nova Lima Group argillaceous rocks; place the blocks on the southwest side of those faults to at the south it veers abruptly to the east in the Mariana the northwest (Maxwell, written commun., 1964). quadrangle as a tear or strike fault, placing in juxta­ Farther southeast, in the Sao Bartolomeu quadran­ position the Nova Lima Group with the Piracicaba gle, the Fundao fault is a thrust and continues thus to Group and cutting out at least the middle and lower the axis of the Mariana anticline and around the nose, Minas Series. Its east end lies beyond the napped area. where outliers (klippen) of the basal Minas Series on The Agua Quente fault has many branches and splits, the Nova Lima Group demonstrate the low dip of the some of which reunite with the main fault; others wan­ fault plane (fig. 27). der off to disappear in the granitic rocks or to turn into According to the present interpretation, the displace­ strike faults in metasedimentary rocks. Near the pretty ment on the Fundao fault is at least 5 km and is to the hamlet of Agua Quente the f ault is subp&.rallel to the northwest or west-northwest. The fault is a thrust where tightly squeezed Santa Rita syncline; it cuts out the trending across this direction and largely a tear fault west limb of this fold from that place on to the north where trending parallel to it. (Maxwell, written commun., 1962). The ^rarrn spring In the Dom Bosco syncline the thrust fault system is at Agua Quente (about 30°C) may represent the sur­ imbricate, the sheets piled on one another toward the facing of deeply circulating water along this fault. The west and bounded to the north and south by tear faults. displacement cannot be calculated, owing to the com­ No means of deciphering the individual displacement of plex structure, but the fault juxtaposes pre-Minas rocks each fault or the total displacement is known; it is with Minas Series rocks for much of its length and has probably large, perhaps more than 20 km. It was sug­ a minimum displacement of several hundred meters in gested that the displacement at the west end might be the central and northern parts of its extent. If the east in the order of 7 km, but most of the total shortening side of the fault is up, as it is farther to the south and took place in the imbricate system to the east. as is the regional rule, the missing west limb of the Santa The Dom Bosco fault system starts at the east end Rita syncline must be below the present surface, pos­ of the syncline with a fault of great displacement thrust­ sibly thrust beneath the Serra do Caraga or lost in the ing the Nova Lima Group over the Itacolomi Series. granitic rocks at depth. The vertical component of the fault must be at least South of the Serra do Caraga a series of other north- trending high-angle thrust or reverse faults lies to the west of, and parallel to, the Agua Quente fault. These faults constitute a system of interbranching faults which all bend to the southeast southward and to the northeast northward. Those bending to the northeast join other faults until only two master faults are left, both of which bend sharply to the northeast and dis­ appear, one as a strike fault, the other as a cross fault that disappears into the Cambotas Quartzite, offsetting the contact only about 75 m. The influence of the Caraga monolith and the Mariana anticline on the trends of these faults seems clear. Most of these faults are paral­ lel to the strike of the beds, and movement on individual faults is difficult to define, owing to irregular pre-Ita- colomi erosion in this area. Minimum displacement can­ not be less than several hundred meters on most of the FIGURE 27. Thrust fault, Caue Itabirite over Moeda Formation, faults and may be much more. The north-south faults near N. 12,950, E. 6,200, Mariana quadrangle. The Caue Ita­ dip east at moderate to high angles. birite here contains disseminated gold and was being removed Within the Serra do Caraga another intricate thrust- by hydraulicking 1964, when the picture was taken. The happy fault system roughly parallels the system just discussed geologist is A, L. M. Barbosa, who predicted the position of the fault before it was exposed. At that time few other geolo­ but is not continuous with it (Maxwell, written com­ gists believed the fault existed there. mun., 1962). As stated above, the Cambotas Quartzite PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A97

in the Serra do Caraga is intensely deformed, consisting seems preferable because of the space relations, because of recumbent folds sliced by thrust faults striking north it is consistant with the regional structural geology, and and dipping 30°-50° E., by other faults striking north­ because such an unconformity has not been found ebe- west and dipping northeast, and by still other vertical where in the region. faults striking east-northeast and east-southeast. All are To the north and west of the Fundao fault system and best developed in the orthoquartzites of the Cambotas the Dom Bosco syncline, only two significant thnist Quartzite. They extend into the protoquartzite and faults are known: that at the junction of the Serra da subgraywacke of the Maquine Group, but are neither as Moeda and the Serra do Curral, mentioned earlier, and numerous nor as prominent there. Only a few of these the one just discussed. faults extend into the phyllites and schists of the Nova Many small thrust faults are to be found throughout Lima Group or into and across the Minas Series or the the region, faults with displacements measured in centi­ greenstone schist on the east and south sides of the area. meters, meters, or very rarely, tens to hundreds of me- Lithology apparently played a part in the development ters. Most of these resulted from local adjustment of faults, such structures forming in greatest number during the folding. They do not always fall into general in brittle rocks, in fewest number in plastic rocks. It areal patterns, as do those discussed above, and they P re should be added that faults crop out more conspicuously of only local importance. in the bare outcrops of orthoquartzite than in the soil- covered rocks of more mixed composition. TEAR FAULTS On plate 1 a major thrust fault, marked with queries, The Eiigenho tear fault, forming much of the south is shown a short distance east of the Rio das Velhas. boundary of the post-Rio das Velhas Series rocks, is a This fault has not been recognized on outcrop. The north unique structure in the Quadrilatero Ferrifero. It has end, in the Rio Acima quadrangle, was mapped by Gair been traced for about TO km from east to west and turns (1962, PL 2); the part to the south was not mapped to the southwest at its west end (Guild, 195T). The by the geologists who worked in those quadrangles. structure branches in several places. The south side This writer believes that a thrust fault must exist near moved up sufficiently to cut out most of the post-Neva the location shown because of the strong structural dis­ Lima Group stratigraphic section along most of the cordance between the east and west sides of the Rio das trace of the fault; thus, it may be considered to have a Velhas valley and because the necking of the Rio das minimum vertical displacement of 2,000 in. The hori­ Velhas uplift into an hourglass shape creates space zontal component cannot be estimated; there is little problems that can best be explained by the existence of apparent horizontal offset in the Serra do Ouro Branco. such a fault. The structure is younger than folding and thrust fault­ South of the area mapped by Gair the location of this ing. It is also younger than the emplacement of granitic fault is questionable, for the rocks are homogeneous, rocks south of Ouro Preto and possibly younger than horizon markers are absent, and weathering is deep. some intrusive granitic rocks near Congonhas. It is cer­ The fault is visualized as generally following the bed­ tainly older than the diabase dikes, for Johnson (1962) ding but cutting across the dip at a low angle, with the considered that the long dike cutting the Serra do Ouro east side moving west. It was located on the map by Branco was emplaced along a branch of this fault. persistant lateral erosion features and, in the Rio de A large tear fault is well exposed just west of the Pedras quadrangle (pi. 6), by the presence of old surfi- Mutuca II hematite deposits, a short distance south of cial gold workings which can be followed across that Belo Horizonte in the Macacos quadrangle. This fault quadrangle for more than 15 km. The gold deposition is part of the rather enigmatic system at the north end was at first believed to have been controlled by a par­ of the Moeda syncline (p. A88) and limits the Mutuca II ticular stratigraphic horizon; no evidence of this was ore deposit. The fault plane has been exposed in several found in the field, and it seems at least equally probable underground workings both in the Mutuca II deposit that the gold, which was associated with sulficle min­ and the Zoroastro Passos deposit about 4 km to the eralization (O'Rourke, written conimun., 1960), may south and in surface workings at these and other inter­ have been deposited in a permeable zone along the trace mediate localities. These workings were opened after of the postulated fault. In the rest of the region gold mapping of the area by Pomerene. mineralization is posttectonic. This north-south fault is postore, has produced a Conceivably, the structural discordance is caused by notable unhealed breccia zone in hematite ore, and 1 as an unconformity within the Nova Lima Group; such an juxtaposed the Nova Lima Group against the Itabira unconformity would be difficult to locate in the homo­ Group. At its north end, the fault trace swings to the geneous rocks of that group. The thrust fault solution east and passes into a steep reverse fault with Minas A98 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL Series thrust over Rio das Velhas Series (fig. 28). In faults is instead controlled by the direction of local fold the Mutuca II area and to the south the fault is nearly trends. vertical; evidently the east side moved north relative Thus, most of the cross faults in the west half of to the west side. The horizontal displacement is un­ the Serra do Curral are nearly at right gugles to the known, but this writer suspects that the Mutuca II ore strike of the range, whereas a few others are at about deposit may once have been contiguous to, or part of, 45° to that trend. In the east half of that Serra, most the Capao Xavier deposit, on the west side of the fault cross faults are between 30° and 45° to the trend, and a some 3 km to the south. That deposit continues under few are at right angles. In the Serra do Itabirito most Tertiary and Eecent cover not less than 900 m east of cross faults are at right angles to the local strike, even the position shown by Pomerene (1964, pi. 2), as though that trend varies nearly 90°. In the northern revealed by subsequent drilling. Although this fault part of the Serra da Moeda, most cross faults approach was considered by Pomerene to swing to the west as the perpendicular to the trend, whereas in the southern shown on plate 1; the writer thinks that it continues to part, most approach 45° to the trend. Few cross faults the south as a strike fault, gradually dying out to the were mapped in the north edge of the Don Bosco syn- southward. It is not exposed south of Feixos. cline; they are at right angles to the fold structure, and As discussed in previous pages, part of the Fundao they change direction around the even bend at the fault should be considered a tear fault, and other thrust southwest corner of the Bio das Velhas uplift. The east faults also merge into tear faults for part of their side of the Santa Rita syncline is cut by cross faults extension. normal to the trend. The same is true in the Monlevade CROSS FAULTS district, where two major cross faults are nearly at right The regional structures are cut by many high-angle angles to a third, and all are roughly normal to the local cross faults, most of them small but a few of them large. structural orientation. The fault systems and cross-fold The cross faults seem to fall into two major orientations: system in the Santa Rita syncline are possibly the most those striking nearly at right angles to local structural complex because of the varying strikes of the rocks. trends, and those striking between 30° and 45° to those Displacements of most cross faults range from a few trends. Although some authors have related cross faults meters to perhaps 100 in. Two cross faults in the to regional stress patterns and have even assigned spe­ Monlevade area have displacements of more than 1,000 cific ages to such faults on the basis of their orientation, m; the third more than 2,000 m (Reeves, 1966). Other it is evident from plate 1 and from the individual major cross faults are a fault in the northern part of quadrangle maps of the region that the direction of cross the Catas Altas quadrangle, on the east side of the Quadrilatero Ferrlfero, which offsets th-5 greenstone schist about 700 m; three major cross faults in the Serra do Itabirito that have displacements of perhaps 400- 800 m (Wallace, 1965); and two in the Serra do Curral south of Belo Horizoiite, both with displacements in the order of 1,000 m. One of the Serra do Curral cross faults was chosen as a site for a water aqueduct through the mountain despite the presence on outcrop of one of the major fault breccias in the region and an obvious major offset of the rocks. The tunnel had been under construction for about 8 years and still ^as not com­ pleted in 1968; the original estimated cost has already been exceeded many times. Because they are closely related to the structural trends of the major post-Itacolomi folds and in many places cut intrusive granite, the cross faults are the FIGURE 28. Tear fault in the Mutuca II iron-ore deposit, near youngest structures to have formed and represent the N. 11,200 E. 3,200, Macacos quadrangle. Iron ore in the Cau§ final adjustment to stresses at the end of orogeny. There Itabirite is to left of hammer; Nova Lima Group phyllite is to right. At depth this fault dips steeply in the opposite is no evidence that they represent regionally directed direction. stress related to a later orogeny. PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A99

GEOLOGIC EVOLUTION OF THE throughout the Quadrilatero. This involves all P~e- QUADRILATERO FERRlFERO cambrian rocks of the region and indicates that the cen­ ter of the post-Itacolomi orogeny lay to the southeast. Because some of the reports resulting from this proj­ This orogenic zone has been traced roughly parallel to ect, as well as other papers on the region, discuss several the coast of Brazil for hundreds of kilometers by many orogenies, the philosophy underlying the discussion geologists, although the absolute age of the deformation herein must be stated, for the writer only discusses two is not well established either in the Quadrilatero Fer­ orogenies in the region, a post-Bio das Velhas orogeny rifero or elsewhere. and a post-Itacolomi orogeny. The terms "orogeny" and The question of whether there was one or more ore 7- "epeirogeny" are used in the sense defined by G. K. Gil­ enies in post-Itacolomi time is subject to more than one bert and I. C. Abbe, Jr., as quoted in the glossary of interpretation. The writer conceives of the orogeny in the American Geological Institute (1960, p. 206). the Quadrilatero Ferrifero as the result of gradual Orogeny is "the process of formation of mountain building up of directed forces in the subcrust and cru^, ranges by folds, faults, upthrusts, and overthrusts, af­ possibly by deep convection cells. These driving forces fecting comparatively narrow belts * * * while the were periodically relieved by folding, thrust faultirg, epeirogenic movements of the earth's crust produce and shearing, and recrystallization of the rocks. Stress f

change in source and a fundamental tectonic change in At that time granitic rocks occupied a much smaller the borderland. part of the total area of the Quadrilatero Ferrif ero tic an Here the record of Minas sedimentation ends in this today. The whole area to the east, northeast, and south­ region, for the next recorded event was strong uplift, east of the Quadrilatero Ferrifero was probably occu­ warping, and erosion of the Minas sediments to depths pied by Rio das Velhas Series rocks overlain by Minas at least 1,000 m into the last surface of sedimentation. Series and probably shaly Itacolomi Series rocks. The On this surface was deposited the Itacolomi Series. existence of granitic rocks in the Caete-Cocais uplift or The Itacolomi Series is a paralic group of sediments, elsewhere to the east, northeast, or southeast is doubtful, changing rapidly in nature in space and time. Coarse although a major granitic body some tens of kilometers lenticular conglomerates were deposited with fine silts farther to the north must have been already exposed, as and clean and dirty sandstones. These sediments shown by the nature of the preorogenic sediments to the possibly also came from the west and southwest, for north outside the Quadrilatero Ferrifero in the Serra they are generally finer to the northeast, but so little is do Cipo and the relations of the granitic rocks with the left in this area that no definite statement as to proven­ nearly flatlying postorogenic sediments which now roof ance is possible. them near Lagoa Santa and Vespaziano. A great thickness of sediments that has now been The first event in the post-Itacolomi deformation v^as completely removed from the region was undoubtedly probably a partial uplift of the Bagao complex and the deposited upon the rocks we now find, for the youngest granitic areas to the south and north which had boen Precambrian rocks now exposed are metamorphosed formed during the post-Rio das Velhas orogeny. THs to the same grade as the oldest. The great erosion that may even have started during the uplift that caured stripped this cover must have taken place during and the deposition of the Itacolomi and the younger, now after the post-Itacolomi orogeny, not only because of missing, rocks; and it is believed to have occurred very the metamorphic evidence cited above, but also because early in the major deformation, because later stresses the plastic failure of the pre-Minas, Minas, and Ita­ were all from the eastern quadrants, and the open Dom colomi rocks during that orogeny indicate considerable Bosco syncline could hardly have been formed by stress depth of burial at the beginning of that orogeny. Herz from this direction. The southern part of the Moeda (written commun., 1966) believes, on the basis of the syncline, contemporaneous with the Dom Bosco syn­ metamorphic minerals present in the surviving rocks, cline, probably started to form at this time also, caught that their depth of burial must have been 10 km or between the Bagao granitic rocks and those to the west. more, and this writer agrees. These synclines were open folds at this time. Following At the beginning of the post-Itacolomi orogeny, the this event, the gradual buildup of stress from the south­ rocks in the Quadrilatero Ferrifero were quite different east and east began, and the rest of the Rio das Veil as in nature and distribution than now. The Itacolomi uplift took shape, channelized at the north end by the Series rocks were sandstones, conglomerates, and sandy S. 60° W.-trending structure at the north side of the shales. The Minas Series rocks were sandstone, shale, region. The Santa Rita syncline, then possibly trending shaly and cherty limestones, limestones, jaspery iron- north-northwest or northwest, was outlined, and the formation, siltstone, tuffs, and the wackes. Whether the eastern part of the Gandarela syncline and the Con- Rio das Velhas Series rocks had been regionally meta­ ceigao anticline began to form, their location goverrQ.d morphosed during the pre-Minas orogeny is moot; by the thick quartzite mass now exposed in the Serras regardless, there is no unequivocal evidence that on the das Cambotas and do Tamaiidua. The west end of of the whole they were generally more highly metamorphosed Gandarela syncline was part of the east flank of the Pio than now. The main masses of the rock may have been das Velhas uplift structure, and the west end of the shales or argillites and slightly metamorphosed silt- syncline was probably a fold open toward the east, stones and wackes, with the Maquine Group silty sand­ shielded from the west and northwest-trending stress stone, silty shales, and subgraywacke. Thus, none of by the great mass of quartzite now forming the Serra these rocks, except conceivably the Rio das Velhas Series do Carac,a. rocks, were then as competent as they are today, and Stress from the east became more severe, folds be­ notable thicknesses were composed of incompetent and came tighter and steeper, and finally many folds w^re relatively incompetent sediments. Interlayering of overturned. Synclinal keels projected deep into the widely varying rock types makes the sediments aniso- crust; the recumbent folds of the Cambotas Quartzite tropic in their response to extenal force. Massive (then a sandstone?) formed in the Serra do Cara°.a. abruptly lenticular quartzose sedimentary bodies also The Vargem do Lima syncline was tightly folded in this contribute to structural anisotropism. part of the deformation, probably before the full de- A102 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL velopment of the Conta Historia syncline. Tlie Santa Velhas uplift. The isoclinal folding of the Conceigao Rita syncline became steep and was tightly folded uplift and further relative vertical movement of the against the Caraga monolith at the north end and southeast and east side of the Caraga mrss occurred, severly overfolded south of that mass. Cross folding, possibly with underthrusting at depth of the tightly caused by varying directions of stress and strain, began squeezed north end of the Santa Rita synclinal zone. and modified strongly the original fold and the south­ The ultramafic dikes and sills, now greenstone schist ern part of the Caraga uplift. Cross folding in the and talc schist, were probably intruded abcnt this time, Moeda syncline, less intense than that in the Santa Rita and metasomatic hematite bodies in the Caue Itabirite Syncline, also occurred. The east end of the Gandarela had formed in part during and shortly before this part syncline was tightly compressed against the thick of the orogeny. quartzite of the Serras do Tamandua and das Cambotas. The major direction of strain seems to have been to­ The west end, in the stress lee of the Caraga mass, re­ ward the west, and later the north of west, varying from mained open. The east end of the Conceigao anticline area to area with local lithologic and structural condi­ was tightly folded also. Crumpling of less competent tions. Thus, when the main epoch of thrust faulting rocks reached its apogee, and major rock flowage to started, the direction of movement in the south area was relative low stress parts of folds occurred. A massive east to west, guided by the granitic rocks of the Bagao compression of the crust toward the west and northwest complex and those to the south. Between the salients resulted. formed by the Mariana anticline and the Caraga-Con- Temperatures of the rocks increased, in part owing ceigao uplifts, the principal movements were east to to downfolding, in part to the deformation, and in west. North of the Caraga-Conceigao uplift the great large part probably to increasing temperatures caused Fundao thrust fault system had a principally north­ by the subcrustal activity driving the orogeny. Great west movement. The Fundao fault formed after the east mountains may have covered the whole area, only the side of the Gandarela syncline could yield no more by roots remaining today. Recrystallization and the begin­ folding; it broke and the southeast side moved upward nings of the change of the rocks to metamorphic rocks and to the northwest. The west end of th?, Gandarela accompanied increasing temperatures. With compaction syncline had been protected by the monolithic Caraga and recrystalHzation and unloading by erosion of over­ mass and had escaped most of the folding stress; it burden, quartzite became more brittle and massive, the moved northwestward only enough to bifurcate the axis plastic iron-formation became brittle itabirite, and the of the fold, connecting with the Ouro Fino and Conta argillaceous rocks slate, phyllite, and schist. Failure by Historia synclinal structures by relatively low trans­ folding gradually ceased as the rocks changed their verse anticlines which formed saddle folds like that at physical characteristics and yielding began to be by the north end of the Moeda syncline. One saddle fold fracturing rather than folding. was formed by the Conceigao anticlinal axis; the other The northwestward diminution of metamorphic by a lower northeasterly trending anticline, probably grade in the east side of the region shows that the higher formed early in the folding. temperatures were to the southeast, toward the erogenic The monolith of the Serra do Caraga and the iso- center, but there is no evidence that a significant tem­ tropic rocks of the Bagao complex seem to have channel­ perature gradient affecting regional metamorphism ized the strain to the northwestward on the west side existed in the central and western parts of the region, of the Serra do Caraga. Intense shearing in the south­ which are at the same metamorphic grade except where eastern part of the Vargem do Lima syncline parallel modified by intrusive activity. With many local excep­ to the axis of the fold, the formation of northwest- tions, stress seems to have diminished to the northwest­ trending strike faults parallel to the axis of the Conta ward, for foliation and shearing diminish in that direc­ Historia syncline, and the formation of the system of tion, and the complexity and tightness of folding also tear faults related to the thrusts to the south and to the diminished in that direction. Fundao fault attest rock failure on a larpre scale with More brittle rocks such as quartzite faulted first, and the east side of the faults moving northwest. Clearly, imbricate systems of minor thrust and tear faults such the whole block bounded to the north by the Fundao as those in the Serra do Caraga were formed, probably fault and to the west by the tear-thrust system moved to at the same time that nearby less brittle rocks were still the northwest. The rootless Ouro Fino syncline, at the failing by folding. Shearing of quartzitic rocks such edge of the upper plate, was twisted to the northeast and as the Maquine rocks in the Vargem do Lima syncline refolded on itself toward the low-stress sheltered area began, probably synchronously with the plastic bowing behind the Caraga monolith as the upper plate moved toward the northwest and overturning of the Rio das to the northwest. Strong imbricate thrust faulting oc- PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A103

curred in the sole of the plate where the Fundao fault rocks, with a rubidium-strontium age from feldspars merged with the southeast-trending tear fault. This of about 1,000 million years, were also intruded. northwest couple imposed strong cross folding in the A puzzling feature of the granitic rocks of the Qu^d- Caue Itabirite at the southwest corner of the Gandarela rilatero Ferrifero is the common presence, particu­ syncline. larly in the east side of the region, of biotite with a The Agua Quente thrust fault may well mark the west potassium-argon age of about 475-500 million ye^.rs, limb of another large mobile block, in which domi- although the strontium-rubidium age of muscovite and nantly pre-Minas rocks were thrust over Minas Series feldspar in the same rocks may be twice as old or older. rocks. There is now no way of measuring the displace­ No granitic rock with a strontium-rubiduim age of the ment along this fault, but the writer believes the fault feldspars less than about 1,000 million years is yet to be one of the larger faults of the region. known. This period of extreme stress and great crustal short­ The geologic map of Brazil published in 1960 by the ening undoubtedly further depressed the rocks of the Divisao de Geologia e Mineralogia of the DNPM, com­ synclinal belts, raised rock temperatures still further, piled by A. R. Lamego, shows that the coastal Precam- and palingenesis may have occurred at depth. Heat foci, brian shield area is composed of rocks which tr^nd perhaps caused by the movement of fluid magma from about N. 60° E. from Sao Paulo to southern Espirito depth, perhaps by metasomatizing fluids at depth, per­ Santo, a distance of about 750 km. A north west-trending haps both, developed; and metasedimentary rocks of fold axis runs from about Cabo Sao Thome on the coast suitable composition were gneissified and granitized in through the Quadrilatero Ferrifero and disappears un­ the region of the Bagao complex, forming thermal der the posttectonic Sao Francisco Series (Late Pre- aureoles. Mobile granitic magma was also emplaced here cambrian?) sedimentary rooks to the northwest. North and elsewhere. It is probable that the eastern and north­ of this fold axis the Precambrian rocks trend to the eastern granites and granitic gneisses formed, largely north for about 1,500 km. The age of this megastructure from Nova Lima Group rocks but also from others, at is not known, but the youngest Precambrian roekr of this stage. The granitic gneiss at the north side of the Brazil (believed by some to be Cambrian in age), the Bonfim dome and the mobile and other young granites Lavras Series of northern Minas Gerais and southern on the north side of the Serra do Curral were also Bahia, are involved. formed and emplaced, as included and intruded rocks The writer suggests as a working hypothesis that show crumpling and deformation similar to those unaf­ this major flexure may date from about 500 million fected by the granite. years ago. At the time of this deformation, although Perhaps owing to increased plasticity of the subcrust no granitic rocks were formed or intruded in the Quad­ and increasing grain size in the metasedimentary rocks, rilatero Ferrifero, the rocks may have been hep.ted perhaps owing to some other cause such as a natural enough to cause leakage of argon from biotite but not ending of the erogenic cycle by dying out of subcrustal enough to greatly disturb the strontium-rubidium ratio activity, the generally east to west strain seems to have in the more stable muscovite and feldspars. ended after relief by the large scale thrust faulting, and Guimaraes has shown that a major period of pegma­ no further crustal shortening is recorded in the rocks tite formation occurred at about 450 million years ago, after the thrust-faulting cycle. None of the major faults which he related to a "Caledonian orogeny"; conceiva­ seem to be significantly folded. The thrust faults are bly this may represent the period of heating and loes of themselves faulted only along the south border, where argon from the biotite of the older granitic rocks. the great Engenho fault cut off the thrust plates to the Toward the end of the Paleozoic or early in Meso,xoic south, and others of the many thrust plates in the time, epeirogenic uplift of this region occurred, and southern part of the Dom Bosco syncline were cut by the peneplaned surface began to be dissected. Four ma- east-west faults. During the relaxation of the stresses jar uplifts are here recorded by relict erosion surfrces, and in the later period of adjustment, normal faults correlated by King with the Early Cretaceous, Late developed perpendicular to the trends of the major folds Cretaceous, middle Tertiary, and late Tertiary. Many and at 30°-45° to those trends. A few cross faults with minor uplifts occurred between the major uplifts, leav­ major displacement developed at this time or later, after ing intermediate erosion surfaces. In this area there is the major folding. no evidence for faulting or folding during this interval Although much of the granitic rock formed during or to the present. the post-Itacolomi orogeny was foliated generally par­ Although it is conventional to relate the diabase dikes allel to the foliation of the adjacent metasediments (pos­ with the Jurassic plateau lavas of Sao Paulo-Parana, sibly a relict foliation), posttectonic unfoliated granitic dated by potassium-argon methods as about 135 mil- A104 REGIONAL GEOLOGY OF THE QUADRILATERO FERRIFERO, MINAS GERAIS, BRAZIL

lion years old (Amaral and others, 1966), there is, as 1964, Supergene iron ores of Minas Gerais, Brazil: Econ. of 1966, no firm evidence for this. Because the diabasic Geology, v. 59, no. 7, p. 1203-1240. -1965, Nature and origin of the high-grade hematite ores rocks are locally intimately related to the major gab- of Minas Gerais, Brazil: Econ. Geology, v. 60 no. 1, p. 1-46. broic intrusives, also undated, it seems probable to this Dorr, J. V. N., 2d, and Barbosa, A. L. M., 1963, Geology and ore writer that they may be somewhat older, as the gab- deposits of the Itabira district, Minas Gerais Brazil: U. S. broic rocks are in many places very coarse grained and Geol. Survey Prof. Paper 341-C, 110 p. may well have been intruded at notable depth. Dorr, J. V. N., 2d, Coelho, I. S., and Horen, Arthur, 1956, The manganese deposits of Minas Gerais, Brazil: Internat. Geol. REFERENCES Cong., 20th, Mexico City, 1956. Symposium sobre yacimien- tos de nianganeso, v. 3, p. 279-346. Dorr, J. V. N., 2d, Gair, J. E., Pomerene, J. B., and Rynearson, Aldrich, L. T., Hart, S. R., Tilton, G. R., Davis, G. L., Rama, G. A., 1957, Revisao da estratigrafia pre-cambriana do Quad- S. N. L, Steiger, R., Richards, J, R., and Gerken, J. S., 1964, rilatero Ferrifero: Brazil Dept. Nac. Prod'icao Mineral, Isotope Geology in Ann. Kept. Director Dept. Terrestrial Div. Fomento Produgao Mineral, Avulso 81, "1 p. Magnetism Carnegie Inst. Washington Yearbook 63: p. Dorr, J. V. N., 2d, Herz, Norman, Barbosa, A. L. M., and Sim- 331-340. mons, G. C., Esbogo geologico do Quadriiatero Ferrifero de Alves, B. P., 1961, Sumario sobre estratigrafia e estrutura das Minas Gerais Brasil: Brazil Dept. Nac. Pro'lucao Mineral quadriculas de Caete e Serra da Piedade: Soc. Intercambio Pub. Espec. 1, 120 p. [Portuguese and English] [1961]. Cultural e Estudos Geologicos (Ouro Preto), Pub. 1, p. 257- Ebert, Heinz, 1963, The manganese-bearing Lafaiete-Forma- 260. tion as a guide-horizon in the Pre-Cambrian of Minas Ger­ Amaral, G., Cordani, U. G., Kawashita, K, and Reynolds, J. H., ais: Acad. Brasileira Cienc. Anais, v. 35, no. 4, p. 545-559. 1966, Potassium-argon dates of basaltic rocks from south­ Freyberg, Bruno ^on, 1932, Ergebnisse geologischer Forschun- ern Brazil: Geochim. et Cosniochim. Acta v. 30, no. 2, p. gen in Minas Geraes (Brasilien) : Neues Jahrb. Mineralo- 159-190. gie, Geologie u. Palaontologie, Sonderband 2, v. 2, 403 p. American Geological Institute, 1960, Glossary of geology and Gair, J. E., 1958, The Sahara Formation: Soc. Brasileira Geolo­ related sciences, with supplement [2d ed.] : Washington, gia Bol., v. 7, no. 2, p. 68-69. D.C., Am. Geol. Inst., 397 p. 1962, Geology and ore deposits of the Nova Lima and Barbosa, Octavio, 1949, Contribuigao a geologia do centro de Rio Acirua quadrangles, Minas Gerais, Braz'l: U. S. Geol. Minas Gerais: Mineragao e Metalurgia, v. 14, no. 79, p. 3-19. Survey Prof. Paper 341-A, 67 p. 1954, Evolution du geosynclinal Espinhaco: Internat. Goldschmidt, V. M., 1954, Geochemistry: London, Oxford Univ. Geol. Cong., 19th, Algiers 1952, Comptes rendus, sec. 13, Press, 730 p. pt. 2, f asc. 14, p. 17-36. Gorceix, Henri, 1884, Bacias terciaria de agua doce nos arredores Becker, G. F., 1895, Reconnaissance of the gold fields of the de Ouro Preto (Gandarella e Fonseca), Mina^ Geraes, Bra­ southern Appalachians: U.S. Geol. Survey 16th Ann. Rept, sil : Ouro Preto, Escola de Minas, Annaes, no. 3, p. 95-114. pt. 3, p. 251-331. Gordon, Mackenzie, Jr., Tracey, J. I., Jr., and Elli*, M. W., 1958, Bolau, E., 1952, Geology of the Caete-uplift, central Minas Ger­ Geology of the Arkansas bauxite region: U.S. Geol. Survey ais, Brazil: Geol. Fo'ren. Stockholm Fb'rh., v. 74, no. 4, p. Prof. Paper 299, 268 p. 475-493. Guild, P. W., 1957, Geology and mineral resources of the Congon- Brajnikov, Boris, 1947, Essai sur la tectonique de la region a Test has district, Minas Gerais, Brazil: U.S. Geol. Survey Prof. de Belo Horizonte, Minas Gerais, Bresil: Soc. Geol. France Paper 290, 90 p. Bull., 5th ser., v. 17, p. 321-335. Guimaraes, Djalma, 1931, Contribuigao a geologia do Estrdo de Buchi, James, 1961. Geologia da Fazenda da Alegria : Soc. In­ Minas Geraes: Brazil, Service Geologico e Mineralogico, tercambio Cultural e Estudos Geologicos (Ouro Pr§to), Pub. Bol. 55. 36 p. 1, p. 127-148. 1935, Contribuigao ao estudo da origem dos depositos Burton, Capt. R. F., 1869, Explorations of the highlands of the de minerio de ferro e manganez do centro de Minas Geraes: Brazil: London, Tinsley Bros., vols. 1 and 2, 921 p. Brazil Service Fomento Produgao Mineral, Bol. 8, 70 p. Oastano, J. R., and Garrels, R. M., 1950, Experiments on the 1953, Notas & margem da critica: Belo Horizonte, Inst. deposition of iron with special reference to the Clinton Tecnologia Indus. Avulso 16, 39 p. iron ore deposits: Econ. Geology, v. 45, no. 8, p. 755-770. 1958, Geologia estratigrafica e economica do Brasil: Belo Correns, C. W., 1952, Mineralogische Unterschungen an Sedi- Horizonte, Brasil, Estab. Graficos Santa Mar'a S. A., 450 p. mentaren Eisenerz: Internat, Geol. Cong., 19th, Algiers 1961, Fundamentos da metalogenese e os dep<5sitos 1952, Symposium sur les gisements de fer du monde, v. 2, p. minerals do Brazil: Brazil Dept. Nac. Produgao Mineral, 28-30. Div. Fomento Produgao Mineral, Bol. 109, 441 p. Harder, E. C., and Chamberlin, R. T., 1915, The geology of cen­ Derby, O. A., 1906, The Serra do Espinhago: Jour. Geology, v. tral Minas Gerais, Brazil: Jour. Geology, v. 23, no. 4, 14, no. 3, p. 374-401. p. 341-378, no. 5, p. 385-424. 1910, The iron ores of Brazil: Internat. Geol. Cong., llth Henwood, W. J., 1871, Observations on metalliferous deposits: Stockholm 1910, The iron ore resources of the world, v. On the gold mines of Mines Gerais in BrazP : Royal Geol. 2, p. 813-822. Soc. Cornwall Trans., v. 8 pt. 1, p. 168-370. Dorr, J. V. N., 2cl, 1958a, The Caue Itabirite: Soc. Brasileira Herz, Norman, 1962, Chemical composition of Precambrian Geologia Bol., v. 7, no. 2, p. 61-62. pelitic rocks, Quadriiatero Ferrifero, Miuas Gerais, Brazil, 1958b, The Gandarela Formation: Soc. Brasileira Geolo­ in Short papers in geology and hydrology : U.F Geol. Survey gia Bol., v. 7, no. 2, p. 63-64. Prof. Paper 450-C, p. C75-C78. PHYSIOGRAPHIC, STRATIGRAPHIC, AND STRUCTURAL DEVELOPMENT A105

Herz, Norman, and Dutra, C. V., 1964, Geochemistry of some Bol. 226, 55 p. Also published in 1965, as Zur geologie der kyanites from Brazil: Am. Mineralogist, v. 49, p. 1290-1305. sudlichen Espinhago Zone und ihrer prakambrischen Dia- Hurley, P. M., 1961, The basement of Central and South mantvorkommen, Minas Gerais, Brasilien: Deutsch, G?ol. America, or, how not to date a continent [with discussion] : Gesell. Zeitschr, Jahrg. 1963, v. 115, p. 177-215. New York Acad. Sci. Annals, v. 91, art, 2, p. 571-575. Pomerene, J. B., 1958a, The Cercadinho Formation: Soc. Bra­ James, H. L., 1954, Sedimentary facies of iron formations: Econ. sileira Geologia Bol., v. 7, no. 2, p. 64-65. Geology, v. 49, no. 3, p. 235-293. 1958b, The Tabooes Quartzite: Soc. Brasileira Geologia 1955, Zones of regional metamorphism in the Precam- Bol., v. 7, no. 2, p. 66-67. brian of northern Michigan: Geol. Soc. America Bull., v. 1958c, The Barreiro Formation: Soc. Brasileira Geologia 66, no. 12, pt. 1, p. 1455-1487. Bol., v. 7, no. 2, p. 67-68. Johnson, R. F., 1962, Geology and ore deposits of the Cachoeira 1964, Geology and ore deposits of the Belo Horizonte, do Campo, Dom Bosco, and Ouro Branco quadrangles, Minas Ibirite, and Macacos quadrangles, Minas Gerais, Bra zil: Gerais, Brazil: U.S. Geol. Survey Prof, Paper 341-B, 39 p. U.S. Geol. Survey Prof. Paper 341-D, 84 p. King, L. C., 1956, A geomorfologia do Brasil Oriental: Revista Reeves, R. G., 1966, Geology and mineral resources of the Mon- Brasileira Geografla, Ano 18, no. 2, p. 147-265. levade and Rio Piracicaba quadrangles, Minas Gerais, Bra­ Lacourt, Fernando, 1936, Resumo do geologia da folha de Ouro zil : U.S. Geol. Survey Prof. Paper 341-E, 58 p. Preto: Brasil Univ. Escola Nac. de Minas e Metalurgia Rynearson, G. A., Pomerene, J. B., and Dorr, J. V. N., 2d, 1954. (Ouro Preto), Annaes, No. 27, 48 p. Contacto basal da Serie de Minas na parte ocidental do Lamego, A. R., compiler, 1960, Mapa geologico do Brasil: Brazil Quadrilatero Ferrifero, Minas Gerais, Brasil: Brazil Dept. Dept. Nac. Produgao Mineral, Div. Geologia e Mineralogia, Nac. Produgao Mineral, Div. Geologia e Mineralogia, Avulso scale 1: 5,000,000. 34, 18 p. Matheson, A. F., 1956, The St. John del Rey Mining Company, Scott, H. K., 1900, The manganese ores of Brazil: Iron and S^eel Limited, Minas Geraes, Brazil; history, geology, and min­ Inst. Jour. 1. 57, p. 179-208 (with discussion by O. A. Derby eral resources: Canadian Mining Metall. Bull., v. 49, no. 525, and others, p. 209-218). p. 37-^3. Simmons, G. C., 1958, The Fgcho do Funil Formation: Soc. Mawe, John, 1816, Travels in the interior of Brazil: Philadel­ Brasileira Geologia Bol., v. 7, no. 2, p. 65-66. phia, M. Carey, 366 p. 1968a, Geology and iron deposits of the western Se^ra Maxwell, C. H., 1958, The Batatal Formation: Soc. Brasilerira do Curral, Minas Gerais, Brazil: U.S. Geol. Survey P"of. Geologia Bol., v. 7, no. 2, p. 60-61. Paper 341-G, 57 p. Moore, S. L., 1969, Geology and ore deposits of the Antonio dos 1968b, Geology and mineral resourses of the Barao de Santos, Gongo Soco, and Conceigao do Rio Acirna quad­ Cocais area, Minas Gerais, Brazil: U.S. Geol. Survey P~of. rangles, Minas Gerais, Brazil: U.S. Geol. Survey Prof. Paper 341-H. 46 p. Paper 341-1, 50 p. Simmons, G. C., and Maxwell, C. H., 1961, Grupo Tamandur da Moraes, L. J. de, and Barbosa, Octavio, 1939, Ouro no centro de Serie Rio das Velhas: Brazil Dept. Nac. Produgao Mineral, Minas Gerais: Brazil Dept. Nac. Produgao Mineral, Div. Div. Geologia e Mineralogia, Bol. 211, 30 p. [1962]. Fomento Produgao Mineral, Bol. 38,186 p. Teixeira da Costa, Manoel, 1961, Sedimentagao e orogenese da Moraes Rego, L. F., 1932, As jazidas de ferro no Centro de Serie de Minas: Ouro Preto, Soc. Intercambio Cultural e Minas Gerais: Belo Horizonte, Campanha Economica de Estudos Geol6gicos, no. 1, p. 55-61. Minas Gerais, 81 p. Tolbert, G. E., 1964. Geology of the Raposos gold mine, Minas Oliveira, A. I. de, 1956, Brazil, in Jenks, W. F., ed., Handbook Gerais, Brazil: Econ. Geology, v. 59, no. 5, p. 775-798. of South American geology: Geol. Soc. America Mem. 65, Tyler, S. A., 1948, Itabirite of Minas Gerais, Brazil: Jour. Sed. p. 1-62. Petrology, v. 18, no. 2, p. 86-87. Petti John, F. J., 1957, Sedimentary rocks [2d ed.] : New York, Wallace, R. M., 1958, The Moeda Formation: Soc. Brasileira Harper & Bros., 718 p. Geologia Bol., v. 7, no. 2, p. 59-60. Pflug, Reinhard, 1965, A geologia do parte meridional da Serra 1965, Geology and mineral resources of the Pico de do Espinhago e zonas adjacentes, Minas Gerais: Brazil, Itabirito district, Minas Gerais, Brazil: U.S. Geol. Survey Dept. Nac. Produgao Mineral, Div. Geologia e Mineralogia, Prof. Paper 341-F, 68 p.

INDEX

[Italic irage numbers indicate major references]

Page D Accessibility...... A7 Carbonate-facies, iron-formation..__... A18,19,26 Page Acknowledgments... ___ -__...... 6 Carbonate-magnetite facies, iron-formation... 18 Deformation, and metamorphism, regional. .. A102 AguaQuente fault___...... __.._... 96,103 Casa Forte Formation___....__.....__ 23 major periods _ ...... 82 Alegria syneline, Cau6 Itablrite...... __.. 40 Cau6 Itabirite.______...... _.....__ 40 Depositional environment, Cambotas Quart­ Alluvium...... 8,75 associated cleavage quartz...... 81 zite - 100 Aluminum, production_..._....__.--.-. 4 contact relations__------_....- 40 Caraca Group-- --..-..------. 38,100 Analyses, chemical, amphibolitie itabirite.... 42 gold...... 82 Itabira Group. __ . . 47 chemical, dolomitie itabirite.------..--.- 42 physiographic expression....-.----.-...-- 43 Itacolomi Series..------68 itabirite..-...-.-....--..-...... -..... 42 source of water...... _._...... 39 Maquine" Group __ . _. . . 2' 27,99 marble, Fe~eho do Funil Formation... 53 Cercadinho Formation..__-..-.-...... - 49 Nova Lima Group 20,30,99 mudstone-.....------.---.---.-....-. 70 erosion surface_...-.--..-..-. .-...-.- 13 Piracieaba Group.. ____ 58 Tabooes Quartzite___._-__--.._. 54 quartz-kyanite veins.....----.-.--...--.- 82 Sahara Formation.. . __ ___ .. .. . 57,58 Andalusite...-__-..._._...... _.. 18 weathering..------.------.------...--.- 52 Tamandua Group.------_ 33,100 Argillaceous rocks, mechanical properties..... 9 Chapadade Canga, pediment-.--.----.----.- 13 Derby, O. A., quoted.-- -- 14 Chlorite . . ..._--_--.. .-. - 43 Diabase.. ______. _____ 79 B Chloritoid 23,25,52 dikes.. 103 Bacao complex._....__...... ___.... 8,101 Cities of region. __. _...... ____. __ 4 D iamantina facies, Minas Series ... _ .... 67 eflect on faulting...... 87 Climate.__..__._...... __...... 7 Dikes... , . 80 pegmatite dikes...... 80 CoUuvium.-.--.--.---..---.---.---.-.---..-- 76 diabase.. ___ ...-.. . __ -... 103 relation to folding...... __._. 86,87 Companhia Vale do Rio Doce....__------6 ultramafle----.-... ------79 relation to Rio das Velhas uplift...... 9% Coneeifao anticline.. . _ -...._..... 91,93,101,102 Dolomite, Cereadinho Formation. _ ... 52 Barbacena Series...... 64 Conceicao-Caraca uplift...... _..-.---....-. 85,93 metallurgical and refractory ... . 39 Barbosa, A. L. M., and Dorr, J. V. N., 2d, Conglomerate, basal, Cercadinho Formation.. 49,50 splash rock ___ . __ . ______. __ . 72 quoted...... _._ 40 basal, Fonseea Formation..._.-...... 68 weathering. _..-... ___ ... ____ .. 10,72 Barreiro Formation._...... 54 Minas S ries.....--.-----.------.- 36 Dolomitization, Gandarela Formation. __ ... 46 equivalents...... _..._...... _...... 49 Palmital Formation...... -..-._._.--. 22 Dom Bosco fault-...... ______. _ 96 Batatal Formation...... 37,39 Sahara Formation.....-.------.----- 56 Dom Bosco syneline __ ..--.. ---....---.-- 85, 86 Bauxite...... __..._...._...... __ 4,10 Casa Forte Formation...__-...... 23 cross faults.. ____ . ... 98 Gandarela Formation...... --.-...._._. 39 intraformational, Gandarela Formation.. 46 drainage.----...-- -. . -. 8 laterite deposits...... 73 Itacolomi Series..._...... __..... 60,61 thrust faults.... 96 relation to erosion surfaces..-.-----....--. 14 Maquine Group undivided...-----.._---- 24 Dorr, J.V.N., 2d,quoted.------45 Bicas Gneiss Member, Elefante Formation... 49 Moeda Formation...... 36 and Barbosa, A. L. M., quoted-. Jfl Bonflm dome_..-....._--.-...... _---....._. 85 Nova Lima Group..------20 Drainage of region...... -...--..-. ... 7 Borraehudos Granite__..__....-_...__ 9,77,94 Sahara Formation---.---.------.----- 57 Building stone_. .,____-.-..-____ 39,53 Conta Historia syneline...... - 8,85,91,93,96 E Conta Historia area, Cau6 Itabirite--..-.__ 40 Elefante Formation.. ... ___ .. . 49 Contact relations, Barreiro Formation..--...- 55 Engenho fault... -...... -- - 87,97 Caete-Cocais uplift...... 101 Batatal Formation. -.--.-. ------37 Epeirogeny, definition _ ------99 Caete" complex, relation to Rio das Velhas Cambotas Quartzite-.---.------.- 31 Erosion, post-Minas..-.- . . 60 uplift___....___._...... 93 Casa Forte Formation .---.------23 pre-Itaeolomi__ .... - _ - _ ... 84 Caeteuplift-.. . . 85 Cau6 Itabirite- .------40 pre-Minas.-..-....- - .-- 28 Cambotas Quartzite - ---, - - . SO Cercadinho Formation.------49 Erosion surfaces. . - . 10 dikes...... 80 Fecho do Funil Formation.-.-___ .-. 52,54 between post-Gondwana and Sul-Amer- equivalents._...... _...... 32 Fonseea Formation.- ---. ------68 icano surfaces _ -. ..-...-.... - 13 faulting_---...... ----.....___--.-.. 97 Gandarela Formation...... -. 45 folding...._------_..__-----.._ 94 F quartz veins...... 81 Itacolomi Series-- .- . . ..- . 60 Faults, cross ___ __ . sedimentary structures.._...... 31 Maquine Group..------21,27 Dom Bosco syneline _ _ Canga, Cau6 Itabirite..-----...... 43 Maquine Group undivided----.. ... 24 influence on mineralization. ... 97 chemical__..._..___._...... 45,75 Minas Series 27,36 Moeda syneline _ ...... _ --- 88 cat by Bio Piraeieaba.--...... 13 Nova Lima Group--. - 16,27 pre-Itacolomi... .--...... -. --.- 84 definition ...... 73 Palmital Formation------22 relation tolithology..---- - _ ---- 97 Serrado Caraca... __ -.--.. 94 normal...---...... 74 Piracieaba Group.... -._. . .. . 48 on erosion surfaces_.---.--__---.-...._-. 14 tear...... 97 origin...... _...... 75 Rio das Velhas Series-_....___.. ... 36 thrust-.-----.-.---.-.------95 relation to erosion surfaces...-._.----....- 11,13 SabaraFormation --- . -- - 55 FSchodo Funil Formation.. _ ... . 49,5# types___...... _ 74 Tabooes Quartzite_..-.-.. . . 54 Feldspar, in Cambotas Quartzite 31 weathering...... _._...... _...... _..... 10,74 Tamandua Group ------.. 66 Fluorite lenses --. . _ ------46 Canga rica...-----.....--..-.....-----..----. 75 unnamed formation..._ . ... 32 Flyseh suite .... . 58 CapaoXavier deposit...... _..... 98 Coordinate system...... --- 5 Folds, Isoclinal.... 94 major __ . .._...... -.-. ...-- 85 Capivara syneline...____._...... _.... 94 Copper sulfldes___ .______-. -.-. --.- 46 Caraca Group, Batatal Formation_...... 37 Fonseea basin------68 definition.----..__...... S3 Cross faults.--.--. _.. .__ 98 Fonseea Formation.-. ___.- ._. -- 68 depositional environment_..--_--..-.--. 38 Cross folds, Moeda syneline. _- .. _--. - 87,88 covering pediment - . . - 13 metamorphism.-..-._---.------_-.--. 39 Santa Rita syneline..... ------85 Fossils, plant, Fonseea Formation. ------68 Moeda Formation...._...... __._. SA sitesof ore bodies - -._ __ 86 Freyberg, Bruno von, quoted_ .. . 63 A107 A108 INDEX

Page Iron formation Continued Page Fuchsite . -- -- A23 unnamed formation__. . . A32 Metachert, Batatal Formation___.__..... A38 Fundao fault.. 95,98,102 volcanic origin_....-----.----...-.....- 79 Casa Forte Formation______.__ 23 Gandarela syncline-- - - 90 weathering...-..------. 10 Gandarela Formation...... __.....__ 46 Ouro Fino syncline__.__. . 91 Itabira, u-onproduction.-.-.------_.....---. 4 Nova Lima Group....._.______20,26 Itabira district, structure.------.------9l Sabara Formation.__._....._..__ 57 G Itabira facies, Minas Series------....-...--- 67 Metamorphism, and deformation, regional__ 102 Gabbrolc rocks...____.__..-. -.--... 79 Itabira Group, Cau6 Itabirite-.. -...... --- 40 Caraca Group...... _....______39 Gair, J. E., quoted.. 17,18 definition..... _ .- __. 39 confusion caused by variation. _____ 64 Gandarela basin ... . 68 depositional environment ------47 contact, Nova Lima Group..______17 Gandarela Formation...___.__...... 44 difficulty in division ... .--.... 39 Itabira Group...._.______48 contact relations .. ...--.....---..-.- 45 economic resources_._-_--- _.__. 39 Maquing Group....__.__....____. 25 economic resources..-.------.---.-.--.-.- 39 Gandarela Formation___-...-__.-... 44 Minas Series...____._..._____ 28 marble....--.--.------...-.-...... Jf metamorphism..__.-..----..--.- --- 48 Nova Lima Group._.._.______21 physiographic expression.. ------46 Itabirite, amphibolitic__...... - 42 Piracicaba Group..__.______59 sinkholes.------.------45,47,10 chemical analyses___.-..-..-.--....- . 42 quartz segregation______81 Gandarelasyncline..---.------85,90,93,101 definition.._...-..--..-.-.-.-..-----.--. If) Rio das Velhas Series....______100,101 Batatal Formation.------.----- 38 dolomitlc .- .. . -. 42,45,46 Metasomatization, Cambotas Quartzite___ 31 Cau6 Itabirite - - - 40 folded...... -..-...... -...... ---.- 60 Metavolcanic rocks, Nova Lima Group___ 18 Cercadlnho Formation. ------49 Gandarela Formation.--. ______. 45,46 Methane___.______.______20 Gandarela Formation..- -_ ------44 talus, commercial_-....---.. .--...-.. 76 Minas Series . . . 14 Garnet.--..-..------...... ------18,56,59 weathering.. ..-.-.--__._.__... 10 age relative to granitic rocks______77 Geologic history of region...._._ ..... 99 Itacolomi Quartzite_.__._..._____ 59 age relative to ultramafic rocks...--___ 79 Gneiss, granitic, relation to Rio das Velhas Itacolomi Series__...... _...... ___._. 59 basal, correlation. ______33 Series...... 16 deformation..--... .---. ---___..... 84 contact relations-___.____..___ 27,36 Gneissicrock, reconstituted^.------.... 15 dikes...... __ . 80 deformation______84 Gold, Cau6 Itabirite....------39 evolution of nomenclature... _...... 64 dikes..-... --.-...... _.. 80 deposition, relation to faults..------.. 97 nature of sediments_ ...--...--...-- 101 distribution.... . 28 lode, host rocks__-___ _.___ .. 19 division.. ______.... 27 quartz veins... -- .-.-.. 82 evolution of nomenclature.___ ..- 64 placer.....------..----- 3,76,82 metamorphism--______28 production.._.._._.. .__------3,82 King, L. C., study of erosion surfaces____ 10 rotation during deformation. __ 83 Gondwana erosion surface--.------.------11 Kyanite...... 23,25,31,52 superposition or lateral equivalence of Gongo Soco deposit, palladium__-....----.- 39 Boticamine, production ...... 82 units...- 67 Granite, cobbles__---...------.------57 Kyanite-quartz veins__...... 82 Tamandua Group______._ 88 Granitic rocks..-.-._.--.... -.- - ... 76 weathering..__.______.______28 age------_____._..._.__ --- 103 Mineral deposits, relation to erosior surfaces.. 14 age relative to Minas Series-...- . 77 Mining history...... _...... 3 Moeda Formation....______....._ 34 emplacement ------100 Lagoa das Antas, sinkhole lake --- ..... 47 deposition, synchronous witv Batatal radiogenic age..------77 Lagoa das Coutas, sinkhole lake ...... 47 Formation___....__...___ 39 weathering---..------8 Lagoa de Metro, sinkhole lake. .- .___.._ 69 physiographic expression.___...... 36 Granitization. -_-..-.-.-..-. .... -. . 16 Lagoa S6ca, sinkhole lake.. __ _ 47 quartz veins.___...__.._...____ 81 Graywacke, Nova Lima Group. ___. -_ -_- 20,26 Landslides---..-...... -.------.. 9,76 Moeda syncline__.______...__ 85, W, 101 Palmital Formation__...._...... 23 Lapaseca. _ ...... -...--...--.-... _ 19 Sahara Formation---..-----...--.---.---- 56 Laterite.. 72 Cau« Itabirite- - --- 40 weathering--...------.------9 Lignite-.. . .. 13,68 drainage . ..____...- 8 Greenschist, Gandarela Formation...... -.-.. 46 Limonite, canga.-.- . - --- - 74 junction with Serra do Curral. _.. __ 88,95 Greenstone..-..------56 Lithology, relation to faults--. .------97 Monlevade area, Batatal Formatter. _.__ 38 Grit, Cercadinho Formation...... _-...----... 49,50 Location.... ___ .. __ ...... _ 3 cross faults- _____.__ .__-__ 98 Moeda Formation.______.._.__- 34,36 Monlevade district, structure___.__...... 91 rice__...-----.--..-.------..---- . 51 M Morro do Chapeu, Cau6 Itabirite..-__._.. 40 type Itacolomi___-...... --.--- 59 Magnetite....-.-...... 19 Morro Velho gold mine.___.__.___... 4 GuanhSes facies, Minas Series...-.__...... 67 initabirite-. 40,42 Guild, P. W., Quoted...-.....-....-.---.----- 81 placer.... ______... __ .. 76 methane ------.---. 20 Gullying .- - .--- . 8 Manganese---...------_ ------4 Mudstone..... _.____.____.___ 69 associated with itabirite... ------43 Mutuca II deposit, faulting______...... 97 H Cercadinho Formation.--- . .. 52 Gandarela Formation...... 45 N Hematite, Cercadinho Formation....------.- 50 in cross folds.. __ 86 in alluvium_ .....- --...-...... 76 Itabira Group...------39 Nomenclature, stratlgraphic.__....___... 63 Santo Antonio facies, Itacolomi Series... - 61 possible commercial occurrence. -.. -. 73 Nova Lima Group.._____.__._._... 16 weathering.-..-_-__----.------10 relation to erosion surfaces. ... - 14 contact relations_.___...... _ _.. 16,27 Hematite-quartz veins.._____.... . 82 Tabooes Quartzite------54 deformation...... ___..... 83 Mapping of region. _ _ . 5 depositional environment_....___.. 20 MaquinS Group... _ ..-. _ . _ . _____ .. SI dikes. . . - 80 Industries of area.. _ .- . - 4 contact relations _ ....-.---.-.----.---.--'21,27 flysch suite...... _..._...._ 26 Iron, in cross folds. _.-.-._-.------. 86 Itabira Group. _..---.._--.--- 39 depositional environment ...-. ... 24, 27 gold production.....____...... 82 mining.------___ - - 4 metamorphism... ------25 iron-formation ..______._...... 17 supergene, relation to erosion surfaces.- 14 molasse suite..--.------27 iron-formation facies...... _...---...... 78 Iron-formation, correlation with Caufi Ita­ quartz veins.--.. . _ ... _ ...... 81 llthology-. - . . 16 birite 40 undivided ___ .-.- -. -..- S4 metamorphism-..-.______21 lenses, evidence of folding. _. . _ ------94 Maquine" Quartzite, dikes...... _ . 80 metasomatism____....__._____ 32 lenticular ...-_------__-__ 21 Marble, Caufi Itabirite 43 metavolcanic rocks_.______18 Nova Lima Group..--__...... _. 17, 18 F6cho do Funil Formation. _ _..______53 parent rocks....-.-...... 20 occurrence outside Quadrilatero Ferrlfero. 40 Gandarela Formation.. __ ...... __ 46 possible subdivision-...---. _....._. 16 origin___-__-_-.._--_-----.-.....--___-- 47 Mariana anticline...... -.------.- 8,86,93,96 thickness__.______..______._ 16 pebbles in conglomerate..------__----- 23 Maxwell, C. H., quoted.--.- -.-. -.-- 31 ultramafic rocks.___...._____._. 79 Sabara Formation...... ___ . 57 Mechanical properties, argillaceous rocks ..... 9 weathering . _____ .... 17 INDEX A109

Page Page Page O Quartzite, Batatal Formation.._.--....._. A38 Serra das Gaivotas, thrust fault... __. __ A88 Ornamental stone------.----..-..--- _-.- A39 Cambotas Quartzite.---.-.---.---..--_.. 31 Serra do Caraca, alluvium to north _ __ _ 8 Orogeny, definition___.__ _--_-_ -_ ___._ 99 Casa Forte Formation__...... ___... 23 Batatal Formation.._...____.__ 38 Orogenic activity -_ _ .-.- - ---.-_. 84 Cercadinho Formation...... ____ 60 erosion surface______.______11,13 Ouro Finosyncline.------91 Maquine" Group undivided__..-.__.. 24 isoclinal folds.___...... ______94 Ouro PrSto region, Caraca Group...---.. __ 34 Moeda Formation_...--_-.. ..-..__ 34,36 metamorphism.___...... ______25 Nova Lima Group..._...... _..... 19 thrust faults.-...... ___.__ % Palmital Formation__-.. ______22 type locality of Caraca Group.._____ 34 SabarS Formation__...._.___.__. 57 Serra do Curral__.___....____.__ §8 Palladium, Itabira Group.___ ----- 39 Santo Antonio facies, Itacolomi Series. - - - 61 Batatal Formation.__.....______37 Palmital Formation__. ____._ -___ 22 Tabooes Quartzite__------____. 54 cross faults. ______98 Pantame Member, Elefante Formation...... 49 type Itacolomi__...... _.--...--. 69 erosion surfaces._.----..._._____. 12,13 Paving stone..------.------.--- 39,53 weathering.-__------__. 9 junction with Moeda syncline______88,95 Pebbles, elongation.___...-. ....- 23 water gaps._ ____...______8 Pediments.... - 10,13 R Serra do Itabirito...... ____... 87 Pegmatite dikes.. _. 80 References____...______...__. 104 cross faults...._.____.______gg Penninite.-..------.------...-..-.. - 43 Rio Carmo.-..______._.___.__ 7 faulting___ _.____.______88 Peti granite...--...... - 77 RiodasVelhas-.---- . _- - 7,8 Moeda Formation....____._____ 36 Peti phase, Borrachudos Granite. . 94 Rio das Velhas Series, age...___-____ 26 Serra do Ouro Branco. ______32 Pettijohn, F. J., quoted.____.-.-.-.. 26,58 canyons.-..------.-__.___...___ 8 erosion surface..______12 Pfiug, Reinhard, interpretation of Minas contact relations...... ___.....__ 36 Serra do Ouro PrSto, Moeda Formation. . __ 37 definition.------...... --..-.-...-...... - IB Serra do Tamandua, iron ore..______32 Phyllite, carbonaceous, Nova Lima Group. .. evolution of nomenclature. _--- .--_----- 64 Serra Geral, north side, correlation problems.. 66 carbonaceous, Batatal Formation ...... folding...._...... __-..----- 83 Sinkholes, Gandarela Formation_____ 10,45,47 Cercadinho Formation ...... --. interpretation.-..------.__._...-.-. 25 Sftio Largo Amphibolite ...._._.___ 44 dolomitic, Gandarela Formation. --, .-. ultramafic rocks..-----.--..---..------.- 79 Soapstone...... __.______. 78 ferrunginous, Nova Lima Group. ... - Rio das Velhas uplift...... 8,85,90,92 association with steatite...... ______81 graphitic, Nova Lima Group ... 20,26 Rio das Velhas valley, thrust fault.... _... - 97 Soil, dolomite....-.---...-...... -____ 10 Barreiro Formation.... -..-.. .. 55 Rio Doce...... 7 granitic rocks _____.__ 8,71 Batatal Formation...... 38 Rio Gualaxo._____...... __..... 7 quartzose rocks...... _.....__... 9 Itacolomi Series.... _.. - 59,61 Rio MaranhSo_____...___...._----- 8 Splash Rock.....-.--.--.- .--.___... 72 Moeda Formation...... -. - 36 Rio Mata Porcos, captured stream. ___. __ 8 Staurolite______..______. 56 Nova Lima Group. .. . -- 17 Rio Paraopeba______... 7,8 Sabara Formation-L______59 Palmital Formation...... 22 Rio Piracicaba.--.--.----_...... 7 Steatite.__...... __...... ___ 78 Santo Antonio facies, Itacolomi Series ... - 61 River silting...... 8 association with quartz ___.___. 81 silver.---.---..---.------51 Rivers in region._.__.___.____._._._--_ -- 7 Stibnite...... __..... 46 weathering.-.------.-.-- _ ---.--.-- 9 Road metal..____...... ___..... 39 Stratigraphic coi-relatton, history and prob- Pico do Frazao, erosion surface ------13 Roads..---..---.---..------.----.-- 7 lems ...... __ 63 Piedadesyncline...- _ ------..-. ... 89 Rutile ...... 81 Stratigraphy, problems in Precambrian rocks. 15 Piracicaba Group, Barreiro Formation ------64 Stream piracy. ______7 Cercadinho Formation. ----- . ... 49 Subgraywacke, Sahara Formation _____. 56 contact relations _ ------48 Sabara Formation------.---.-.-----.-----.- 55 weathering.. --... __ _.______9 definition.-.- ____ ....-...... 48 St. John del Rey Mining Co.-..--.----..-- - 6,16 Sul-Americano erosion surface.._....___ 13 depositions! en vironment ..-.------.--... 58 Sand, economic aspect_ ---.__ - - 54 diabase dikes.- ____ ...... ----- 80 glass .-.--.---...-...... -.-..----. 76 FScho do Funil Formation... _ ------52 Santa Barbara Gneiss..----..---..--...- 93,94 metamorphism.- -- -_ - - 59 Santa Ritasyncline...... _..___ 85,93,98,101 Tabooes Quartzite ... _ . __ . 10,54 Tabooes Quartzite - ...... ----- 54 Santana mine...... _...... 82 Talus ...... 76 Planation, pre-Minas- . -- -. ------28 Santo Antonio facies, Itacolomi Series..... _ 60 Tamandua Group, assignment- 64,65 Porphyry dikes...... -..-..-... 80 Sao Francisco...._._ ___.....__.. 7 Cambotas Quartzite------. - SO Post-Gondwanna erosion surface. ..------.-.- IS Sao Francisco Series.___..._...__..._ 77 contact relations... __ 66 relation tomudstone ...-.------. ... 70,71 Saprolite, argillaceous rocks..... __ -...-.... 9,71 definition-.--- - 28 Post-Itacolomi orogeny __ .-... .- - -...84, 101 definition- . 71 depositions! environment- 33 R io das Velhas uplift - 92,93 granitic rocks..-._-.--.--- . .- 8 equivalents- ___ ... 82 Precambrian rocks _ -.------.--..--.-- 14 Sabara Formation..--. ..----.--...-.... 55 unnamed formations __ -.- SI parent rocks.. .. __ ...... _ ..-.-. ..- 101 Schist, carbonaceous_.-----....----_..._ 20 Tear faults. _ -_..-._----. _ 97 spatial relations with mudstone _....- 70 garnet___.__..___...... __.....-- 21 Tertiary deposits.-----. -. 68 stratigraphic problems . . . . ------15 graphitic-.-.---...-.-..---...--..-.-...- 20 Thrust faults..------.-.-- 95 structural development ------82 Nova Lima Group..._...__._...... 17 20,57 Pre-Itacolomi deformation _ -...-.-.---...-- -84, 101 Sabara Formation..------.----.-.. 56 Titanium, in mudstone. . ____ . _ . __ _ 71 Pre-Muias deformation ------83, 100 staurolite.-_ ._-----____ -_ - -. 21 Topaz, imperial-.------.------. ___ . 52 Rio das Velhas uplift. ------92,93 talc-chlorite.--__.-...... ______78 Topography, relation to rock types .. - 8 Project geologists ... _ ...... - 6 unnamed formation.....-...... --.-.-.. 32 Tourmaline.....--- .... --... 18 36,39 weathering.--..------.....-. .___- 9 Project history.... _ . --..--...-. -.- 4 black . 80 .------.----. - 61 Sedimentary structures, Cambotas Quartzite. 31 Tourmalinlte - - - - 34,80 Pseudoitabirite __ ..... Sedimentary suites, significance.-- -- -.- 63 Pyrite ...... __ 46 Turbidltes---.- -.-.-- 57,62 ....--.-----.--- 55 Serpentinite______._...... ___.... 78 Barreiro Formation Serra da Moeda, Caraca Group,__...... 34 Pyrophyllite .. .-...... -..-..... -...... 52 Caueitabirite. 40 U cross faults__...... __.___ 98 Ultramafic rocks.. _ . ____ .. ___ . ___ 78 Q Moeda Formation..._...... _... 36 Ultrametamorphism. ___ .... __ __ .. 16 Quartz...... -...... --..- 81 water gap.--.-....----.--.-.-..-----..--- 8 Unconformities, significance. - _ ...... 63 Quartz-ankerite, Nova Lima Group ------_ 19 Serra da Eola Moca, erosion surface__------12 Unconformity, between Minas and Rio das Quartz-dolomite, Nova Lima Group. _ ------19 Serra das Cambotas_--.___...... ____. 8 Velhas Series.-- ..--.. .- 27,94 Quartz veins, mineral associations. --.--. . 81 erosion surface_---.---...... --.-.... 13 Uplift, epeirogenic, effect on physiography. . . 10 A110 INDEX

W Weathering Continued Page Page Moeda Formation...... A37 Vargem do Lima syncline... _ _ A22,23,24, 25, 93, 101 Water, Cau6 Itablrite...... -..-.. A39 products-.------.----..-----.----..-. 71 Vegetation of region.-_-_.__.______.._. 4 Weathering..------__-__------. 8 Velhas erosion surface--.------.--.------13 depth, argillaceous rocks...... 9 Volcanic ash, mudstone---.------.----.- 71 granitic rocks...... -.....-.----.-.- 8 Nova Lima Group...... 20 Minas Series...------28 Zoroastro Passos deposit, faulting-

U.S. GOVERNMENT PRINTING OFFICE : 1969 O 334-413